An insight into the role of branched-chain α-keto acid dehydrogenase (BKD) complex in branched-chain fatty acid biosynthesis and virulence of Listeria monocytogenes

ABSTRACT

Listeria monocytogenes is a foodborne bacterial pathogen that causes listeriosis. Positive regulatory factor A (PrfA) is a pleiotropic master activator of virulence genes of L. monocytogenes that becomes active upon the entry of the bacterium into the cytosol of infected cells. L. monocytogenes can survive and multiply at low temperatures; this is accomplished through the maintenance of appropriate membrane fluidity via branched-chain fatty acid (BCFA) synthesis. Branched-chain α-keto acid dehydrogenase (BKD), which is composed of four polypeptides encoded by lpd, bkdA1, bkdA2, and bkdB, is known to play a vital role in BCFA biosynthesis. Here, we constructed BKD-deficient Listeria strains by in-frame deletion of lpd, bkdA1, bkdA2, and bkdB genes. To determine the role in in vivo and in vitro, mouse model challenges, plaque assay in murine L2 fibroblast, and intracellular replication in J744A.1 macrophage were conducted. BKD-deficient strains exhibited defects in BCFA composition, virulence, and PrfA-regulon function within the host cells. Transcriptomics analysis revealed that the transcript level of the PrfA-regulon was lower in ΔbkdA1 strain than those in the wild-type. This study demonstrates that L. monocytogenes strains lacking BKD complex components were defective in PrfA-regulon function, and full activation of wild-type prfA may not occur within host cells in the absence of BKD. Further study will investigate the consequences of BKD deletion on PrfA function through altering BCFA catabolism.

IMPORTANCE

Listeria monocytogenes is the causative agent of listeriosis, a disease with a high mortality rate. In this study, we have shown that the deletion of BKD can impact the function of PrfA and the PrfA-regulon. The production of virulence proteins within host cells is necessary for L. monocytogenes to promote its intracellular survival and is likely dependent on membrane integrity. We thus report a link between L. monocytogenes membrane integrity and the function of PrfA. This knowledge will increase our understanding of L. monocytogenes pathogenesis, which may provide insight into the development of antimicrobial agents.

INTRODUCTION

Listeria monocytogenes is a foodborne facultative intracellular bacterial pathogen and is the causative agent of listeriosis, a disease with a high mortality rate (1). Invasive listeriosis typically manifests as bacteremia with or without central nervous system infection, including severe meningitis, meningoencephalitis, and brain abscess in immunocompromised people. Listeriosis complications in pregnant women result in miscarriage or stillbirth as well as neonatal infection (2). In healthy individuals, L. monocytogenes can colonize the gastrointestinal tract, resulting in febrile gastroenteritis (3). Although listeriosis is not as common as other foodborne diseases, the significant mortality rate associated with L. monocytogenes makes it a major public health concern (4). L. monocytogenes is notable as it survives as an environmental bacterium associated with groundwater and decaying vegetation yet maintains the ability to force entry into various eukaryotic cells, survive intracellularly, elude the immune system, and disseminate throughout host tissues via a sophisticated mechanism involving a number of virulence factors (5).

L. monocytogenes causes disease following the consumption of contaminated food products, and the bacterium is recognized for its impressive ability to adapt to stressful conditions encountered in various habitats including food processing plants, which in turn leads to its widespread dissemination (6). For example, L. monocytogenes is exceptional in ability to survive and multiply in a wide range of temperatures, including refrigeration temperatures, and this characteristic of the organism leads to frequent and fatal food-associated outbreaks (5). Maintaining membrane fluidity with a high proportion of branched-chain fatty acids (BCFA), particularly anteiso-C_15:0 and anteiso-C_17:0 fatty acids, is a key mechanism that contributes to L. monocytogenes adaptation to low temperatures and other environmental stresses (including pH, salt, and/or CO₂) (7–9). BCFAs are synthesized from the degradation of branched-chain amino acids (BCAAs; Fig. S1). In the initial step, BCAAs are catabolized into branched-chain α-keto acids by branched-chain amino acid transaminase, and subsequently, branched-chain α-keto acids are oxidatively decarboxylated into short-chain CoA precursors by branched-chain α-keto acid dehydrogenase (BKD) (10).

BKD is a multi-subunit enzyme complex that is composed of four polypeptides: a dihydrolipoamide dehydrogenase (E3), a dehydrogenase (E1α), a decarboxylase (E1β), and a dihydrolipoamide acyltransferase (E2), which are encoded by lpd, bkdA1, bkdA2, and bkdB genes, respectively. The function of BKD in BCFA biosynthesis is well documented in L. monocytogenes and several other Gram-positive pathogens (9, 11, 12). Previous work has shown that transposon (Tn) insertion mutations into lpd and bkdB (originally defined as cld-1 and cld-2) cause cold-sensitivity phenotypes due to lower levels of anteiso-C_15:0 and anteiso-C_17:0 (13, 14). The growth defects of cld-1 and cld-2 mutants were reversed upon the addition of 2-methylbutyrate (2MB), the precursor for anteiso-BCFA (15). BKD thus appears important for the maintenance of L. monocytogenes membrane integrity in response to environmental changes. Other studies have reported that BKD and subsequent anteiso-BCFA synthesis are required for intracellular replication and L. monocytogenes pathogenicity (16, 17).

The success of L. monocytogenes as an intracellular pathogen also relies on the secretion of a number of virulence-associated proteins whose gene expression is regulated by the transcriptional activator Positive Regulatory Factor A (PrfA); also relevant are genes induced by conditions of stress that are regulated by the alternative sigma factor σ^B (18, 19). Within the cytosol of infected-host cells (Fig. S2), PrfA becomes highly active and induces key virulence factors required for adhesion and invasion of L. monocytogenes into host cells, phagosomal escape, intracellular replication inside the cytosol, actin-based bacterial motility, and spread to adjacent cells (20, 21). Genes induced by PrfA include inlA and inlB which encode internalins, plcA which encodes a phosphatidylinositol phospholipase, plcB which encodes a lecithinase with broad-substrate specificity, mpl which encodes a metal-dependent protease involved in PlcB maturation, and actA which encodes an actin-based motility protein ActA (22). The mechanism by which PrfA becomes activated inside the host cells is not fully understood; however, recently, it has been recognized that PrfA activation within the cytosol occurs following PrfA binding to glutathione (23). Here, we described the impact of each component of the BKD complex in the modulation of membrane integrity and fluidity as well as its relation with the virulence factors of L. monocytogenes.

