Food Microbiology 48 (2015) 72e82
Contents lists available at ScienceDirect
Food Microbiology journal homepage: www.elsevier.com/locate/fm
A novel gene, lstC, of Listeria monocytogenes is implicated in high salt tolerance Laurel S. Burall, Alexandra C. Simpson, Luoth Chou, Pongpan Laksanalamai, Atin R. Datta* Center for Food Safety and Applied Nutrition, U.S. Food and Drug Administration, Laurel, MD 20708, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Received 14 July 2014 Received in revised form 1 December 2014 Accepted 16 December 2014 Available online 24 December 2014
Listeria monocytogenes, causative agent of human listeriosis, has been isolated from a wide variety of foods including deli meats, soft cheeses, cantaloupes, sprouts and canned mushrooms. Standard control measures for restricting microbial growth such as refrigeration and high salt are often inadequate as L. monocytogenes grows quite well in these environments. In an effort to better understand the genetic and physiological basis by which L. monocytogenes circumvents these controls, a transposon library of L. monocytogenes was screened for changes in their ability to grow in 7% NaCl and/ or at 5 C. This work identified a transposon insertion upstream of an operon, here named lstABC, that led to a reduction in growth in 7% NaCl. In-frame deletion studies identified lstC which codes for a GNAT-acetyltransferase being responsible for the phenotype. Transcriptomic and RT-PCR analyses identified nine genes that were upregulated in the presence of high salt in the DlstC mutant. Further analysis of lstC and the genes affected by DlstC is needed to understand LstC's role in salt tolerance. Published by Elsevier Ltd.
Keywords: Listeria monocytogenes Salt growth Stress adaptation Food preservation
1. Introduction Listeria monocytogenes is a foodborne pathogen that causes invasive listeriosis and gastroenteritis. Invasive listeriosis typically affects the elderly, pregnant women and immune-compromised individuals and causes meningitis, encephalitis, miscarriage and stillbirth (Datta, 2003). These infections are associated with a high mortality rate of approximately 20% making L. monocytogenes responsible for many of the foodborne fatalities, despite a relatively low incidence of disease (Silk et al., 2012). Among pregnant Hispanic women, nearly a third of these infections lead to fetal loss or neonatal death (Silk et al., 2012). Additionally, these infections have a high percentage of hospitalizations and protracted recoveries due to the neurological damage associated with them (Havelaar et al., 2012). L. monocytogenes has also been shown to cause gastroenteritis characterized by nausea, abdominal cramping and diarrhea, though the prevalence of these infections is unknown as gastroenteritis cases are not routinely screened for L. monocytogenes (Barbuddhe and Chakraborty, 2009; Norton and Braden, 2007). Listeria contamination is a problem in foods considered readyto-eat as they are not subjected to a final kill step prior to
* Corresponding author. Tel.: þ1 240 402 3422; fax: þ1 301 210 7975. E-mail address: [email protected] (A.R. Datta). http://dx.doi.org/10.1016/j.fm.2014.12.008 0740-0020/Published by Elsevier Ltd.
consumption. These ready-to-eat (RTE) foods include soft cheeses and deli meats, as well as fresh produce including salad greens, sprouts and cantaloupe. RTE foods with long shelf-life, high water activity and neutral pH are at greater risk for becoming a vehicle of foodborne listeriosis because of the high growth of L. monocytogenes. The standard practices for controlling microbial growth in such foods, e.g. refrigeration and high salt, are not adequate as Listeria survive and grow relatively well under such conditions (Lado and Yousef, 2007). Listeria has been isolated from various foods containing high levels of NaCl and several studies have indicated extended survival of Listeria in high salt concentrations. However, under experimental conditions, Listeria growth appears to be limited to 7e9% of NaCl (Lado and Yousef, 2007). The ability to survive and grow in foods containing high salt is further complicated by the presence of naturally occurring compatible solutes, also known as osmoprotectants, e.g. glycine betaine, carnitine and glycerol in many foods. Although it is known that high salt reduces microbial growth by reducing water activity, changing electrochemical potential across cell membrane, a thorough understanding of the genetics and underlying adaptation mechanisms is required for development of effective control measures. Several studies have identified genes that affect Listeria's ability to grow at low temperature and high salt environments through modification of compatible solute transport, membrane lipid composition alteration and changes in ribosomal protein
L.S. Burall et al. / Food Microbiology 48 (2015) 72e82
(Laksanalamai et al., 2010). Previous studies have already shown the applicability of transposon mutagenesis to aid in the identification of genes associated with Listeria's ability to grow in what would normally be considered restrictive conditions (Burall et al., 2012; Chassaing and Auvray, 2007; Ells et al., 2009; Madeo et al., 2012). Previous work in our laboratory identified a role for iap in salt and cold tolerance and verified a previously established role for the gbu operon in cold and salt tolerance (Angelidis and Smith, 2003a, 2003b; Burall et al., 2012; Ko and Smith, 1999). In this study, we applied the same techniques used before to screen for novel genes implicated in the cold and salt tolerance phenotypes of L. monocytogenes. This work identifies and characterizes the phenotype and expression alterations associated with the deletion of a previously unknown gene, encoding a putative acetyltransferase, with a role in salt tolerance. 2. Materials & methods 2.1. Strains, media and reagents
2.2. Generation of transposon mutants Competent LS402 was electroporated with pLTV3 as described previously (Burall et al., 2012; Park and Stewart, 1990). Transformants were selected on BHI agar containing tetracycline (12.5 mg/mL), erythromycin (1 mg/mL) and lincomycin (25 mg/mL). The presence of pLTV3 in the transformants was confirmed by isolation of the plasmid from Listeria using the QiaPrep miniprep plasmid isolation kit (Qiagen, Valencia, CA) with the addition of 10 mg/mL lysozyme (SigmaeAldrich, St. Louis, MO) in P1 buffer and incubation of the resuspended cells for 10 min at room temperature before proceeding with the manufacturer's protocol. The plasmid recovered from L. monocytogenes was verified via restriction digest with BamHI. A library of LS402 Tn917 mutants was then generated using the procedures described previously (Burall et al., 2012). The library, consisting of approximately 5200 TetSEryRLinR isolates, was assembled in 96-well microtiter plates into groups of 60 isolates each for screening purposes. 2.3. Screening of transposon mutants
L. monocytogenes LS402 is a serotype 4b strain that was isolated during the Italian corn salad gastroenteritis outbreak (Aureli et al., 2000). L. monocytogenes LS411 is a serotype 4b food strain from the 1985 Los Angeles Jalisco cheese invasive listeriosis outbreak (Linnan et al., 1988). These two strains served as parent strains for the various mutants generated in this study. A full list of L. monocytogenes strains used in this study is provided in Table 1. E. coli HB101 was used to maintain pLTV3 and its culture conditions are described elsewhere (Burall et al., 2012; Camilli et al., 1990). As previously described (Burall et al., 2012), L. monocytogenes strains were grown in brain heart infusion (BHI) broth or BHI agar. Salt growth was assessed in BHI broth or BHI agar supplemented with 7% NaCl. Antibiotics used for isolation, characterization and maintenance of the mutants and antibiotic usage during spectrophotometric screens are described elsewhere (Burall et al., 2012). E. coli TOP10 (Invitrogen, Carlsbad, CA) was used for maintenance of pKSV7 (Smith and Youngman, 1992) with chloramphenicol (10 ug/ mL) used for maintenance and selection of it and its derived constructs. Antibiotics were obtained from SigmaeAldrich (St. Louis, MO). All media were obtained from BD (Franklin Lakes, NJ). Restriction enzymes and DNA ligase were purchased from New England Biolabs (Ipswich, MA), and used as per the manufacturer's protocols.
Table 1 Listeria monocytogenes strains used in this study. Strain Genetic background
Source
LS402 LS411 LS401 LS661 LS667 LS654 LS592 LS593 LS652 LS653 LS725 LS723 LS832 LS824 LS950 LS951 LS952 LS953
Martin Wiedmann Martin Wiedmann Lab collection Lab collection Lab collection This study This study This study This study This study This study This study This study This study This study This study This study This study
Wildtype, 1997 Italian Corn Salad Outbreak, 4b Wildtype, 1985 Jalisco Cheese Outbreak, 4b Wildtype, EGDe, serotype 1/2a Wildtype, 2010 Chicken Salad Outbreak, 1/2a Wildtype, 2011 Cantaloupe Outbreak, serotype 1/2b LS402 Tn917:LMOf2365_2170-lstA intergenic LS402 DlstA LS411 DlstA LS402 DlstABC LS411 DlstABC LS402 DlstC LS411 DlstC LS401 DlstC LS667 DlstC LS661 DlstC LS402 pIMK2 LS725 pIMK2 LS725 pIMK2-lstC
73
The ability of the mutants to grow in three different conditions was examined in both BHI broth or agar cultures. Growth was assessed using a Spectramax M2 (Molecular Devices, Downingtown, PA) and a Growth Curves Bioscreen-C (Growth Curves USA, Piscataway, NJ) measuring absorbance at 600 nm for BHI broth cultures or visual assessment for BHI agar cultures. The growth of each mutant was assessed individually but analyzed against the averaged results of all the prospective mutants. The three conditions were unmodified BHI broth at 37 C (control), unmodified BHI broth at 5 C (cold), and BHI broth supplemented with 7% NaCl at 37 C (salt). A starter plate containing 60 isolates, each in its own respective well, was used to inoculate all of the trial conditions as described elsewhere (Burall et al., 2012). Growth on agar plates was assessed daily while growth in the microtiter plates was assessed using absorbance at 600 nm at intervals previously used (Burall et al., 2012). Mutants identified as having altered growth in either of the stress conditions with no alteration in nonstress conditions were further tested (secondary screen) for their altered phenotypes. These secondary screens allowed the use of replicate cultures of each mutant to verify reproducibility. Putative cold-altered mutants were assessed using the Bioscreen-C set at 5 C from 24 h to 120 h post inoculation. A third screen was performed in quadruplicate without antibiotics on any putative mutants still showing altered growth during secondary screens, using LS402 as the control strain. Cold-growth was measured for at least eight days to allow entry into stationary phase to determine maximum growth levels. The absorbance measurements were used to determine exponential growth rate, maximum absorbance and, where possible, the length of lag and exponential phases. Any outliers, defined as ±1 standard deviation from the average, were selected for further screening. In the third screening, individual strains were compared to the averaged results for LS402 and outliers were defined as before. Growth data was also plotted to visually compare alterations more readily observed in a growth curve which allowed more nuanced analysis of the growth, for example a variable growth rate that was not observable in a single averaged number. 2.4. Analysis of growth phenotypes Candidate Tn917 mutants, consistently showing altered growth either in 7% salt or at 5 C, and subsequent deletion mutants were assessed via colony forming units (CFU)/mL in both the cold and salt conditions to confirm the changes in their growth phenotype.
