Malonate utilization by Pseudomonas aeruginosa affects quorum-sensing and virulence and leads to formation of mineralized biofilm-like structures

Abstract

Pseudomonas aeruginosa is an opportunistic pathogen that uses malonate among its many carbon sources. We recently reported that, when grown in blood from trauma patients, P. aeruginosa expression of malonate utilization genes was upregulated. In this study, we explored the role of malonate utilization and its contribution to P. aeruginosa virulence. We grew P. aeruginosa strain PA14 in M9 minimal medium containing malonate (MM9) or glycerol (GM9) as a sole carbon source and assessed the effect of the growth on quorum sensing, virulence factors, and antibiotic resistance. Growth of PA14 in MM9, compared to GM9, reduced the production of elastases, rhamnolipids, and pyoverdine; enhanced the production of pyocyanin and catalase; and increased its sensitivity to norfloxacin. Growth in MM9 decreased extracellular levels of N-acylhomoserine lactone autoinducers, an effect likely associated with increased pH of the culture medium; but had little effect on extracellular levels of PQS. At 18 h of growth in MM9, PA14 formed biofilm-like structures or aggregates that were associated with biomineralization, which was related to increased pH of the culture medium. These results suggest that malonate significantly impacts P. aeruginosa pathogenesis by influencing the quorum sensing systems, the production of virulence factors, biofilm formation, and antibiotic resistance.

Graphical Abstract

Pseudomonas aeruginosa utilizes malonate as a carbon source, and upon its growth in blood from trauma patients, P. aeruginosa increased the expression of malonate utilization genes. Here we report that malonate influences the production of several P. aeruginosa virulence factors, with part of this effect through its reversible inactivation of two of the P. aeruginosa cell-to cell communication molecules. Additionally, after 16 hours of growth, malonate induces P. aeruginosa to form biomineralized aggregates.

1 |. INTRODUCTION

Pseudomonas aeruginosa is a ubiquitous Gram-negative opportunistic pathogen that survives in a wide range of environments (Botzenhart & Döring, 1993). P. aeruginosa causes serious infections in immunocompromised patients, including bacteremia in burn and trauma patients, urinary tract infections in hospitalized patients, and respiratory infections in cystic fibrosis patients (Bodey et al., 1983). These infections are difficult to eradicate due to the ability of P. aeruginosa to form persistent biofilms as well as its resistance to commonly used antibiotics (Costerton et al., 1999). The World Health Organization listed P. aeruginosa as a major bacterial threat for which new antibiotics are urgently needed (WHO, 2017). Damage to the host caused by P. aeruginosa is due to its ability to produce numerous cell-associated and extracellular factors, most of which are controlled by the quorum sensing systems (Gellatly & Hancock, 2013). Extracellular virulence factors include exotoxin A, elastases (LasB and LasA), alkaline protease, type III secretion system effector molecules, and pyocyanin, while cell-associated virulence factors include the flagellum, type IV pili, exopolysaccharides, and lipopolysaccharides (Gellatly & Hancock, 2013). At different infection sites, P. aeruginosa exists within biofilms; sessile, complex, and highly structured communities that are surrounded by exopolymeric substances (Costerton et al., 1999). Within the biofilm, bacterial cells are protected from the effects of the host immune response (Domenech et al., 2013). Additionally, within the biofilm, bacterial resistance to antibiotics is considerably higher than that of the planktonic free-living states (Høiby et al., 2010). Besides the typical surface-attached biofilms, P. aeruginosa forms aggregates within the cystic fibrosis lung (Bjarnsholt et al., 2009, Singh et al., 2000). These aggregates have characteristics of typical biofilms in that they increase resistance to host defenses and antibiotics, have a composition of host (mucin, DNA) and microbial (DNA, alginate) substances (Bjarnsholt et al., 2009, Palmer et al., 2005). Growth of P. aeruginosa in sputum from cystic fibrosis patients induces formation of these aggregates (Palmer et al., 2005). Using artificial sputum medium, we showed that production of the aggregates, or biofilm like structures (BLS) required mucin, extracellular DNA, and at least 10% oxygen (Haley et al., 2012). Recently, it has been postulated that such aggregates/BLS may be the forerunners of typical surface-attached biofilms (Kragh et al., 2016). Additionally, BLS/aggregates formed by E. coli, P. aeruginosa, and other bacteria have been found within human cells, providing another means of bacterial evasion of antibiotic therapy (Mirzaei et al., 2020).

We recently examined the effect of trauma-induced changes in blood on the P. aeruginosa transcriptome by growing P. aeruginosa strain UCBPP-PA14 (PA14) in whole blood from either healthy volunteers or trauma patients and comparing the difference in gene expression. Results showed that in comparison with whole blood from healthy volunteers, the growth of P. aeruginosa in whole blood from trauma patients significantly altered the expression of numerous genes (Elmassry et al., 2019). Among those genes whose expression was significantly induced were five genes that code for the malonate decarboxylase, referred to as the malonate-utilization operon, which includes two additional genes mdcABCDEGH (https://www.pseudomonas.com/, accessed April 2021) (Maderbocus et al., 2017, Winsor et al., 2016). In bacteria, there are two categories of malonate decarboxylases; a biotin-dependent membrane enzyme and a biotin-independent cytosolic enzyme (Chohnan & Takamura, 2004). P. aeruginosa, as an aerobic/facultatively anaerobic species, carries only the biotin-independent cytosolic enzyme (Chohnan & Takamura, 2004). The P. aeruginosa malonate decarboxylase consists of four subunits – MdcC is the acyl-carrier protein (ACP), MdcA is an acetyl-ACP:malonate ACP transferase, and MdcD and MdcE have decarboxylase activity (Maderbocus et al., 2017). Together, MdcB, MdcG, and MdcH activate the MdC component for catalysis (Maderbocus et al., 2017).

Through its involvement in the fatty acid biosynthesis, malonate, a dicarboxylic acid, is essential for carbon metabolism in bacteria (Gueguen et al., 2000, Jordan et al., 1986, Schweizer & Hofmann, 2004). P. aeruginosa utilizes malonate as a carbon source by decarboxylating it to acetate (Gray, 1952, Maderbocus et al., 2017). Malonate also influences P. aeruginosa function by inhibiting cell respiration and the activity of succinate dehydrogenase (Bowman et al., 2017). Currently, very little is known regarding the effect of malonate utilization on P. aeruginosa pathogenesis. Recent studies suggested that, through its inhibition of cell respiration, malonate increases the resistance of P. aeruginosa to aminoglycosides (Elmassry et al., 2019, Meylan et al., 2017). Other studies showed that malonylation, a recently discovered post-translational modification in which a malonyl group is added to lysine residues of proteins, may influence the functions of these proteins (Bowman et al., 2017, Gaviard et al., 2019, Peng et al., 2011, Qian et al., 2016). Additionally, malonate may influence the function of the quorum sensing (QS) systems, which strongly regulate the production of numerous virulence factors including LasB. Despite the potential relationship of malonate to virulence, the only recent report on malonate decarboxylase in PA14 described its crystal structure (Maderbocus et al., 2017).

In this study, we show that malonate utilization by P. aeruginosa impacts quorum sensing, virulence factor production, and antibiotic resistance. It influences P. aeruginosa virulence by significantly altering the expression of different QS genes and the production of QS-associated virulence factors. Additionally, when grown to a late stage of growth in minimal medium containing malonate as a sole carbon source, P. aeruginosa forms structures that resemble biofilms (biofilm-like structures, BLS) that are associated with mineral precipitation, also known as biomineralization (Heim, 2011, Weiner, 2003).

2 |. RESULTS

2.1 |. Malonate as a sole carbon source supports the growth of P. aeruginosa at different stages of growth

To investigate the effect of trauma-induced changes in human whole blood on the expression of P. aeruginosa genes, we grew P. aeruginosa strain UCBPP-PA14 (PA14) in whole blood from either trauma patients or healthy volunteers and compared the expression of different genes using RNA-seq analysis. Among the genes whose expression was significantly increased upon the growth of PA14 in blood from trauma patients were the five genes (mdcA-E) of the malonate-utilization operon (Elmassry et al., 2019). Using publicly available data at the PATRIC (Pathosystems Resource Integration Center, https://www.patricbrc.org/ accessed February 2021), an online resource that stores and integrates a variety of data types (Wattam et al., 2017) we analyzed the average expression of the five genes in 63 transcriptomic experiments in P. aeruginosa. Comparisons of 247 different experimental conditions utilized in these 63 experiments revealed that the expression of mdcA-E in P. aeruginosa is induced or repressed by several factors (Figure S1) (Wattam et al., 2017). Among the factors that upregulated the expression of the malonate-utilization operon were phosphatidylcholine, cobalamin, norepinephrine, cyclamate, mucin, and cystic fibrosis sputum (Figure S1). In contrast, the following factors downregulated the expression of the operon: azithromycin, vanadate, and growth within a biofilm (Figure S1). Despite these findings, little is known about the role of malonate in influencing the virulence of P. aeruginosa. Therefore, in this work, we investigated the effect of malonate utilization on the production of virulence factors, the function of the QS systems, and antibiotic resistance of the P. aeruginosa strain PA14.