RESULTS

Analysis of the BKD-operon and its predicted promoter regions

We performed in silico analyses of the BKD-operon and its promoter region using online tools (BacPP, BPROM, and PePPER) (24, 25). The promoter region appears to be 137 bp in size and located upstream of buk (Fig. 1A). Interestingly, we identified two putative independent promoters: one is σ^A-dependent (rpoD, or the housekeeping sigma factor), and the second one is σ^B-dependent (sigB, or the stress response associated sigma factor; Fig. 1B). These two promoters are located within the same DNA region located upstream of the buk gene. By comparing different L. monocytogenes strains, we found that the σ^A- and σ^B-dependent promoter sequences [GTTT-N12-17-GGGWAW (W is A or T)] are well conserved in the BKD-operon of a broad range of L. monocytogenes isolates, including EGD-e and 10403S strains. Moreover, we found two sigma B stressosome components, sbrA (sigmaB-dependent small RNA A) and prli42 (encoding a membrane protein), located downstream of the BKD-operon (26, 27). A GC-rich DNA region (24 bp) within the 3'-end of buk (AGCGGAGTGCAACGCGTGCTCGCT) was also identified that has the potential to form a stem-and-loop structure for putative processing of a long mRNA covering an operon extending from ptb to bkdB.

Fig 1.

Schematic representation of BKD-operon in L. monocytogenes F2365 strain. (A) BKD-operon consists of six genes, including ptb, buk, and four genes encoding BKD. Downstream of the BKD-operon is sbrA and prli42. The region upstream of the buk contains two putative σ^A- and σ^B-dependent promoters. The putative PrfA-binding site is located within buk gene. Two transposon insertional mutants cld-1 and cld-2 constructed by disrupting bkdB and lpd genes are marked with a triangle (14). Putative σ^A- and σ^B-dependent promoters in the BKD-operon are marked with arrows. (B) The putative −35 and −10 boxes of sigA and sigB recognized promoters in the BKD-operon.

BKD complex components play a role in BCFA biosynthesis

L. monocytogenes strains containing in-frame deletions in the genes encoding four BKD subunits (lpd, bkdA1, bkdA2, and bkdB) were constructed in strain F2365 (serotype 4b) using allelic exchange (28). A total of 1,428 bp, 996 bp, 984 bp, and 1,251 bp of lpd, bkdA1, bkdA2, and bkdB open-reading frames were deleted and confirmed by PCR and sequencing. We then analyzed the BCFA composition of total lipid extracts from the BKD-deficient strains and wild-type strains grown in a defined minimal medium (MM) by gas chromatography-mass spectrometry (GC-MS) (29). The analysis revealed a pronounced decrease in anteiso-C_15:0 and anteiso-C_17:0, across all BKD-deficient strains (Δlpd, ΔbkdA1, ΔbkdA2, and ΔbkdB) compared to the wild-type strain F2365 (Table 1), indicating the pivotal role of the BKD complex in the biosynthesis of anteiso-BCFAs. The anteiso-BCFAs, derived from isoleucine, are critical for the optimal function and fluidity of bacterial membranes (30). The reduction of anteiso-C_15:0 was most pronounced in the ∆bkdA1, ∆bkdA2, and ∆bkdB strains, where the levels were reduced from 75.05% ± 3.99% in the wild-type to 6.21% ± 1.44%, 2.26% ± 0.88%, and 7.06% ± 2.90%, respectively. In addition, a significant decrease in the odd-numbered iso-BCFA (iso-C_15:0 and iso-C_17:0, leucine-derived) was detected in the four BKD-deficient strains compared to the wild-type. Moreover, a significant increase in the even-numbered iso-BCFA (iso- C_14:0 and iso-C_16:0, valine derived) was detected in the Δlpd, ΔbkdA1, ΔbkdA2, and ΔbkdB strains compared to wild-type (Table 1). This change highlights the BKD complex’s involvement in the metabolism of leucine and valine as precursors for the synthesis of iso-BCFAs. In contrast, there was a notable increase in the total percent of straight-chain fatty acids (SCFAs) in the Δlpd, ΔbkdA1, ΔbkdA2, and ΔbkdB strains compared to wild-type (26.13% ± 1.37%, 59.32% ± 2.22%, 71.47% ± 12.88%, and 62.70% ± 8.54% vs 12.38% ± 1.21%). More specifically, C_14:0 (myristic acid), C_15:0 (pentadecanoic acid), and C_16:0 (palmitic acid) were significantly higher in Δlpd (5.46% ± 1.13%, 0.49% ± 0.09%, and 0.14% ± 2.60%), ΔbkdA1 (17.66% ± 1.91%, 4.47% ± 1.16%, and 37.09% ± 0.87%), ΔbkdA2 (10.73% ± 3.64%, 1.38% ± 0.54%, and 59.27% ± 8.61%), and ΔbkdB (6.34% ± 2.28%, 1.04% ± 0.40%, and 55.30% ± 5.84%) compared to wild-type stains (0.61% ± 0.08%, 0.13% ± 0.01%, and 11.66% ± 1.39%). This elevation in SCFAs further supports that modification in membrane fatty acids composition is a mechanism of cold adaptation of L. monocytogenes. Complementation showed that the introduction of the corresponding wild-type gene restored the BCFA profile in BKD-deficient strains to wild-type levels.

TABLE 1.

The total fatty acid composition of L. monocytogenes F2365, BKD-deficient strains, and their complemented strains^d

Strains	Percentage (wt/wt) of total fatty acids (mean ± SD)
	SCFAs					Total SCFA	Odd-number anteiso-BCFA		Odd-number iso-BCFA		Even-number iso-BCFA
	C12:0	C14:0		C15:0	C16:0		Anteiso-C15:0	Anteiso-C17:0	Iso-C15:0	Iso-C17:0	Iso-C14:0	Iso-C16:0	Total BCFA
F2365	0.03 ± 0.0	0.61 ± 0.08		0.13 ± 0.01	11.66 ± 1.39	12.38 ± 1.21	75.05 ± 3.99	0.24 ± 0.07	5.60 ± 2.21	1.23 ± 0.38	0.01 ± 0.00	1.34 ± 0.50	83.46 ± 7.15
∆lpd	0.04 ± 0.0	5.46 ± 1.13a		0.49 ± 0.09a	20.14 ± 2.6a	26.13 ± 1.37c	56.28 ± 8.11c	0.04 ± 0.02	3.87 ± 1.01	0.73 ± 0.02	0.02 ± 0.00	7.54 ± 1.50	68 ± 10.64b
∆bkdA1	0.09 ± 0.02	17.66 ± 1.9c		4.47 ± 1.16a	37.09 ± 0.87c	59.32 ± 2.22c	6.21 ± 1.44c	0.01 ± 0.00	1.82 ± 0.41	0.78 ± 0.09	0.05 ± 0.01	24.76 ± 5.16c	33.64 ± 7.13c
∆bkdA2	0.14 ± 0.0	10.73 ± 3.64a		1.38 ± 0.54	59.27 ± 8.61c	71.47 ± 12.88	2.26 ± 0.88c	0.01 ± 0.00	1.68 ± 0.64	1.21 ± 0.00	0.05 ± 0.01	12.00 ± 5.05	16.81 ± 7.28c
∆bkdB	0.04 ± 0.0	6.34 ± 2.28a		1.04 ± 0.40	55.30 ± 5.84c	62.70 ± 8.54b	7.06 ± 2.90c	0.02 ± 0.01	1.14 ± 0.40	1.03 ± 0.00	0.05 ± 0.01	16.87 ± 7.16b	25 ± 11.07c
∆lpd_Com	0.02 ± 0.00	1.45 ± 0.30		0.23 ± 0.06	12.44 ± 0.53	14.13 ± 0.17	76.33 ± 3.91	0.14 ± 0.03	2.59 ± 0.66	0.47 ± 0.06	0.01 ± 0.00	3.23 ± 0.83	82 ± 5.50
∆bkdA1_Com	0.01 ± 0.00	0.32 ± 0.05		0.05 ± 0.00	5.80 ± 0.67	6.17 ± 0.76	88.85 ± 0.05	0.19 ± 0.01	2.22 ± 0.54	0.33 ± 0.04	ND	0.97 ± 0.23	92.56 ± 1.41
∆bkdA2_Com	0.01 ± 0.00	0.34 ± 0.04		0.05 ± 0.00	7.09 ± 0.47	7.46 ± 0.54	86.53 ± 1.45	0.20 ± 0.01	2.74 ± 0.61	0.44 ± 0.04	0.01 ± 0.00	0.96 ± 0.18	90.88 ± 2.28
∆bkdB_Com	0.01 ± 0.00	0.33 ± 0.00		0.04 ± 0.00	6.97 ± 0.00	7.35 ± 0.00	89.28 ± 1.93	0.20 ± 0.02	2.53 ± 0.43	0.41 ± 0.01	0.01 ± 0.00	0.99 ± 0.14	93.41 ± 1.37