74
L.S. Burall et al. / Food Microbiology 48 (2015) 72e82
Table 2 Primers used in this investigation. Primer
Table 2 (continued )
Sequence
Tn917 insertion site identification Tn917L1 ACGGTTGAAAACTGTACC Tn917L2 GATACAAATTCCTCGTAG Tn917F GGAGCATATCACTTTTCTTGGAGAG qPCR of Tn917 mutant 37G5: lstA forward GTGGAGGAACTTTGCGACTC lstA reverse GAGTTCTGCTGCACTATGGC lstB forward TCCCGAATGTCGAAACTGTA lstB reverse TGTCGGAATATACGAACCGA lstC forward ACCAATAAACACGTCGCAGA lstC reverse TAGAACACGCAAGGCAATTT lmof2365_2170 forward GGTTGTTGGCACTGAAGAAA lmof2365_2170 reverse ATTTGCCGCACAATAAACTG SOE Deletion lstA: a SOE-A CTACTGCAGAACCACGCCCCTTTTCAG SOE-B GAGTATTTTTCCTGGTTCTACTCGA SOE-C TCGAGTAGAACCAGGAAAAATACTCGG TGTTTTAAATGTTCAAACCCAC a SOE-D TAAGAATTCCAAATCCGTCTATCTTACCGTCT XF ACCCAACATAATAATCTTCATACAGC XR CGGCGCTCATCACATCTT lstABC: a SOE-A AGCTGCAGTGATATCCTTTCTTCCAGTTGTCTC SOE-B GCTTACTTCTTTGCCGTCCC SOE-C GGGACGGCAAAGAAGTAAGCTTCAT TAAAATTGCCTTGCGT a SOE-D AAGAGCTCCCCGTAACTTGCTTAAATAGTGG XF ATTTTTTCACGTCTTTCTGTTGG XR GGCGAATTCTTTGTCTTGCTAC lstC: a SOE-A GTTCTAGAGTCCCGAATGTCGAAACT a SOE-A2 GTTCTAGAGTCCCGAATGTGGAAACTGTGG SOE-B TGGCGTCCAAACTAAATGTTCT SOE-B2 TGGTGTCCAAACTAAATGTTCT SOE-C AGAACATTTAGTTTGGACGCCAAT TCAAGGCGAATTTGTGGAT SOE-C2 AGAACATTTAGTTTGGACACCAATT CAAGGCGAATTTGTTGAT a SOE-D TGAAGCTTCGTAACTTGCTTAAATAG XF TCCCAGCGTACTACCTTGTGC XF2 TCCCAGCCTACTACCTTGTGC XR ACGGTGAGGCATTTTCAAACA SOE lstC Complementation: a SOE-A TTCTGCAGGTCTTTCTGTTGGTGTATAGCCA SOE-B GTATTTTTCCTGGTTCTACTCGA SOE-C TCGAGTAGAACCAGGAAAAATACG GAAAACCTCGCTATCTTACG a SOE-D GTGTCGACCGCATCACCTTTTATTTGTTCGT RT-qPCR of LS723: LMOf2365_0002_F TTTGTTATTGAGCGTGATCGTCTTGT LMOf2365_0002_R GGATGCTTGGCCAGAACTAATG LMOf2365_0008_F CAAGTGATTTACGGCGGACCATA LMOf2365_0008_R ATACATTTTCGCGCCTGCTTCAA LMOf2365_0238_F GTTACGAAGATCAAGCGGCATTTTTA LMOf2365_0238_R GATGTGGGATTTTCAGCTTTTCTGTATC LMOf2365_0841_F CGCCCATACGATTCGCTATTAT LMOf2365_0841_R CCGGCGGTACATATCTTGTTG LMOf2365_0871_F GTAGCAGGTTTATGCGAAATGGTC LMOf2365_0871_R CGCCGGCAATAATCATCGTAAT LMOf2365_0993_F AACAGTATAACGCGTGGGATTTTG LMOf2365_0993_R AGGTGTTTGAACGCCGAGTAAGTA LMOf2365_1105_F TATGGCACATTTGATTTGATTCAC LMOf2365_1105_R CGTGGTAAATAAACGACTTCACA LMOf2365_1427_F TGGTTATGAACTCGCAAACAGG LMOf2365_1427_R TCATCTCGTGTGACAGGGATTTT LMOf2365_1506_F AACCGCTTTGCTACATGATTATGC LMOf2365_1506_R ATATACGCCGGGAAATGTTCTTC LMOf2365_1563_F AGATGTGGTAGAAGGTCGCTGTAA LMOf2365_1563_R TGGTGGCTCAATATCTTCTTTTTC LMOf2365_2067_F GTGGTGGCGGTACTGGTGGAC LMOf2365_2067_R GCGGCTGCATAAACGACTGGA LMOf2365_2238_F ATCTTGGCGCTTGAACGAACG
Primer
Sequence
LMOf2365_2238_R LMOf2365_2303_F LMOf2365_2303_R LMOf2365_2304_F LMOf2365_2304_R LMOf2365_2398_F LMOf2365_2398_R
ATGCGCCGCGATGACTACTC AAGCTATGAGGGACGCAAAGAG CTAGGTTCAGCAGCAAGAGTGGT CTGCATCCACAATAAGCAAGTATGAA CAATCGCATCTGCAATTCGTTTTA AAGCCATGTGGTTTGTAGTGTCG TCGGGTTGCAGCATGATAAGAATA
a Underlined text indicates restriction sites introduced into the primer to facilitate cloning.
Growth trials were performed as in the previous study (Burall et al., 2012). Bacterial counts at each time point were compared and statistical analyses were performed using InStat v3 using a standard ANOVA and a Tukey post test (GraphPad, La Jolla, CA). Mutants were screened by colony count to verify normal growth in BHI broth at 37 C for 24 h. Additionally, LS411 mutants were tested for growth in BHI broth supplemented with 7% NaCl at 5 C (cold-salt) with counts determined at defined intervals for 1e2 weeks, depending on observed growth rate. The parent strain of the respective mutant, either LS402 or LS411, served as the controls in all experiments and all strains were tested in replicates of six in at least two independent experiments. 2.5. Identification of Tn917 insertion site Single primer PCR was used to identify the flanking sequences largely as previously described (Karlyshev et al., 2000). Three primers, Tn917L1 (annealing, 48 C), Tn917L2 (annealing, 43 C), and Tn917F (annealing, 54 C) (Table 2) were used to amplify sequences present at either end of the transposon (Liu et al., 2006). DNA was isolated using DNeasy blood and tissue kit (Qiagen, Valencia, CA) using the manufacturer's protocol for Gram positive bacteria with one hour incubations in the enzymatic lysis buffer and in proteinase K. The PCR reaction was performed as described in an earlier study (Burall et al., 2012). The PCR products were then run on a 1% TBE-agarose gel and dominant bands were excised and purified using the QiaQuick Gel Extraction Kit (Qiagen). The product was cloned into pCR2.1 using Invitrogen's TOPO cloning kit and transformed into DH5a Electromax (Invitrogen). Plasmids with the expected insert size were sent for sequencing (Macrogen USA Corporation, Rockville, MD) using T7 and M13 primers. The sequence was then analyzed using BLAST to identify the Tn917:F2365 genome junction (Altschul et al., 1990; Nelson et al., 2004). The F2365 sequence was then compared to the genome annotation via BLAST and the Comprehensive Microbial Resource (CMR) to identify the location of the insertion site (Altschul et al., 1990; Peterson et al., 2001). 2.6. RT-qPCR for gene expression analysis The parent and mutant strains were grown in BHI broth at 37 C to mid-log phase (Abs600 ¼ 0.045e0.055), as determined by absorbance. The bacterial cells were collected and RNA was isolated using the Ambion RiboPure™ RNA extraction kit (Invitrogen, San Diego, CA) following the manufacturer's protocol but using a 20 min cell disruption step. RNA preparations were validated for the absence of DNA contamination using a BioAnalyzer 2100 with the RNA 6000 Nano Kit (Agilent, Santa Clara, CA). In those instances where DNA was present, the sample was treated with DNase I (Ambion/Life Technologies, Grand Island, NY) according to the manufacturer's protocol. Primers were designed to the target genes to yield an amplicon between 90 and 120 bp (Table 2). Gene
L.S. Burall et al. / Food Microbiology 48 (2015) 72e82
expression was analyzed by quantitative RT-PCR using the Bio-Rad iQ5 system. The resulting amplicons were generated and detected using iScript™ One-Step RT-PCR Kit with SYBR Green (Bio-Rad, Hercules, CA). Independent duplicate trials were performed and gene expression was measured in triplicate for each RNA sample. The Ct value was determined and averaged, and the average Ct value was used to determine the fold change in expression, using gap as the reference gene.
75
Technical Manual (Affymetrix, 2009). 4 mg of the cDNA were fragmented by 2.4 unit DNaseI (Promega, Madison, WI) at 37 C for 10 min in a 20 ml reaction containing 1X One-Phore-All buffer (GE Healthcare, Waukesha, WI), followed by heat-inactivation at 95 C for 10 min. The fragmented cDNA was 30 -end labeled with 1 nmole of biotin-11-ddATP (ABI) and 30 units of terminal transferase (Promega, Madison, WI) at 37 C for 4 h. The labeled product was then used for hybridization onto the Listeria GeneChip® (Laksanalamai et al., 2012).
2.7. Creation of in-frame deletion mutants The results from the RT-qPCR analysis identifying which genes were affected by the Tn917 insertion, were compared with annotation analyses (Peterson et al., 2001) to identify which genes were likely to be implicated in the observed phenotype. Based on these results, three deletion mutant constructs were made using splicing by over extension (SOE) deletion. These mutations were made in LS402 and LS411, to verify which gene was actually implicated in the observed phenotypes (Aureli et al., 2000; Linnan et al., 1988). Primers for SOE deletion were designed to develop two ~500 bp amplicons (AeB and CeD) at the 50 and 30 ends of the targeted genes (Chan et al., 2008; Wiedmann et al., 1998) (Table 2). The overlapping primer sequence was then used to amplify the ~1000 bp hybrid product, which lacked the intervening sequence. This product was cloned into the temperature sensitive plasmid, pKSV7, using the indicated restriction sites (Table 2) and the initial plasmid ligations were transformed into E. coli TOP10 cells and verified before being transformed into LS402, LS411 and LS667, as described earlier (Burall et al., 2012). Slight differences in sequence appeared to affect the ability of the original pKSV7-DlstC construct to be used in the 1/2a strains so a second set primers (Table 2) and new plasmid constructs were used for amplification of LS401 and LS661 sequences to account for sequence differences when compared to the F2365 genome. For the generation of both 1/2a mutants, the second set of primers was used to amplify the required sequences from each respective parent strain to make the required pKSV7 deletion construct. The ligation products were transformed into E. coli TOP10 cells and verified before being transformed into LS401 and LS661, as previously described. For all of the transformation reactions, the CamR transformants were then passed several times at 40e42 C with chloramphenicol (10 mg/mL) to select cells with the plasmid integrated into the chromosome. Further passage without antibiotic selection at 30 C led to the isolation of cells with the in-frame deletions without the plasmid backbone. The deletion event was verified using the XF and XR primer set developed for that mutant, which flanked the two initial amplicons (AeB and CeD), using a crude lysis method, as described elsewhere (Burall et al., 2011).