To test the ability of malonate as a sole carbon source to support the growth of PA14, we grew PA14 in M9, a minimal medium, containing malonate as a sole carbon source (MM9) (Experimental Procedures). The growth of PA14 in MM9 was compared to its growth in M9 containing glycerol as a sole carbon source (GM9) over a 48-h period by determining the growth index (OD₆₀₀) (Figure 1a). The density of the PA14 culture in MM9 was higher than that in GM9 through 20 h post inoculation (hpi) (Figure 1a). At 22–48 hpi (late stationary phase), the growth of MM9 was lower than that in GM9 (Figure 1a). Further inspection of the culture showed that, by 48 hpi, PA14 cultured in MM9 formed aggregates that were resistant to dispersal by vortexing (Figure S2a). As the formation of these aggregates may have prevented the accurate assessment of the PA14 growth index by OD₆₀₀ measurement, we sonicated the remaining volumes of both GM9 and MM9 cultures and determined the OD₆₀₀. The OD₆₀₀ of the sonicated samples was higher in MM9 from 12 hpi onward until 48 hpi when the OD₆₀₀ was higher in GM9 (Figure 1b). Comparison of the pre- and post-sonication OD₆₀₀ values showed no significant difference for PA14 grown in GM9 (Figure S2b). While there was no difference in the OD₆₀₀ for PA14 grown in MM9 through 16 hpi, from 18 hpi through 48 hpi, the OD₆₀₀ values for PA14 grown in MM9 were much higher after sonication (Figure S2c). We also determined the colony forming units (CFU) present in the cultures at each time point. The CFU of PA14 grown in MM9 was higher than that of PA14 grown in GM9 at most time points until 22 h, when the CFU were significantly lower in MM9 than GM9 (Figure 1c). We also calculated the aggregation index at each time point by dividing the post-sonication OD₆₀₀ values by the pre-sonication OD₆₀₀ values as previously described (Kharadi & Sundin, 2019). This aggregation index was much higher in MM9 from 18 hpi through 48 hpi (Figure 1d). For further experiments to analyze the expression of QS and QS-related genes and the production of virulence factors, PA14 was growth in either MM9 or GM9 for 16 hpi, the time point prior to the formation of the aggregates and the last time point where there was no difference in the sonicated and unsonicated OD₆₀₀ values for PA14 grown in MM9 (Figure S2c).

FIGURE 1.

The growth of PA14 in MM9 and GM9 varies. An aliquot of overnight PA14 culture in lysogeny broth (LB) was inoculated into GM9 or MM9 to a starting OD₆₀₀ of 0.003. Cultures were incubated at 37°C with shaking. Samples were obtained at the indicated intervals and the OD₆₀₀ was determined. (a) PA14 growth index is higher in MM9 until 20 h post-inoculation (hpi). The samples for this graph were not sonicated. Visible aggregates of cells were noted in the MM9 cultures beginning at 18 hpi. (b) Aggregation of cells in the MM9 cultures altered the OD₆₀₀. The remaining volumes of both GM9 and MM9 cultures at each time point were sonicated for 30 s and the OD₆₀₀ was determined. (c) Colony forming units (CFU) of PA14 in MM9 were higher than those in GM9 at most time points until 22 hpi. The sonicated samples were diluted tenfold and plated to determine the CFU present in each sample. Data were log-transformed prior to graphing. (d) The aggregation index of PA14 in MM9 jumps sharply at 18 hpi. The aggregation index was calculated as the ratio of OD₆₀₀ post-sonication to OD₆₀₀ pre-sonication (Kharadi & Sundin, 2019). In each panel, values represent the means of three independent experiments ± standard error of the mean (SEM). Statistical significance was determined by two-way ANOVA with Šídák’s multiple comparisons test; *P < 0.05, **P < 0.01, ***P <0.001, ****P < 0.0001.

2.2 |. Malonate utilization influences the expression of QS and QS-related genes and the production of virulence factors

Preliminary visual inspection of the overnight PA14 cultures in MM9 or GM9 suggested that the growth of PA14 in MM9 induces pyocyanin production; compared to GM9, the MM9 culture was dark blue (Figure S2d). Together with several other virulence factors, pyocyanin production is stringently controlled by the QS systems. Thus, to examine the effect of malonate as a sole carbon source on the expression of the QS genes lasI, lasR, rhlI, and rhlR, we utilized PA14 strains containing plasmids that carry lacZ transcriptional fusions in each of these genes (Table 1). We grew PA14/pPCS1001 (lasR-lacZ), PA14/pPCS223 (lasI-lacZ), PA14/pPCS1002 (rhlR-lacZ), and PA14/pLPRI (rhlI-lacZ) in either MM9 or GM9 for 16 hpi and analyzed the level of β-galactosidase activity as previously described (Miller, 1972, Stachel et al., 1985). Compared with its growth in GM9, the growth of PA14 in MM9 significantly reduced the expression of lasR, lasI, rhlR, and rhlI (Figure 2a and 2b). Using qRT-PCR, we also examined the level of expression of the pqsC gene as a representative of the pqsA-E operon. Compared with its growth in GM9, the growth of PA14 in MM9 significantly increased pqsC expression (Figure 2c).

Table 1.

Strains and plasmids used in this study

Strains	Description	Reference
PA14	P. aeruginosa strain UCBPP-PA14; protrotroph; clinical isolate from burn patient	(Rahme et al., 1995)
PA103	P. aeruginosa hypertoxigenic strain, nonproteolytic; protrotroph	(Liu, 1973)
PAK	P. aeruginosa strain K; hyperpiliated; sensitivity to Pf1 phage; protrotrophic	(Bradley, 1972, Takeya & Amako, 1966)
PAO1	P. aeruginosa protrotrophic strain; clinical isolate from a wound	(Holloway et al., 1979)
PA14/MrT7/insertions in  PA14_24480 (pelA)  PA14_72380 (algB)  PA14_39970 (phzA2)  PA14_09400 (phzS)  PA14_09490 (phzM)  PA14_50360 (flgK)  PA14_05380 (pilK)  PA14_17530 (recA)  PA14_25150 [sulA]†  PA14_17470 [nlpD]†  PA14_51430 (pqsA)  PA14_45940 (lasI)  PA14_19130 (rhlI)  PA14_40260 [bapA]†  PA14_55810 [pprB]†  PA14_31290 (pa1L)  PA14_09520 (mexI)	PA14 with MAR2xT7 mariner transposon insertion within the specific gene; Gmr	(Liberati et al., 2006); http://pa14.mgh.harvard.edu/cgi-bin/pa14/home.cgi accessed February 2021
PAO1pqsA CTX-lux::pqsA	Reporter strain for PQS and HHQ; pqsA mutant of PAO1 containing a copy of the pqsA promoter linked to the luxCDABE genes and inserted into a neutral site in the chromosome	(Fletcher et al., 2007)
E. coli JM109 (psB1142)	Report strain for 3OC12-HSL; Escherichia coli JM109 carrying lasR and the las promoter of P. aeruginosa fused to the luxCDABE cassette from Photorhabdus luminescens; responds to long-chain AHLs (C10-C14); Tcr (Tc 20 μg/ml)	(Flynn et al., 2016, Winson et al., 1998)
E. coli JM109 (pSB536)	Reporter strain for C4-HSL; JM109 carrying ahyR and the ahyI promoter fused to luxCDABE; responds to short chain AHLs (C4); Apr (25 μg/ml)	(Flynn et al., 2016, Winson et al., 1998)
Plasmids	Description	Reference
pPCS1001	lasR-lacZ transcriptional fusion in pLP170; Apr, Cbr	(Pesci et al., 1997)
pPCS223	lasI-lacZ transcriptional fusion in pLP170; Apr, Cbr	(Van Delden et al., 1998)
pPCS1002	rhlR-lacZ transcriptional fusion in pLP170; Apr, Cbr	(Pesci et al., 1997)
pLPRI	rhlI-lacZ transcriptional fusion in pLP170; Apr, Cbr	(Van Delden et al., 1998)
ppvdD	pvdD-lacZ transcriptional fusion in pMP190; Smr	(Rombel et al., 1995)
pMP220::PpvdS	pvdS::lacZ translational fusion in pMP220; Tcr	(Ambrosi et al., 2002)
pSW228	toxA::lacZ translational fusion in pSW205; Apr, Cbr	(West et al., 1994)
pRL88	regA(P1/P2)::lacZ translational fusion in pSW205; Apr, Cbr	(Storey et al., 1990)
pDH10	pchR::lacZ translational fusion; Smr, Cmr	(Heinrichs & Poole, 1996)

^†Genes names in brackets are names of the PAO1 orthologs for the PA14 gene. Orthologs were obtained from the Pseudomonas Ortholog Database within the Pseudomonas Genome Database (Winsor et al., 2016).

FIGURE 2.

Malonate as a sole carbon source affected the expression of QS genes in PA14 at a late stage of growth. PA14 alone (for pqsC) or carrying lacZ transcriptional fusion plasmids for lasR, lasI, rhlR, or rhlI (Table 1) was grown overnight in LB and inoculated into GM9 or MM9 to OD₆₀₀ 0.04. All cultures were harvested at 16 hpi. The level of β-galactosidase activity was determined and each value was normalized to the OD₆₀₀ of the respective culture. Expression of the las genes (a) and the rhl genes (b) was reduced in PA14 grown in MM9. (c) Expression of pqsC, representative of the PQS synthesis operon, was determined by qRT-PCR in PA14. In contrast to las and rhl, pqsC expression was enhanced in PA14 grown in MM9. Gene expression was normalized using the 16S ribosomal RNA gene PA14_08570 as an internal control. Values in all panels represent the means of three independent experiments ± SEM. Statistical significance was calculated by two-tailed unpaired t-test; *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Since QS was affected by the growth of PA14 in MM9, we assessed the effect of PA14 growth in MM9, in comparison with GM9, on the production of QS-related factors LasB, LasA, rhamnolipids, pyocyanin, and catalase. Compared with its growth in GM9, the growth of PA14 in MM9 significantly reduced the level of LasB elastolysis in the elastin Congo red assay (Figure 3a) (Hamood et al., 1996). The percentage of LasA staphylolysis in cultures of PA14 grown in MM9 was significantly lower than that in cultures grown in GM9 (Figure 3b) (Diggle et al., 2002). Similarly, the growth of PA14 in MM9 resulted in a significantly reduced level of rhamnolipids (Figure 3c) (Pinzon & Ju, 2009). In contrast, the amount of pyocyanin extracted from the supernatant of PA14 grown in MM9 was significantly higher than that when it was grown in GM9 (Figure 3d) (Essar et al., 1990, Schaber et al., 2004). The amount of catalase breakdown of hydrogen peroxide by PA14 grown in MM9 was also significantly higher than in GM9 (Figure 3e) (Iwase et al., 2013). We also examined the effect of PA14 growth in MM9 on the expression of pa1L, which codes for PA-IL galactophilic lectin (a virulence-associated protein) (Gilboa-Garber, 1972, Grishin et al., 2015, Winzer et al., 2000). Using qRT-PCR, we detected comparable levels of pa1L expression in PA14 that was grown in either GM9 or MM9 (Figure 3f).