^a Indicates a significant difference at P < 0.05.

^b Indicates a significant difference at P < 0.01.

^c Indicates a significant difference at P < 0.001.

^d The Student's t test was used to measure the statistical significance between mutants and wild-type.

BKD complex components play a vital role in cold tolerance of Listeria

Given the impact of the BKD complex on BCFA composition in the membrane of L. monocytogenes, we set out to test phenotypes associated with the loss of BKD to determine the growth in low temperature. Growth kinetics and survival (CFU/mL) of BKD-deficient strains and wild-type strains were first compared by measuring OD₆₀₀ in brain-heart infusion (BHI) at 4°C, 25°C, and 37°C. The growth was then measured in defined MM at 25°C. At 25°C, the four BKD-deficient strains (Δlpd, ΔbkdA1, ΔbkdA2, and ΔbkdB) exhibited significant reductions in growth as measured by optical density and the numbers of viable bacteria (CFU/mL) in comparison to the wild-type strain when grown in BHI (Fig. 2) or in defined MM (Fig. S3A). At refrigeration temperature (4°C), BKD-deficient strains exhibited more severe growth defects in comparison to the parent F2365 strain (Fig. S3B). At 37°C, the growth of BKD-deficient strains was almost identical to the wild-type strain during the exponential phase; however, in the stationary phase, there was a small statistically insignificant reduction in growth compared to the wild-type in BHI (Fig. S3C and D), overall indicating that a functional BKD is required for cold tolerance in L. monocytogenes as previously reported but not for bacterial growth at 37°C (15). The growth and viability of complemented strains were similar to the wild-type in BHI and MM at all tested temperatures (Fig. 2A and B).

Fig 2.

BKD-deficient strains exhibit pronounced growth defects at 25°C compared to parent strain F2365 and their corresponding complemented strains restored the growth. (A) Bacterial growth curves were determined by OD₆₀₀ measurement at 1-hour interval. Bacterial overnight cultures were diluted with fresh BHI in a 48-well plate and incubated at 25°C for 40 hours in Cytation 5 cell imaging multimode reader (Agilent BioTek). The experiment was repeated twice with four replicates each. Error bars are the SEM. Brown-Forsythe and Welch ANOVA with Dunnett’s T3 multiple comparison test were used to measure the statistical significance between mutants and wild-type. *** Indicates a significant difference at P < 0.001. (B) Bacterial viabilities were determined by measuring CFU/mL after overnight culture in BHI at 25°C. The experiment was performed three times with four replicates each. Error bars are the SEM. Welch’s t test was used to measure the statistical significance between mutants and wild-type. ** Indicates a significant difference at P < 0.01, and *** indicates a significant difference at P < 0.001.

Components of the BKD complex are required for full phospholipase activity and listeriolysin-O protein expression levels in L. monocytogenes

Given the remarkable increases in PrfA-dependent protein secretion that occur following entry of L. monocytogenes into the cytosol and the function of PrfA-regulon were downregulated in the intracellular infection process, we investigated the relative secretion of phospholipase activity and amount of listeriolysin-O (LLO) protein in BKD-deficient strains relative to the parent strain. Expression of the two phospholipases-encoding genes (plcA and plcB) and listeriolysin-encoding hly gene is controlled by PrfA following its activation in host cells (21). To characterize phospholipase activity, lecithinase activity was determined on Brilliance Listeria agar (BLA) supplemented with lecithin. We investigated the influence of BKD deletion on the amount of LLO protein relative to the parent strain using western blot analysis. After 48 hours of incubation at 37°C, phospholipase activity as visualized by an opaque zone was significantly reduced for Δlpd, ΔbkdA1, ΔbkdA2, and ΔbkdB strains compared to wild-type (Fig. 3A; Fig. S4A through E). The complementation of BKD-deficient strains resulted in similar phospholipase activity to the wild-type. Western blot analysis revealed that Δlpd, ΔbkdA1, ΔbkdA2, and ΔbkdB strains produced approximately twofold less LLO protein compared to the wild-type (Fig. 3B; Fig. S4F).

Fig 3.

