Production of both l‐ and d‐ N‐acyl‐homoserine lactones by Burkholderia cepacia and Vibrio fischeri

Abstract Quorum sensing (QS) is a complex process in which molecules, such as l‐N‐acyl‐homoserine lactones (l‐AHLs), are produced as essential signaling molecules allowing bacteria to detect and respond to cell population density by gene regulation. Few studies have considered the natural production and role of the opposite enantiomers, d‐AHLs. In this work, production of d,l‐AHLs by Burkholderia cepacia and Vibrio fischeri was monitored over time, with significant amounts of d‐AHLs detected. Bioluminescence of V. fischeri was observed with maximum bioluminescence correlating with the maximum concentrations of both l‐ and d‐ octanoyl‐homoserine lactones (l‐ and d‐OHL). l‐Methionine, a precursor to l‐AHLs, was examined via supplementation studies conducted by growing three parallel cultures of B. cepacia in M9 minimal media with added l‐, d‐, or d,l‐methionine and observing their effect on the production of d,l‐AHL by B. cepacia. The results show that addition of any methionine (l‐, d‐, or d,l‐) does not affect the overall ratio of l‐ to d‐AHLs, that is d‐AHL production was not selectively enhanced by d‐methionine addition. However, the overall AHL (l‐ and d‐) concentration does increase with the addition of any methionine supplement. These findings indicate the possibility of a distinct biosynthetic pathway for d‐AHL production, possibly exposing a new dimension within bacterial communication.

& Greenberg, 2002). The biological precursor for AHLs is Sadenosylmethionine (SAM). Figure 1 shows the general structure of AHLs which consists of a γ-butyrolactone moiety with an acylated α-amino group. The acyl chain length varies from 4 to 14 carbons. However, longer chain lengths up to 18 carbons have been reported (Mohamed et al., 2008). The α-carbon on the lactone ring is a stereogenic center. Chemically, AHLs can exist in either the l-or d-configuration. It is well known that chiral proteogenic amino acids in all organisms are primarily l-amino acids. Until recently, bacteria were believed to produce only the l-form of AHLs for QS. This belief led researchers to conduct studies using, almost exclusively, achiral methods (Liu et al., 2018). Most AHL analyses from bacterial cultures are currently conducted using nonenantioselective methods, such as: thin-layer chromatography combined with biosensor strains, GC-MS, LC-MS, or SFC-MS (Charlton et al., 2000;Gao et al., 2020;Hoang et al., 2020;Zhou et al., 2020). To the best of our knowledge, only one study has mentioned the presence of d-AHL, that is, d-decanoyl-homoserine lactone detected in a culture of B. cepacia LA3. The experiment used a single drop microextraction technique with chiral GC-MS (Malik et al., 2009). They postulated that bacteria could biosynthesize d-AHL as well as precursor d-methionine derivatives. Other studies have shown the biosynthesis of d-amino acids (Armstrong et al., 1993a;Armstrong et al., 1993b;Hernández & Cava, 2016;Weatherly et al., 2017).
A thorough investigation of the natural production of all manner of d-AHLs and l-AHLs is necessary to better understand the QS process. In this study, wild-type V. fischeri (ES114) and B. cepacia (25416) were cultured, and the production of both enantiomers of AHLs was monitored over time using different nutrient conditions. Vibrio fischeri classically has been a point of reference in QS studies. It plays an active role in oceanic environments, having a symbiotic relationship with the bobtail squid (Koehler et al., 2019). Previously, the production of l-AHLs from bacterial cultures was monitored over time and compared to bacterial growth (OD 600 ) (Fekete et al., 2010;Ravn et al., 2001;Widdel, 2007). These studies did not apply chiral methods for the analysis of AHLs as it was incorrectly assumed that only l-AHLs were produced. This study will present the first growth versus time studies that focus on the production of both l and d-AHLs from V. fischeri and B. cepacia. Additionally, nutrient conditions are altered, for B. cepacia, by supplementation with l-methionine, d-methionine, or racemic methionine, to investigate whether or not different stereochemical forms of this amino acid affects the production of l and d-AHLs. Deionized water was obtained from an in-house Millipore system producing water of 18.2 MΩ•cm resistivity.