2.9. Microarray hybridization, washing, staining and scanning Hybridizations of the cDNA were performed according to the Affymetrix GeneChip® Expression Analysis Technical Manual (Affymetrix, 2009). Hybridization reactions containing 4 mg of labeled fragmented cDNA, 100 mM MES, 1M Naþ, 20 mM EDTA, 0.01% Tween-20, 50pM control oligoB2 (Affymetrix, Santa Clara, CA), 0.1 mg/mL herring sperm DNA (Promega), 7.8% dimethylsulfoxide (DMSO) (Sigma, St. Louis, MO), were heated at 95 C for 1 min followed by incubation at 45 C for 5 min, prior to hybridizing onto the Affymetrix Listeria GeneChip® at 45 C with rotation (60 rpm) for 16 h in a hybridization oven. The wash and staining procedures were carried out on an Affymetrix FS-450 fluidics station using the mini_prok2v1_450 fluidics script as described by GeneChip® Expression Analysis Technical Manual (Affymetrix, 2009). Arrays were scanned using a GeneChip® Scanner 3000 7G with GCOS v1.4 software. 2.10. Parsing CEL files, probe set summarization methods and data analysis tools All Affymetrix CEL files generated in this study were parsed and analyzed using Robust Multi Array (RMA) algorithm for summarized probe-set intensities implemented by the Affy package of R and Bioconductor (Bolstad et al., 2003; Gentleman et al., 2004; Irizarry et al., 2003, 2006). 2.11. Complementation of lstC SOE primers were designed to amplify the sequences upstream of lstA and coding sequences of lstC while deleting the majority of lstAB from the final construct. The restriction sites inserted via the SOE-A and SOE-D primers were used to clone the construct into pIMK2 (Monk et al., 2008). The resulting plasmid was cloned into E. coli for isolation and verification and then transformed into LS725. Empty vector controls were also transformed into LS402 and LS725. 3. Results & discussion
2.8. Preparation of cDNA for microarray analysis 3.1. Screening of the Tn917 mutant library L. monocytogenes strains LS411 and LS723 were grown in brain heart infusion (BHI) broth at 37 C to use as a starter culture. The starter cultures were diluted to absorbance at 600 nm of 0.1 and were inoculated into 15 ml fresh BHI broth or BHI broth with 7% NaCl. The cultures were incubated at 37 C in a shaking incubator at 170 rpm until they reach an absorbance at 600 nm of 0.5 ± 0.05. The cultures were collected by centrifugation at 4 C for 10 min and immediately processed for RNA extraction using the Ambion RiboPure Bacterial kit (Invitrogen, San Diego, CA), following the manufacturer's protocol with the exception of a 20 min cell disruption step. The quality of the RNA preparation was verified as described for RT-qPCR RNA preparations. Complementary DNA was generated from 10 mg of total RNA according to the Affymetrix GeneChip Expression Analysis
Primary screening of the LS402 Tn917 library (n ¼ 5220) identified 1005 salt-altered mutants and 935 cold-altered mutants that were subjected to a secondary screen that pared these numbers to 93 salt-altered mutants and 98 cold-altered mutants. These mutants were tested in a third round of screening that ultimately identified ten salt and ten cold mutants for verification by colony count analysis. With one exception, none of the prospective candidate mutants were confirmed by colony count. A single saltattenuated Tn917 mutant, 37G5, showed no alteration in growth relative to LS402 in BHI at 37 C; however, it showed significantly lower growth in BHI with 7% NaCl at 37 C as compared to the parent strain LS402 at 24 and 32 h (Fig. 1). Growth studies at 5 C indicated that the mutant has no cold growth defect when
76
L.S. Burall et al. / Food Microbiology 48 (2015) 72e82
Fig. 1. Growth of the Tn917 mutant, 37G5 (◊), relative to its parent strain, LS402(), at 37 C in unsupplemented BHI (A), BHI with 7% NaCl (B) and at 5 C in unsupplemented BHI (C). No reduction in growth is observed in regular BHI at 37 C or at 5 C but a ten-fold significant reduction in final population density (*, P < 0.001) is observed in the presence of 7% additional NaCl at 24 and 32 h. Standard deviations are indicated by the error bars.
compared to its parent strain, LS402, suggesting that the role of the mutation is limited to salt adaptation (Fig. 1). 3.2. Identification of the Tn917 insertion site in 37G5 Single primer PCR was used to amplify the DNA directly flanking the transposon. Sequence analysis identified the transposon insertion site in an intergenic region between two divergently transcribed genes, identified here by their F2365 and EGD-e locus tags (in parenthesis), LMOf2365_2170 (lmo2138) and LMOf2365_2171 (lmo2139). The latter gene appeared to be part of an operon of three genes, LMOf2365_2171, LMOf2365_2172 (lmo2140) and LMOf2365_2173 (lmo2141). In order to determine which genes were affected by the transposon insertion, RT-qPCR was performed using primers specific to the four candidate genes, as well as gap (Table 2). This data showed that there was little to no change in expression of LMOf2365_2170 but that all three of the latter, divergently transcribed genes were down-regulated, implicating the possibility of the various combinations of LMOf2365_2171, LMOf2365_2172 and LMOf2365_2173 in the observed phenotype (Fig. 2). This observation also confirmed that these three genes are most likely part of an operon. Previously, Toledo-Arana et al. analyzed the transcriptome of L. monocytogenes EGD-e and identified the existence of a homologous similar operon (Toledo-Arana et al., 2009). We named these genes lstA, lstB and lstC for listerial salt tolerance. BLAST analysis of the first two genes, lstA and lstB, indicated the presence of domains consistent with classification as an ABC transporter and had homology to the two gene ABC transporter of Bacillus subtilis, natAB, which has been implicated in sodium efflux (Cheng et al., 1997). Sodium efflux ABC transporters actively shuttle sodium out of a cell; therefore, cells lacking this functionality would be impaired in their ability to transport sodium from the cell. In high salt concentrations, this could lead to a toxic accumulation of intracellular sodium, loss of osmotic balance and ultimately cell death. The third gene, lstC, had no homology to any previously
characterized genes and was only identified in members of the Listeriaceae family. Analysis of this gene revealed that it contains a GCN5-related N-acetyltransferase (GNAT) domain. This domain uses acetyl-coenzyme A as a donor to acetylate a primary amine on the acceptor (Dyda et al., 2000). GNAT family proteins have a wide range of substrates used and functions, ranging from aminoglycosides for aminoglycoside resistance to histones, metabolic substrates, and even larger acyl groups to proteins and other compounds (Dyda et al., 2000; Vetting et al., 2005). While many of GNAT proteins have been identified and three dimensional structures are available for many of these proteins, this information has not been correlated to substrate specificity, making it difficult to predict the function of LstC (Kuhn et al., 2012). Homologs of lstC are found in Listeria innocua, Listeria marthii, Listeria seeligeri, Listeria ivanovii and Listeria welshimeri with 85% or greater identity over
Fig. 2. Semiquantitative RT-PCR shows downregulation of the genes flanking the 37G5 Tn917 insertion site. The schematic provides the local genetic organization of the region flanking the Tn917 insertion site, indicated by the arrow. Bar graph results show the observed downregulation of the indicated ORFs from two independent trials.
L.S. Burall et al. / Food Microbiology 48 (2015) 72e82
most of the protein (91e100% coverage); however, Listeria grayii homolog has a markedly reduced identity to the F2365 protein (57%) which is not surprising given the fact that L. grayii is genetically very distinct compared to other Listeria species. The closest homolog outside of the Listeriaceae family is found in Paenibacillus with 53% identity over the entirety of the protein. This suggests that the protein is unique to Listeria and, based on the phenotype observed, may have developed as part of the repertoire that allows Listeria to grow in high salt conditions. As three genes were implicated in the phenotypic alterations observed in the 37G5 mutant, RT-qPCR and deletion mutagenesis were performed to determine which gene(s) were involved in the high salt growth phenotype. 3.3. Growth profiles of the in-frame deletion mutants LS411 and LS402 were used to construct the deletion mutants to identify the gene responsible for the observed phenotype. To verify that the transposon effect was related to the three genes, lstABC, a deletion construct was made removing the entire operon. This mutant was found to have a similar growth reduction in the presence of high salt as seen with the transposon mutant, 37G5:Tn917, and the effects were very similar in both LS402 and LS411 strains (Figs. 3b and 4b). Based on the BLAST information available on the genes, which suggested that LstA provided the energy necessary for the sodium efflux transporter to function, a deletion of lstA was made but showed no growth differences in the presence of high salt from the parent strain in either of these strains (Figs. 3b and 4b). These results led us to believe that the sodium efflux transporter was not involved in the salt attenuation observed in the transposon and operon deletion mutants. In-frame deletions of lstC in LS402 and LS411 showed similar growth defect in 7% salt as that seen for both the transposon and lstABC deletion mutants. The growth defect in presence of 7% salt appeared to be more pronounced in the lstABC and lstC deletion mutants compared to the Tn917induced mutant 37G5. It is possible that this difference was caused by residual expression in the Tn917 mutant as the insertion
77
was in the upstream regulatory sequences. Although the absolute growth differences seen between the LS411 DlstABC and DlstC mutants varied, the differences between the parent strains and the mutants were consistent. The growth of all the mutants and their parent strains were found to be comparable at 37 C and 5 C (Figs. 3a, c and 4a, c) suggesting that lstC plays no detectable role in cold growth and doesn't affect growth in non-stress conditions. Based on this data, we believe that the loss of lstC was the cause by which high salt growth was reduced in 37G5. To further evaluate this phenotype, we examined the growth of LS411 and its DlstC mutant in BHI supplemented with 7% NaCl at 5 C (Fig. 4d). The mutant continued to show significant reduction in growth relative to the parent, though it wasn't as pronounced as at 37 C which may be due to a variety of physiological factors including differences in growth dynamics at this condition. Further studies were not undertaken in this condition as there was no apparent added effect on growth. In order to determine whether this role was more broadly attributable to L. monocytogenes or unique to serotype 4b strains, lstC deletion mutants were made in LS401 (EGD-e, 1/2a), LS661 (1/ 2a) and LS667 (1/2b) (Table 1). As seen with LS411 and LS402, the LS667 DlstC mutant (LS824), was attenuated for growth in BHI with 7% NaCl at 37 C but not when grown in BHI broth (Fig. 5). On the other hand, neither the DlstC mutants of LS401 nor the LS661 showed any attenuation in 7% salt growth when compared to their parent strains (Fig. 5). Interestingly, both of the 1/2a parent strains (LS401 and LS661) showed reduced growth in the presence of 7% salt when compared to the other three parent strains, LS402, LS411 and LS667 and the growth defect was greater than that of the lstC deletion mutants of serotype 4b (LS402, LS411) and serotype 1/2b (LS667) strains. Other work has shown deficiencies in growth of lineage II strains during salt stress, in agreement with our observation that salt stress has a greater effect on these strains (Bergholz et al., 2010). These serotypic or lineage differences suggest that at least 1/2a strains used in this study, and perhaps most 1/2a strains, have other deficiencies in high salt growth mechanisms and these differences could mask the role of lstC in salt growth or it may be
Fig. 3. LS402 and its respective deletion mutants grown at 37 C in unsupplemented BHI (A), BHI with 7% NaCl (B) and at 5 C in unsupplemented BHI (C). The strains are LS402 (), LS402 37G5 (◊), LS402DlstA (B), LS402 DlstABC ( ), and LS402 DlstC ( ). (*, P < 0.001) Standard deviations are indicated by error bars.