BKD complex components played vital role phospholipase activity, LLO protein level expression, cell-to-cell spread, and intracellular replication of L. monocytogenes. (A) Phospholipase activity of the L. monocytogenes strains was detected on BLA supplemented with lecithin. Bacteria were spotted onto the BLA plate and incubated for 48 hours at 37°C. After 2 days, the zone of opaqueness was measured to determine phospholipase activity of strains. All experiments were performed three independent times with at least three replicates. Error bars are the SEM. ∆prfA and NF-L1777 (prfA*) strains used as a negative and positive control, respectively. Welch’s t test was used to compare BKD-deficient strains with wild-type. (B) Immunoblots of LLO expression and P60 of the indicated strains examined using western blots. LLO protein was normalized to P60 abundance and measured as a percentage of wild-type. Strains grown in BHI plus 0.1% activated charcoal were used for protein extraction and then separated using SDS-PAGE, transferred to PVDF membrane followed by blocking with anti-LLO, graph depiction derived from immune-blot images showing that the amount of LLO protein was lower in the Δlpd, ΔbkdA1, ΔbkdA2, and ΔbkdB strains than the parent strain F2365. ∆prfA and NF-L1777 strains used as a negative and positive control, respectively. (C and D) Plaque number and size were determined as percentages compared to the wild-type (set as 100%). Monolayers of L2 fibroblasts were infected with the indicated strains for 96 hours. The experiment was repeated three independent times. At least 10 plaques were measured for each experiment. ∆prfA and NF-L1777 strains used as a negative and positive control, respectively. Welch’s t test was used to compare BKD-deficient strains with wild-type. (E) Intracellular growth kinetic in J774A.1 macrophage cells. Macrophages were infected with indicated strains for 1 hour. After infection, monolayers were washed, and a medium containing gentamicin was added. At 1-, 3-, and 5-hour post-infection, macrophages were lysed, and released intracellular bacteria were quantified by determining CFU/mL. ∆prfA and NF-L1777 strains used as a negative and positive control, respectively. The experiment was repeated twice with four replicates each. Two-way ANOVA with Dunnett’s multiple comparison test was used to compare BKD-deficient strains with wild-type. * Indicates a significant difference at P < 0.05, ** indicates a significant difference at P < 0. 01, and *** indicates a significant difference at P < 0.001. (F) Real-time PCR (RT-PCR) analyses of PrfA-regulon in ΔbkdA1 compared to wild-type. The data represent means ± SEs from three biological replicates.

BKD complex components are required for cell-to-cell spread and intracellular replication of L. monocytogenes

The ability of L. monocytogenes to survive and replicate intracellularly in host cells is required for bacterial virulence (31). We determined the ability of mutants lacking BKD to spread from cell-to-cell by assessing plaque formation in murine L2 fibroblasts. The Δlpd, ΔbkdA1, ΔbkdA2, and ΔbkdB exhibited fewer numbers of plaques (an average of 48.5%, 55.44%, 59.41%, and 57.43% reduction in plaque numbers) and a modest reduction in plaque size (6.86%, 8.05%, 10.83%, and 14.19% reduction in plaque size) compared to those cells infected with wild-type strain (Fig. 3C and D). Complemented strains exhibited normal patterns of plaque formation. At 1-, 3-, and 5-hour post-infection, bacterial replication in J744A.1 macrophages revealed that intracellular numbers of Δlpd, ΔbkdA1, ΔbkdA2, and ΔbkdB strains were decreased compared to the parent strain (Fig. 3E). There was no significant difference at 1 and 3 hours but highly significant at 5 hours of post-infection.

BKD complex components are needed for virulence of L. monocytogenes

To assess the impact of the loss of BKD during infection, Swiss Webster mice were intravenously (IV) infected with L. monocytogenes mutant and wild-type strains, and the bacterial burdens were compared in the spleen and liver following 72 hours of infection. Mice infected with the Δlpd, ΔbkdA1, ΔbkdA2, and ΔbkdB strains exhibited significantly reduced bacterial burdens in livers and spleens in comparison to mice infected with wild-type, indicating that BKD is required for full L. monocytogenes virulence (Fig. 4A and B). Compared to mice infected with the wild-type strain, L. monocytogenes strains with deletions of lpd, bkdA1, bkdA2, and bkdB exhibited approximately 2.5-, 4.5-, 4.9-, and 4.9-log₁₀ reduction in bacterial loads in livers and −2, 3.7-, 3.7-, and 3.6 -log₁₀ lower bacterial burdens in spleens (Fig. 4A and B). Complementation of the mutant strains with functional gene copies fully restored the virulence defect associated with the loss of BKD.

Fig 4.

BKD-deficient L. monocytogenes strains are attenuated in mice. Each animal group (five mice per cage) was IV infected with the indicated strain. Bacterial loads were enumerated in spleens (A) and livers (B) at 72 hours post-infection. Data were analyzed using a nonparametric Mann-Whitney test. The horizontal dot line represents the minimum detection limit, and some mice bacterial loads were below the detection level. The data represent the mean ± SEM of five mice in each group. ** Indicates a significant difference at P < 0.01. (C and D) Growth with the BCFA precursor (2 MB) did not restore virulence to BKD-deficient L. monocytogenes strains. Each animal group (five mice per cage) was IV infected with the indicated strains grown overnight with 2 MB. At 24 and 72 hours post-infection, spleens and livers were harvested, and bacterial loads were enumerated. Data were analyzed using a nonparametric Mann-Whitney test. The horizontal dot line represents the minimum detection limit, and some mice bacterial loads were below the detection level. The data represents the mean ± SEM of five mice in each group. ** Indicates a significant difference at P < 0.01, and ns indicates not significant.

Bacterial growth in the presence of BCFA-precursor 2MB did not restore virulence to BKD-deficient strain

2MB is a precursor for the biosynthesis of odd-numbered anteiso-BCFAs, particularly anteiso-C_15:0, under conditions when endogenous BCFA synthesis is compromised (13, 32). The utilization of 2 MB as an exogenous substrate was shown restoration of the proportion of anteiso-C_15:0 and growth at low temperatures in L. monocytogenes strains lacking a functional BKD (15). While BKD catabolizes BCAAs, especially isoleucine, to generate acyl-CoA, 2MB is metabolized by acyl-CoA synthetase into its acyl-CoA ester, thereby bypassing the need for BCAA catabolism mediated by BKD. BCFA composition and growth at 4°C in BKD-deficient strains were partially restored by supplementation of the growth medium with 2MB (Fig. S3E). Given that comprise of anteiso-BCFA in the membrane of ΔbkdA1 strain, we investigated whether full virulence could be restored with the addition of 2MB to growth media. L. monocytogenes strains (ΔbkdA1, wild-type, and complemented) were grown in BHI supplemented with 2MB overnight. Then bacterial pellets were harvested, washed with saline, and used to infect Swiss Webster mice. We found that ΔbkdA1 strain grown in the presence of the 2 MB precursor prior to infection still exhibited significantly lower bacterial burdens within the liver (3.1- and 5.2-log₁₀ lower CFUs/g) and spleen (1- and 2.6-log₁₀ lower CFUs/g) in comparison to the wild-type F2365 strain at 24 and 72 hours post-infection (Fig. 4C and D). This result indicated that the addition of the 2 MB precursor to media restored the growth at low temperature but was unable to compensate for the roles of BKD inside of host cells.

Transcriptomic analysis showed that bkdA1 deletion negatively affects the expression of PrfA-regulon

The defect in the cell-to-cell spread and intracellular replication in BKD-deficient strains led us to investigate the expression level of PrfA-regulon during growth in the J774A.1 macrophage cell line. Transcriptomic analyses of the ΔbkdA1 and wild-type strains were compared using RNA-Seq at 4 hours post-infection in the murine macrophage cell line J774A.1 (33). Results showed that the transcript level of PrfA-regulon (prfA, hly, mpl, plcB, plcA, actA, inlA, inlB, plcA, and hpt) was lower in ΔbkdA1 strain than those in the wild-type strain (Table 2; Fig. S5). The expression of the PrfA-regulon was determined by qPCR and confirmed to be consistent with the RNA-seq results (Fig. 3F). Moreover, the expression of bkdA2 and bkdB genes was decreased in the ΔbkdA1 strain. Interestingly, we found that the deletion of bkdA1 caused upregulation of the BCAA biosynthesis operon, ilvDBNC and leuABCD. In addition, acsA gene encoding acetyl-CoA synthetase and fabH encoding beta-ketoacyl-ACP synthase III were upregulated in the ΔbkdA1 strain (Fig. S5). Together, these results suggest that the BKD mutation negatively affects the expression of PrfA-regulon and possibly PrfA activation.