| Analysis of standard N-acyl-homoserine lactones by chiral gas chromatography-mass spectrometry (GC-MS)
A 100 μL aliquot of a 10 ppm standard solution of acyl and 3-oxoacyl-HLs in dichloromethane (C4-C12 acyl; C4-C8 3-oxohexanoyl) was placed in a 1 mL sample vial and dried with a gentle stream of ultrahigh purity (UHP) N 2 . Derivatization of 3-oxoacyl-HLs was conducted as outlined in the literature (Readel et al., 2020). A 100 μL aliquot of BSTFA with 1% TMCS was added to the sample vial with the dried standard solution and sealed with a

| Extraction reproducibility of N-acylhomoserine lactones (AHL) standards from bacterial media
Fresh M9 media and PBB were both spiked with a 2 μg/mL solution of standard AHLs. Only PBB media was spiked with 3-oxohexanoyl-homoserine lactone (OHHL). The spiked media samples were then extracted using 1 g HLB cartridges which were sequentially conditioned with 5 mL of HPLC grade methanol and 5 mL of DI water. Ten mL of the spiked media sample was then loaded on the cartridge bed at a rate of ~2 mL/min. Loaded samples were then washed with a solution of 95:5 v/v% water:methanol at a rate of ~2 mL/min. Elution was done twice with 4 mL of HPLC grade acetonitrile. Eluted samples were dried by rotary evaporation on a BUCHI R-100 Rotavapor unit (BUCHI, Flawil, Switzerland) attached to a vacuum in a heated water bath (50°C). M9 media samples were reconstituted in 100 μL of dichloromethane and analyzed by GC-MS.
PBB samples were processed and analyzed by the above-mentioned method. The method limit of detection (LOD) for acyl was approximately 0.001 μg/mL while for 3-oxoacyl-HLs was 0.006 μg/mL. Good extraction reproducibility of AHLs was seen in M9 minimal (M9) media and photobacterium broth (PBB). The %RSD of recovery for the spiked AHLs standards were all lower than 3%. Only l and d-OHHL showed %RSD of recovery at 6.4% in PBB.

| Analysis of enantiomeric AHLs from bacterial cultures by chiral GC-MS
N-acyl-homoserine lactones were extracted and analyzed from preliminary 24-h cultures of B. cepacia and V. fischeri by using the above-mentioned culturing, extraction, and GC-MS analysis methods. The 3-oxohexanoyl-homoserine lactone (OHHL) and 3-oxooctanoyl-homoserine lactone (OOHL) concentrations were below the LOD for the GC-MS in bacterial samples. Therefore, OHHL and OOHL in the samples were quantified by LC-MS/MS as previously reported (Readel et al., 2020).

| Growth and AHL production by Burkholderia cepacia
Overnight cultures of B. cepacia and V. fischeri were seeded to fresh M9 media and PBB media respectively and grown using the abovementioned conditions. Quadruplicate samples were evaluated for B. cepacia and V. fischeri. Burkholderia cepacia: a 1.5 mL aliquot was taken from the growth cultures at times: 4.5, 15, 20, 30, 45, 54, 69, 93 h, diluted by half with fresh M9 media, and placed into polystyrene cuvettes. OD 600 absorbance measurements were taken to monitor bacterial growth with a Vernier SpectroVis ™ Plus spectrophotometer. Two 10 mL samples were taken from the growth culture and centrifuged, only supernatant was set aside. The supernatant of each sample was spiked with l-dodecanoyl-homoserine lactone (l-dDHL) to a concentration of 8.0 μg/mL and guided through the SPE method described in Section 2.4. The extractant was rotary evaporated to dryness. The leftover residue was dissolved into 200 μL of dichloromethane and transferred to a 2 mL sample vial for analysis by GC-MS.
2.8 | Growth, AHL, and bioluminescence production by Vibrio fischeri A 1.5 mL aliquot was taken from the growth cultures at times: 6, 9, 10, 11, 12, 16, 20, 24, 28, 33.5, 44.5, 52, 60, and 74 h, diluted by half with fresh PBB media and placed in polystyrene cuvettes. OD 600 measurements were taken as described in the previous section. Replicate 10 mL samples were taken per growth culture and processed in the same way as B. cepacia. The supernatant of each sample was spiked with l-heptanoyl-homoserine lactone (l-HpHL) to a concentration of 3 μg/mL and extracted using the SPE method. After rotary evaporation, the residue was dissolved in 200 μL of dichloromethane, transferred to a 2 mL sample vial, and dried by blowing with UHP N 2 . The residue was derivatized via the method described in Section 2.4 and analyzed by GC-MS. The bioluminescence intensities of V. fischeri were analyzed by using Image J from the National Institute of Health, USA (http://imagej.nig.gov.ij). First, the vial image was selected, and a "Plot Profile" option was used to generate an intensity plot of gray values as a function of position on the bioluminescence image. The larger the "gray value," the higher is the intensity of the bioluminescence. Further documentation and calculation of gray values are provided in the software. The highest gray values from the "Plot Profile" are listed in the Results, Section 3.2.

