The Lactococcus lactis KF147 nonribosomal peptide synthetase/polyketide synthase system confers resistance to oxidative stress during growth on plant leaf tissue lysate

Abstract Strains of Lactococcus lactis isolated from plant tissues possess adaptations that support their survival and growth in plant‐associated microbial habitats. We previously demonstrated that genes coding for a hybrid nonribosomal peptide synthetase/polyketide synthase (NRPS/PKS) system involved in production of an uncharacterized secondary metabolite are specifically induced in L. lactis KF147 during growth on plant tissues. Notably, this NRPS/PKS has only been identified in plant‐isolated strains of L. lactis. Here, we show that the L. lactis KF147 NRPS/PKS genes have homologs in certain Streptococcus mutans isolates and the genetic organization of the NRPS/PKS locus is conserved among L. lactis strains. Using an L. lactis KF147 mutant deficient in synthesis of NrpC, a 4′‐phosphopantetheinyl transferase, we found that the NRPS/PKS system improves L. lactis during growth under oxidative conditions in Arapidopsis thaliana leaf lysate. The NRPS/PKS system also improves tolerance of L. lactis to reactive oxygen species and specifically H2O2 and superoxide radicals in culture medium. These findings indicate that this secondary metabolite provides a novel mechanism for reactive oxygen species detoxification not previously known for this species.


| INTRODUCTION
Numerous bacterial species in multiple phyla possess the capacity to synthesize secondary metabolite peptides and carboxy acids by nonribosomal peptide synthetases (NRPS) and polyketide synthases (PKS) (Weissman, 2015). These secondary metabolites can include proteinogenic and nonproteinogenic amino acids and ketides assembled into diverse end-products with functions ranging from antibiotics to immunosuppressants (Siezen & Khayatt, 2008). NRPS/PKS systems have been described in detail for many industrially significant bacteria but there are few reports characterizing these systems and their resulting products by lactic acid bacteria (LAB) (Lin et al., 2015;Wu et al., 2010).
LAB are gram-positive, nonsporulating bacteria in the Firmicutes phylum and constitute a genetically diverse collection of species characterized by their fermentation of mono-and di-saccharides to lactic acid. Strains of Lactococcus lactis are among the most extensively characterized LAB and are used in a variety of applications including starter cultures in cheese production , industrial product synthesis (Mierau et al., 2005), and for the delivery of therapeutics to the human gastrointestinal tract (Cano-Garrido, Seras-Franzoso, & Garcia-Fruitos, 2015;Wells & Mercenier, 2008). Although L. lactis has been most extensively investigated for its adaptation to dairy environments, it is generally regarded that plants are the ancestral habitat for this species Cavanagh, Casey et al., 2015). L. lactis KF147, a strain originally isolated from mung bean sprouts, is genetically distinct from dairy-associated isolates . Its unique traits include the ability to metabolize a wide array of plantderived carbohydrates and synthesize exopolysacharides .
The genome of L. lactis subspecies lactis KF147 contains a hybrid NRPS/PKS gene cluster (Siezen et al., , 2010. These chromosomally-located genes have thus far only been identified in plant-derived lactococci . However, the conditions in which those genes are expressed were not described. Because the production of secondary metabolites poses a significant energy burden on the cell, genes necessary for the production of those cell products are frequently cryptic in standard laboratory media (Rutledge & Challis, 2015). In fact, stringent transcriptional control and regulation of NRPS/PKS production is a frequent challenge that impedes the study of these systems and other secondary metabolites (Rutledge & Challis, 2015). Therefore, it was notable when we found that the L. lactis KF147 genes coding for the hybrid NRPS/PKS are induced during the growth of this strain in Arabidopsis thaliana leaf tissues (Golomb & Marco, 2015). The findings provided an opportunity to study the functional relevance of the NRPS/PKS system in L. lactis. Here, we examine the genetic relatedness of the NRPS/PKS to other lactococci and Streptococcus mutans strains and demonstrate the contribution of the secondary metabolite to L. lactis growth in oxidative conditions, most appreciably on plant tissues.

| Bacterial strains and culture conditions
Lactococcus lactis subsp. lactis KF147 (Kelly, Davey, & Ward, 1998) was provided by NIZO food research (Ede, The Netherlands) and was maintained as a frozen glycerol stock at −80°C. L. lactis was routinely grown at 30°C without agitation on M17 broth (BD, Franklin Lakes, NJ) supplemented with 0.5% d-glucose (GM17). Arabidopsis thaliana Col-1 leaf lysate (ATL medium) was prepared as previously described (Golomb & Marco, 2015). Escherichia coli DH5α was grown at 37°C on LB (Fisher Scientific, Waltham, MA). When necessary, the culture medium was supplemented with ampicillin at 100 μg/ml for E. coli and erythromycin at 5 μg/ml for L. lactis. For reactive oxygen species tolerance experiments, optical density (OD) at 600 nm was used to examine L. lactis growth in GM17 in a Synergy 2 microplate reader (BioTek, Winooski, VT) in 200 μl volumes. L. lactis numbers were also determined by plating serial dilutions of the cultures onto GM17 agar and incubation at 24 hr prior to viable cell enumerations. L. lactis used for growth experiments was prepared from stationary phase cells incubated in GM17 for approximately 18 hr. Cells were collected by centrifugation (10,000g, 3 min) and washed twice in phosphate-buffered saline [PBS, 137 mmol/L NaCl, 2.7 mmol/L KCl, 4.3 mmol/L Na 2 HPO 4 , 1.4 mmol/L KH 2 PO 4 (pH 7)] prior to use.