TABLE 2.

Genes with significantly changed expression in ΔbkdA1 strain

	Locus tag	Protein name	Log2 fold change	P value
Downregulated genes	PrfA-regulon
	LMOf2365_0211	Positive regulatory factor A (prfA)	−1.7	3.4E-09
	LMOf2365_0212	1-phosphatidylinositol phosphodiesterase (plcA)	−1.5	1.21E-04
	LMOf2365_0213	Listeriolysin O (hly)	−1.6	7.71E-05
	LMOf2365_0214	Zinc metalloproteinase (mpl)	−1.9	8.5E-03
	LMOf2365_0215	Actin assembly-inducing protein (ActA)	−1.6	2.9E-03
	LMOf2365_0216	Phosphatidylcholine phospholipase C (plcB)	−1.9	1.4E-13
	LMOf2365_0471	Internalin A (InlA)	−1.8	4.4E-04
	LMOf2365_0472	Internalin B (InlB)	−1.7	9.7E-05
	LMOf2365_0855	Hexose phosphate transporter (hpt)	−1.5	5.5E-26
	BKD operon
	LMOf2365_1390	Alpha-ketoacid dehydrogenase subunit beta (bkdA2)	−4.3	4.1E-147
	LMOf2365_1391	2-oxo acid dehydrogenase subunit E2 (bkdB)	−4.2	1.2E-180
Upregulated genes	Valine, leucine, and isoleucine biosynthesis
	LMOf2365_2006	Dihydroxy-acid dehydratase (ilvD)	1.5	1.2E-04
	LMOf2365_2007	Biosynthetic-type acetolactate synthase large subunit (ilvB)	1.7	1.2E-04
	LMOf2365_2008	Acetolactate synthase small subunit (ilvN)	1.5	6.8E-21
	LMOf2365_2009	Ketol-acid reductoisomerase (ilvC)	1.7	3.2E-13
	LMOf2365_2010	2-isopropylmalate synthase (leuA)	7	6.2E-39
	LMOf2365_2011	3-isopropylmalate dehydrogenase (leuB)	1.8	7.4E-16
	LMOf2365_2012	3-isopropylmalate dehydratase large subunit (leuC)	1.5	3.1E-20
	LMOf2365_2013	3-isopropylmalate dehydratase small subunit (leuD)	1.5	3.8E-23
	LMOf2365_2014	Threonine ammonia-lyase (ilvA)	1.52	1.3E-09
	Fatty acid metabolism
	LMOf2365_2235	3-oxoacyl-ACP synthase III (fabH)	2.3	1.6E-84
	LMOf2365_2700	acyl—CoA ligase (acsA)	4.1	1.7E-182
	Metabolic pathways
	LMOf2365_1387	Butyrate kinase (buk)	1.9	3.8E-32
	LMOf2365_1386	Phosphate acetyltransferase (ptb)	2.9	2.4E-62
	LMOf2365_1388	Dihydrolipoyl dehydrogenase (lpd)	2.0	2.3E-28

DISCUSSION

The bacterial membrane serves as a defensive barrier as well as an environmental sensor and is critical for all facets of bacterial survival and replication (34, 35). Bacterial membranes must maintain integrity while being able to support a variety of activities and functions, including cell division, motility, and protein secretion and transport (36). L. monocytogenes experiences a substantial increase in the secretion of virulence proteins upon entry into mammalian host cells; these proteins are required for vacuolar escape and replication within the cytosol. Translocation of virulence factors across the L. monocytogenes membrane requires an increase in the activity of proteins that contribute to bacterial membrane integrity (34, 35). The goal of this study was to explore how modulation of membrane fluidity and BCFA abundance can impact virulence factors secretion within the cytosol of infected host cells.

We identified two potential independent promoters, sigA (σ^A)-dependent (also known as rpoD) and sigB (σ^B)-dependent, that occur in the promoter region of the BKD-operon. L. monocytogenes prfA gene is another example of a gene that has putative σ^A- and σ^B-dependent binding sites within the same upstream promoter region (named prfAP2). Two sigma B stressosome elements (Prli42 and SbrA) were identified downstream of the BKD-operon (26, 27). In L. monocytogenes and other Gram-positive bacteria, σ^B plays a key role in coordinating gene transcription in response to various environmental stresses, including high osmolarity, low pH, cold exposure, and antimicrobial stress. Since σ^B is involved in regulating various genes associated with stress responses, it seems feasible that σ^B assists L. monocytogenes in maintaining BCFA and membrane fluidity through its regulation of the BKD-operon. The presence of σ^A- and σ^B-dependent promoters may provide flexibility for L. monocytogenes to alter the expression of BKD in response to changes in the surrounding environment. Future studies will decipher the mechanism of BKD-operon regulation through Sigma B.

BKD is an important contributor to the BCFA synthetic pathway. Our data indicate that individual deletions of BKD components (lpd, bkdA1, bkdA2, and bkdB) all result in reduced anteiso-BCFA in the L. monocytogenes membrane and in significant growth defects at low temperature while maintaining normal bacterial growth at 37°C. This finding is consistent with previous reports that transposon insertions within lpd and bkdB (originally referred to as cld-1 and cld-2 mutants) of L. monocytogenes exhibited significant growth defects at low temperatures (13, 15, 37). The modulation of BCFA levels is a well-characterized mechanism for adjusting membrane fluidity in response to changes in environmental conditions (33, 38, 39). This mechanism has been observed in several bacterial species, such as Bacillus subtilis, Staphylococcus aureus, Enterococcus faecalis, and Pseudomonas putida (12, 40–42). BCFAs have lower melting points than equivalent SCFAs which cause lower phase transition temperatures to membrane lipids, thereby preventing the rigidification of membranes at low temperatures (43). This fluidity is crucial for maintaining membrane-bound processes, including nutrient transport and signal transduction, which are essential for bacterial growth and survival in cold environments (44, 45). This result provides further evidence that the observed cold sensitivity phenotypes in BKD-deficient strains are mainly caused by reduced levels of BCFAs and increased levels of SCFAs.