FIGURE 3.

Growth of PA14 in MM9 significantly affected the production of several virulence factors. PA14 was grown as described in Figure 2 and cultures were harvested at 16 hpi. Cells were pelleted and the supernatants separated. Supernatants were used to determine levels of LasB, LasA, rhamnolipids and pyocyanin while the pellets were used to determine levels of catalase and pa1L expression. The supernatants were used directly to determine the levels of (a) LasB release of Congo red from elastin, (b) LasA lysis of staphylococcal cells, and (c) levels of rhamnolipids using the colorimetric methylene blue assay. (d) Pyocyanin was extracted from the supernatant as described in Experimental Procedures. (e) Catalase levels were assessed by the foam test using pelleted cells. Values for the assays described in panels (a) through (e) were normalized to the growth index (OD₆₀₀) of the respective cultures. (f) Expression of pa1L that encodes PA-IL galactophilic lectin was determined by qRT-PCR using mRNA extracted from pelleted cells. Gene expression was normalized using the 16S ribosomal RNA gene PA14_08570 as an internal control. Values in all panels represent the means of three independent experiments ± SEM. Statistical significance was calculated by two-tailed unpaired t-test for all except panel (b) which was done by two-way ANOVA with Šídák’s multiple comparisons posttest; *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

It is possible that the difference in concentrations of malonate (40 mM) and glycerol (110 mM) in MM9 and GM9 affected the production of virulence factors. While the 110 mM (1% v/v) concentration of glycerol has been used extensively since 1980 (Ohman et al., 1980), we found no standard concentration for malonate when used as a sole carbon source. The concentration of 40 mM malonate fell in the middle of the range of reported concentrations used (Janssen & Harfoot, 1990, Maderbocus et al., 2017, Meylan et al., 2017, Minato et al., 2013, Stoudenmire et al., 2017). Therefore, we used increasing concentrations of malonate or decreasing concentrations of glycerol and examined the growth of PA14 and the production of pyocyanin at each concentration. Interestingly, while the changes were not significant, the growth index (OD₆₀₀ of sonicated sample) at 18 hpi decreased with decreasing glycerol concentrations and also decreased with increasing malonate concentration (Figure S3a). A similar pattern was observed with levels of pyocyanin extracted from the supernatants of the cultures grown in different concentrations of glycerol and malonate; however, the level of pyocyanin production was significantly increased in all concentrations of malonate compared to any concentration of glycerol (Figure S3b). Finally, we used equimolar (100 mM) concentrations of glycerol and malonate and examined the growth index of the sonicated samples (OD₆₀₀), the aggregation index (indicative of formation of the mineralized biofilm-like structures [MBLS] described in sections 2.6 and 2.7), and pyocyanin production at 24 hpi. There was no significant difference in the growth index, but as we had seen with the original concentrations of glycerol and malonate, the aggregation index and level of pyocyanin were significantly increased (Figure S4). Finally, MBLS formed after 16 hpi in all molar concentrations of malonate in MM9 and did not form at any concentration of glycerol in GM9 (Table S3). This indicates that the aggregation was not induced by cell starvation in MM9.

These results suggest that, as a sole carbon source and in comparison with glycerol, malonate differentially affects the QS-regulated virulence factors. It significantly reduced the expression of the las and rhl genes but enhanced the expression of pqsC. Consequently, the production of the las/rhl stringently controlled factors, LasB, LasA, and rhamnolipids, was significantly reduced while that of the PQS-controlled pyocyanin was significantly increased.

2.3 |. Alkaline conditions created by malonate utilization affect the stability of 3OC₁₂-HSL and C₄-HSL but not HHQ/PQS within the culture supernatant of PA14

As density-dependent communication molecules, and upon the growth of PA14 in GM9 to late stages of growth (16 hpi), we expected the extracellular levels of autoinducers to be considerably higher than their intracellular levels. First, we compared the levels of the N-acylhomoserine lactone autoinducers N-3-oxododecanoyl homoserine lactone (3OC₁₂-HSL) and N-butanoyl-L-homoserine lactone (C₄-HSL) and the Pseudomonas quinolone signal 2-heptyl-3-hydroxy-4(1H)-quinolone (PQS) and its precursor 2 heptyl-4-quinolone (HHQ) produced by PA14 grown in either GM9 or MM9 (Fletcher et al., 2007, Flynn et al., 2016, Winson et al., 1998). Compared with GM9, the growth of PA14 in MM9 reduced the extracellular levels of 3OC₁₂-HSL from 15,000 to ten relative light units (RLU) and that of C₄-HSL from 9 × 10⁸ to one RLU (Figure 4a and 4b). However, the extracellular level of HHQ/PQS was only reduced by half, from 18,000 to about 9,000 RLU (Figure 4c). In contrast, the influence on the intracellular levels of the three autoinducers was not the same. Compared to its growth in GM9, the growth of PA14 in MM9 significantly reduced the intracellular levels of 3OC₁₂-HSL and C₄-HSL but significantly increased the intracellular level of HHQ/PQS (Figure 4). Second, we compared the relative amounts of extra- and intracellular autoinducers produced upon the growth of PA14 in each medium. Considering the combined extracellular and intracellular levels as 100% (the total level of the synthesized autoinducer), the percentages of the extracellular 3OC₁₂-HSL, C₄-HSL, and HHQ/PQS produced by PA14 grown in GM9 were 91, 96.3, and 97%, respectively (Figure 4d). In contrast, the percentages of the extracellular levels of 3OC₁₂-HSL and C₄-HSL were 1.4 and essentially zero of their synthesized amounts (Figure 4d). In contrast, the extracellular level of the HHQ/PQS was higher than the intracellular one; 65% of the total synthesized HHQ/PQS was extracellular (Figure 4c and 4d). These results suggest that upon the growth of PA14 in MM9, and in contrast to HHQ/PQS, 3OC₁₂-HSL and C₄-HSL are either inefficiently secreted to the extracellular environment or degraded extracellularly.

FIGURE 4.

Growth of PA14 in MM9 alters the intracellular and extracellular levels of autoinducers. Cultures of PA14 were grown as described in Figure 2. At 16 hpi, cultures were harvested and the levels of extracellular (within the supernatant fractions) and intracellular (within lysates from the cell pellets) autoinducers were extracted (Experimental Procedures). Aliquots of the extracts were placed on standardized amounts of freshly diluted overnight cultures of the reporter strains E. coli JM109/pSB1142 for 3OC12-HSL, JM109/pSB536 for C4-HSL, and P. aeruginosa PAO1pqsA CTX-lux::pqsA for HHQ/PQS (Table 1) in wells of a luminometer plate. Following incubation for 3 h, luminescence, which is proportional to the amount of autoinducer specific for each reporter, was measured and the background level subtracted. Luminescence values, reported as relative light units (RLU), were normalized using OD₆₀₀ value of each respective culture. (a) Extracellular and intracellular levels of 3OC₁₂-HSL. (b) Extracellular and intracellular levels of C₄-HSL. (c) Extracellular and intracellular levels of HHQ/PQS. Values for panels (a), (b) and (c) represent the means of three independent experiments ± SEM. Data were log-transformed prior to graphing. Statistical significance was calculated by two-tailed unpaired t-test; *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. (d) The relative proportions of extracellular and intracellular levels of autoinducers in GM9 and MM9.

At 24 hpi, the pH of the PA14 MM9 culture reached 9.5 while that of the GM9 culture remained at 7.5. We hypothesized that the pH of PA14 culture in MM9 influenced the stability of one or all three QS molecules within the extracellular environment. Yates et al. (Yates et al., 2002) previously demonstrated that at an alkaline pH, N-acylhomoserine lactones (such as 3OC₁₂-HSL and C₄-HSL) undergo lactonolysis (reversible opening of the lactone ring at high pH). To test our hypothesis that pH, rather than carbon source, is responsible for the loss of the autoinducers, we grew PA14 in GM9 for 16 hpi and separated the supernatant. One fraction of the supernatant was adjusted to pH 9.5 and another fraction was left at pH 7.5. After 24 h at 37°C, the supernatant at pH 7.5 contained significantly higher levels of 3OC₁₂-HSL and C₄-HSL than the fraction at pH 9.5 (Figure 5). In contrast, the levels of HHQ/PQS, which are not affected by lactonolysis, were comparable (Figure 5). As lactonolysis is reversible by changing the pH to 2.0 (Yates et al., 2002), we then grew PA14 in MM9 for 24 hpi to reach pH 9.5 and harvested the supernatant. Supernatant fractions were adjusted to pH 2, 5, and 7.5 and one was left at 9.5. After 24 h at 37°C, 3OC₁₂-HSL levels recovered in the fractions adjusted to pH 2.0 and 5.0 and C₄-HSL levels recovered at pH 2.0 (Figure S5). These results strongly suggest that the increased pH within the culture medium when PA14 is grown in MM9 inactivates 3OC₁₂-HSL and C₄-HSL through lactonolysis, thereby interfering with the function of both the las and rhl QS systems. As a result, the production of different 3OC₁₂-HSL- and C₄-HSL-related virulence factors is significantly reduced. In contrast, pyocyanin production, which is strongly controlled by the pqs QS system, was significantly increased (Figure 3d).

FIGURE 5.

Alkaline pH within the growth medium affected the stability of 3OC₁₂-HSL and C₄-HSL but not HHQ/PQS. PA14 was grown in GM9 for 16 hpi. The supernatant fraction was separated and divided into two aliquots; one untreated (pH 7.5) and the other adjusted to pH 9.5, the pH of MM9 at 24 hpi with PA14. Both aliquots were incubated overnight at 37°C and the autoinducers were then extracted from the supernatants. Levels of autoinducers were determined by the luminescence assay as described for Figure 4. Values represent the means of three independent experiments ± SEM. Data were log-transformed prior to graphing. Statistical significance was calculated by two-tailed unpaired t-test; ****P < 0.0001.