▵
▫
78
L.S. Burall et al. / Food Microbiology 48 (2015) 72e82
Fig. 4. LS411 and its respective deletion mutants grown at 37 C in unsupplemented BHI (A), BHI with 7% NaCl (B), at 5 C in unsupplemented BHI (C) and at 5 C in BHI with 7% NaCl (D). The strains are LS411 (), LS411DlstA (B), LS411 DlstABC ( ), and LS411 DlstC ( ). (*, P < 0.001) Standard deviations are indicated by error bars.
▵
▫
that one or more of the target gene products affected by lstC is absent or mutated in the 1/2a strains. To further confirm the role of LstC in high salt growth, we performed complementation of the DlstC mutant of LS402 (LS725). Two approaches were tried to provide a promoter to the gene complementation construct. A transcriptional fusion of lstC with pIMK2 (Monk et al., 2008), which provided a promoter upstream of the cloning site, failed to show any complementation (data not shown). A second approach used the SOE approach to create a hybrid construct containing 460 bp upstream of the lstA gene followed by the first 30 codons of lstA, 12 codons of lstB and an intact lstC gene. This construct was cloned into pIMK2 as a transcriptional fusion and transformed into LS725. Empty vector controls were transformed into LS402 and LS725. qRT-PCR was performed to verify presence of the transcript in the parent and complemented strains only (data not shown). Growth trials with these three strains
showed no difference in growth in BHI at 37 C (Fig. 6a). However, growth in BHI with 7% NaCl at 37 C showed a reduction in salt growth for LS725/pIMK2 (DlstC) when compared to LS402/pIMK2. The complemented mutant strain showed growth that was statistically higher than the uncomplemented mutant but lower than the parent strain (Fig 6b). This data suggests partial complementation has occurred and the lack of complete complementation may be due to differences in transcription from the artificial nature of the plasmid construct when compared to the native transcription in the operon. 3.4. Determination of genes affected by lstC Based on the determination that the loss of lstC was responsible for the growth reduction in high salt medium, we performed whole genome transcriptome analysis using a pan-genomic microarray,
Fig. 5. LS401, LS661 and LS667 and their respective deletion mutants grown at 37 C in unsupplemented BHI (A), BHI with 7% NaCl (B). The strains are LS401 (B), LS401DlstC (C), LS661 (◊), LS661 DlstC (A), LS667 ( ), and LS667 DlstC (:). (*, P < 0.001) Standard deviations are indicated by error bars for differences determined to be statistically significant.
▵
L.S. Burall et al. / Food Microbiology 48 (2015) 72e82
79
▵
Fig. 6. Complementation of the DlstC mutant in the LS402 background in BHI borth at 37 C (A) and BHI with 7% NaCl at 37 C (B). The strains are LS402 pIMK2 ( ), LS725 pIMK2 (B), and LS725 pIMK2-lstC (◊). (*, P < 0.001 compared to LS402 pIMK2; #, P < 0.001 compared to both LS402 pIMK2 and LS725 pIMK2) Standard deviations are indicated by error bars.
Listeria GeneChip, described elsewhere (Laksanalamai et al., 2012). This microarray allowed us to look for differences in expression of genes and intergenic regions. The Listeria GeneChip includes intergenic regions that may include sRNAs as annotated in F2365. Use of LS411, an isolate from the same outbreak as F2365, along with its isogenic DlstC mutant, LS723, instead of LS402 allowed us to examine the potential for alterations in expression linked to sRNAs (Nelson et al., 2004). We concentrated on loci unaffected by the loss of lstC in BHI at 37 C but whose expression were significantly up- or down-regulated in the presence of 7% NaCl at 37 C. This analysis identified 16 genes and four intergenic regions that were significantly upregulated in the absence of lstC (Table 3). It is interesting to note that none of the genes was downregulated in the same growth condition. Fifteen of the 16 genes were analyzed via RT-qPCR to verify the differences seen using microarray analysis (Fig. 7). Strains were grown in the presence and absence of 7% salt and the expression levels of the 15 genes were compared to determine whether the genes were specifically upregulated by the mutant in the presence of high salt. The last gene, LMOf2365_1392, wasn't analyzed as its size (117 bp) was too small to allow for primers that fit within the parameters for RT-qPCR. Four of these genes (Table 4) were found to have a greater than two-fold increase expression in the lstC deletion mutant when grown in 7% salt compared to the parent strain, suggesting that the loss of LstC is leading to over-expression of these genes. Two other genes (LMOf2365_2303 and LMOf2365_2304), while upregulated in the high salt condition, appear to have reduced expression when compared to the expression seen in the parent strain. However, the variability between to two trials makes it difficult to make any definitive statement about this data, though further investigation may be warranted as LMOf2365_2303 encodes comK, a gene with some relevance to virulence and environmental survival. Of further note, while LMOf2365_0238 fails to reach the threshold of two-fold upregulation to be considered as showing a pronounced change, the near two-fold upregulation seen in LS723 is interesting given that the parent strain shows pronounced down-regulation of this gene in the presence of high salt. This switch from repression to upregulation is seen for three other genes (LMOf2365_0008, LMOf2365_0993 and LMOf2365_2067) as well. The fact that the 1/2a strains were much reduced for growth in 7% salt and DlstC didn't further affect the growth phenotype in high salt allowed us to speculate that perhaps one or more of these genes (Table 4) product(s) that the DlstC affected were mutated/ non-functional in the 1/2a strains but functional in the 4b and 1/
2b strains. Using the genomes of three 4b strains (F2365, HPB2262, CLIP 80459), four 1/2a strains (EGD-e, 10403S, J2818, F6854) and two 1/2b strains (FSL R2-503, FSL J1-194) and tools from MicrobesOnline (www.microbesonline.org) (Dehal et al., 2010), we analyzed the proteins encoded by the six genes mentioned in Table 4 to see if any of them had any conserved mutations, relative to F2365, that were present in all four 1/2a strains but were absent in the 1/2b and 4b strains. Four of the genes/proteins showed no consistent changes in 1/2a sequences. Two, however, had amino acid substitutions relative to F2365 that were conserved in all four 1/2a strains. Comparison with the other 4b and 1/2b strains showed no alterations of these amino acids in these five strains. LMOf2365_0008 had 14 conserved substitutions in all four 1/2a sequences while LMOf2365_2067 had two such substitutions plus a third amino acid that was identically altered in three of the 1/2a strains and part of a 74aa deletion in the fourth. The first two of these genes is of added interest as LMOf2365_0008 encodes a cardiolipin synthetase which has been shown in some bacteria to play a role in osmotic stress adaptation (Romantsov et al., 2009). The second, LMOf2365_2067, encodes murG and has been implicated in membrane interactions, especially involving cardiolipin (van den Brink-van der Laan et al., 2003), and is part of a pathway that requires a GNAT acetyltransferase, as discussed below. Although evaluation of these six genes does not point to a clear role for lstC in high salt growth, it does provide some scope for speculation. Independent of the serotype-based analysis, some of the genes have been implicated in adaptation to environmental conditions. It was interesting to note that, though the variability is problematic, two adjacent genes were affected by the loss of LstC. A ComK transcriptional regulator and a DNA-binding protein are adjacent to each other in the genome and are convergently transcribed. The comK transcript was the one found to be most affected by the absence of a functional LstC. While it is unclear what, if any, role this ComK protein could have in osmotic stress response as it is not the full length protein, this is not the first observation of comK upregulation in response to environmental factors. Chatterjee et al. found upregulation of this transcript during cytoplasmic growth (Chatterjee et al., 2006). The authors link this increased comK expression with DNA repair genes that also show upregulation. Our study failed to identify any other of the DNA repair genes they identified as being up-regulated in the DlstC mutant, suggesting that DNA repair is not the primary reaction to the loss of lstC. We also evaluated the possible functional roles of these genes by attempting to identify possible sB regulation. Only one gene (Table 4), LMOf2365_2304, has been found to be affected by sB,
80
L.S. Burall et al. / Food Microbiology 48 (2015) 72e82
Table 3 Microarray analysis identifies loci affected by the deletion of lstC. F2365 locus Taga (EGDe homolog)
Probe IDb
BHI fold changec
Salt fold changed
IGLMOf2365_0214 IGLMOf2365_1431 IGLMOf2365_1779
IGLMOf2365_0214_s_at IGLMOf2365_1431_at IGLMOf2365_1779_at IGLMOf2365_1779_s_at IGLMOf2365_1779_x_at IGLMOf2365_2888_s_at LMOf2365_0002_s_at LMOf2365_0008_at LMOf2365_0008_s_at LMOf2365_0238_s_at LMOf2365_0841_s_at LMOf2365_0871_s_at LMOf2365_0993_s_at LMOf2365_1105_at LMOf2365_1105_s_at LMOf2365_1392_at LMOf2365_1392_s_at LMOf2365_1392_x_at LMOf2365_1427_s_at LMOf2365_1427_x_at LMOf2365_1506_s_at LMOf2365_1563_s_at LMOf2365_1563_x_at LMOf2365_2067_s_at LMOf2365_2238_s_at LMOf2365_2303_s_at LMOf2365_2304_s_at LMOf2365_2398_s_at
0.998250943 0.481838894 0.128574562 0.706220848 0.205035419 0.80119854 1.22607478 0.602796261 0.395029167 1.288780436 0.453638985 0.429274341 0.252628568 0.905313975 0.132774082 0.727184811 0.695065926 0.986444003 0.196613735 1.041365508 0.964368959 0.994116043 0.240735829 1.328642363 0.521114311 0.100510711 0.346039348 0.088372391
2.135932459 2.685227699 0.145957488 2.494033212 0.334973232 3.239205564 2.625622838 2.356426346 0.925496807 3.177552996 2.158086779 2.191152459 3.525524065 3.286591658 0.266102076 1.738793215 3.068588469 2.34128786 0.778075897 2.08087062 3.763104456 2.876653301 0.890327217 3.013229901 2.813819011 2.153707122 2.23775842 3.192253545
IGLMOf2365_2888 LMOf2365_0002 (lmo0002) LMOf2365_0008 (lmo0008) LMOf2365_0238 LMOf2365_0841 LMOf2365_0871 LMOf2365_0993 LMOf2365_1105
(lmo0226) (lmo0822) (lmo0854) (lmo0973) (lmo1089)
LMOf2365_1392
LMOf2365_1427 (lmo1408) LMOf2365_1506 (lmo1487) LMOf2365_1563 (lmo1544) LMOf2365_2067 LMOf2365_2238 LMOf2365_2303 LMOf2365_2304 LMOf2365_2398
(lmo2035) (lmo2205) (lmo2270) (lmo2334) (lmo2427)
a
Indicates the affected gene relative to the F2365 genome annotation and, where possible, the closest homolog in EGDe. Indicates the probes utilized to assess expression in a given loci. Some loci have multiple probes due to variability in sequence amongst different genomes. The change in expression was considered significant if the fold change was greater than two, up or down, with a B statistics value >2 and a P value
Contents lists available at ScienceDirect
Food Microbiology journal homepage: www.elsevier.com/locate/fm
A novel gene, lstC, of Listeria monocytogenes is implicated in high salt tolerance Laurel S. Burall, Alexandra C. Simpson, Luoth Chou, Pongpan Laksanalamai, Atin R. Datta* Center for Food Safety and Applied Nutrition, U.S. Food and Drug Administration, Laurel, MD 20708, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Received 14 July 2014 Received in revised form 1 December 2014 Accepted 16 December 2014 Available online 24 December 2014
Listeria monocytogenes, causative agent of human listeriosis, has been isolated from a wide variety of foods including deli meats, soft cheeses, cantaloupes, sprouts and canned mushrooms. Standard control measures for restricting microbial growth such as refrigeration and high salt are often inadequate as L. monocytogenes grows quite well in these environments. In an effort to better understand the genetic and physiological basis by which L. monocytogenes circumvents these controls, a transposon library of L. monocytogenes was screened for changes in their ability to grow in 7% NaCl and/ or at 5 C. This work identified a transposon insertion upstream of an operon, here named lstABC, that led to a reduction in growth in 7% NaCl. In-frame deletion studies identified lstC which codes for a GNAT-acetyltransferase being responsible for the phenotype. Transcriptomic and RT-PCR analyses identified nine genes that were upregulated in the presence of high salt in the DlstC mutant. Further analysis of lstC and the genes affected by DlstC is needed to understand LstC's role in salt tolerance. Published by Elsevier Ltd.