In the present study, SCFAs (C_14:0, C_15:0, and C_16:0) were major components of the fatty acid profile in the BKD-deficient strains. SCFAs are synthesized by fatty acid synthesis (FASII) using acetyl-CoA and malonyl-CoA precursors (46–48). Typically, acetyl-CoA and malonyl-ACP are catalyzed by β-ketoacyl-ACP synthase III (FabH), followed by subsequent rounds of condensation and acyl chain elongation using a subset of different enzymes (47, 49). As mentioned above, the synthesis of BCFA requires an amino acid-derived branched-chain acyl-CoA as a precursor, which is synthesized by BKD. It was reported previously that FabH prefers acyl-CoA than acetyl-CoA in L. monocytogenes (50). Lacking functional BKD is likely to reduce the pool of acyl-CoA precursors and increase the flux toward acetyl-CoA and malonyl-ACP synthesis and utilization by FabH, leading to increased SCFA production. It is also possible that a lack of BKD affects the overall fatty acid synthesis machinery. However, decreased synthesis of antiso-BCFAs in the BKD deletion strains might result in FabH enzyme utilizing more acetyl-CoA in the initiation step leading to reduced anteiso-BCFA compared to the wild-type strain. Finally, elevated levels of C_14:0, C_15:0, and C_16:0 might be compensating for the lack of anteiso-BCFA in the BKD deletion strains. The detailed examination of these alterations offers a valuable framework for future investigations into the regulatory mechanisms governing bacterial lipid metabolism and their implications for pathogenicity.

While all four subunits of the BKD complex are required for BCFA biosynthesis, the deletion of the E3 component (encoded by lpd) had less impact on growth, BCFA composition, and virulence compared to the other three components (E2 encoded by bkdB, E1α encoded by bkdA1, and E1β encoded bkdA2). E1 is a decarboxylase that decarboxylate branched-chain keto acids using the cofactor thiamin pyrophosphate and transfers the remaining acyl group to E2. The action of E2, which has a covalently attached lipoyl moiety, leads to the subsequent transfer of the acyl group to CoA. The E3 component is a flavoprotein in which oxidation of the lipoyl moiety by the FAD coenzyme results in the reoxidation of FADH₂ to FAD and producing NADH (51). Therefore, E3 might not be directly involved in the decarboxylation of branched-chain keto acids and the transfer of the acyl group to CoA, which are critical steps for the formation of BCFAs, but E1α, E1β, and E2 are important enzymes in this decarboxylation step and are thus more directly linked to BCFA synthesis. Together, the role of E3 in electron transfer and reoxidation, while important for the overall functionality of the complex and cellular energy balance, may not be as directly crucial for the synthesis of BCFAs; hence its impact is lesser when deleted.

The concept of functional redundancy suggests that the role of the E3 component might be compensated by other mechanisms or enzymes within the L. monocytogenes genome. KEGG analysis revealed that lpd has some exceptional paralogs gene (enzymes NADH oxidase; Locus tag: LMOf2365_2268) compared to E1α, E1β, and E2 that do not have any paralogs in L. monocytogenes genome. It is possible that NADH oxidase compensates the E3 component or might work as an alternative route for the reoxidation of FADH₂ to FAD and producing NADH, thereby mitigating the impact of its absence.

The reduced levels of LLO and phospholipase in BKD-deficient strains suggest that the in vivo survival defect of the BKD-deficient strains might be, in part, from lack of full PrfA activation and a corresponding reduction in expression of the PrfA-regulon. This concept is supported by immunoblot results as well as RNA-seq analysis findings that the PrfA regulon was downregulated in BKD-deficient strains in intracellular replication. This finding supports a model in which a decrease in anteiso-BCFA in the membrane reduces both PrfA activation and the secretion of PrfA-regulon gene products (Fig. 5). There are several potential metabolic and physiological cues that can explain the mechanism by which elimination of BKD can affect PrfA-regulon expression and/or secretion. The transcription, translation, and activation of PrfA are directly and indirectly regulated by many metabolic and physiological signals (19, 21, 23, 52–56). Notable PrfA activity is relatively low in the external environment, but when L. monocytogenes enters mammalian cells, PrfA undergoes a conformational transition from an “inactive” to an “active” form (57).

Fig 5.

Model representing the mechanism by which the deletion of BKD and alteration of BCAA catabolism can impact the function of PrfA and PrfA-regulon. (A) Under normal conditions, BKD contributes to BCAA catabolism through BCFA synthesis. (B) In the absence of BKD, rates of BCAA consumption slow down, leading to BCAA accumulation. High levels of BCAA accumulation are likely to have consequences for the transcription of PrfA and PrfA-regulon through CodY and the regulation of cysteine uptake. The figure was created with Biorender.

Low BCAA availability is established as a signal for PrfA and PrfA-regulon function through the global metabolic regulator CodY, a sensor of BCAAs (53, 55). Under low BCAA concentrations, CodY directly binds to the regulatory regions of prfA that promote transcription, leading to the upregulation of virulence genes encoding LLO and ActA (53, 58). Under high BCAA concentrations, CodY retains the ability to repress PrfA. It is possible that the absence of BKD leads to the accumulation of BCAAs, which prevents L. monocytogenes ability to sense and trigger the BCAA starvation response that is required to activate PrfA and PrfA-regulon. In line with this observation, several studies in eukaryotic cells linked defective BCAA catabolism and the consequent alteration in metabolism, including the suppression of glucose metabolism and the disruption of pyruvate utilization (43, 59, 60).

Another metabolic signal that mediates the expression and activity of PrfA is reduced glutathione (GSH) which is produced by bacteria or derived from hosts (23, 61). Upon binding to GSH, PrfA experiences a conformational change that results in its activation. L. monocytogenes strain carrying the active form of PrfA was found to bypass the need for GSH or host-derived signaling molecules for PrfA activation (23, 62). GSH is a tripeptide composed of glutamate, cysteine, and glycine (63). L. monocytogenes is auxotrophic to cysteine (the rate-limiting precursor for GSH) and relies on transporter systems for cysteine uptake. The ABC transporter TcyKLMN was recently shown to be a cysteine importer that supplies cysteine for GSH synthesis (55). This transporter is negatively regulated by BCAA concentrations, connecting BCAAs to cysteine uptake and GSH biosynthesis as a mechanism that controls PrfA. It is also possible that the BKD-deficient strain is defective in cysteine uptake that drives the production of GSH, thereby facilitating PrfA activation (Fig. 5). Together, these findings suggest a model in which BKD-deficient strains are defective in their ability to activate PrfA and the PrfA regulon (plcA, plcB, and hly). Future research will elucidate the cues that influence PrfA activity in BKD-deficient conditions.

In conclusion, our results revealed a novel facet of L. monocytogenes physiology that is influenced by the central virulence regulator PrfA. We found that BKD complex components were required to the maintain growth kinetics and membrane fluidity of L. monocytogenes through BCFA. Defective PrfA regulon function may contribute to the attenuation of BKD-deficient L. monocytogenes strains. Future studies will explore signaling pathways involved in reducing the PrfA-regulon function in the BKD-deficient strains.