2.4 |. Malonate utilization reduces the expression of the siderophores genes

Since no supplemental iron was added to GM9 or MM9, both media are iron-deficient, which allowed us to determine if malonate influences siderophore production in PA14 through a potential iron-independent mechanism. Of the two main siderophores produced by P. aeruginosa (pyoverdine and pyochelin), pyoverdine is more readily measured (Stintzi et al., 1996). Compared with its growth in GM9, the growth of PA14 in MM9 significantly reduced pyoverdine production at 16 hpi (Figure 6a). We confirmed these results by analyzing the expression of the pyoverdine operon using the lacZ fusion plasmid ppvdD. Compared with its growth in GM9, the growth of PA14/ppvdD in MM9 significantly reduced the level of β-galactosidase activity (Figure 6b). As the toxA gene and its transcriptional activator regA are negatively regulated by iron (Frank et al., 1989), we determined the effect of malonate on the expression on these two genes using toxA-lacZ and regA-lacZ fusions plasmids in PA14 (Table 1). Both fusion plasmids produced significantly lower levels of β-galactosidase activity when PA14 was grown in MM9 compared with GM9 (Figure 6b). Malonate may influence the expression of the above described iron-regulated genes through its effect on their main transcriptional regulator, pvdS (Cunliffe et al., 1995, Frank et al., 1989). To test this possibility, we utilized a pvdS-lacZ fusion plasmid in PA14 (Table 1). Compared with its growth in GM9, the growth of PA14 in MM9 significantly reduced the level of β-galactosidase activity produced from the pvdS promoter (Figure 6b). If the observed phenomenon is not related to iron regulation, exogenously added iron would further repress the expression of these genes in PA14 grown in MM9. Thus, we grew PA14 either in MM9 or MM9 containing exogenously added iron. Both pyoverdine production and pvdS expression were repressed upon the addition of exogenous iron (Figure S6). To assess the effect of malonate on the siderophore pyochelin, we determined the level of expression of the main regulator of pyochelin synthesis, pchR (Michel et al., 2005), using the pchR-lacZ fusion plasmid pDH10 in PA14 (Table 1). The level of β-galactosidase activity produced by pDH10 during the growth of PA14 in MM9 was significantly lower than that detected when the strain was grown in GM9 (Figure 6b). These results suggest that as a sole carbon source, malonate significantly represses the expression of several iron-regulated genes through a mechanism unrelated to iron.

FIGURE 6.

The growth of PA14 in MM9 significantly reduced the production of pyoverdine and the expression of siderophore genes and other iron-regulated genes. PA14 alone (for pyoverdine assay) or carrying lacZ transcriptional fusion plasmids for pvdD, toxA, regA, pvdS, or pchR (Table 1) was grown overnight in LB and inoculated into GM9 or MM9 to OD₆₀₀ 0.04. All cultures were harvested at 16 hpi. (a) Pyoverdine production was analyzed by spectrophotometry (A₄₀₅) and the values were normalized to the OD₆₀₀ of the respective culture. (b) The level of β-galactosidase activity from each gene fusion was determined and the values were normalized to the OD₆₀₀ of the respective culture. In both panels, values represent the means of three independent experiments ± SEM. Statistical significance was calculated by two-tailed unpaired t-test; **P < 0.01, ***P < 0.001, and ****P < 0.0001.

2.5 |. Malonate utilization increases sensitivity of P. aeruginosa to norfloxacin

Previous studies suggested that malonate induces P. aeruginosa resistance to aminoglycosides (e.g., kanamycin and tobramycin) through its inhibitory effect on cell respiration (Elmassry et al., 2019, Meylan et al., 2017). It is possible that malonate may influence the resistance of P. aeruginosa to other classes of antibiotics that are structurally different from aminoglycosides. To test this possibility, we determined the minimum inhibitory concentration (MIC) of norfloxacin, erythromycin, chloramphenicol, rifampicin, vancomycin, and colistin for PA14 that was grown in either MM9 or GM9. With the exception of norfloxacin, a member of the fluoroquinolone class of antibiotics, the MIC of the tested antibiotics against PA14 was comparable regardless of the medium in which it was grown (Table S1). The MIC of norfloxacin for PA14 that was grown in MM9 was lower than that when the strain was grown in GM9 (one μg/ml vs. eight μg/ml, respectively) (Table S1). These results suggest that while malonate increases P. aeruginosa resistance to aminoglycosides, it decreases its resistance to norfloxacin and possibly other fluoroquinolone antibiotics.

2.6 |. PA14 forms biofilm-like structures in the presence of malonate as a sole carbon source

After 48 h of incubation under shaking conditions at 37°C, the PA14 culture in GM9 was visibly different from that in MM9. The culture in GM9 was homogenous and turbid while the MM9 culture contained aggregates that settled to the bottom of the culture flask leaving the rest of the culture clear (Figure S2a). Such a phenomenon may influence biofilm formation by PA14 in MM9. To examine this possibility, we compared biofilm formation by PA14 in GM9 with that in MM9. Biofilms were formed on the walls of polystyrene tubes and stained with crystal violet using the previously described protocol (Déziel et al., 2001, O’Toole & Kolter, 1998). When grown in GM9, PA14 formed a biofilm with a considerable biomass (Figure 7a). In contrast, PA14 grown in MM9 produced a very limited biofilm (Figure 7a). Rather, the MM9 biofilm medium contained numerous aggregates suggesting that, during its growth in MM9, PA14 failed to attach to the surface of the tube and form a typical biofilm. We previously demonstrated that, in the presence of mucin, P. aeruginosa formed similar floating aggregates that we termed “biofilm-like structures” (BLS) (Haley et al., 2012). Visualization by both confocal laser scanning microscopy and electron microscopy (EM) revealed that those BLS closely mimic typical biofilms (Haley et al., 2012). Our EM analysis of the PA14 MM9-induced BLS revealed their close resemblance to the previously described BLS (Haley et al., 2012). Individual bacteria were embedded in an apparent exopolymer matrix (Figure 7b). In addition, individual bacteria formed long chains by attaching to each other (Figure 7b). This phenomenon is not limited to strain PA14 as other P. aeruginosa strains including, PAO1, PA103, and PAK produced similar BLS when grown in MM9 (in comparison with their growth in GM9) (Table S2).

FIGURE 7.

Growth of PA14 in MM9 significantly reduced biofilm development and led to formation of biofilm-like structures (BLS). (a) When grown in MM9 for 48 hpi, PA14 formed significantly less biofilm on the walls of the polystyrene tubes than those formed when the strain was grown in GM9. Biofilm biomass was measured using the crystal violet assay and normalized using the OD₆₀₀ value of the respective culture. Values represent the means of three independent experiments ± SEM Statistical significance was calculated by two-tailed unpaired t-test; **P < 0.01. (b) Electron micrographs of the BLS that formed when PA14 was grown in MM9. The BLS were collected from 24 h cultures of PA14 and visualized using scanning electron microscopy. Images are representative of three samples. Magnification and scale: top left, × 5,000, 10 μm; top right, × 4,000, 10 μm; bottom left, × 5,000, 10 μm; bottom right, × 15,000, 2 μm. Scale is indicated on each image (black bars).

We further investigated the composition of the BLS. The two major components that significantly contribute to the formation of the biofilm exopolymeric substances (EPS) are the polysaccharides and DNA (Allesen-Holm et al., 2006, Ryder et al., 2007). P. aeruginosa biofilms were dissolved when treated with either glycoside hydrolases, which destroys the polysaccharides, or DNases (Allesen-Holm et al., 2006, Ryder et al., 2007). Therefore, we suspended the BLS in PBS and treated the suspensions with either glycoside hydrolases (a mixture of bacterial α-amylase [from Bacillus subtilis; MP Biomedicals, Irvine, CA, USA] and fungal cellulase [from Aspergillis niger; MP Biomedicals, Irvine, CA, USA]) or RNAse-free DNase I (QIAGEN, Valencia, CA, USA) at a concentration of 0.25% (w/v) and 2.73 Kunitz units/μl respectively. Within 15 min, both treatments dissolved the BLS suggesting that the composition of the BLS is similar to that of typical P. aeruginosa biofilm (Figure 8a). Several P. aeruginosa cell-associated and extracellular products including pyocyanin, alginate, pili, and flagellum significantly contribute to different stages of biofilm development (Laverty et al., 2014). One or more of these products may contribute to the development of BLS when PA14 is grown in MM9. We utilized PA14 specific mutants that carry mutations in the following genes: algB or pelA (exopolysaccharide production); phzA2 or phzM (phenazine production); flgK (flagellar system); pilK (chemotactic system); and recA, sulA, or nlpD (stress response and cell division) genes. We grew each mutant in either GM9 or MM9 and analyzed BLS formation. All mutants produced BLS in MM9 but not GM9, suggesting that none of these genes contribute to the formation of BLS in MM9 (Table S2). We also eliminated the possible association between BLS formation and the P. aeruginosa QS systems. We previously showed that deletion of either rhlI or rhlR gene compromised the mucin-induced BLS formation in P. aeruginosa (Haley et al., 2012). Thus, we assessed BLS formation by different PA14 QS mutants. PA14 defective in lasI, rhlI, or pqsA formed BLS similar to that formed by their PA14 parent strain (Table S2).

FIGURE 8.

The BLS contain DNA, exopolysaccharides, and minerals. (a) The BLS contain DNA. Left: BLS isolated from a 24-h culture of PA14 in MM9 and suspended in PBS show no change. Right: The BLS dissolved upon their treatment with DNase for 15 minutes at room temperature. The same results were obtained using glycoside hydrolases, indicating the presence of exopolysaccharide in the BLS. (b) The BLS are mineralized. Tubes on the left are pellets from 24-h cultures of PA14 in GM9 and tubes on the right are pellets from 24-h cultures of PA14 in MM9. The BLS stained purple (appears dark gray in grayscale photograph) with alizarin red S (right) indicating the presence of biomineralized magnesium and/or calcium.