| Construction of the L. lactis nrpC deletion mutant BAL1
Standard molecular biology techniques were performed as previously described (Sambrook, Fritsch, & Maniatis, 1989). Double-crossover homologous recombination was used to generate a markerless nrpC deletion mutant of KF147, using the suicide vector pRV300 (Leloup, Ehrlich, Zagorec, & Morel-Deville, 1997). All primers used for mutant construction are shown in Table S1. The upstream flanking region of nrpC was PCR amplified, using primers A and B and the downstream flanking region was PCR amplified, using primers C and D. The resulting amplicons were combined by splicing-by-overlap extension (SOEing) PCR (Horton, Cai, Ho, & Pease, 1990). The resulting PCR product was digested with SalI and SacII, cloned into pRV300, and transformed into E. coli DH5α to yield pNRPC-KO. pNRPC-KO was next introduced into KF147 by electroporation as previously described (Golomb & Marco, 2015). Single-crossover mutants were selected on GM17 agar plates supplemented with erythromycin. A single-crossover mutant grown on GM17 broth in the absence of erythromycin for approximately eight passages was sufficient to facilitate excision of the plasmid. Double-crossover mutants were identified by replica plating onto GM17 agar supplemented with erythromycin. Erythromycin-sensitive colonies were screened by PCR for a nrpC deletion. PCR products with primers A and D that resulted in a product size of 1125 bp indicated the absence of nrpC and pRV300 from the chromosome (Table S1).
The PCR products were also sequenced (UC Davis DNA Sequencing Facility (http://dnaseq.ucdavis.edu) for confirmation. A single nrpC deletion mutant, strain BAL1, was used in subsequent experiments.

| ROS tolerance
The capacity of L. lactis to tolerate Reactive oxygen species (ROS) was examined in multiple ways including (1) growth in H 2 O 2, (2) para- where indicated, 20 mmol/L paraquat.

| Statistical analysis
Significant differences in L. lactis growth rates were determined using an unpaired, two-tailed Student t-test. Differences were considered significant if p < .05.

| The NRPS/PKS locus contains distinct functional features and is conserved among plant-associated L. lactis strains
The L. lactis KF147 genes coding for the NRPS/PKS system (llkf_1211 to llkf_1222) span over 40 kb (ca. 1.5% of the chromosome) and consist of a two-component response regulator ( Du et al., 2013). Genes coding for the NRPS/PKS system were also detected in L. lactis Li-1, KF134, KF146, and KF196 according to comparative genome hybridization  and confirmed by BLASTp analysis (data not shown). Each of these strains was isolated from (fermented) plant foods and are classified to the L. lactis subspecies lactis. Due to the high sequence conservation (>99%) amongst the NRPS/PKS systems in these strains, we expect them to synthesize the same secondary metabolite.
Inspection of the genes flanking the NRPS/PKS locus in L. lactis strains KF147 and NCDO2118 showed that they are inserted relative to other L. lactis strains between genes encoding acetolactate synthase (als, upstream) and two hypothetical LPxTG membrane proteins (lpxAB) followed by the kdpDEABC potassium transport system (downstream) (Figure 1 and Table 1). L. lactis strains that lack the NRPS/PKS system but share the same flanking gene arrangement as KF147 were isolated from other sources including strain IO-1 (a kitchen sink isolate) (Kato et al., 2012), CV56 (a vaginal isolate) (Gao et al., 2011), and KLDS 4.0325 (a dairy isolate) (Yang, Wang, & Huo, 2013) (Figure 1). Importantly, even within the lactis subspecies, there is considerable variation in the NRPS/PKS flanking regions because not all strains display this genetic composition. In this regard, the dairy-associated F I G U R E 1 The hybrid NRPS/PKS from L. lactis KF147. Genes on the genomic island are colored as follows: two-component transcriptional regulator (yellow), secondary metabolite biosynthesis enzymes (green), ABC transporter (orange), macrolide biosynthetic protein AvrD (purple), phytoene dehydrogenase (pink), DNA integration/recombination/inversion protein (red), and hypothetical proteins (white). Insertion of the genomic island is shown relative to L. lactis strains originating from different sources (strains IO-1, CV56, and KLDS 4.0325) genes in blue. The fold-induction of each gene in ATL relative to GM17 is shown in parentheses as determined previously (Golomb & Marco, 2015). N/C indicates no change in expression was observed. Gene expression results were not reported for llkf_RS06220, llkf_RS06235, and llkf_RS06240 because they are newly annotated open reading frames L. lactis strain IL1403 does not share the same gene arrangement.
In IL1403, als is located on a distal part of the chromosome and this strain lacks kdpDEABC.
The L. lactis KF147 NRPS/PKS system is most similar to S. mutans LJ23, a strain for which the complete genome sequence is available (Aikawa et al., 2012). None of the genes downstream of llkf_1224 had homologs in S. mutans LJ23, with the exception of the gene encoding the putative macrolide biosynthetic protein AvrD (llkf_1224).
This gene is in the same location relative to the NRPS/PKS system in S. mutans LJ23 and shares 61% amino acid identity with 96% coverage (Table 1).
Although the function of the NRPS/PKS in S. mutans LJ23 has not been directly investigated, the same system in S. mutans UA140 was recently characterized (Wu et al., 2010). This NRPS/PKS product was the first to be characterized for a LAB species and shown to confer pigmentation and improve oxygen and hydrogen peroxide stress tolerance (Wu et al., 2010). The S. mutans UA140 NRPS/PKS locus is identical to that of S. mutans LJ23 (personal communication, Fengxia Qi) and therefore is also conserved with L. lactis KF147 (Figure 1 and Table 1).