MATERIALS AND METHODS

Bacterial strains and culture

Bacterial strains and plasmids used in this study are listed in Table S1. Escherichia coli Top10 (Invitrogen) was grown in Luria-Bertani (LB; Difco Laboratories) broth and agar. L. monocytogenes strains were grown in BHI broth (Difco) or defined MM (64) as required. Murine L2 fibroblast cell lines (CRL-2648; ATCC) and J744A.1 macrophages (TIB-67; ATCC) were grown in Dulbecco’s modified Eagle’s medium (DMEM; ATCC, Manassas, VA) supplemented with 10% fetal bovine serum (Atlanta Biologicals, Norcross, GA) and 1% glutamine. Cultures were maintained at 37°C with 5% CO₂ under humidified conditions. When necessary, erythromycin (10 µg/mL), ampicillin (100 µg/mL), neomycin (50 µg/mL), kanamycin (50 µg/mL), and/or gentamicin (10 µg/mL) were used.

Construction of deletion and complemented strains

The four subunits BKD enzyme complex encoded by lpd (LMOf2365_1388), bkdA1 (LMOf2365_1389), bkdA2 (LMOf2365_1390), and bkdB (LMOf2365_1391) were targeted for in-frame deletion using splice overlap extension and allelic exchange with pHoss1 (28). The sequences of primers used for the construction and validation of mutant strains are listed in Table S1. In brief, approximately 1 kb fragments upstream and downstream of target genes were PCR amplified followed by overlap extension PCR using flanking primers (A and D). After digestion with SalI and SmaI endonucleases, overlap extension fragments were cloned in pHoss1 and transformed into E. coli Top10. The cloned plasmids transformed into L. monocytogenes strain F2365 by electroporation for integration and homologous recombination. PCR analysis and Sanger sequencing were used to further confirm deletions. For complementation assays, a DNA fragment containing the lpd, bkdA1, bkdA2, or bkdB was amplified from the F2365 genome and ligated into pIMK2 integration vector (65). The resulting plasmids were electro-transformed into Δlpd, ΔbkdA1, ΔbkdA2, and ΔbkdB to obtain the complemented strain. Successful complementation was verified by PCR analysis and confirmed by sequencing.

Membrane fatty acid analysis

Bacterial samples for fatty acid methyl esterification (FAME) analysis were performed in GC-MS (29). In brief, 250 mL cultures of each L. monocytogenes strain were grown in MM at 25°C up to the stationary phase. Bacteria pellets were harvested, washed with cold phosphate-buffered saline (PBS), and weighed. About 20–25 mg of bacterial pellets was placed in Thermo Scientific 9 mm Clear Glass Vials with Teflon-lined screw-cap. Tridecanoate methyl ester, 10 N KOH, and methanol were used for saponification at 55°C in a water bath for 90 minutes. After cooling the vial, 24 N H₂SO₄ was added into the vial for direct transesterification at 55°C in the water bath for 90 minutes. The FAME formed during esterification was extracted in hexane and transferred into a 2 mL amber GC vial with a Teflon-lined screw-cap. The fatty acid composition was determined by an Agilent 7890A GC system equipped with an HP-88 capillary column (30 m × 0.25 mm × 0.20 µm), an autosampler, a split/splitless injector, and an Agilent 5975C inert XL MSD with triple-axis mass detector. The FAME was separated in a 20-minute temperature-gradient program with hydrogen as the carrier gas flowing at a constant rate of 1.5 mL/min. The transfer line, ion source, and quadrupole were heated at 250°C, 230°C, and 150°C, respectively. Ionization was performed in an electron impact mode at 70 eV. Ions were detected in a selected ion monitoring mode optimized for saturated, monounsaturated, and polyunsaturated fatty acids. FAME was identified by comparing their retention times, target ions, and ratios of target ions to qualifier ions with those of authentic FAME standards. The fatty acid concentration was calculated by an internal standard calibration method.

Determination of bacterial growth and survival

Growth of BKD-deficient strains in BHI broth and MM was compared at 4°C, 25°C, and 37°C. The wild-type F2365 was included as a control. Strains were grown overnight at 37°C in BHI broth with shaking and then normalized based on the optical density at 600 nm (OD₆₀₀) for each strain. Bacterial strains were washed with PBS and subcultured by diluting cultures 1:200 into fresh BHI broth or MM in a 48-well plate. The plates were incubated on Cytation 5 Cell Imaging Multi-Mode Reader (BioTek) under static growth conditions. The OD₆₀₀ was monitored automatically every hour for 40 hours with moderate shaking (160 rpm) for 15 seconds before reading. The growth assays were conducted in three independent experiments, and each experiment was run with four replicates. Counting of CFU was performed by 10-fold serial dilution from overnight culture followed by spreading in BHI agar and incubation for 24 hours at 37°C. For experiments requiring supplementation with 2MB, L. monocytogenes strains were grown overnight in BHI and then subcultured into BHI supplemented with 2MB (100 µM) and incubated at 4°C. Bacterial viability at the indicated time points was determined by quantifying bacterial CFU/mL.

Detection of phospholipase activity

BLA supplemented with lecithin (Oxoid) was used to detect phospholipase activity of the L. monocytogenes strains as previously described (66). Briefly, 10 µL of normalized overnight cultures grown at 37°C in BHI were placed onto the BLA plate and incubated at 37°C. After 48 hours, white halos surrounding the bacterial colony (zones of opaqueness) were measured to determine phospholipase activity of strains. All experiments were performed three independent times with at least three replicates.

Western blots for LLO protein

LLO protein was detected in the whole bacterial lysate of L. monocytogenes wild-type strain F2365, BKD-deficient strains, and their complemented strains as previously described (67). ∆prfA and NF-L1777 (prfA*) strains used as a negative and positive control, respectively. Briefly, L. monocytogenes strains were grown overnight at 37°C in BHI broth containing 0.1% activated charcoal. Bacterial cultures were all normalized to equivalent densities using OD₆₀₀. The bacterial pellets were harvested by centrifugation, washed with ice-cold PBS, and suspended in bacterial lysis buffer (50 mN Tris HCL pH 8.0, 5% glycerol, 0.5% triton ×100, 2 mM PMSF, and 1.5 mM EDTA) at 4°C for 30 minutes followed by sonication. DNase (20 µg/mL) was added to bacterial suspension followed by incubation with shaking at 4°C for 1 h. Bacterial proteins were precipitated by centrifugation. Proteins were suspended into 4× Laemmli sample buffer with β-mercaptoethanol and heated at 100°C for 10 minutes prior to separation on SDS-PAGE followed by transfer onto PVDF membranes. LLO protein was detected using rabbit anti-LLO (ab43018; Abcam) and horseradish peroxidase (HRP)-conjugated goat anti-rabbit secondary antibodies (as 09602; Agrisera) (54). Housekeeping gene P60 antibody (primary antibody) and HRP-conjugated rabbit anti-mouse secondary antibodies (Dako P0260) were included as a control. Densitometry was measured using ImageJ software.

