An advanced bioinformatics approach for analyzing RNA-seq data reveals sigma H-dependent regulation of competence genes in Listeria monocytogenes

Background Alternative σ factors are important transcriptional regulators in bacteria. While σB has been shown to control a large regulon and play important roles in stress response and virulence in the pathogen Listeria monocytogenes, the function of σH has not yet been well defined in Listeria, even though σH controls a large regulon in the closely related non-pathogenic Bacillus subtilis. Results Using RNA-seq characterization of a L. monocytogenes strain with deletions of all 4 genes encoding alternative σ factors (ΔBCHL), which was further modified to overexpress sigH (ΔBCHL::Prha-sigH), we identified 6 transcription units (TUs) that are transcribed from σH-dependent promoters. Five of these TUs had not been previously identified. Identification of these promoters was facilitated by use of a bio-informatics approach that compared normalized RNA-seq coverage (NRC), between ΔBCHL::Prha-sigH and a ΔBCHL control, using sliding windows of 51 nt along the whole genome rather than comparing NRC calculated only for whole genes. Interestingly, we found that three operons that encode competence genes (comGABCDEFG, comEABC, coiA) are transcribed from σH-dependent promoters. While these promoters were highly conserved in L. monocytogenes, none of them were found in all Listeria spp. and coiA and its σH-dependent promoter were only found in L. monocytogenes. Conclusions Our data indicate that a number of L. monocytogenes competence genes are regulated by σH. This σH-dependent regulation of competence related genes is conserved in the pathogen L. monocytogenes, but not in other non-pathogenic Listeria strains. Combined with prior data that indicated a role of σH in virulence in a mouse model, this suggests a possible novel role of σH-dependent competence genes in L. monocytogenes virulence. Development and implementation of a sliding window approach to identify differential transcription using RNA-seq data, not only allowed for identification of σH-dependent promoters, but also provides a general approach for sensitive identification of differentially transcribed promoters and genes, particularly for genes that are transcribed from multiple promoter elements only some of which show differential transcription. Electronic supplementary material The online version of this article (doi:10.1186/s12864-016-2432-9) contains supplementary material, which is available to authorized users.


Background
At the transcriptional level, bacterial gene expression under rapidly changing environmental conditions is controlled by changes in associations between different alternative σ factors and the catalytic core of RNA polymerase. Alternative σ factors are important contributors to gene expression under stress conditions in both Gram-positive and Gram-negative bacteria [1]. In the foodborne pathogen Listeria monocytogenes, four alternative sigma factors (σ B , σ H , σ L , σ C ) play a role in transcriptional regulation. σ B and the σ B regulon have been well defined in this pathogen through a variety of different approaches [2][3][4]. A number of different studies have indicated that σ B controls a large regulon that plays important roles in both stress response and virulence. The regulons controlled by the other alternative σ factors in L. monocytogenes have been less well defined.
While σ H specifically has been shown to play an important role in regulating spore formation and competence in B. subtilis [5], the function of σ H in Listeria has not yet been well defined. In B. subtilis, σ H has been reported to regulate >400 genes (approx. 240 genes positively and approx. 180 negatively) [6]. B. subtilis σ H is also involved in the transition from exponential phase to stationary phase, nutrient transport, and the regulation of many other transcription factors and cell-wall-binding proteins [7]. In L. monocytogenes characterization of a sigH null mutant suggested a role for σ H in growth on minimal medium and under alkaline conditions as well as a role in virulence, as assessed in an intraperitoneal inoculated mouse model [8]. While previous microarray experiments identified 56 genes as being directly upregulated by σ H , these experiments also found that the largest category of co-regulated genes were represented by genes that showed transcript levels affected by both the σ B and σ H deletions [9]. Based on this considerable overlap between the σ B and σ H regulon, we surmised that different approaches are needed to further define σ Hdependent genes in L. monocytogenes. In order to eliminate redundancies between transcriptional regulation by σ H and other alternative σ factors, we chose to perform RNA-seq based comparisons of transcript levels between a L. monocytogenes mutant with deletions of all 4 genes encoding alternative σ factors (ΔBCHL), and one with the same background that was modified to overexpress sigH (ΔBCHL::P rham -sigH). We also implemented an advanced bio-informatics approach where we compare normalized RNA-seq coverage (NRC), between these two strains using a sliding window of 51 nt and 25 nt overlap along the whole genome, rather than comparing NRC calculated only for whole genes; a similar approach was detailed independently in a recent study [10]. We surmised that this approach would allow for more sensitive identification of differentially regulated genes and gene fragments, particularly for genes that may be preceded by multiple promoters (e.g., a σ H and a σ A dependent promoter), where differential transcript levels may only be detectable downstream of the σ H -dependent promoter, but not in the actual open reading frame (ORF) where transcription from the σ A -dependent promoter may obscure the differential transcription from the σ H -dependent promoter.