Keywords: Listeria monocytogenes Salt growth Stress adaptation Food preservation
1. Introduction Listeria monocytogenes is a foodborne pathogen that causes invasive listeriosis and gastroenteritis. Invasive listeriosis typically affects the elderly, pregnant women and immune-compromised individuals and causes meningitis, encephalitis, miscarriage and stillbirth (Datta, 2003). These infections are associated with a high mortality rate of approximately 20% making L. monocytogenes responsible for many of the foodborne fatalities, despite a relatively low incidence of disease (Silk et al., 2012). Among pregnant Hispanic women, nearly a third of these infections lead to fetal loss or neonatal death (Silk et al., 2012). Additionally, these infections have a high percentage of hospitalizations and protracted recoveries due to the neurological damage associated with them (Havelaar et al., 2012). L. monocytogenes has also been shown to cause gastroenteritis characterized by nausea, abdominal cramping and diarrhea, though the prevalence of these infections is unknown as gastroenteritis cases are not routinely screened for L. monocytogenes (Barbuddhe and Chakraborty, 2009; Norton and Braden, 2007). Listeria contamination is a problem in foods considered readyto-eat as they are not subjected to a final kill step prior to
* Corresponding author. Tel.: þ1 240 402 3422; fax: þ1 301 210 7975. E-mail address: [email protected] (A.R. Datta). http://dx.doi.org/10.1016/j.fm.2014.12.008 0740-0020/Published by Elsevier Ltd.
consumption. These ready-to-eat (RTE) foods include soft cheeses and deli meats, as well as fresh produce including salad greens, sprouts and cantaloupe. RTE foods with long shelf-life, high water activity and neutral pH are at greater risk for becoming a vehicle of foodborne listeriosis because of the high growth of L. monocytogenes. The standard practices for controlling microbial growth in such foods, e.g. refrigeration and high salt, are not adequate as Listeria survive and grow relatively well under such conditions (Lado and Yousef, 2007). Listeria has been isolated from various foods containing high levels of NaCl and several studies have indicated extended survival of Listeria in high salt concentrations. However, under experimental conditions, Listeria growth appears to be limited to 7e9% of NaCl (Lado and Yousef, 2007). The ability to survive and grow in foods containing high salt is further complicated by the presence of naturally occurring compatible solutes, also known as osmoprotectants, e.g. glycine betaine, carnitine and glycerol in many foods. Although it is known that high salt reduces microbial growth by reducing water activity, changing electrochemical potential across cell membrane, a thorough understanding of the genetics and underlying adaptation mechanisms is required for development of effective control measures. Several studies have identified genes that affect Listeria's ability to grow at low temperature and high salt environments through modification of compatible solute transport, membrane lipid composition alteration and changes in ribosomal protein
L.S. Burall et al. / Food Microbiology 48 (2015) 72e82
(Laksanalamai et al., 2010). Previous studies have already shown the applicability of transposon mutagenesis to aid in the identification of genes associated with Listeria's ability to grow in what would normally be considered restrictive conditions (Burall et al., 2012; Chassaing and Auvray, 2007; Ells et al., 2009; Madeo et al., 2012). Previous work in our laboratory identified a role for iap in salt and cold tolerance and verified a previously established role for the gbu operon in cold and salt tolerance (Angelidis and Smith, 2003a, 2003b; Burall et al., 2012; Ko and Smith, 1999). In this study, we applied the same techniques used before to screen for novel genes implicated in the cold and salt tolerance phenotypes of L. monocytogenes. This work identifies and characterizes the phenotype and expression alterations associated with the deletion of a previously unknown gene, encoding a putative acetyltransferase, with a role in salt tolerance. 2. Materials & methods 2.1. Strains, media and reagents
2.2. Generation of transposon mutants Competent LS402 was electroporated with pLTV3 as described previously (Burall et al., 2012; Park and Stewart, 1990). Transformants were selected on BHI agar containing tetracycline (12.5 mg/mL), erythromycin (1 mg/mL) and lincomycin (25 mg/mL). The presence of pLTV3 in the transformants was confirmed by isolation of the plasmid from Listeria using the QiaPrep miniprep plasmid isolation kit (Qiagen, Valencia, CA) with the addition of 10 mg/mL lysozyme (SigmaeAldrich, St. Louis, MO) in P1 buffer and incubation of the resuspended cells for 10 min at room temperature before proceeding with the manufacturer's protocol. The plasmid recovered from L. monocytogenes was verified via restriction digest with BamHI. A library of LS402 Tn917 mutants was then generated using the procedures described previously (Burall et al., 2012). The library, consisting of approximately 5200 TetSEryRLinR isolates, was assembled in 96-well microtiter plates into groups of 60 isolates each for screening purposes. 2.3. Screening of transposon mutants
L. monocytogenes LS402 is a serotype 4b strain that was isolated during the Italian corn salad gastroenteritis outbreak (Aureli et al., 2000). L. monocytogenes LS411 is a serotype 4b food strain from the 1985 Los Angeles Jalisco cheese invasive listeriosis outbreak (Linnan et al., 1988). These two strains served as parent strains for the various mutants generated in this study. A full list of L. monocytogenes strains used in this study is provided in Table 1. E. coli HB101 was used to maintain pLTV3 and its culture conditions are described elsewhere (Burall et al., 2012; Camilli et al., 1990). As previously described (Burall et al., 2012), L. monocytogenes strains were grown in brain heart infusion (BHI) broth or BHI agar. Salt growth was assessed in BHI broth or BHI agar supplemented with 7% NaCl. Antibiotics used for isolation, characterization and maintenance of the mutants and antibiotic usage during spectrophotometric screens are described elsewhere (Burall et al., 2012). E. coli TOP10 (Invitrogen, Carlsbad, CA) was used for maintenance of pKSV7 (Smith and Youngman, 1992) with chloramphenicol (10 ug/ mL) used for maintenance and selection of it and its derived constructs. Antibiotics were obtained from SigmaeAldrich (St. Louis, MO). All media were obtained from BD (Franklin Lakes, NJ). Restriction enzymes and DNA ligase were purchased from New England Biolabs (Ipswich, MA), and used as per the manufacturer's protocols.
Table 1 Listeria monocytogenes strains used in this study. Strain Genetic background
Source
LS402 LS411 LS401 LS661 LS667 LS654 LS592 LS593 LS652 LS653 LS725 LS723 LS832 LS824 LS950 LS951 LS952 LS953
Martin Wiedmann Martin Wiedmann Lab collection Lab collection Lab collection This study This study This study This study This study This study This study This study This study This study This study This study This study
Wildtype, 1997 Italian Corn Salad Outbreak, 4b Wildtype, 1985 Jalisco Cheese Outbreak, 4b Wildtype, EGDe, serotype 1/2a Wildtype, 2010 Chicken Salad Outbreak, 1/2a Wildtype, 2011 Cantaloupe Outbreak, serotype 1/2b LS402 Tn917:LMOf2365_2170-lstA intergenic LS402 DlstA LS411 DlstA LS402 DlstABC LS411 DlstABC LS402 DlstC LS411 DlstC LS401 DlstC LS667 DlstC LS661 DlstC LS402 pIMK2 LS725 pIMK2 LS725 pIMK2-lstC
73
The ability of the mutants to grow in three different conditions was examined in both BHI broth or agar cultures. Growth was assessed using a Spectramax M2 (Molecular Devices, Downingtown, PA) and a Growth Curves Bioscreen-C (Growth Curves USA, Piscataway, NJ) measuring absorbance at 600 nm for BHI broth cultures or visual assessment for BHI agar cultures. The growth of each mutant was assessed individually but analyzed against the averaged results of all the prospective mutants. The three conditions were unmodified BHI broth at 37 C (control), unmodified BHI broth at 5 C (cold), and BHI broth supplemented with 7% NaCl at 37 C (salt). A starter plate containing 60 isolates, each in its own respective well, was used to inoculate all of the trial conditions as described elsewhere (Burall et al., 2012). Growth on agar plates was assessed daily while growth in the microtiter plates was assessed using absorbance at 600 nm at intervals previously used (Burall et al., 2012). Mutants identified as having altered growth in either of the stress conditions with no alteration in nonstress conditions were further tested (secondary screen) for their altered phenotypes. These secondary screens allowed the use of replicate cultures of each mutant to verify reproducibility. Putative cold-altered mutants were assessed using the Bioscreen-C set at 5 C from 24 h to 120 h post inoculation. A third screen was performed in quadruplicate without antibiotics on any putative mutants still showing altered growth during secondary screens, using LS402 as the control strain. Cold-growth was measured for at least eight days to allow entry into stationary phase to determine maximum growth levels. The absorbance measurements were used to determine exponential growth rate, maximum absorbance and, where possible, the length of lag and exponential phases. Any outliers, defined as ±1 standard deviation from the average, were selected for further screening. In the third screening, individual strains were compared to the averaged results for LS402 and outliers were defined as before. Growth data was also plotted to visually compare alterations more readily observed in a growth curve which allowed more nuanced analysis of the growth, for example a variable growth rate that was not observable in a single averaged number. 2.4. Analysis of growth phenotypes Candidate Tn917 mutants, consistently showing altered growth either in 7% salt or at 5 C, and subsequent deletion mutants were assessed via colony forming units (CFU)/mL in both the cold and salt conditions to confirm the changes in their growth phenotype.