2.7 |. Biomineralization induced during the growth of PA14 in MM9 is critical for the development of BLS

Besides their typical biofilm-like features (i.e., the encasing of the bacteria within the EPS matrix), the BLS may contain components of MM9 that are essential for the development of BLS. These components would be included during the aggregation of the bacteria by the process of biomineralization, which is known to occur when microorganisms are grown in the presence of minerals such as magnesium, calcium, or manganese (Heim, 2011, Weiner, 2003). Since M9 is a chemically defined medium (Miller, 1972), we tested this possibility by preparing different variants of MM9 that contained malonate but lacked one of the components of MM9 (Table S3). When grown to a late stationary phase (24 hpi), PA14 formed the BLS in every MM9 variant except the one that lacked magnesium sulfate (Table S3). However, the growth of PA14 in the magnesium sulfate-deficient MM9 was delayed, with the culture not reaching the density observed after 16 hpi until 48 hpi. This slower growth without magnesium is not surprising as M. Webb showed in 1951 that gram-negative bacteria deprived of magnesium do not divide, although six subcultures in magnesium-deficient medium were required to deplete the cells of stored magnesium (Webb, 1951). Furthermore, Wolfe et al. showed in 1954 that magnesium is required for malonate decarboxylation (Wolfe et al., 1954); thus, the lack of magnesium would likely prevent utilization of malonate. Due to the growth defect caused by the lack of magnesium, we cannot at this time state whether magnesium sulfate is required for MBLS formation. While no BLS formed in the magnesium sulfate-deficient MM9, we cannot explicitly state that magnesium is required for formation of the BLS because of the growth defect caused by the lack of magnesium.

Biomineralization can occur when bacteria growing in an artificial or an environmental medium containing either citrate or malonate as a carbon source produce alkaline byproducts that increase the pH of the culture, which then causes the minerals to precipitate (Heim, 2011, McCutcheon et al., 2019, Purgstaller et al., 2019, Weiner, 2003). EM examination of the BLS revealed crystal-like structures present within the EPS (Figure S7). We utilized alizarin red S, which stains minerals such as magnesium or calcium bluish red or purple (Lievremont et al., 1982), to assess the presence of mineralization in the MM9 PA14 cultures. After 24 hpi, cell pellets from PA14 grown in GM9 did not stain with alizarin red S, while cell pellets from PA14 grown in MM9 were stained purple (dark gray in grayscale) (Figure 8b). Therefore, the structures are mineralized BLS (MBLS). As we observed the MBLS only at increased pH (9.5), prevention of the pH increase may interfere with both the formation of the BLS and the biomineralization. Therefore, we conducted two experiments; one in which the pH of the MM9 culture was not allowed to reach 9.5 and one in which the pH of the GM9 culture was adjusted to 9.5 The MBLS formed in MM9 that reached pH 9.5 and in GM9 adjusted to pH 9.5 (Table S3). However, no MBLS formed in the pH-adjusted MM9 culture or the unadjusted GM9 culture (Table S3). These results show that the rise in pH is critical for the formation of the MBLS.

3 |. DISCUSSION

Malonate, a ubiquitous C3-dicarboxylic acid, is one of the most abundant organic acids present in the human body, soil, and the roots and nodules of leguminous plants (Li & Copeland, 2000, Vreanova et al., 2013, Wishart et al., 2018). Malonate plays a critical role in several metabolic processes. It is central to carbon metabolism and fatty acid biosynthesis (Gueguen et al., 2000, Jordan et al., 1986, Schweizer & Hofmann, 2004). In addition, malonate is an inhibitor of succinate dehydrogenase and the Krebs tricarboxylic acid (TCA) cycle (Bowman et al., 2017, Pardee & Potter, 1949) and plays a role in nitrogen metabolism (Kim, 2002). It has recently been shown that certain P. aeruginosa proteins, including the elastase protein LasB, are post-translationally modified by malonylation of the amino acid residue lysine (Bowman et al., 2017, Gaviard et al., 2019, Peng et al., 2011, Qian et al., 2016). Protein malonylation is non- enzymatic; just the presence of malonate as a carbon source in the medium is predicted to lead to the malonylation of numerous proteins (Taherzadeh et al., 2018, Wang et al., 2017, Xu et al., 2016). We used the recently developed in silico “Mal-Lys” program (Xu et al., 2016) to interrogate the P. aeruginosa proteome. Over 700 proteins, including LasB, were predicted to undergo malonylation (Figure S8). Such modification may modulate the function of these proteins but further work will be required to determine the relevance of these findings.

Malonate is utilized as a carbon source by several bacterial genera in the Proteobacteria phylum, including Acinetobacter, Klebsiella, and Pseudomonas (Koo & Kim, 1999, Maderbocus et al., 2017). As a carbon source, malonate is converted into acetate and CO₂ by the malonate decarboxylase enzyme, which is encoded by the malonate-utilization operon (Gray, 1952, Maderbocus et al., 2017). We recently showed that when utilized as a sole carbon source, malonate increased the resistance of P. aeruginosa to the aminoglycoside kanamycin (Elmassry et al., 2019). Despite its importance, few studies have investigated the potential role of malonate metabolism in the pathogenesis of P. aeruginosa.

P. aeruginosa contains three major QS systems: the las, rhl, and pqs (de Kievit & Iglewski, 2000, Pesci et al., 1999). The las system consists of LasI, an autoinducer synthesis protein, which synthesizes 3OC₁₂-HSL and LasR, its transcriptional activator (Passador et al., 1993, Pearson et al., 1994). Similarly, the rhl system comprises RhlI that synthesizes C₄-HSL and RhlR, its transcriptional activator (Brint & Ohman, 1995, Winson et al., 1995). The stage of growth of P. aeruginosa affects the stability of the secreted 3OC₁₂-HSL and C₄-HSL molecules (AHLs). Yates et al. (Yates et al., 2002) previously observed that the AHLs accumulate within the exponential phase of P. aeruginosa in lysogeny broth (LB) but disappear during the stationary phase when the pH of the medium reaches ~8.0. Their disappearance is due to pH-dependent inactivation by alkaline hydrolysis or lactonolysis, a process that can be reversed by acidifying the growth medium to pH 2.0 (Yates et al., 2002). Lactonolysis is influenced by the length of the acyl side chain which provides stability to the AHL molecules; the short acyl molecule C₄-HSL was more readily inactivated than the longer acyl chain molecule 3OC₁₂-HSL and required a stronger acidification for its reactivation (Yates et al., 2002). Similarly, we found that the levels of both AHLs were reduced at pH 9.5 (Figure 4). The level of 3OC₁₂-HSL was restored at pH 5.0, while restoration of C₄-HSL required acidification to pH 2.0 (Figure S5). Our results suggest that lactonolysis of the extracellular AHLs influenced the activity of the las and rhl QS systems, causing a significant reduction in the expression of the QS genes lasI, lasR, rhlI, and rhlR (Figure 2). This reduction in expression, which was observed only when PA14 was grown with malonate as the sole carbon source, led to significant reduction in production of the virulence factors LasB, LasA, and rhamnolipids, whose production is stringently controlled by las and rhl, respectively (Figure 3). In contrast to the las and rhl genes, the expression of pqsC (and likely the entire pqsA-E operon) was significantly increased in PA14 grown in the presence of malonate (Figure 2), which would account for the increase in pyocyanin (Figure 3). The PQS biosynthesis genes are known to constitute an operon (https://www.pseudomonas.com/, accessed April 2021) (Winsor et al., 2016). We have previously used a P. aeruginosa PAO1 mutant strain in which the lacZ gene was fused inframe with pqsC in the chromosome; β-galactosidase activity resulting from this strain is due transcriptional activation of the promoter upstream of pqsA (Carty et al., 2006). Similarly, our recent RNA-seq analyses showed that expression of all five genes of the pqs operon was upregulated when PA14 was grown in whole blood from trauma patients compared to its growth in whole blood from healthy volunteers (Elmassry et al., 2019), while expression of all five genes was downregulated in PAO1 grown in blood from healthy volunteers compared to LB (Beasley et al., 2020).

In the autoinducer assay, which measures both PQS and HHQ (Fletcher et al., 2007), the reduction in the level of HHQ/PQS within the supernatant of PA14 MM9 culture was less than twofold (compared to over a thousand-fold for the AHLs) (Figure 4). Furthermore, the extracellular level of HHQ/PQS, which are not inactivated by lactonolysis, was unaffected when we adjusted the pH from 7.5 to 9.5 (Figure 5). Intracellularly, PQS (and HHQ) bind to the transcriptional regulator MvfR and activate the transcription of the PQS biosynthesis genes as well as several QS-controlled virulence genes including genes within the phzA-G operon that codes for the pyocyanin synthesis proteins (Diggle et al., 2006, Pesci et al., 1999). The pyocyanin level within the supernatant of PA14 culture in MM9 was increased rather than reduced (Figure 3d). It is possible that the increase in pyocyanin is due to the increased intracellular level of HHQ/PQS seen in PA14 grown in MM9 (~30-fold increase over that in GM9) (Figure 4c). This increased intracellular level of PQS may enhance MvfR activation, thereby increasing the expression of the pqsA-E operon as well as the phzA-G operon. Alternatively, the phenomenon may be related to the role that malonate plays in the synthesis of 2,4-dihydroxy quinolone (DHQ), an abundant metabolite secreted during the stationary phase of growth of P. aeruginosa at concentrations comparable to those of PQS (Zhang et al., 2008). Malonyl-CoA and/or the malonyl-acyl carrier protein perform an essential intermediate step in the formation in the biosynthesis of DHQ by producing a short-lived intermediate that undergoes intermolecular rearrangement to form DHQ (Gruber et al., 2016, Zhang et al., 2008). DHQ binds to MvFR and activates pqsA-E transcription, sustaining pyocyanin production (Gruber et al., 2016).