| ATL and paraquat in combination delay growth of the NRPS/PKS deletion mutant
We previously found significant increases in L. lactis KF147 NRPS/ PKS gene transcript levels during growth in Arabidopsis leaf tissue lysate (ATL) compared to in GM17 (Golomb & Marco, 2015).
Therefore, we investigated the importance of the NRPS/PKS system in conferring optimal growth rates to L. lactis in ATL by constructing strain BAL1, a markerless, nrpC deletion mutant of L. lactis KF147. The nrpC gene was targeted for deletion because it encodes the PPTase responsible for transferring the 4′-phosphopantetheine moiety to the carrier protein domain of nonribosomal peptide synthetases and polyketide synthases. Removal of the PPTase prevents substrate attachment and synthesis of the NRPS/PKS product (Beld, Sonnenschein, Vickery, Noel, & Burkart, 2014). Static incubation in ATL revealed that the NRPS/PKS system did not provide a specific advantage to L. lactis during exponential phase growth in this medium, as both KF147 and BAL1 exhibited the same growth rates and no differences in final cell yields were observed (data not shown).
Because the NRPS/PKS system from L. lactis KF147 is homologous to a system in S. mutans strains UA140 and LJ23, we reasoned that the requirement for the NRPS/PKS product could depend on the extent to which reactive oxygen species are released by the plant tissues.
To study the possible function of the NRPS/PKS system in tolerance for both) (Figure 3a). The growth impairment was greater for BAL1 than for wild-type KF147 in the presence of paraquat (p < .0005).
After 2.5 hr, the growth rates of both strains slowed considerably (μ = 0.225 ± 0.013 hr −1 and μ = 0.144 ± 0.008 hr −1 for KF147 and BAL1, respectively) and the deletion mutant was again more impaired (p < .0005) (Figure 3b). Similarly, exposure of L. lactis KF147 and BAL1 to the superoxide-generating reagents xanthine oxidase (XO) and xanthine under static conditions in GM17 also resulted in a subtle but significant decrease in growth rate for BAL1 compared to KF147 ( Table 2).
The addition of H 2 O 2 to GM17 also resulted in a significantly decreased growth rate for BAL1 as compared to KF147 (  (Lin et al., 2015) and S. mutans UA159 (Joyner et al., 2010), and UA140 (Wu et al., 2010).
Remarkably, the L. lactis KF147 NRPS/PKS system is similar in both gene structure and function to the NRPS/PKS in S. mutans UA140 and LJ23 and not to orthologous systems in other lactic acid bacteria. Both of the latter species are members of the Streptococcaceae, however, they inhabit distinct ecological niches and it is unclear as to in which species the NRPS/PKS system originated (or perhaps a third as of yet unknown species). In addition to the NRPS/PKS system, the genetic locus also contains a gene annotated as avrD (llkf_1224). It is notable that this protein is also produced by certain plant pathogens and is recognized by plant host resistance (R) proteins to result in plant production of an oxidative burst and generation of superoxide radicals (Lamb & Dixon, 1997;Spoel & Dong, 2012). This finding is consistent with our results showing that the production of the NRPS/ PKS appears to be dispensable for L. lactis during growth in standard laboratory media, but relevant to the plant environment, possibly either to counter responses to pathogenic bacteria (oxidative burst) or to survive generally, aerobic conditions on leaf tissues. In conclusion, tolerance to ROS compounds due to expression of an NRPS/PKS product represents a novel strategy for L. lactis ROS detoxification, possibly during growth on living plants, and with potential application to support the viability of this organism in culture production and preservation.

ACKNOWLEDGMENTS
We thank Pujita Munnangi for her assistance with this project. We also acknowledge Dr. Fengxia Qi at the University of Oklahoma for providing S. mutans UA140 sequence data and Dr. Roland Siezen at Radboud University Nijmegen for insightful discussion. This research was supported by start-up funds provided to Dr. Marco and the Agricultural Experiment Station at the University of California, Davis.
T A B L E 2 Growth rates (hr −1 ) of L. lactis KF147 and BAL1 exposed to ROS