Results and discussion
A sliding window method for identification of differentially regulated genes and promoters provides a sensitive approach for identification of σ H dependent genes Traditionally, analysis of RNA-seq data is performed by calculating the normalized RNA-seq coverage (NRC) for a given annotated gene and comparing NRCs between strains with different genetic backgrounds or strains grown under different conditions. When this approach was used here, we initially identified 5 genes that showed significantly higher transcript levels (FDR < 0.05; FC > 2.0) in the L. monocytogenes strain overexpressing sigH (ΔBCHL::P rha -sigH), as compared to the control strain ΔBCHL::P rha , which does not contain sigH. These 5 genes represented LMRG_00908, LMRG_00935 (comEA), LMRG_00937 (comEC), LMRG_01629 (lytG), and LMRG_01643 (coiA) (see Additional file 1: Table S1). A total of four σ Hdependent promoters were identified upstream of these five genes as comEA and comEC share the same promoter. Considering these promoters and previously reported operon structures in L. monocytogenes [11,12], these 5 σ Hdependent genes were found to represent 2 multigene operons, including (i) LMRG_00908, LMRG_00907 (dnaG) and LMRG_00906 (rpoD); and (ii) LMRG_00935 (comEA), LMRG_00936 (comEB) and LMRG_00937 (comEC)) and 2 monocistronic genes, coiA and lytG, for a total of 8 genes that are positively regulated by σ H (Fig.1a-d). We did not find any genes that were significantly downregulated by σ H (FDR < 0.05; FC < 0.5).
Subsequent sliding window analysis, using the same RNA-seq data analyzed with the standard BaySeq approach (described in Methods), identified two additional σ H -dependent promoters, in addition to also identifying the promoters described above. The two newly identified promoters, along with previously reported operon structures in L. monocytogenes [11,12], identified two additional multigene operons, including (i) LMRG_00797 to LMRG_00791 and (ii) LMRG_01010 to LMRG_01005) as σ H -dependent (Table 1). Overall, the new σ Hdependent operons identified using the sliding window approach represent a total of 13 genes that are positively regulated by σ H (Fig. 1e-f ). The LMRG_01010-LMRG_01005 operon, which was identified only with the sliding window approach, showed both a σ H -and Black arrows indicate the direction in which the RNA-seq reads were mapped to the chromosome. Genes significantly differentially expressed by the standard approach are labeled with a * next to the gene name. Position of significant fragments is shown as dotted lines underneath the genes with their maximum sliding window fold change shown underneath. Stem loop symbols indicate transcriptional terminators. Genes colored in gray are part of the operons found to be significantly differentially expressed by the sliding window approach and are drawn to scale. Genes colored in magenta are not part of the significant operon and are not drawn to scale. Promoters are indicated by stemmed arrows. A window size of 100 nt was used for the smoothing method. Values on the graph represent the center of these 100 nt windows two σ A -dependent promoters (Fig. 1f ). These data indicate that the sliding window RNA-seq analysis approach described here provides a superior sensitivity for identification of differentially transcribed genes, particularly when multiple promoters are found upstream a gene or operon.
