A single point mutation in the listerial betL σ^A-dependent promoter leads to improved osmo- and chill-tolerance and a morphological shift at elevated osmolarity

Abstract

Betaine uptake in Listeria monocytogenes is mediated by three independent transport systems, the simplest of which in genetic terms is the secondary transporter BetL. Using a random mutagenesis approach, based on the E. coli XL1 Red mutator strain, we identified a single point mutation in a putative promoter region upstream of the BetL coding region which leads to a significant increase in betL transcript levels under osmo- and chill-stress conditions and a concomitant increase in stress tolerance. Furthermore, the mutation appears to counter the heretofore unreported “twisted” cell morphology observed for L. monocytogenes grown at elevated osmolarities in tryptone soy broth.

Introduction

A characteristic feature of the intracellular foodborne pathogen Listeria monocytogenes is its ability to thrive in a variety of stressful environments.^1,2 This phenotypic robustness can be attributed at least in part to the ability of the organism to accumulate a variety of protective compounds, termed compatible solutes, which help to buffer the cell from the detrimental effects of a range of environmental insults.^3-5

Previous work in our labs^6-8 (and others^9,10) led to the identification and characterization of the principal compatible solute uptake/synthesis systems in L. monocytogenes. The first of these loci to be identified, betL, encodes a high affinity (Km 7.9 mM; Vmax 134 nmol/min/mg of protein) sodium-motive force dependent secondary betaine uptake system (BetL),⁶ which is a member of the BCCT family of osmolyte transport systems.³ Detailed in silico analysis revealed the presence of two putative promoter regions; a σ^A-like promoter and a σ^B-dependent promoter, respectively.¹¹ However, while transcriptional control through alternative sigma factors is important, the final yield of BetL protein product is also likely determined by translation efficiency.⁴ In E. coli, and to a lesser extent Bacillus subtilis, for example, the use of non-ATG initiation codons has previously been shown to modulate expression at the translational level.¹² Given that betL is initiated with an alternative TTG start codon, it is likely that the locus is also regulated to some degree at the level of translation.^6,11

Furthermore, in addition to transcriptional and translational control, detailed biochemical analyses revealed that the BetL protein is itself activated in response to changes in osmolarity. Rapid activation of pre-existing BetL protein [half-life (t1/2), 2 min] in response to relatively low NaCl concentrations (1 to 2% NaCl) suggests that BetL is one of the primary responders to rapid fluxes in external osmolarity.¹¹

We outline the use of a random mutagenesis strategy, employing the Epicurian coli^® mutator strain XLI-Red, to screen for mutations in the betL gene which result in improved stress resistance. The deletion of a single thymine residue (from a string of seven thymines), within the spacer region between the -10 and -35 binding sites of the σ^A-like promoter, resulted in a dramatically improved osmo- and chill-tolerance phenotype when expressed against both E. coli MKH13 and L. monocytogenes LO28BCG_SOE backgrounds. Furthermore, we also report for the first time an unusual “twisted-cell” morphology exhibited by L. monocytogenes when grown at elevated osmolarity (> 7% NaCl). Interestingly, this twisted cell morphology is not observed in L. monocytogenes LO28BCG_SOE cells expressing the betL* gene.

Results and Discussion

Shotgun cloning of the L. monocytogenes LO28 genome, followed by heterologous complementation of the compatible solute uptake mutant Escherichia coli MKH13, led to the identification of betL—the first genetic element linked to listerial osmotolerance.⁶ In silico analysis revealed the presence of a consensus σ^B-dependent promoter-binding site downstream of a putative σ^A-like promoter, suggesting that in addition to being regulated at the protein level, betL is also likely regulated at the level of transcription.¹¹ RNA slot blot and reverse transcription analysis proved that this is indeed the case, with the gene showing a 1.6-fold increase in the level of transcription following 15 min exposure to 4% NaCl.¹³ Furthermore, deleting betL from the listerial chromosome resulted in a dramatic reduction in the ability of the mutant to accumulate betaine (~19% that of the wild type), with an associated drop in the growth rate of the mutant at elevated osmolarities.⁶

In the current study, the complete betL gene (under the transcriptional control of its native promoters) was amplified from pCPL1 (using the primer pair betLFPstΙ / betLRXbaΙ) and cloned into pCI372, a shuttle vector capable of replicating in both E. coli and L. monocytogenes. As expected, the resulting construct, designated pRS1, reversed the salt sensitive phenotypes of the osmotically challenged strains E. coli MKH13 and L. monocytogenes LO28BCG_SOE¹⁴ (data not shown).