| Methionine enantiomers as nutrients for Burkholderia cepacia
Overnight cultures of B. cepacia were grown as described previously and supplemented with one of three amounts of methionine: (i) 6.8 mmol of l-methionine (ii) 6.8 mmol of d-methionine and (iii) 3.4 mmol each of l and d-methionine, grown at 30°C (200 rpm).
Collection of OD 600 and N-acyl-homoserine lactone growth curves was done following the method for B. cepacia in Section 2.7.

| Presence of both enantiomeric forms of AHLs from bacterial cultures
The AHL standards shown in Table 1 were selected because the occurrence of such l-AHLs is well-documented in cultures of B. cepacia and V. fischeri (Kuo et al., 1994;Lewenza et al., 1999;Venturi et al., 2004). The focal point of this study is the stereochemical production of these AHLs over time. An initial 24-h culture of B. cepacia was tested for the presence and identification of any AHLs produced. The detected AHLs were l-HHL, l-and d-OHL, and l-DHL (Figure 2a,b). The amounts of l-HHL and l-DHL were near the limit of detection (LOD, Experimental Procedures). In a 24-h preliminary culture of V. fischeri, the extracts contained substantial amounts of land d-HHL and l-and d-OHL (Figure 3a,b). The 3-oxo-AHLs, OHHL and OOHL, were detected at much smaller amounts for V. fischeri, and only by LC-MS/MS. Table 2 summarizes the screening results.  were detected by LC-MS/MS (Readel et al., 2020) at concentrations of approximately 0.0001 μg/mL for both enantiomers.
Bioluminescence was monitored throughout the growth of V. fischeri ( Figure 6). The initial observation of bioluminescence by V.
fischeri was at 11 h with maximum luminescence (light production) seen at approximately 24 h. A gradual dimming (decreasing light production) was observed starting at 28 h and continued to the end of the experiment (Figure 6). The maximum luminescence correlated with the maximum production of d-OHL and l-OHL at ~24-28 h ( Figure 5a). It is important to note that the most produced AHL was l-OHL and the second most produced was its enantiomer d-OHL. All other l-AHLs were of lower concentrations than d-OHL for both B.

| Enantiomeric N-acyl-homoserine lactone production by Burkholderia cepacia in methioninesupplemented M9 media
There has only been one instance where a d-AHL has been reported from a bacterial culture. d-decanoyl-homoserine lactone was detected in a culture of B. cepacia. (Malik et al., 2009) The authors postulated that the production of this molecule "... could be derived from its biosynthesis since it forms from amino acid derivatives and the bacteria were shown to be able to produce dmethionine derivatives." (Malik et al., 2009)