74
L.S. Burall et al. / Food Microbiology 48 (2015) 72e82
Table 2 Primers used in this investigation. Primer
Table 2 (continued )
Sequence
Tn917 insertion site identification Tn917L1 ACGGTTGAAAACTGTACC Tn917L2 GATACAAATTCCTCGTAG Tn917F GGAGCATATCACTTTTCTTGGAGAG qPCR of Tn917 mutant 37G5: lstA forward GTGGAGGAACTTTGCGACTC lstA reverse GAGTTCTGCTGCACTATGGC lstB forward TCCCGAATGTCGAAACTGTA lstB reverse TGTCGGAATATACGAACCGA lstC forward ACCAATAAACACGTCGCAGA lstC reverse TAGAACACGCAAGGCAATTT lmof2365_2170 forward GGTTGTTGGCACTGAAGAAA lmof2365_2170 reverse ATTTGCCGCACAATAAACTG SOE Deletion lstA: a SOE-A CTACTGCAGAACCACGCCCCTTTTCAG SOE-B GAGTATTTTTCCTGGTTCTACTCGA SOE-C TCGAGTAGAACCAGGAAAAATACTCGG TGTTTTAAATGTTCAAACCCAC a SOE-D TAAGAATTCCAAATCCGTCTATCTTACCGTCT XF ACCCAACATAATAATCTTCATACAGC XR CGGCGCTCATCACATCTT lstABC: a SOE-A AGCTGCAGTGATATCCTTTCTTCCAGTTGTCTC SOE-B GCTTACTTCTTTGCCGTCCC SOE-C GGGACGGCAAAGAAGTAAGCTTCAT TAAAATTGCCTTGCGT a SOE-D AAGAGCTCCCCGTAACTTGCTTAAATAGTGG XF ATTTTTTCACGTCTTTCTGTTGG XR GGCGAATTCTTTGTCTTGCTAC lstC: a SOE-A GTTCTAGAGTCCCGAATGTCGAAACT a SOE-A2 GTTCTAGAGTCCCGAATGTGGAAACTGTGG SOE-B TGGCGTCCAAACTAAATGTTCT SOE-B2 TGGTGTCCAAACTAAATGTTCT SOE-C AGAACATTTAGTTTGGACGCCAAT TCAAGGCGAATTTGTGGAT SOE-C2 AGAACATTTAGTTTGGACACCAATT CAAGGCGAATTTGTTGAT a SOE-D TGAAGCTTCGTAACTTGCTTAAATAG XF TCCCAGCGTACTACCTTGTGC XF2 TCCCAGCCTACTACCTTGTGC XR ACGGTGAGGCATTTTCAAACA SOE lstC Complementation: a SOE-A TTCTGCAGGTCTTTCTGTTGGTGTATAGCCA SOE-B GTATTTTTCCTGGTTCTACTCGA SOE-C TCGAGTAGAACCAGGAAAAATACG GAAAACCTCGCTATCTTACG a SOE-D GTGTCGACCGCATCACCTTTTATTTGTTCGT RT-qPCR of LS723: LMOf2365_0002_F TTTGTTATTGAGCGTGATCGTCTTGT LMOf2365_0002_R GGATGCTTGGCCAGAACTAATG LMOf2365_0008_F CAAGTGATTTACGGCGGACCATA LMOf2365_0008_R ATACATTTTCGCGCCTGCTTCAA LMOf2365_0238_F GTTACGAAGATCAAGCGGCATTTTTA LMOf2365_0238_R GATGTGGGATTTTCAGCTTTTCTGTATC LMOf2365_0841_F CGCCCATACGATTCGCTATTAT LMOf2365_0841_R CCGGCGGTACATATCTTGTTG LMOf2365_0871_F GTAGCAGGTTTATGCGAAATGGTC LMOf2365_0871_R CGCCGGCAATAATCATCGTAAT LMOf2365_0993_F AACAGTATAACGCGTGGGATTTTG LMOf2365_0993_R AGGTGTTTGAACGCCGAGTAAGTA LMOf2365_1105_F TATGGCACATTTGATTTGATTCAC LMOf2365_1105_R CGTGGTAAATAAACGACTTCACA LMOf2365_1427_F TGGTTATGAACTCGCAAACAGG LMOf2365_1427_R TCATCTCGTGTGACAGGGATTTT LMOf2365_1506_F AACCGCTTTGCTACATGATTATGC LMOf2365_1506_R ATATACGCCGGGAAATGTTCTTC LMOf2365_1563_F AGATGTGGTAGAAGGTCGCTGTAA LMOf2365_1563_R TGGTGGCTCAATATCTTCTTTTTC LMOf2365_2067_F GTGGTGGCGGTACTGGTGGAC LMOf2365_2067_R GCGGCTGCATAAACGACTGGA LMOf2365_2238_F ATCTTGGCGCTTGAACGAACG
Primer
Sequence
LMOf2365_2238_R LMOf2365_2303_F LMOf2365_2303_R LMOf2365_2304_F LMOf2365_2304_R LMOf2365_2398_F LMOf2365_2398_R
ATGCGCCGCGATGACTACTC AAGCTATGAGGGACGCAAAGAG CTAGGTTCAGCAGCAAGAGTGGT CTGCATCCACAATAAGCAAGTATGAA CAATCGCATCTGCAATTCGTTTTA AAGCCATGTGGTTTGTAGTGTCG TCGGGTTGCAGCATGATAAGAATA
a Underlined text indicates restriction sites introduced into the primer to facilitate cloning.
Growth trials were performed as in the previous study (Burall et al., 2012). Bacterial counts at each time point were compared and statistical analyses were performed using InStat v3 using a standard ANOVA and a Tukey post test (GraphPad, La Jolla, CA). Mutants were screened by colony count to verify normal growth in BHI broth at 37 C for 24 h. Additionally, LS411 mutants were tested for growth in BHI broth supplemented with 7% NaCl at 5 C (cold-salt) with counts determined at defined intervals for 1e2 weeks, depending on observed growth rate. The parent strain of the respective mutant, either LS402 or LS411, served as the controls in all experiments and all strains were tested in replicates of six in at least two independent experiments. 2.5. Identification of Tn917 insertion site Single primer PCR was used to identify the flanking sequences largely as previously described (Karlyshev et al., 2000). Three primers, Tn917L1 (annealing, 48 C), Tn917L2 (annealing, 43 C), and Tn917F (annealing, 54 C) (Table 2) were used to amplify sequences present at either end of the transposon (Liu et al., 2006). DNA was isolated using DNeasy blood and tissue kit (Qiagen, Valencia, CA) using the manufacturer's protocol for Gram positive bacteria with one hour incubations in the enzymatic lysis buffer and in proteinase K. The PCR reaction was performed as described in an earlier study (Burall et al., 2012). The PCR products were then run on a 1% TBE-agarose gel and dominant bands were excised and purified using the QiaQuick Gel Extraction Kit (Qiagen). The product was cloned into pCR2.1 using Invitrogen's TOPO cloning kit and transformed into DH5a Electromax (Invitrogen). Plasmids with the expected insert size were sent for sequencing (Macrogen USA Corporation, Rockville, MD) using T7 and M13 primers. The sequence was then analyzed using BLAST to identify the Tn917:F2365 genome junction (Altschul et al., 1990; Nelson et al., 2004). The F2365 sequence was then compared to the genome annotation via BLAST and the Comprehensive Microbial Resource (CMR) to identify the location of the insertion site (Altschul et al., 1990; Peterson et al., 2001). 2.6. RT-qPCR for gene expression analysis The parent and mutant strains were grown in BHI broth at 37 C to mid-log phase (Abs600 ¼ 0.045e0.055), as determined by absorbance. The bacterial cells were collected and RNA was isolated using the Ambion RiboPure™ RNA extraction kit (Invitrogen, San Diego, CA) following the manufacturer's protocol but using a 20 min cell disruption step. RNA preparations were validated for the absence of DNA contamination using a BioAnalyzer 2100 with the RNA 6000 Nano Kit (Agilent, Santa Clara, CA). In those instances where DNA was present, the sample was treated with DNase I (Ambion/Life Technologies, Grand Island, NY) according to the manufacturer's protocol. Primers were designed to the target genes to yield an amplicon between 90 and 120 bp (Table 2). Gene
L.S. Burall et al. / Food Microbiology 48 (2015) 72e82
expression was analyzed by quantitative RT-PCR using the Bio-Rad iQ5 system. The resulting amplicons were generated and detected using iScript™ One-Step RT-PCR Kit with SYBR Green (Bio-Rad, Hercules, CA). Independent duplicate trials were performed and gene expression was measured in triplicate for each RNA sample. The Ct value was determined and averaged, and the average Ct value was used to determine the fold change in expression, using gap as the reference gene.
75
Technical Manual (Affymetrix, 2009). 4 mg of the cDNA were fragmented by 2.4 unit DNaseI (Promega, Madison, WI) at 37 C for 10 min in a 20 ml reaction containing 1X One-Phore-All buffer (GE Healthcare, Waukesha, WI), followed by heat-inactivation at 95 C for 10 min. The fragmented cDNA was 30 -end labeled with 1 nmole of biotin-11-ddATP (ABI) and 30 units of terminal transferase (Promega, Madison, WI) at 37 C for 4 h. The labeled product was then used for hybridization onto the Listeria GeneChip® (Laksanalamai et al., 2012).
2.7. Creation of in-frame deletion mutants The results from the RT-qPCR analysis identifying which genes were affected by the Tn917 insertion, were compared with annotation analyses (Peterson et al., 2001) to identify which genes were likely to be implicated in the observed phenotype. Based on these results, three deletion mutant constructs were made using splicing by over extension (SOE) deletion. These mutations were made in LS402 and LS411, to verify which gene was actually implicated in the observed phenotypes (Aureli et al., 2000; Linnan et al., 1988). Primers for SOE deletion were designed to develop two ~500 bp amplicons (AeB and CeD) at the 50 and 30 ends of the targeted genes (Chan et al., 2008; Wiedmann et al., 1998) (Table 2). The overlapping primer sequence was then used to amplify the ~1000 bp hybrid product, which lacked the intervening sequence. This product was cloned into the temperature sensitive plasmid, pKSV7, using the indicated restriction sites (Table 2) and the initial plasmid ligations were transformed into E. coli TOP10 cells and verified before being transformed into LS402, LS411 and LS667, as described earlier (Burall et al., 2012). Slight differences in sequence appeared to affect the ability of the original pKSV7-DlstC construct to be used in the 1/2a strains so a second set primers (Table 2) and new plasmid constructs were used for amplification of LS401 and LS661 sequences to account for sequence differences when compared to the F2365 genome. For the generation of both 1/2a mutants, the second set of primers was used to amplify the required sequences from each respective parent strain to make the required pKSV7 deletion construct. The ligation products were transformed into E. coli TOP10 cells and verified before being transformed into LS401 and LS661, as previously described. For all of the transformation reactions, the CamR transformants were then passed several times at 40e42 C with chloramphenicol (10 mg/mL) to select cells with the plasmid integrated into the chromosome. Further passage without antibiotic selection at 30 C led to the isolation of cells with the in-frame deletions without the plasmid backbone. The deletion event was verified using the XF and XR primer set developed for that mutant, which flanked the two initial amplicons (AeB and CeD), using a crude lysis method, as described elsewhere (Burall et al., 2011).