In an effort to improve BetL mediated osmotolerance, a random mutagenesis strategy was employed to introduce point mutations into the cloned listerial betL gene. Plasmid pRS1 (harbouring betL) was transformed into the E. coli mutator strain XL1-Red. Mutations in three of the primary DNA repair pathways of XL1-Red result in a mutation rate which is ~5,000-fold higher than that of the wild type; hence, pRS1 replication within this strain led to the introduction of random point mutations throughout the plasmid, including the target betL gene. The randomly mutated pRS1 “bank,” designated pRS1^mut, was subsequently transformed into E. coli MKH13, and transformants were selected on LB agar plates containing 7% added NaCl (a salt concentration above the growth limit for MKH13::pRS1). No colonies were obtained following a control transformation with unmutated pRS1, but transformation efficiencies of 55 CFU/μg of DNA were achieved from pRS1^mut; with colonies appearing after 48 h at 37°C. Following overnight growth at elevated osmolarities, plasmids from selected osmotolerant transformants were extracted and the cloned insert sequenced. In each case, the same mutation was observed; i.e. a deletion of one of a string of seven thymines within the spacer region between the -10 and -35 binding sites of the betL σ^A-like promoter (Fig. 1). Proof that the observed phenotype was the result of the single point mutation in the cloned betL gene (as opposed to random mutations in the pCI372 plasmid backbone) was obtained by re-complementation studies, in which the mutated betL (designated betL*) was cloned into a pPL2 backbone, creating pRS2. Further proof that the observed “hyper-osmotolerance” phenotype was the result of the betL* mutation was obtained using the QuikChange^®XL Site-Directed Mutagenesis Kit—replacing the missing thymine reversed the observed phenotype from “hyper-osmotolerance” to normal. Confirmation that the increased osmotolerance phenotype of betL* is the direct result of improved betaine mediated osmoprotection was obtained following growth in defined medium at elevated osmolarity (5% added NaCl) in the presence and absence of betaine. While no growth was observed for any of the E. coli strains in the absence of betaine (data not shown), the strain expressing betL* grew significantly better than the control strains in the presence of betaine (Fig. 2).

Figure 1. DNA sequence upstream of the betL coding region. Inverted repeats, indicated by pairs of arrows, delineate a likely rho-independent transcription terminal signal (ΔG -13 kcal). The predicted ribosome-binding site (RBS) and the putative σ^A and σ^B-dependent -10 and -35 sites are underlined. The first 20 amino acids of the predicted BetL protein are also presented; beginning with the alternative initiation codon TTG.

Figure 2. Growth of E. coli MKH13::pRS2 (▼; betL*), E. coli MKH13::pRS3 (○; betL) and E. coli MKH13::pPL2 (●; negative control) in M9 minimal medium containing 2 mM betaine and 5% added NaCl.