Newly identified σ H -dependent genes include three σ Hdependent operons that encode competence proteins A comparison of the σ H -dependent genes identified here with σ H -dependent genes and operons that were previously found with microarray based characterization of a L. monocytogenes ΔsigH strain [9] showed that we identified a number of σ H -dependent genes and operons that had not been previously identified, including LMRG1629 (lytG), LMRG_00937 (comEA), LMRG_00935 (comEC), LMRG001643 (coiA) ( Table 1). The previous microarray study used for comparison used a ΔsigH strain in a wildtype background and used a standard probe based microarray (not a tiling array); no additional transcriptomic studies characterizing the σ H regulon in Listeria were available for comparison. Interestingly, among the 56 genes identified as directly regulated by σ H in the previous microarray based study of a ΔsigH strain [9], only 16 showed a FC >2.0 (the other 40 genes showed FCs between 1.5 and 2.0). Only one of these previously identified genes (LMRG_00908) was also found to be σ H -dependent here. This may suggest that a number of genes recognized as σ H -dependent in this previous study only show differential σ H -dependent transcription in the presence of the other three alternative σ factors, which were not present in the L. monocytogenes ΔBCHL strain that was used here. These observations further highlight the importance of the use of alternative approaches (e.g., use of a ΔBCHL strain overexpressing sigH) we have taken here to allow for identification of additional σ Hdependent genes. Ectopic and artificial induction of sigma factors and other regulatory proteins has been used successfully in the past to study those regulators where physiological induction signals have not been found [13] and/or where mutants cannot be made [14]. In some cases, large periods of time have passed since linking a set of genes to a regulator (using artificial induction) and discovering a physiological condition where a given regulator and regulated genes are induced. For example in S. aureus, σ H dependent transcription of competence genes after overexpression was first reported in 2003 [15], while functional confirmation of σ H contributions to competence were only reported almost 10 years later [16]. While our data reported here thus provide important evidence for contributions of σ H to regulation of competence genes in L. monocytogenes, future experiments will be needed to probe the phenotypic importance of this regulatory pathway and to define under which conditions competence gene expression is upregulated by σ H in wildtype L. monocytogenes.
Importantly, we identified three σ H -dependent transcription units that encode competence proteins, including (i) comGABCDEFG, (ii) comEABC and (iii) coiA ( Fig. 1). In B. subtilis, a naturally competent bacterium, ComG proteins are required for exogenous DNA to reach the membrane bound receptor ComEA during transformation [17,18]. Among the genes in the comEABC operon, only comEA and comEC (encoding a a Each fragment is composed of one or more 51 nt overlapping windows b This indicates which genes were identified as differentially transcribed with the analysis approach that calculated the normalized RNA-seq coverage for the complete gene (rather than sliding windows) (for full results see Additional file 1: Table S1) polytopic membrane protein that forms the membrane translocation channel) have been shown to be required for transformation in B. subtilis [19]. B. subtilis coiA has also been shown to be involved in the establishment of DNA transport [20]. Even though competence traditionally implies uptake of DNA, competence genes have been shown to also play roles in survival [21] and virulence [22], processes which do not necessarily require DNA uptake (reviewed in [23]). In L. monocytogenes, competence is not well understood. While this organism seems to possess the machinery for competence, the gene encoding a key regulator of competence in B. subtilis, comK, is often interrupted by prophages in L. monocytogenes strains [24]. Excision of this prophage and restoration of an intact comK has been shown to be involved in phagosomal escape and virulence [22] but competence was not tested in this study. The one published study that experimentally tested for competence in L. monocytogenes did not find evidence for competence even among two L. monocytogenes strains that carried an intact comK [25].
Interestingly, S. aureus σ H has also been shown to regulate competence-related genes [15]. Although S. aureus was originally thought to not be competent, it has now been reported that, facilitated by a complex regulatory mechanism, S. aureus cells are able to uptake exogenous DNA, such as antibiotic resistance determinants, through horizontal gene transfer [16]. Even though the specific conditions for competence in L. monocytogenes have not been found yet, the fact that this pathogen has conserved σ H -dependent mechanisms of regulation of competence genes suggests a specific physiological role of these genes in L. monocytogenes.
Among other genes identified as being directly regulated by σ H are the housekeeping genes, rpoD, encoding σ A and dnaG, encoding the DNA primase (Fig. 1). Another σ H -dependent gene, lytG, encodes an enzyme with a mannosyl-glycoprotein endo-beta-Nacetylglucosamidase-like domain, which is found in enzymes such as lysozymes and the flagellar protein J that can hydrolyze peptidoglycan [26]. Two genes involved in oxidative stress, msrA and msrB were also found to be directly regulated by σ H , suggesting a possible role of σ H in oxidative stress response [27]. Moreover, the σ H -dependent promoter upstream this operon has a long 5' UTR (270 nt), which overlaps with the 5' UTR of another gene, LMRG_01011, encoded in the opposite strand. LMRG_01011 encodes a hemolysin III protein (Hly III) and the expression of this protein, could therefore, be σ H regulated posttranscriptionally through RNA interference caused by the 5' UTR from P LMRG_01010 .