| DISCUSS ION
Burkholderia cepacia (25416) and V. fischeri (ES114) were found to produce d-AHLs throughout their growth period. All d-AHLs detected were enantiomers of the expected l-AHLs that are produced by these two species (Fuqua et al., 1994;Kuo et al., 1994;Lewenza et al., 1999;Venturi et al., 2004). It is known that the AHL responsible for the regulation of the lux operon and bioluminescence in V. fischeri is OHHL (Fuqua et al., 1994;Kuo et al., 1994). However, the amounts of OHL produced were ~1500 times higher than OHHL for this strain of V. fischeri during this study. It has been observed that ES114 produces OHL at higher amounts than OHHL (Stabb et al., 2008). Nevertheless, bioluminescence of V. fischeri correlated with the most produced AHLs, l-and d- OHL. The role of l-OHL in V. fischeri and its involvement in bioluminescence has been explored by others. These studies conclude that the role of l-OHL is as an initial inducer of bioluminescence and regulator of OHHL production in the lux system of V. fischeri (Collins et al., 2005;Lupp & Ruby, 2004;Miyashiro & Ruby, 2012;Sun et al., 2018).
d-OHL was the second most produced AHL both by V. fischeri and by B. cepacia cultured in M9 media ( Figure 4). All other AHLs were produced in much lower amounts. Although pH-mediated chemical hydrolysis could decrease the total AHLs, this would not affect the relative amounts of AHL enantiomers that have identical chemical hydrolysis rates. Further, chemical hydrolysis at the pH encountered in these studies was minimal (pH 7.2 to 7.4) (see Materials and Methods, Section 2.2). Questions arise as to both the mechanism of production and the possible biological significance of the ubiquitous d-AHLs. One suggested biosynthetic pathway was that d-AHLs are produced via the same process as the l-AHLs, but perhaps with a d-methionine derivative starting material.
Another possibility is that, initially, the AHLs produced are of the l-configuration and they are subsequently acted upon by a racemase enzyme. Amino acid racemases, including ones for serine and methionine, are well-known throughout many biological systems (Radkov & Moe, 2014), but no AHL racemases have been reported, to our knowledge. F I G U R E 7 Production of (a) l-octanoylhomoserine lactone (l-OHL) and (b) d-octanoyl-homoserine lactone (d-OHL) with biomass (Abs OD 600) curves during growth of B. cepacia supplemented with 6.8 mmol of l-methionine (l-Met), 6.8 mmol of d-methionine (d-Met) or supplemented with 3.4 mmol of lmethionine and 3.4 mmol of d-methionine (d,l-Met). Data represent averages with standard deviations calculated from four replicates 2014). It appears that any added methionine (ie, l-, d-or d,l-) to the B. cepacia cultures was mainly used as an alternative carbon and nitrogen source in the M9 minimal media. The fact that l-Met supplemented cultures produced lower amounts of all AHLs than d-Met cultures was surprising. Perhaps l-Met has immediate utility in the organism's other metabolic pathways (Tanaka et al., 1985), while d-Met is more broadly useful as a carbon and nitrogen source. d-AHL production was not exclusively enhanced by the addition of any of the methionine stereoisomers. In the biosynthesis of l-AHLs, l-methionine is used to make S-adenosylmethionine, with this metabolite transforming into various l-AHLs when acylated by an acyl-carrier protein. It appears that d-AHLs may not be directly produced by the conventional biosynthetic pathway of l-AHLs. dmethionine does not seem to play a direct role in the production of d-AHLs. Auto-racemization studies have been applied to AHLs, with findings suggesting there is negligible physico-chemical racemization occurring in the bacterial culture nor during the chromatographic time frame (Hodgkinson et al., 2011;Malik et al., 2009).
There are at least two other possibilities for the origin of d-AHLs, another unidentified synthetic pathway, distinct from that of l-AHL synthesis, or the aforementioned possibility of chiral interconversion between l-and d-AHLs. Collins et al. (2005) studied the directed evolution of V. fischeri LuxR as a means to broaden the spectrum and sensitivity to different l-AHLs. They found that broadening a given protein's specificity is the easiest evolutionary path to obtaining higher selectivity to new signaling molecules. However, their focus was on mutations that affected or were affected by the nature of the acyl group, but not on the stereochemistry of the homoserine lactone group. This is not to say that an analogous mechanism may not be relevant for d-AHLs, only that it hasn't been considered or studied. d-AHLs are known to have acted as autoinducer molecules . However, it also is possible that they play some other, as yet undetermined, regulatory or feedback role in quorum sensing.

| CON CLUS IONS
This work is the first to detect d-AHLs in the wild-type strain of V. fischeri (ES114) and the first to monitor the production of d-AHLs over time for B. cepacia (25416) and V. fischeri (ES114). The AHL screening and production curves demonstrated that d-OHL and d-HHL are produced in significant amounts by both bacteria. In V. fischeri, maximum bioluminescence correlated with maximum concentrations of both d-OHL and l-OHL. The addition of any enantiomer of methionine, into M9 minimal media, augments the production of all AHLs (l-and d-) by B. cepacia. However, the d-to l-ratios of AHLs are not affected. The supplementation of the B. cepacia growth media with d-Met enhanced the production of both l and d-AHLs to a greater extent than the addition of equivalent amounts of l or d,l-Met. The study indicates that another biosynthetic pathway is likely involved in d-AHL production. Alternatively, the initially produced l-AHL is converted to the d-AHL. The reasons for the ubiquitous presence of d-AHLs are considered but are not yet well understood.

ACK N OWLED G EM ENTS
The authors are grateful to Len Sidiski at Millipore Sigma (USA) for providing the SPE and BSTFA derivatizing agents. The assistance of Dr. M. Farooq Wahab and Dr. Jeongjae Yu is acknowledged. Dr.
Kevin Schug is acknowledged for providing access to culturing and growth curve monitoring equipment.

CO N FLI C T O F I NTE R E S T
None declared.

E TH I C S S TATEM ENT
None required.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data generated and analyzed during this study are provided in this publication.