In order to assess the effect of the betL* mutation against the native listerial background, the constructs were transformed into LO28ΔBCG_SOE; a strain which is devoid of betaine uptake.¹⁵ When expressed against the Listeria background, a significant advantage of pPL2 (the pRS2 and pRS3 backbone) over the multi-copy cytoplasmic pCI372 plasmid, is that pPL2 integrates as a single copy in the listerial chromosome (specifically the tRNA^Arg-attBB′ in both serotype 1/2 and 4b strains of Listeria^16,17), thereby removing plasmid copy number as a variable in assessing the true contribution of betL* to L. monocytogenes osmotolerance. Given the location of the thymine deletion, between the -10 and -35 promoter binding sites, it seemed plausible that the increased osmotolerance (Fig. 3A) was the result of increased transcription in betL* relative to the betL wild type. RT-PCR analysis of the mutant and wild type genes suggests that this is indeed the case (Fig. 3B). While no significant difference in transcript levels were observed between betL and betL* at 0% NaCl, at both log and stationary phase growth, expression levels do appear to differ at 4% NaCl, in both growth phases. Similarly, betL* appears to exhibit slightly higher transcript levels under cold stress conditions (particularly at stationary phase Fig. 4B) and a concomitant increase in growth at refrigeration temperatures (Fig. 4A); a finding which is not altogether surprising given that BetL has previously been linked to chill tolerance in L. monocytogenes.¹⁸ Promoter variants in genes encoding compatible solute uptake is not a new phenomenon and has been reported previously in the literature for a number of microorganisms; including E. coli,¹⁹ Salmonella typhimurium²⁰ and B. subtilis.^21,22 While the mechanism of promoter activation in the current study is still unclear, it is tempting to speculate that the deleted thymine may affect DNA topology, thereby facilitating improved promoter binding and activation. In support of this hypothesis, promoters of osmotolerance genes in different organisms have previously been shown to be induced only when DNA is highly supercoiled.^23,24 Interestingly, proU (encoding betaine uptake in E. coli) is controlled by changes in DNA topology²⁵; while the osmoregulated promoter for opuA, encoding a betaine uptake system in B. subtilis, is likely subject to similar control . Indeed, both the osmoregulated proU and opuA promoters deviate from the consensus 17-bp in the length of their -10 and -35 spacer regions; with sub-optimal spacing of 16 and 18-bp respectively^24,26,27—the latter being the predicted distance between the -10 and -35 binding sites of the betL σ^A-like promoter (Fig. 1). Given that RNA polymerase makes specific contacts with both the -10 and -35 regions; the relative orientation of these sequences is likely an important determinant for efficient transcription initiation.²⁸ Promoters with sub-optimal spacer regions, like the betL σ^A-like promoter, are thus likely to respond sensitively to environmentally controlled alterations in DNA topology and as such belong to a special class of DNA twist-sensitive promoters.²⁹ Indeed, Alice and Sanchez-Rivas³⁰ observed a direct link between osmotolerance and DNA supercoiling in B. subtilis, while Grau et al.,³¹ noted similar fluctuations in DNA supercoiling during cold adaptation.

Figure 3.(A) Growth of L. monocytogenes LO28BCG_SOE::pRS2 (▼; betL*), LO28BCG_SOE::pRS3 (○; betL) and LO28BCG_SOE::pPL2 (●; negative control) in TSB with 7% added NaCl. (B) Transcript levels of betL and betL* against the L. monocytogenes LO28BCG_SOE background, following exposure to 4% NaCl for 15 min. Control PCRs were performed with the rrn primers U141 and L142. For each RT-PCR, results are presented after 22 cycles and are representative of three biological replicates.

Figure 4.(A) Growth of L. monocytogenes LO28BCG_SOE::pRS2 (▼; betL*), LO28BCG_SOE::pRS3 (○; betL) and LO28BCG_SOE::pPL2 (●; negative control) in TSB at 4°C. (B) Transcript levels of betL and betL* against the L. monocytogenes LO28BCG_SOE background, following exposure to 10°C for 30 min. Control PCRs were performed with the rrn primers U141 and L142. For each RT-PCR, results are presented after 22 cycles and are representative of three biological replicates.

In addition to dramatically improving the growth of L. monocytogenes at elevated osmolarities, the betL* mutation reverses a previously unreported “twisted-cell” morphology for Listeria grown in complex media (BHI or TSB) at 7% NaCl (Fig. 5). While the existence of elongated listerial cells at elevated osmolarities is not a new phenomenon^32,33; the “twisted-cell” phenotype has not previously been reported. While the exact role of this twisted morphology is unclear, it may function as a survival strategy reducing the cellular surface area exposed to the bathing solution, thereby reducing the severity of the stress. Indeed, bacterial “huddling”—the observed close association of individual bacterial cells—has previously been reported by Corcoran et al.,³⁴ in probiotic lactobacilli subjected to in both spray and freeze-drying. The lack of a twisted morphology for L. monocytogenes LO28BCG_SOE expressing betL* suggests that the cells are less stressed and thus better equipped for growth and survival at elevated osmolarities.

Figure 5. Scanning electron micrograph (SEM) at × 20k magnification of L. monocytogenes LO28BCG_SOE::pRS2 expressing betL* (column 1) and LO28BCG_SOE::pRS3 expressing betL WT (column 2), with no added NaCl (row 1) and 7% added NaCl (row 2). Inset—transmission electron microscopy (TEM) of L. monocytogenes LO28BCG_SOE::pRS2 expressing betL* (column 1) and LO28BCG_SOE::pRS3 expressing betL WT (column 2), with 7% added NaCl (row 2)

Finally, the betL* mutation in L. monocytogenes represents a double edged sword; from a food safety perspective, a single point mutation with the potential to induce such dramatic shifts in cell growth and survival at low a_w and temperatures—making an already dangerous pathogen even more formidable—raises significant food-safety concerns which need to be addressed.^1,35 However, from a synthetic biology point of view, a boosted-stress resistance locus, such as betL*, represents a useful BioBrick³⁶ for the design of more physiologically robust pharmabiotic strains.^37-42