2.9. Microarray hybridization, washing, staining and scanning Hybridizations of the cDNA were performed according to the Affymetrix GeneChip® Expression Analysis Technical Manual (Affymetrix, 2009). Hybridization reactions containing 4 mg of labeled fragmented cDNA, 100 mM MES, 1M Naþ, 20 mM EDTA, 0.01% Tween-20, 50pM control oligoB2 (Affymetrix, Santa Clara, CA), 0.1 mg/mL herring sperm DNA (Promega), 7.8% dimethylsulfoxide (DMSO) (Sigma, St. Louis, MO), were heated at 95 C for 1 min followed by incubation at 45 C for 5 min, prior to hybridizing onto the Affymetrix Listeria GeneChip® at 45 C with rotation (60 rpm) for 16 h in a hybridization oven. The wash and staining procedures were carried out on an Affymetrix FS-450 fluidics station using the mini_prok2v1_450 fluidics script as described by GeneChip® Expression Analysis Technical Manual (Affymetrix, 2009). Arrays were scanned using a GeneChip® Scanner 3000 7G with GCOS v1.4 software. 2.10. Parsing CEL files, probe set summarization methods and data analysis tools All Affymetrix CEL files generated in this study were parsed and analyzed using Robust Multi Array (RMA) algorithm for summarized probe-set intensities implemented by the Affy package of R and Bioconductor (Bolstad et al., 2003; Gentleman et al., 2004; Irizarry et al., 2003, 2006). 2.11. Complementation of lstC SOE primers were designed to amplify the sequences upstream of lstA and coding sequences of lstC while deleting the majority of lstAB from the final construct. The restriction sites inserted via the SOE-A and SOE-D primers were used to clone the construct into pIMK2 (Monk et al., 2008). The resulting plasmid was cloned into E. coli for isolation and verification and then transformed into LS725. Empty vector controls were also transformed into LS402 and LS725. 3. Results & discussion
2.8. Preparation of cDNA for microarray analysis 3.1. Screening of the Tn917 mutant library L. monocytogenes strains LS411 and LS723 were grown in brain heart infusion (BHI) broth at 37 C to use as a starter culture. The starter cultures were diluted to absorbance at 600 nm of 0.1 and were inoculated into 15 ml fresh BHI broth or BHI broth with 7% NaCl. The cultures were incubated at 37 C in a shaking incubator at 170 rpm until they reach an absorbance at 600 nm of 0.5 ± 0.05. The cultures were collected by centrifugation at 4 C for 10 min and immediately processed for RNA extraction using the Ambion RiboPure Bacterial kit (Invitrogen, San Diego, CA), following the manufacturer's protocol with the exception of a 20 min cell disruption step. The quality of the RNA preparation was verified as described for RT-qPCR RNA preparations. Complementary DNA was generated from 10 mg of total RNA according to the Affymetrix GeneChip Expression Analysis
Primary screening of the LS402 Tn917 library (n ¼ 5220) identified 1005 salt-altered mutants and 935 cold-altered mutants that were subjected to a secondary screen that pared these numbers to 93 salt-altered mutants and 98 cold-altered mutants. These mutants were tested in a third round of screening that ultimately identified ten salt and ten cold mutants for verification by colony count analysis. With one exception, none of the prospective candidate mutants were confirmed by colony count. A single saltattenuated Tn917 mutant, 37G5, showed no alteration in growth relative to LS402 in BHI at 37 C; however, it showed significantly lower growth in BHI with 7% NaCl at 37 C as compared to the parent strain LS402 at 24 and 32 h (Fig. 1). Growth studies at 5 C indicated that the mutant has no cold growth defect when
76
L.S. Burall et al. / Food Microbiology 48 (2015) 72e82
Fig. 1. Growth of the Tn917 mutant, 37G5 (◊), relative to its parent strain, LS402(), at 37 C in unsupplemented BHI (A), BHI with 7% NaCl (B) and at 5 C in unsupplemented BHI (C). No reduction in growth is observed in regular BHI at 37 C or at 5 C but a ten-fold significant reduction in final population density (*, P < 0.001) is observed in the presence of 7% additional NaCl at 24 and 32 h. Standard deviations are indicated by the error bars.
compared to its parent strain, LS402, suggesting that the role of the mutation is limited to salt adaptation (Fig. 1). 3.2. Identification of the Tn917 insertion site in 37G5 Single primer PCR was used to amplify the DNA directly flanking the transposon. Sequence analysis identified the transposon insertion site in an intergenic region between two divergently transcribed genes, identified here by their F2365 and EGD-e locus tags (in parenthesis), LMOf2365_2170 (lmo2138) and LMOf2365_2171 (lmo2139). The latter gene appeared to be part of an operon of three genes, LMOf2365_2171, LMOf2365_2172 (lmo2140) and LMOf2365_2173 (lmo2141). In order to determine which genes were affected by the transposon insertion, RT-qPCR was performed using primers specific to the four candidate genes, as well as gap (Table 2). This data showed that there was little to no change in expression of LMOf2365_2170 but that all three of the latter, divergently transcribed genes were down-regulated, implicating the possibility of the various combinations of LMOf2365_2171, LMOf2365_2172 and LMOf2365_2173 in the observed phenotype (Fig. 2). This observation also confirmed that these three genes are most likely part of an operon. Previously, Toledo-Arana et al. analyzed the transcriptome of L. monocytogenes EGD-e and identified the existence of a homologous similar operon (Toledo-Arana et al., 2009). We named these genes lstA, lstB and lstC for listerial salt tolerance. BLAST analysis of the first two genes, lstA and lstB, indicated the presence of domains consistent with classification as an ABC transporter and had homology to the two gene ABC transporter of Bacillus subtilis, natAB, which has been implicated in sodium efflux (Cheng et al., 1997). Sodium efflux ABC transporters actively shuttle sodium out of a cell; therefore, cells lacking this functionality would be impaired in their ability to transport sodium from the cell. In high salt concentrations, this could lead to a toxic accumulation of intracellular sodium, loss of osmotic balance and ultimately cell death. The third gene, lstC, had no homology to any previously
characterized genes and was only identified in members of the Listeriaceae family. Analysis of this gene revealed that it contains a GCN5-related N-acetyltransferase (GNAT) domain. This domain uses acetyl-coenzyme A as a donor to acetylate a primary amine on the acceptor (Dyda et al., 2000). GNAT family proteins have a wide range of substrates used and functions, ranging from aminoglycosides for aminoglycoside resistance to histones, metabolic substrates, and even larger acyl groups to proteins and other compounds (Dyda et al., 2000; Vetting et al., 2005). While many of GNAT proteins have been identified and three dimensional structures are available for many of these proteins, this information has not been correlated to substrate specificity, making it difficult to predict the function of LstC (Kuhn et al., 2012). Homologs of lstC are found in Listeria innocua, Listeria marthii, Listeria seeligeri, Listeria ivanovii and Listeria welshimeri with 85% or greater identity over
Fig. 2. Semiquantitative RT-PCR shows downregulation of the genes flanking the 37G5 Tn917 insertion site. The schematic provides the local genetic organization of the region flanking the Tn917 insertion site, indicated by the arrow. Bar graph results show the observed downregulation of the indicated ORFs from two independent trials.
L.S. Burall et al. / Food Microbiology 48 (2015) 72e82
most of the protein (91e100% coverage); however, Listeria grayii homolog has a markedly reduced identity to the F2365 protein (57%) which is not surprising given the fact that L. grayii is genetically very distinct compared to other Listeria species. The closest homolog outside of the Listeriaceae family is found in Paenibacillus with 53% identity over the entirety of the protein. This suggests that the protein is unique to Listeria and, based on the phenotype observed, may have developed as part of the repertoire that allows Listeria to grow in high salt conditions. As three genes were implicated in the phenotypic alterations observed in the 37G5 mutant, RT-qPCR and deletion mutagenesis were performed to determine which gene(s) were involved in the high salt growth phenotype. 3.3. Growth profiles of the in-frame deletion mutants LS411 and LS402 were used to construct the deletion mutants to identify the gene responsible for the observed phenotype. To verify that the transposon effect was related to the three genes, lstABC, a deletion construct was made removing the entire operon. This mutant was found to have a similar growth reduction in the presence of high salt as seen with the transposon mutant, 37G5:Tn917, and the effects were very similar in both LS402 and LS411 strains (Figs. 3b and 4b). Based on the BLAST information available on the genes, which suggested that LstA provided the energy necessary for the sodium efflux transporter to function, a deletion of lstA was made but showed no growth differences in the presence of high salt from the parent strain in either of these strains (Figs. 3b and 4b). These results led us to believe that the sodium efflux transporter was not involved in the salt attenuation observed in the transposon and operon deletion mutants. In-frame deletions of lstC in LS402 and LS411 showed similar growth defect in 7% salt as that seen for both the transposon and lstABC deletion mutants. The growth defect in presence of 7% salt appeared to be more pronounced in the lstABC and lstC deletion mutants compared to the Tn917induced mutant 37G5. It is possible that this difference was caused by residual expression in the Tn917 mutant as the insertion
77
was in the upstream regulatory sequences. Although the absolute growth differences seen between the LS411 DlstABC and DlstC mutants varied, the differences between the parent strains and the mutants were consistent. The growth of all the mutants and their parent strains were found to be comparable at 37 C and 5 C (Figs. 3a, c and 4a, c) suggesting that lstC plays no detectable role in cold growth and doesn't affect growth in non-stress conditions. Based on this data, we believe that the loss of lstC was the cause by which high salt growth was reduced in 37G5. To further evaluate this phenotype, we examined the growth of LS411 and its DlstC mutant in BHI supplemented with 7% NaCl at 5 C (Fig. 4d). The mutant continued to show significant reduction in growth relative to the parent, though it wasn't as pronounced as at 37 C which may be due to a variety of physiological factors including differences in growth dynamics at this condition. Further studies were not undertaken in this condition as there was no apparent added effect on growth. In order to determine whether this role was more broadly attributable to L. monocytogenes or unique to serotype 4b strains, lstC deletion mutants were made in LS401 (EGD-e, 1/2a), LS661 (1/ 2a) and LS667 (1/2b) (Table 1). As seen with LS411 and LS402, the LS667 DlstC mutant (LS824), was attenuated for growth in BHI with 7% NaCl at 37 C but not when grown in BHI broth (Fig. 5). On the other hand, neither the DlstC mutants of LS401 nor the LS661 showed any attenuation in 7% salt growth when compared to their parent strains (Fig. 5). Interestingly, both of the 1/2a parent strains (LS401 and LS661) showed reduced growth in the presence of 7% salt when compared to the other three parent strains, LS402, LS411 and LS667 and the growth defect was greater than that of the lstC deletion mutants of serotype 4b (LS402, LS411) and serotype 1/2b (LS667) strains. Other work has shown deficiencies in growth of lineage II strains during salt stress, in agreement with our observation that salt stress has a greater effect on these strains (Bergholz et al., 2010). These serotypic or lineage differences suggest that at least 1/2a strains used in this study, and perhaps most 1/2a strains, have other deficiencies in high salt growth mechanisms and these differences could mask the role of lstC in salt growth or it may be
Fig. 3. LS402 and its respective deletion mutants grown at 37 C in unsupplemented BHI (A), BHI with 7% NaCl (B) and at 5 C in unsupplemented BHI (C). The strains are LS402 (), LS402 37G5 (◊), LS402DlstA (B), LS402 DlstABC ( ), and LS402 DlstC ( ). (*, P < 0.001) Standard deviations are indicated by error bars.