In its reduced form, pyocyanin, a redox active phenazine metabolite, interferes with electron transport, cellular respiration, and energy metabolism within eukaryotic and prokaryotic cells causing them to generate reactive oxygen species (ROS), especially superoxide and hydrogen peroxide (Hassett et al., 1992, Jo et al., 2020, Rada & Leto, 2013). The diversion of the electron transport flow leads to cell death (Hassett et al., 1992, Rada & Leto, 2013). In contrast, P. aeruginosa uses pyocyanin to mediate extracellular electron transport at a distance of hundreds of μm to oxidants outside its cells (Hassett et al., 1992, Rada & Leto, 2013). While it is resistant to the effects of pyocyanin (Hassett et al., 1992), P. aeruginosa must respond to the ROS generated by the host. This effect is most likely mediated by PQS, which plays a complex role in the P. aeruginosa response to oxidative stress, serving as both a pro-oxidant and an antioxidant, (Bredenbruch et al., 2006, Häussler & Becker, 2008). As a pro-oxidant, PQS sensitizes the cells to the effects of hydrogen peroxide (Bredenbruch et al., 2006, Häussler & Becker, 2008). In its role as antioxidant, PQS acts directly (comparably to ascorbic acid) to reduce levels of intracellular ROS (Häussler & Becker, 2008). Besides this effect, PQS upregulates genes related to the oxidative stress response, including genes for catalases and superoxide dismutase (Bredenbruch et al., 2006, Häussler & Becker, 2008). Hassett et al. (Hassett et al., 1992) observed that catalase is often upregulated under growth conditions that increase levels of pyocyanin. Besides the increase in pyocyanin, we also observed a threefold increase in catalase breakdown of hydrogen peroxide (Figure 3e). Thus, the roles of PQS, pyocyanin, and catalase are entwined with respect to oxidative stress.

Our results suggest that malonate utilization as a sole carbon source represses the expression of P. aeruginosa iron-regulated genes. Compared to its growth in GM9, the growth of PA14 in MM9 significantly reduced the expression of the pyoverdine-related genes pvdS and pvdD, the pvdS-regulated genes toxA, and regA, and the pyochelin regulatory gene pchR (Figure 6). Although the exact mechanism of this effect is not known at this time, several points are worth considering. First, the lack of iron in either GM9 or MM9 suggests that the observed phenomenon is iron-independent. We supported this possibility by additional experiments in which we demonstrated that, upon the addition of exogenous iron to the iron-free MM9 medium, both pyoverdine production and pvdS expression were significantly reduced; that is, iron still repressed the expression of these genes in MM9 (Figure S6). Second, while it was previously suggested that pH influences pyoverdine production by P. aeruginosa through its effect on iron solubility within the growth medium (Butaitė et al., 2018), the increased pH observed in MM9 was unlikely to play a role as MM9 does not contain iron. Third, the repression of the key regulators, pvdS and pchR, during the growth of PA14 in MM9 suggests a global effect for malonate that includes all or most of the P. aeruginosa iron-regulated genes including the ferric uptake regulator, FUR (Ochsner et al., 1995). One possible explanation for our findings is that, through its manipulation of the P. aeruginosa intracellular level of iron, malonate influences FUR function. Upon its activation by iron, FUR represses the expression of different iron-regulated genes (Ochsner et al., 1995). In this study, we grew PA14 overnight in the iron-sufficient LB medium and then subcultured it to GM9 or MM9. Consequently, the intracellular levels of iron within PA14 were gradually depleted upon its growth in GM9. However, during growth of PA14 in MM9, malonate may either interfere with the depletion of the intracellular iron or enhances the association of FUR with iron, thereby allowing for more FUR activation and an increased repression of the above described genes. To examine this possibility, we may grow PA14 in the iron-deficient medium, chelex-treated trypticase soy broth dialysate (TSB-DC) instead of LB and then subculture it in either GM9 or MM9. This would deplete PA14 of its intracellular iron prior to its inoculation in either GM9 or MM9.

Among different tested antibiotics, malonate influenced PA14 sensitivity to norfloxacin only. In comparison with its growth in GM9, the growth of PA14 in MM9 reduced the MIC of norfloxacin by eightfold. This sensitivity may be related to stress-induced accumulation of PQS intracellularly. Häussler & Becker (Häussler & Becker, 2008) previously demonstrated a similar sensitivity of P. aeruginosa to another fluoroquinolone antibiotic, ciprofloxacin. Upon its exposure to five mg/ml of ciprofloxacin, the fall in the viability of the P. aeruginosa strain PAO1, as measured by the reduction in both the optical density as well as the CFU/ml, was steeper than those of its pqsA or pqsH mutants (Häussler & Becker, 2008). Compared to its growth in GM9, growth in MM9 increased the intracellular level of PQS (Fig. 4c). At this time, we do not know if PA14 mutants are more resistant than their PA14 parent strain to ciprofloxacin and possibly other fluoroquinolone antibiotics.

We previously observed the formation of BLS when P. aeruginosa was grown in a rich artificial mucin-containing medium (Haley et al., 2012). Subsequently, comparing BLS formation in nutrient-rich LB and LB containing mucin, the BLS were shown to be dependent on mucin (Haley et al., 2014). When grown in the presence of malonate as a sole carbon source, PA14 failed to attach to the surface and form a typical biofilm; rather, at late stages of growth, the bacteria formed free-floating BLS maintained by DNA and exopolysaccharides (Figures 7b and 8). At this time, the molecular mechanism through which malonate facilitates the formation of BLS rather than typical biofilm in a chemically-defined medium is not known. It is clear than none of the previously described P. aeruginosa products that play a role in biofilm development, such as pelA, algB, phzS, phzM, and others, contributes to the BLS formation. PA14 mutants defective in the production of these proteins still formed BLS (Table S2). BLS development was associated with mineral precipitation or biomineralization. The BLS stained with alizarin red S, indicating presence of magnesium or calcium (Figure 8b). We identified magnesium sulfate as the component of MM9 critical for the formation of the mineralized BLS (MBLS); no MBLS developed when PA14 grew in magnesium sulfate-deficient MM9 while MBLS formed in the absence of calcium chloride (Table S3). Furthermore, using EM, we visualized these mineral precipitates which are similar to previously described structures (Figure S7) (Anbu et al., 2016). An alkaline microenvironment is essential for the mineral precipitation as we prevented their formation when we maintained the pH of the MM9 culture medium at 7.5. Thus, when malonate is the sole carbon source and in the presence of magnesium sulfate, PA14 formed MBLS at late stages of growth when the culture reached highly alkaline pH.

Biomineralization is defined as a microbially-induced chemical alteration that results in the precipitation of inorganic minerals (Konhauser & Riding, 2012). Biomineralization occurs through three specific mechanisms; biologically-controlled mineralization, biologically-influenced mineralization, and biologically-induced mineralization (Benzerara et al., 2011, De Muynck et al., 2010, Phillips et al., 2013). In the biologically-induced mineralization, an environment is chemically modified by biological activity that results in supersaturation and precipitation of minerals (Bosak, 2011, De Muynck et al., 2010, Phillips et al., 2013). During this process, identified as microbially-induced calcite precipitation (MICP), metabolic products (such as CO₃²⁻) secreted by the microorganisms react with ions (such as Ca²⁺) in the environment leading to the subsequent precipitation of minerals; calcium carbonates are formed from the supersaturated solution (Frankel & Bazylinski, 2003). Also, in the biologically-induced mineralization, minerals are passively precipitated by the presence of cell surface organic matter associated with the biofilm such as EPS (Braissant et al., 2003). Negatively charged groups on bacterial surfaces function as scavengers for divalent cations such as Ca²⁺ and Mg²⁺ and bind to the cell surface, producing nucleation sites for calcium deposition (Ramachandran et al., 2001, Stocks-Fischer et al., 1999). The bound cation consequently reacts with the anions to form carbonate in an insoluble form (Anbu et al., 2016, Stocks-Fischer et al., 1999). Bacterial metabolic activity during MICP, which produces an alkaline microenvironment, results in the formation of different-sized particles with no set morphology (Frankel & Bazylinski, 2003, Konhauser & Riding, 2012). Biomineralization in biofilm models suggest that mineral precipitation primarily occurs at the biofilm surface (Dupraz et al., 2009, Shiraishi et al., 2008). The heterogeneity within the biofilm environment, such as variations in levels of oxygen and pH, likely leads to supersaturation within the biofilm and facilitates biomineralization (Decho, 2010, Hunter & Beveridge, 2005, Wimpenny et al., 2000). Thus, the most likely scenario to explain our findings is that, in the presence of malonate as a sole carbon source and at late stages of growth, PA14 may produce alkaline byproducts (e.g., sodium hydroxide, carbonate, and bicarbonate), thereby raising the pH of the medium and precipitating magnesium (Heim, 2011, Manzoor et al., 2018b, Purgstaller et al., 2019). Accordingly, the precipitates formed at pH 9.5 but not 7.5. Negative charges on the surface of P. aeruginosa as well as the heterogeneity within the MBLS may also contribute to Mg²⁺ precipitation. Mg²⁺ rather than Ca²⁺ was essential for the formation of the mineral precipitates. This is possibly due to the difference in the concentrations of the two minerals within MM9; compared with Ca²⁺ concentration (0.2 mM), the Mg²⁺ concentration within MM9 is ten times higher (2 mM). It is also possible that the MBLS contain struvite (magnesium ammonium phosphate). P. aeruginosa has been shown to precipitate struvite crystals in alkaline urine (Manzoor et al., 2018b) and that this process can be prevented by acidification of the urine with vitamin C (Manzoor et al., 2018a). Furthermore, Li et al. (Li et al., 2016) showed that ureolytic biomineralization reduces the susceptibility of Proteus mirabilis biofilms to ciprofloxacin. Future experiments including ones designed to analyze the mineral content of the MBLS formed by PA14 growth in MM9 containing variable concentrations of Ca²⁺ and Mg²⁺ will be required to examine this possibility.