Materials and Methods

Bacterial strains, plasmids and culture conditions

Bacterial strains and plasmids used in this study are listed in Table 1. E. coli MKH13 was grown at 37°C in either LB medium or M9 minimal medium (GIBCO/BRL, Eggenstein) containing 0.5% glucose, 0.04% arginine, 0.04% isoleucine and 0.04% valine. Glycine betaine (Sigma) was added to M9 as a filter-sterilised solution to a final concentration of 1 mM.. L. monocytogenes strains were grown in brain heart infusion (BHI) broth or in tryptone soy broth (TSB) supplemented with 0.6% yeast extract (Sigma Chemical Co.) or in chemically defined minimal medium⁴⁴ (DM). Antibiotics, when needed, were made up as concentrated stocks and added to media at the required levels. Where indicated, media osmolarity was adjusted by the addition of NaCl.

Table 1. Bacterial strains and plasmids.

Strain or plasmid	Relevant genotype or characteristicsa	Source or reference
Strains
L. monocytogenes LO28	Serotype 1/2c	P. Cossart, Institut Pasteur
LO28ΔBCGSOE	LO28 ΔbetL, ΔopuC, Δgbu	Wemekamp-Kamphuis et al.,15
E. coli MKH13	MC4100D(putPA)101D(proP)2D(proU)	Haardt et al.,43
XL1-Red	endA1 gyrA96 thi-1 hsdR17 supE44 relA1 lac mutD5 mutS mutT Tn10 (Tetr)	Stratagene
Plasmids
pUC18	ColE1 ori, Apr	Vieira and Messing 43
pCPL1	pUC18 containing 2.5 kb of L. monocytogenes genomic DNA	Sleator et al.,6
pCI372	E. coli/L. lactis shuttle vector, 5.7 kb, Cmr	Sleator et al.,6
pPL2	Site-specific listerial integrative vector, 6.1 kb, Cmr	Lauer et al.,16,17
pRS1	pCI372::betL	This study
pRS1mut	Randomly mutated pRS1 from E. coli XL1-Red	This study
pRS2	pPL2::betL*	This study
pRS3	pPL2::betL	This study

^a Ap^r ampicillin resistance, Cm^r chloramphenicol resistance, Tet^r tetracycline resistance.

DNA manipulations and sequence analysis

Plasmid DNA was isolated using the Qiagen QIAprep Spin Miniprep Kit (Qiagen). E. coli was transformed by standard methods while electro-transformation of L. monocytogenes was achieved by the protocol outlined by Park and Stewart.⁴⁵ Polymerase chain reaction (PCR) reagents (Taq polymerase and deoxynucleoside triphosphates dNTPs) were purchased from Boehringer GmbH and used according to the manufacturer’s instructions with a Hybaid PCR express system. Oligonucleotide primers, listed in Table 2, were synthesized on a Beckman oligo 1000M DNA synthesizer (Beckman Instruments Inc.). Nucleotide sequence determination was performed on an ABI 373 automated sequencer using the BigDye™ Terminator sequence kit (Lark Technologies, Inc.). Nucleotide sequence analysis was performed using BioMapper (nSilico Lifescience Ltd.).

Table 2. PCR primers used in this study.

Primer	Sequence (5′-3′)
betLFPstΙ	CATCTGCAGGCTTTCTCCCCCTTTTTCCTC
betLRXbaΙ	CATTCTAGAGCTCTATTCCAATTACCGCCATTTC
XbaIKO	TAAGCGCCACTCTAGACC
EcoRIKO	GCACGAATTCACCAAGTA
U141	TTGCTCTTCCAATGTTAG
L142	GAGTGCTTAATGCGTTAG

^aNucleotides introduced to create restriction sites are underlined

Generation and screening of betL*

Random mutagenesis was performed using the strategy outlined previously.⁴⁶ Plasmid pRS1 harbouring the listerial betL gene was transformed into the mutator strain Epicurian coli^® XL1-Red (Stratagene) and transformants were selected on LB plates containing chloramphenicol (30 μg/ml). Transformants were then pooled and grown overnight at 37°C in LB broth. Randomly mutated plasmid DNA extracted from this culture was then used to transform the osmolyte uptake mutant E. coli MKH13. Mutations leading to enhanced osmotolerance were selected by plating transformants on LB medium containing 7% added NaCl (a salt concentration which does not permit the growth of MKH13 expressing pRS1). Plasmids isolated from the resultant osmotolerant MKH13 clones were then used to transform L. monocytogenes LO28BCG_SOE (Δbet, ΔopuC, Δgbu).