Plaque assay in murine L2 fibroblasts

Plaque assays using murine L2 fibroblast monolayers were performed (31). Briefly, L2 fibroblasts were grown in DMEM in six-well tissue culture plates until they reached confluence. Confluent monolayers were infected with L. monocytogenes strains at a multiplicity of infection (MOI) of approximately 1–10. One hour post-infection, monolayers were washed with DMEM-containing solid agar overlay, and gentamicin (20 µg/mL) was added to each well. The infected monolayer was incubated for 72 hours at 37°C and 5% CO₂, followed by a second agar overlay containing neutral red (Sigma-Aldrich) to allow visualization of zones of clearing or plaques. The number and diameter of plaques were measured using ImageJ from at least three independent experiments. The average plaque size and number of the wild-type F2365 strain in each experiment was defined as 100%. ∆prfA and NF-L1777 strains used as a negative and positive control, respectively.

Infection of macrophages

Macrophage J774A.1 cells were seeded into a 24-well plate and infected with L. monocytogenes strains at MOI of 1–10 as described (68). After 1 hour of infection, infected monolayers were washed with PBS, and extracellular bacteria were then eliminated by adding DMEM-containing gentamicin (10 µg/mL). After 1-, 3- and 5-hour post-infection, infected cells were washed and lysed, and lysate was spread on BHI agar plates to quantify intracellular bacteria. All experiments were performed with three independent experiments with a minimum of three replicates each. ∆prfA and NF-L1777 (prfA*) strains used as a negative and positive control, respectively.

Virulence in mice

Female Swiss Webster mice (8–10-week-old) from Charles River Labs were randomly housed in cages (5 mice/cage) and kept under specific pathogen-free conditions. All L. monocytogenes strains were grown in BHI at 37°C overnight and then normalized to equivalent cell densities based on OD₆₀₀ nm. Each mouse was injected IV with 2 × 10⁴ CFU of bacterial strains diluted in 200 µL normal saline as previously described (69). Mice infected with L. monocytogenes wild-type F2365, and sterile normal saline were included as controls. Mice were kept under observation with regular food and water. Clinical signs of mice were recorded at 8-hour intervals. To determine bacterial load, mice were euthanized by CO₂ inhalation at 72 hours post-infection. Spleen and liver tissues were aseptically removed, weighed, and homogenized in sterile PBS. Serial dilutions of homogenized tissues were spread on BHI agar plates to determine bacterial concentration in the tissues (CFU/g). For experiments examining the potential rescue of virulence following bacteria growth in the presence of 2 MB, L. monocytogenes strains (ΔbkdA1, wild-type, and complemented) were grown in BHI supplemented with 2MB overnight. The next day, bacterial pellets were harvested, washed, and used to infect Swiss Webster mice (five per group) through IV injection (2 × 10⁴ CFU/mL) (68). At 24- and 72-hour post-infection, livers and spleens were harvested to determine bacterial burdens.

RNA-seq analysis and verification of PrfA-regulon expression by quantitative real-time PCR

A murine macrophage cell line, J774A.1 was infected with L. monocytogenes wild-type F2365 and ΔbkdA1 (three independent biological replicates) at a range of MOIs from 1 to 10 CFU. After 4 hours post-infection, bacterial pellets were collected and used for RNA extraction using a FastRNA SPIN Kit for Microbes with a FastPrep-24 (MP Biomedicals, Santa Ana, CA). On-column DNase treatment with an RNase-Free DNase Set (QIAGEN, Hilden, Germany) was applied to eliminate genomic DNA from total RNA. The quantity and quality of total RNA were analyzed using a NanoDrop ND-1000 spectrophotometer (Thermo Scientific, USA). RNA samples were shipped to Novogene Corporation for quality control, library preparation, and sequencing using the HiSeq platform (Illumina) (31). Bioconductor Edge R was used to identify differentially expressed genes of the wild-type compared to the mutant strain. To validate PrfA-regulon expression and RNA-seq data, quantitative real-time PCR (qRT-PCR) was performed on the same RNAs as those used for transcriptome experiments. cDNA was synthesized from RNA using the SuperScript cDNA synthesis kit (Thermo Scientific). The product of first-strand cDNA synthesis was diluted, and qRT-PCR was performed in a 20 µL reaction containing SYBR Green Real-time PCR master mix (Roche Diagnostic GmbH, Mannheim, Germany). Amplification and detection of specific products were performed with the Mx3000P real-time PCR system (Stratagene). The expression of each gene was normalized against the expression of 16S rRNA. For each gene, triplicate assays were performed, and transcription levels were quantified by 2^-△△CT method (69).

Statistical analysis

All statistical analysis and the dot plots and median values were generated using GraphPad Prism 9.2.0. For the in vivo experiments, we employed the non-parametric Mann-Whitney test to compare bacterial loads in the spleen and liver tissues from infected mice. This test was chosen due to its robustness in situations where the data did not follow a normal distribution and were not homoscedastic, ensuring a reliable comparison of medians across our experimental groups. When comparing the growth of mutant strains to the wild-type, we used the unpaired Brown-Forsythe and Welch ANOVA tests followed by Dunnett’s T3 multiple comparison test. This approach was selected because it does not assume equal variances among groups, making it a more appropriate choice for our data, which exhibited variance heterogeneity. Dunnett’s T3 multiple comparisons test was particularly useful in comparing each mutant with the wild-type while controlling for Type I error across multiple tests. For metrics such as bacterial viability (CFU/mL), phospholipase activity, plaque number, and plaque size, where the data were normally distributed, the unpaired Welch’s t test was utilized. This test is specifically designed to compare the means of two groups and is particularly powerful when the variances may not be equal, which was indicative of our experimental design. The two-way ANOVA with Dunnett’s multiple comparison test was employed to discern the statistical significance between mutants and wild-type strains over different time points during intracellular replication and bacterial viability assessments. This choice was informed by our data structure, which required the analysis of two independent variables (strain type and time), allowing us to observe both individual and interactive effects. The Student’s t test was used to measure the statistical significance between mutants and wild-type in mebrane fatty acid composition. All data were presented as the mean ± SEM. P values of < 0.05 were considered statistically significant in all analyses. * Indicates a significant difference at P < 0.05, ** indicates a significant difference at P < 0.01, and *** indicates a significant difference at P < 0.001.

DATA AVAILABILITY

The data used to support this study are available from the corresponding author upon request.

ETHICS APPROVAL

All animal experiments were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) at Mississippi State University (IACUC 18-508). All methods were carried out in accordance with relevant guidelines and regulations. All methods are reported in accordance with ARRIVE guidelines. Euthanasia was administered by CO₂ inhalation in accordance with the procedures outlined in the American Veterinary Medical Association (AVMA) Panel on Euthanasia.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/jb.00033-24.

Supplemental material. jb.00033-24-s0001.docx.

Table S1; Figures S1 to S5.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.