Visaggio et al. (Visaggio et al., 2015) previously suggested that aggregation is an important cue that triggers the expression of pyoverdine-related genes in P. aeruginosa. When grown in an iron-deficient medium (TSB-DC or M9), and in comparison with its parent strain, the PAO1Δpel/Δpsl double mutant produced significantly lower levels of pyoverdine and the pyoverdine-related virulence factors exotoxin A and PrpL (Visaggio et al., 2015). Neither PAO1Δpel nor PAO1Δpsl showed a similar defect (Visaggio et al., 2015). In addition, and when grown in TSB-DC as a static culture at 37°C for 14 h, PAO1 but not PAO1Δpel/Δpsl formed aggregates of planktonic cells (Visaggio et al., 2015). However, when cell aggregation was artificially induced in PAO1Δpel/Δpsl by growing the strain either as colonies on a solid surface or in liquid culture in the presence of aggregating agents, the production of pyoverdine and pyoverdine-dependent virulence factors was restored (Visaggio et al., 2015). Our study differs from that of Visaggio et al. in several aspects. First, we grew all cultures at 37°C under shaking conditions. Under these conditions, and when grown in TSB-DC or GM9, neither PAO1 nor PA14 produced MBLS (aggregates) (Tables S2 and S3). Both strains produced MBLS when grown in MM9 (Table S2). Second, we analyzed pyoverdine production as well as the expression of toxA and the pyoverdine genes in PA14 MM9 cultures at 16 hpi, prior to the formation of the MBLS. Third, similar to PA14 and when grown in MM9, PA14Δpel formed the MBLS (Table S2). PA14 does not carry psl genes. Fourth, the growth of PA14 in MM9 significantly reduced, rather than increased, the expression of several iron-regulated genes (Figure 6b). Even after the development of the MBLS at 24 hpi, pyoverdine production was not restored (data not shown). Thus, in our study, malonate metabolism may specifically repress the expression of pyoverdine and pyoverdine-related genes through a yet to be determined mechanism.

Using the murine model of thermal injury, we previously tested the potential role of malonate metabolism in P. aeruginosa bacteremia/sepsis (Elmassry et al., 2019). Both PA14 and the malonate utilization mutant PA14ΔmcdA produced mortality in thermally injured/infected mice at 48 hours post infection (Elmassry et al., 2019). However, the bioburden of PA14ΔmcdA within the liver and spleen of infected mice was significantly lower than that of PA14 (Elmassry et al., 2019). Further experiments will be conducted to explore the significance of this reduction in PA14ΔmcdA bioburden.

In summary, our results suggest that as a sole carbon source, malonate differentially influences the three P. aeruginosa QS systems; las, rhl, and PQS. Malonate metabolism as a sole carbon source repressed the expression of the las and rhl genes as well as the production of las and rhl-regulated virulence factors LasB, LasA, and rhamnolipid; probably by increasing the extracellular pH, which reversibly inactivated 3OC₁₂-HSL and C₄-HSL. In contrast, the level of the pH-independent PQS increased thereby increasing the expression of the phenazine genes as well as the production of pyocyanin and catalases. Additionally, as a sole carbon source, malonate repressed the expression of iron-regulated genes possibly through an iron-independent mechanism. Furthermore, the growth of PA14 in MM9 induced the formation of BLS and facilitated the BLS-associated biomineralization.

4 |. EXPERIMENTAL PROCEDURES

4.1 |. Bacterial strains, plasmids, and media

Bacterial strains and plasmids used in this study are described in Table 1. We used the strain PA14 and its isogenic mariner transposon mutants (Table 1) for the majority of the experiments. Bacterial strains were routinely grown overnight in lysogeny broth (LB) (Becton, Dickinson and Company, Franklin Lakes, NJ). When needed, antibiotics were added at the following concentrations: 300 μg/ml of carbenicillin, 50 μg/ml of ampicillin, 15 μg/ml of gentamicin, 60 μg/ml of tetracycline, 200 μg/ml of streptomycin, or 30 μg/ml of chloramphenicol. Antibiotics were obtained through Fisher Scientific, Waltham, MA.

For analysis of the effect of malonate as a sole carbon source on the growth and virulence of PA14, we used the established M9 minimal medium (6.0 g Na₂HPO₄, 3.0 g KH₂PO₄, 0.5 g NaCl, 1.0 g NH₄Cl per liter supplemented with 0.2 mM CaCl₂ and 2 mM MgSO₄) (Fisher Scientific) as a basal medium (Miller, 1972). No iron source was added to the M9 minimal medium; however, a small but potentially biologically relevant amount of iron may be present in the medium originating from the water, chemicals, or glassware used to prepare the medium. For the control medium, we modified the M9 by adding 1% glycerol (v/v; 110 mM) as a sole carbon source (GM9). Glycerol has been used as a main carbon source for P. aeruginosa in studies using TSB-DC (Hamood et al., 1992, Ohman et al., 1980, Suh et al., 1999) and as the sole carbon source in M9 minimal medium (Kruczek et al., 2012, Nikel et al., 2014) and in MOPS (morpholinepropanesulfonic acid) medium (Dolan et al., 2020). Both glycerol and malonate are present in healthy human airways (Farne et al., 2018) and glycerol was found to be abundant in the CF airway (Dolan et al., 2020). For the test medium containing malonate as a sole carbon source (MM9), we added 40 mM malonate, as sodium malonate dibasic (MilliporeSigma, St. Louis, MO), instead of glucose. To exclude the effect of the concentration of glycerol or malonate on the observed phenotypes, we tested decreasing the concentration of glycerol to one-half and one-third of the original concentration (110, 55, and 36.7 mM) and increasing the concentration of malonate two and three times the original concentration (40, 80, and 120 mM) as well as equimolar (100 mM) concentrations of glycerol and malonate in the media. To evaluate the repressive effect of iron on gene expression in MM9, we added 20 μg Fe³⁺/ml (Ohman et al., 1980). PA14, or other strains to be tested, were grown overnight in LB and subcultured into GM9 and MM9 to a starting OD₆₀₀ of 0.04. For analysis of gene expression and virulence factor production, PA14 was routinely grown in GM9 and MM9 for 16 hpi. For analysis of biofilm formation and MBLS formation, GM9 and MM9 cultures were grown for 24 to 48 hpi.

To assess the effect of pH on the formation of the MBLS, we checked the pH of the medium at different times post-inoculation using pH paper strips (Hydrion, Fisher Scientific). The pH of the MM9 culture was 9.5 at 24 hpi while that of the GM9 culture was 7.5. To determine if preventing the rise in pH in MM9 would prevent formation of the MBLS, two sets of PA14-inoculated MM9 cultures were prepared. The control set was left unmanipulated and grown for 24 h at 37°C; the pH of the other set was maintained at pH 7–7.5 throughout the 24 h growth period. The pH of the culture was monitored every four h and readjusted to 7–7.5 using diluted sulfuric acid as necessary when the pH rose above 8. To determine if alkalinization of GM9 would result in MBLS formation, two sets of PA14-inoculated GM9 were prepared. The control set was left unmanipulated and grown for 24 h at 37°C; at 16 hpi, sodium carbonate (40 mM) was added to the culture and incubation continued until 24 hpi.

4.2 |. Determination of the growth index, CFU, and aggregation index

Growth index.

To determine the growth index, PA14 was grown for up to 48 hpi in GM9 and MM9. Briefly, cells were grown at 37°C in ten ml of medium in 250-ml flasks, one flask for each time point to be harvested. A one-ml sample of the culture was obtained at specific time points from two hpi to 48 hpi and the OD₆₀₀ was measured using a spectrophotometer. After 16 hpi, PA14 began to form aggregates (mineralized BLS) in the MM9 cultures. Whether grown in GM9 or MM9, the remaining culture volume (nine ml) from each flask was sonicated for 30 s with a pulse of 20% amplitude using a Model 120 Sonic Dismembrator with a CL-18 probe (Fisherbrand, Fisher Scientific) prior to obtaining the OD₆₀₀.

CFU analysis.

To determine the numbers of CFU present in the cultures, cultures were inoculated and grown as described for the growth index. One-ml samples were obtained at specific time points from two hpi to 48 hpi and the samples were sonicated. Each sonicated sample was serially diluted tenfold and ten-μl aliquots of each dilution were spotted on LB agar plates. Following overnight incubation at 37°C, the CFU were calculated by the formula: (CFU counted × dilution factor) × 100 = CFU/ml.

Calculation of the aggregation index.

The aggregation index was calculated as the ratio of post-sonication OD₆₀₀ to pre-sonication OD₆₀₀ (Kharadi & Sundin, 2019).

4.3 |. Analysis of gene expression

β-galactosidase activity.

Expression of the lasR, lasI, rhlR, rhlI, pvdD, toxA, regA, pvdS, and pchR genes was measured using transcriptional and translational β-galactosidase reporter fusion plasmids transformed into PA14 (Table 1). The assay for β-galactosidase activity was performed as previously described (Miller, 1972, Stachel et al., 1985). Briefly, cultures were incubated at 37°C with shaking. At 16 hpi, cells in one-ml aliquots of the cultures were pelleted, lysed, and the units of β-galactosidase activity were determined as previously described (Miller, 1972, Stachel et al., 1985). The calculation for units of activity includes the amount of growth (OD₆₀₀).

qRT-PCR.

PA14 GM9 and MM9 cultures inoculated as described above were harvested at 16 hpi. Cells were pelleted by centrifugation at 5,000 × g for ten min and resuspended in fresh GM9 or MM9 containing RNAprotect (QIAGEN, Valencia, CA). Cells were pelleted again, the supernatant was discarded, and pellets were stored at −80°C. RNA was extracted from the pellets using the RNAeasy Mini Kit (QIAGEN). Genomic DNA was digested from the samples using RNase-free DNase (QIAGEN). Purified RNA was quantified by NanoDrop spectrophotometer (NanoDrop Technologies, Wilmington, DE) and integrity of the RNA was analyzed using RNA Nano Chip on an Agilent 2100 Bioanalyzer (Agilent, Palo Alto, CA). RNA samples with 1.8–2.2 ratios of absorbance at 260/280 nm were converted to cDNA using the QuantiTect Reverse Transcription Kit (QIAGEN) for qRT-PCR as previously described (Kruczek et al., 2016). For qRT-PCR, equal amounts of cDNA were mixed with iQ SYBR Green Supermix (Bio-Rad, Hercules, CA) together with 2 μM of specific primers for each gene examined. Amplification and detection was done using CFX96 Deep Well Real-Time PCR System (Bio-Rad) and analysis of gene expression was done using CFX Manager 3.1 software (Bio-Rad). Each experiment consisted of three biological replicates analyzed in triplicate. Quantity of cDNA in the samples was normalized using the 16S ribosomal RNA gene PA14_08570.