Site directed mutagenesis

Single nucleotide additions and deletions within the putative betL σ^A-like promoter region were achieved using the QuikChange^®XL Site-Directed Mutagenesis Kit (Stratagene) in accordance with the manufacturer’s instructions.

Transcriptional analysis

Reverse transcriptase (RT)-PCR analysis was performed as previously described.¹³ Essentially L. monocytogenes cells were grown, at 37°C with shaking, to mid-exponential phase in BHI. Ten ml of culture were centrifuged and re-suspended in 1 ml of BHI with 4% added NaCl, for analysis of salt stress, and BHI pre-chilled to 10°C for chill stress. BHI at 37°C with no added salt was used as a control. After 15 and 30 min incubation, cells were harvested by centrifugation and flash-frozen at -80°C with liquid nitrogen. Total RNA was extracted using the hot acid phenol procedure described by Ripio et al.,⁴⁷ and cDNA was synthesized by adding 1 μg of total RNA to 4 μl of 5× RT buffer (Roche), 2 μl of 100 mM dithiothreitol, 0.5 μl of a deoxynucleoside triphosphate mix, 0.25 μl of RNasin, 100 ng of the random primer p(dN)₆ and 1 μl of Expand reverse transcriptase (Roche). The reaction mixture was incubated at 42°C for 9 h.

In all cases control PCR reactions were used to ensure the complete removal of DNA from RNA preparations prior to reverse transcription and to ensure that levels of cDNA for samples to be compared were equal. Oligonucleotide primer pairs XbaIKO / EcoRIKO and U141 / L142 were used for PCR amplification of a 551-bp fragment from the center of betL and a 806 bp rrnA DNA fragment respectively. Each (RT)-PCR reaction was performed in triplicate from three biological replicates.

All glass and plastic-ware used in RNA analysis was first treated with 2% sodium dodecyl sulfate (SDS) for 15 min, before rinsing with 1:10 in diethyl-pyrocarbonate (DEPC) treated water.

Microscopy

Microscopy was performed essentially as described by Considine et al.⁴⁸ Safranin (0.25%) stained samples of individual strains were viewed using Bright-field light microscopy at 1000 × magnification to determine cell morphology. For scanning electron microscopy (SEM), overnight cultures in TSB ± 7% added NaCl (5 mL in a 30 mL round-bottomed tube) were centrifuged at 5,000 rpm × 10 min. Pellets were immediately re-suspended in 2% glutaraldehyde, 2.5% paraformaldehyde in 0.165M phosphate buffer (pH 7.3) and left to fix overnight at 4°C. After fixation, the cells were subjected to three 10 min washes with 0.165M phosphate buffer. The cells were post-fixed with 2% osmium tetroxide for 2 h before being washed again in buffer. Cells were dehydrated by to successive 25%, 50%, 75% and 100% (100% x four times) solutions of acetone for 10 min, before being transferred to tetramethylsilane (TMS) for 15 min and air-dried in a fume hood. The samples were mounted on metal stubs using double-sided carbon tape and sputter coated with ~28 nm layer of gold using a BioRAD Polaron Sputter Coating System. The samples were viewed with a Joel JSM-5510 scanning electron microscope. For transmission electron microscopy (TEM), overnight cultures were fixed and post-fixed as for the SEM preparation. Following a buffer rinse, the cells were dehydrated in a series of solutions of ethanol: 10% for 15 min (three times), 50% for 30 min, 100% for 15 min (three times) and finally in epoxy propane for 30 min. The dehydrated cells were infiltrated with an epoxypropane:araldite mixture (3:1) overnight and then embedded in araldite and cured for 48 h in molds at 50°C. Thin (70 nm) sections were obtained from the polymerized blocks using a Reichert-Jung Ultracut E ultramicrotome and collected on formvar-coated copper grids (100 mesh). The sections were stained with 2% uranyl acetate and Reynolds lead citrate stain and examined using a Joel JEM-2000FXII electron microscope. Electron micrographs were obtained for areas of interest with a Megaview-III digital camera and analySIS software.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.