▵
▫
78
L.S. Burall et al. / Food Microbiology 48 (2015) 72e82
Fig. 4. LS411 and its respective deletion mutants grown at 37 C in unsupplemented BHI (A), BHI with 7% NaCl (B), at 5 C in unsupplemented BHI (C) and at 5 C in BHI with 7% NaCl (D). The strains are LS411 (), LS411DlstA (B), LS411 DlstABC ( ), and LS411 DlstC ( ). (*, P < 0.001) Standard deviations are indicated by error bars.
▵
▫
that one or more of the target gene products affected by lstC is absent or mutated in the 1/2a strains. To further confirm the role of LstC in high salt growth, we performed complementation of the DlstC mutant of LS402 (LS725). Two approaches were tried to provide a promoter to the gene complementation construct. A transcriptional fusion of lstC with pIMK2 (Monk et al., 2008), which provided a promoter upstream of the cloning site, failed to show any complementation (data not shown). A second approach used the SOE approach to create a hybrid construct containing 460 bp upstream of the lstA gene followed by the first 30 codons of lstA, 12 codons of lstB and an intact lstC gene. This construct was cloned into pIMK2 as a transcriptional fusion and transformed into LS725. Empty vector controls were transformed into LS402 and LS725. qRT-PCR was performed to verify presence of the transcript in the parent and complemented strains only (data not shown). Growth trials with these three strains
showed no difference in growth in BHI at 37 C (Fig. 6a). However, growth in BHI with 7% NaCl at 37 C showed a reduction in salt growth for LS725/pIMK2 (DlstC) when compared to LS402/pIMK2. The complemented mutant strain showed growth that was statistically higher than the uncomplemented mutant but lower than the parent strain (Fig 6b). This data suggests partial complementation has occurred and the lack of complete complementation may be due to differences in transcription from the artificial nature of the plasmid construct when compared to the native transcription in the operon. 3.4. Determination of genes affected by lstC Based on the determination that the loss of lstC was responsible for the growth reduction in high salt medium, we performed whole genome transcriptome analysis using a pan-genomic microarray,
Fig. 5. LS401, LS661 and LS667 and their respective deletion mutants grown at 37 C in unsupplemented BHI (A), BHI with 7% NaCl (B). The strains are LS401 (B), LS401DlstC (C), LS661 (◊), LS661 DlstC (A), LS667 ( ), and LS667 DlstC (:). (*, P < 0.001) Standard deviations are indicated by error bars for differences determined to be statistically significant.
▵
L.S. Burall et al. / Food Microbiology 48 (2015) 72e82
79
▵
Fig. 6. Complementation of the DlstC mutant in the LS402 background in BHI borth at 37 C (A) and BHI with 7% NaCl at 37 C (B). The strains are LS402 pIMK2 ( ), LS725 pIMK2 (B), and LS725 pIMK2-lstC (◊). (*, P < 0.001 compared to LS402 pIMK2; #, P < 0.001 compared to both LS402 pIMK2 and LS725 pIMK2) Standard deviations are indicated by error bars.
Listeria GeneChip, described elsewhere (Laksanalamai et al., 2012). This microarray allowed us to look for differences in expression of genes and intergenic regions. The Listeria GeneChip includes intergenic regions that may include sRNAs as annotated in F2365. Use of LS411, an isolate from the same outbreak as F2365, along with its isogenic DlstC mutant, LS723, instead of LS402 allowed us to examine the potential for alterations in expression linked to sRNAs (Nelson et al., 2004). We concentrated on loci unaffected by the loss of lstC in BHI at 37 C but whose expression were significantly up- or down-regulated in the presence of 7% NaCl at 37 C. This analysis identified 16 genes and four intergenic regions that were significantly upregulated in the absence of lstC (Table 3). It is interesting to note that none of the genes was downregulated in the same growth condition. Fifteen of the 16 genes were analyzed via RT-qPCR to verify the differences seen using microarray analysis (Fig. 7). Strains were grown in the presence and absence of 7% salt and the expression levels of the 15 genes were compared to determine whether the genes were specifically upregulated by the mutant in the presence of high salt. The last gene, LMOf2365_1392, wasn't analyzed as its size (117 bp) was too small to allow for primers that fit within the parameters for RT-qPCR. Four of these genes (Table 4) were found to have a greater than two-fold increase expression in the lstC deletion mutant when grown in 7% salt compared to the parent strain, suggesting that the loss of LstC is leading to over-expression of these genes. Two other genes (LMOf2365_2303 and LMOf2365_2304), while upregulated in the high salt condition, appear to have reduced expression when compared to the expression seen in the parent strain. However, the variability between to two trials makes it difficult to make any definitive statement about this data, though further investigation may be warranted as LMOf2365_2303 encodes comK, a gene with some relevance to virulence and environmental survival. Of further note, while LMOf2365_0238 fails to reach the threshold of two-fold upregulation to be considered as showing a pronounced change, the near two-fold upregulation seen in LS723 is interesting given that the parent strain shows pronounced down-regulation of this gene in the presence of high salt. This switch from repression to upregulation is seen for three other genes (LMOf2365_0008, LMOf2365_0993 and LMOf2365_2067) as well. The fact that the 1/2a strains were much reduced for growth in 7% salt and DlstC didn't further affect the growth phenotype in high salt allowed us to speculate that perhaps one or more of these genes (Table 4) product(s) that the DlstC affected were mutated/ non-functional in the 1/2a strains but functional in the 4b and 1/
2b strains. Using the genomes of three 4b strains (F2365, HPB2262, CLIP 80459), four 1/2a strains (EGD-e, 10403S, J2818, F6854) and two 1/2b strains (FSL R2-503, FSL J1-194) and tools from MicrobesOnline (www.microbesonline.org) (Dehal et al., 2010), we analyzed the proteins encoded by the six genes mentioned in Table 4 to see if any of them had any conserved mutations, relative to F2365, that were present in all four 1/2a strains but were absent in the 1/2b and 4b strains. Four of the genes/proteins showed no consistent changes in 1/2a sequences. Two, however, had amino acid substitutions relative to F2365 that were conserved in all four 1/2a strains. Comparison with the other 4b and 1/2b strains showed no alterations of these amino acids in these five strains. LMOf2365_0008 had 14 conserved substitutions in all four 1/2a sequences while LMOf2365_2067 had two such substitutions plus a third amino acid that was identically altered in three of the 1/2a strains and part of a 74aa deletion in the fourth. The first two of these genes is of added interest as LMOf2365_0008 encodes a cardiolipin synthetase which has been shown in some bacteria to play a role in osmotic stress adaptation (Romantsov et al., 2009). The second, LMOf2365_2067, encodes murG and has been implicated in membrane interactions, especially involving cardiolipin (van den Brink-van der Laan et al., 2003), and is part of a pathway that requires a GNAT acetyltransferase, as discussed below. Although evaluation of these six genes does not point to a clear role for lstC in high salt growth, it does provide some scope for speculation. Independent of the serotype-based analysis, some of the genes have been implicated in adaptation to environmental conditions. It was interesting to note that, though the variability is problematic, two adjacent genes were affected by the loss of LstC. A ComK transcriptional regulator and a DNA-binding protein are adjacent to each other in the genome and are convergently transcribed. The comK transcript was the one found to be most affected by the absence of a functional LstC. While it is unclear what, if any, role this ComK protein could have in osmotic stress response as it is not the full length protein, this is not the first observation of comK upregulation in response to environmental factors. Chatterjee et al. found upregulation of this transcript during cytoplasmic growth (Chatterjee et al., 2006). The authors link this increased comK expression with DNA repair genes that also show upregulation. Our study failed to identify any other of the DNA repair genes they identified as being up-regulated in the DlstC mutant, suggesting that DNA repair is not the primary reaction to the loss of lstC. We also evaluated the possible functional roles of these genes by attempting to identify possible sB regulation. Only one gene (Table 4), LMOf2365_2304, has been found to be affected by sB,
80
L.S. Burall et al. / Food Microbiology 48 (2015) 72e82
Table 3 Microarray analysis identifies loci affected by the deletion of lstC. F2365 locus Taga (EGDe homolog)
Probe IDb
BHI fold changec
Salt fold changed
IGLMOf2365_0214 IGLMOf2365_1431 IGLMOf2365_1779
IGLMOf2365_0214_s_at IGLMOf2365_1431_at IGLMOf2365_1779_at IGLMOf2365_1779_s_at IGLMOf2365_1779_x_at IGLMOf2365_2888_s_at LMOf2365_0002_s_at LMOf2365_0008_at LMOf2365_0008_s_at LMOf2365_0238_s_at LMOf2365_0841_s_at LMOf2365_0871_s_at LMOf2365_0993_s_at LMOf2365_1105_at LMOf2365_1105_s_at LMOf2365_1392_at LMOf2365_1392_s_at LMOf2365_1392_x_at LMOf2365_1427_s_at LMOf2365_1427_x_at LMOf2365_1506_s_at LMOf2365_1563_s_at LMOf2365_1563_x_at LMOf2365_2067_s_at LMOf2365_2238_s_at LMOf2365_2303_s_at LMOf2365_2304_s_at LMOf2365_2398_s_at
0.998250943 0.481838894 0.128574562 0.706220848 0.205035419 0.80119854 1.22607478 0.602796261 0.395029167 1.288780436 0.453638985 0.429274341 0.252628568 0.905313975 0.132774082 0.727184811 0.695065926 0.986444003 0.196613735 1.041365508 0.964368959 0.994116043 0.240735829 1.328642363 0.521114311 0.100510711 0.346039348 0.088372391
2.135932459 2.685227699 0.145957488 2.494033212 0.334973232 3.239205564 2.625622838 2.356426346 0.925496807 3.177552996 2.158086779 2.191152459 3.525524065 3.286591658 0.266102076 1.738793215 3.068588469 2.34128786 0.778075897 2.08087062 3.763104456 2.876653301 0.890327217 3.013229901 2.813819011 2.153707122 2.23775842 3.192253545
IGLMOf2365_2888 LMOf2365_0002 (lmo0002) LMOf2365_0008 (lmo0008) LMOf2365_0238 LMOf2365_0841 LMOf2365_0871 LMOf2365_0993 LMOf2365_1105
(lmo0226) (lmo0822) (lmo0854) (lmo0973) (lmo1089)
LMOf2365_1392
LMOf2365_1427 (lmo1408) LMOf2365_1506 (lmo1487) LMOf2365_1563 (lmo1544) LMOf2365_2067 LMOf2365_2238 LMOf2365_2303 LMOf2365_2304 LMOf2365_2398
(lmo2035) (lmo2205) (lmo2270) (lmo2334) (lmo2427)
a
Indicates the affected gene relative to the F2365 genome annotation and, where possible, the closest homolog in EGDe. Indicates the probes utilized to assess expression in a given loci. Some loci have multiple probes due to variability in sequence amongst different genomes. The change in expression was considered significant if the fold change was greater than two, up or down, with a B statistics value >2 and a P value