4.4 |. Analysis of virulence factors

Elastin Congo red assay.

The level of LasB elastase within the supernatant fraction of tested isolates was determined using the previously described elastin Congo red assay (Schaber et al., 2004, Schad et al., 1987). PA14 cells were grown for 16 hpi in either GM9 or MM9 broth at 37°C. Supernatant fractions were isolated and a 100-μl sample of each fraction was added to 2.0 ml 10 mM NaHPO₄ containing 30 mg elastin Congo red (MilliporeSigma). After 14 h of incubation at 37°C, the reactions were centrifuged and the Congo red released by LasB elastolytic cleavage of the elastin::Congo red bond was measured at 495 nm. Values were normalized by dividing by the OD₆₀₀ of each culture.

Staphylolytic assay.

The staphylolytic assay for LasA was conducted as previously described (Diggle et al., 2002, Hamood et al., 1996). Briefly, an overnight culture of Staphylococcus aureus ATCC 25923 was boiled for ten min and then centrifuged for ten min at 10,000 × g. The pellet was resuspended in 10 mM phosphate buffer, pH 7.5, to OD₆₀₀ of 0.8. A 100-μl aliquot of the culture supernatant (PA14 grown for 16 hpi in either GM9 or MM9) was added to 900 μl of the S. aureus suspension. Staphylolysis by LasA within the samples was determined spectrophotometrically by monitoring the decrease in the absorbance at 600 nm. The absorbance was recorded every five min for two h. Values were normalized by dividing by the OD₆₀₀ of each culture. Staphylolysis was presented as percent lysis by further normalizing the calculated values to the absorbance of S. aureus suspension at zero time point for easier interpretation of the results.

Quantification of rhamnolipids.

The level of rhamnolipids within the culture supernatants was determined using the methylene blue assay (Pinzon & Ju, 2009). PA14 was grown for 16 hpi in either GM9 or MM9 and the culture supernatants were separated and acidified to a pH of ~2.3 using 1 N HCl. The acidified supernatants were extracted with chloroform, then carefully removed and mixed with 200 μl of methylene blue reagent (one g/l). After vigorous mixing for four min, the mixtures were left to stand for 15 min. The chloroform phase was then separated and the absorbance was measured at 638 nm. Values were normalized by dividing by the OD₆₀₀ of each culture.

Quantification of pyocyanin.

This was performed following previously described protocols (Essar et al., 1990, Schaber et al., 2004). Five-ml samples of the supernatant fractions of cultures of PA14 grown in GM9 or MM9 for 16 hpi were isolated, mixed with chloroform (1:1), and the lower layer was separated and mixed with two ml of 0.2 N HCl. The pyocyanin-rich organic upper layer was then extracted, and the absorbance at 520 nm was determined. Values were normalized by dividing by the OD₆₀₀ of each culture prior application of the following formula. The amount of pyocyanin, in μg/ml, was calculated using the following formula: OD₅₂₀ × 17.072 = μg of pyocyanin/ml (Essar et al., 1990).

Catalase assay.

The levels of catalase within the cultures were assayed using a previously described protocol (Iwase et al., 2013). Briefly, 100-μl aliquots of PA14 cultures grown in GM9 or MM9 for 16 hpi were mixed with 100 μl 1% Triton X-100 and 100 μl 30% undiluted hydrogen peroxide in a 13 × 100 mm glass tube, vortexed, and the mixture was allowed to settle at room temperature for 30 min. The height of the stable foam was measured in mm. To normalize the levels of catalase, they were divided by the OD₆₀₀ values of the corresponding culture.

Pyoverdine assay.

Pyoverdine levels within the supernatant of PA14 cultures grown in GM9 or MM9 for 16 hpi were determined by measuring the A₄₀₅ values of each sample (Stintzi et al., 1996). Values were adjusted by dividing A₄₀₅ readings by the corresponding OD₆₀₀ values of the culture.

4.5 |. Detection of 3OC₁₂-HSL, C₄-HSL, and HHQ/PQS

PA14 was grown in either GM9 or MM9 for 16 hpi for all experiments involving the autoinducers except the reversal of lactonolysis, for which the cultures were grown for 24 hpi. The growth index (OD₆₀₀) for each sample was determined as above; the samples collected at 24 hpi were sonicated before measuring the OD₆₀₀. The supernatant fraction was collected and passed through a 0.2-μm syringe filter to remove any remaining bacterial cells (VWR, Arlington Heights, IL). Autoinducers were separated from a two-ml of the supernatant via three successive rounds of acidified ethyl acetate (high-performance liquid chromatography [HPLC] grade of ethyl acetate with 0.01% glacial acetic acid) extraction (Pearson et al., 1994). The final extracts were then dried for one hour at 30^oC in a centrifugal vacuum concentrator model 5301 (Eppendorf, Hauppauge, NY). Dried extracts were stored at −20°C and resuspended in 30 μl of methanol (HPLC grade) before use. For the autoinducer assays, overnight cultures of the reporter strains E. coli JM109/pSB1142 for 3OC12-HSL, JM109/pSB536 for C4-HSL, and P. aeruginosa PAO1pqsA CTX-lux::pqsA for HHQ/PQS (Table 1) were diluted to an OD₆₀₀ of 0.5 in fresh LB. Aliquots (100 μL) of the diluted reporter strains were pipetted into separate wells of a 96-well black polystyrene plate (Costar, Corning, Corning, NY), five μl of the methanol-dissolved extract was quickly added to each well, and the plate was incubated statically in the dark at 37°C for three h. Luminescence (reflecting the level of autoinducer activation of its cognate transcriptional factor to promote luxCDABE) was measured using a luminometer (Modulus, Turner Biosystems, Promega, Madison, WI). For each autoinducer assay, we include a negative control in which the reporter strain was incubated with the solvent only (ethyl acetate) to determine levels of background. The value of the specific negative control was subtracted from the luminescence reading for each relevant autoinducer. Luminescence values, reported as relative light units (RLU), were normalized using OD₆₀₀ value of each respective culture.

4.6 |. Determination of the minimum inhibitory concentration (MIC) for antibiotics

The MIC of the tested antibiotics was determined using the broth macrodilution technique described in the Clinical and Laboratory Standards Institute guidelines (CLSI, 2018). An overnight culture of PA14 was inoculated into series of 16 mm × 100 mm tubes containing two ml of either GM9 or MM9 per tube to produce an initial inoculum of about 1 × 10⁵ CFU/ml. Serial twofold dilutions of each tested antibiotic were added and the tubes were incubated for 16–18 h at 37°C with shaking. After incubation, the cultures were visually inspected for growth. The lowest concentration of the antibiotic that inhibited visible growth was considered the MIC.

4.7 |. Development and analysis of biofilms and mineralized BLS

Biofilm development and crystal violet assay for biomass.

Biofilms were developed on a plastic surface as previously described (Déziel et al., 2001, O’Toole & Kolter, 1998, Qaisar et al., 2013). Briefly, an aliquot of an overnight PA14 culture in LB was inoculated into 17 mm × 100 mm polystyrene tube containing either GM9 or MM9 to obtain an initial OD₆₀₀ of 0.02. The tubes were then incubated at 37°C for 48 h with gentle shaking. The planktonic growth cultures were decanted and the adherent biofilms were washed with distilled water to remove non-adherent cells. Adhering biofilms were then stained by the addition of 1% crystal violet solution (w/v). After an hour of incubation at room temperature, the biofilms were then washed twice with distilled water and 95% ethanol was added to elute the stained biofilms. The biofilm mass was measured at an absorbance of 590 (A₅₉₀) and normalized using the OD₆₀₀ value of the respective culture.

Electron microscopy (EM).

PA14 cultures grown in GM9 or MM9 for 24 hpi were centrifuged and the cell pellets were washed with phosphate buffer three times. One ml of each culture was added to a slide pretreated overnight with poly-L-lysine and left for one h to achieve proper adhesion. Slides were fixed in primary fixative solution (2.5% glutaraldehyde, 2.0% paraformaldehyde solution in 0.05 M sodium cacodylate buffer, pH 7.4) for one h at 4°C. Then, slides were washed in 0.05 M sodium cacodylate buffer three times for ten min each. Slides were placed in secondary fixative solution (1% osmium tetroxide in 0.5 M sodium cacodylate buffer) for 30 min, followed by washing in 0.05 M sodium cacodylate buffer. Then, slides were dehydrated with increasing concentrations of ethanol, and dried using a CPD 030 critical point dryer (Leica Microsystems, Buffalo Grove, IL). Slides were mounted on SEM stubs and gold-coated before viewing with a Hitachi S-4300 scanning electron microscope (Hitachi High-Technologies America, Inc., Pleasanton, CA, USA). This procedure was performed using the Texas Tech University College of Arts and Sciences Microscopy facilities.

Alizarin red S staining for biomineralization.

PA14 cultures grown in GM9 or MM9 for 24 hpi were centrifuged to pellet cells and the pellets were washed three times with PBS. One ml of 40 mM alizarin red S (pH 4.1) (MilliporeSigma) was added to the centrifuge tubes containing the pellets and the tubes were incubated at room temperature for 15 minutes. The pellets were washed five times with distilled water to remove any unincorporated dye and examined macroscopically for color. Alizarin red S stains calcium and magnesium precipitates characteristic of biomineralization bluish red to purple (Lievremont et al., 1982).

4.8 |. Statistical analyses

Statistical significance was calculated using GraphPad Prism version 9.0.2 (161) (GraphPad Software, San Diego, CA, www.graphpad.com). For experiments across time (growth index, aggregation index, CFU, and staphylolytic assay), two-way ANOVA with Šídák’s multiple comparisons posttest was used. To compare the effect of different pH levels on the reversal of lactonolysis, data were log-transformed and statistical significance was determined by two-way ANOVA Šídák’s multiple comparisons posttest using pH 2 (the condition for expected maximum recovery of functional autoinducer) as the control value. One-way ANOVA with Tukey’s multiple comparisons posttest was done to compare all pairs of pairs among the different concentrations of malonate and glycerol. Two-tailed unpaired t-test was used to compare results from the two media in a single assay (e.g. individual gene expression or virulence factor level).