σ H -dependent promoters are highly conserved among different L. monocytogenes genomes The six σ H -dependent promoters identified in L. monocytogenes 10403S were all found to be present in the 23 additional L. monocytogenes genomes analyzed here (Additional file 2: Table S2). In general, −35 and −10 promoter regions were highly conserved (Fig. 2, Table 2). For both the comGA and the lytG promoter sequence, the −35 region was completely conserved across all 24 genomes, while the −10 region presented two variants for each of these two promoters. The comEA promoter showed a completely conserved −10 region, while the −35 region showed three variants. On the other hand the −35 region and −10 regions for the three other σ Hdependent promoters (upstream of LMRG_00908, LMRG_01010 and coiA) were each completely conserved across all 24 genomes. By comparison, sigH was found to be present in all 24 genomes with a high level of conservation (only 4 polymorphic amino acid residues over 201 amino acids). Overall, these data indicate considerable conservation of σ H -dependent promoters identified, even though some promoters show lineage or strain specific sequence features. Comparative analysis of the frequency of nucleotide changes within the −35 and −10 signal regions compared to the non-functional sequences (assumed here to be under neutral selection) between these two functional regions suggests that both −35 and −10 signals are under selective pressure among the L. monocytogenes strains analyzed (Table 3). This suggests that the σ H -dependent promoters identified in this study are functional across all L. monocytogenes strains and, therefore, are probably important for L. monocytogenes physiology at certain conditions that remain to be established. Future studies are needed to explore the functional importance of the polymorphic sequence features. Interestingly, a previous study [28] found evidence for positive selection of a promoter region that regulates virulence gene expression in L. monocytogenes. The sites found under positive selection in this previous study generated a putative σ B -dependent promoter in some lineage I strains, which supports that strain and lineage specific promoter region polymorphisms may affect virulence gene expression in L. monocytogenes.
σ H promoters and the σ H regulon have diverged among different Listeria species Further analysis of the six σ H -dependent promoters identified in L. monocytogenes showed that only the LMRG_00908 and LMRG_01010 promoters were found in all 5 Listeria species analyzed (i.e., L. monocytogenes, L. innocua, L. ivanovii, L. welshimeri and L. seeligeri). The comGA promoter was not found in the L. welshimeri genome (Additional file 2: Table S2). The comEA promoter was not found in the L. ivanovii and L. welshimeri genomes and the coiA and lytG promoters were only found in the L. monocytogenes genomes. Analyses of the genes transcribed by these promoters revealed the respective genes were also missing in the genomes where a promoter could not be found. Therefore, the competence genes represented by the comGABCDEFG operon, the comEABC operon and the coiA gene are not found across all Listeria species and the σ H regulon is not conserved across these species. Taking into consideration that a sigH mutant has been shown to have reduced virulence in a mouse model [8], the absence of these σ H -dependent genes in non-pathogenic Listeria spp. could indicate that these genes may have evolved to contribute σ H -dependent virulence related functions in L. monocytogenes. By comparison, sigH was present in all Listeria species genomes; in addition to the 4 polymorphic amino acid residues found in L. monocytogenes, a further 18 polymorphic amino acid residues were found in the Listeria spp. genomes.
Among the promoters found in other Listeria species besides L. monocytogenes, only the LMRG_00908 operon was highly conserved (Fig. 3). This promoter showed a completely conserved −35 region, while two variants of the −10 region were found. The comEA promoter, which was only found in L. monocytogenes, L. innocua and L. seeligeri, had a perfectly conserved −10 region across all three species. However, besides the two variants of the −35 region found among the L. monocytogenes strains, each the L. innocua and the L. seeligeri strain presented distinct −35 sequences. The comGA promoter showed variable −35 and −10 regions across the Listeria species. The LMRG_01010 -35 sequence in L. innocua, L. ivanovii, L. welshimeri and L. seeligeri matched that of the L. monocytogenes strains. The only −10 region that diverged from the others was that found in the L. ivanovii strain. Therefore, in addition to a σ H regulon that is not conserved across all Listeria species, there is also some variation in the σ H -dependent promoters that are present in other Listeria species. While the σ Hdependent promoters present generally seem to be conserved enough to suggest conservation of σ H -dependent regulation, future experimental studies are needed to explore the potential functional importance of these polymorphic sequence features.

Conclusions
Combined with prior data that indicated a role of σ H in virulence in a mouse model, identification of a number of σ H -dependent L. monocytogenes competence genes that are conserved in the pathogen L. monocytogenes,  but not in other non-pathogenic Listeria strains, suggests a possible novel role of σ H -dependent competence genes in L. monocytogenes virulence. The development and use of a sliding window approach to identify differential transcription using RNA-seq data not only allowed for identification of new σ H -dependent promoters in L. monocytogenes, but also provides a general approach for sensitive identification of differentially transcribed promoters and genes. We predict that this approach will be particularly valuable for identification of differentially transcribed genes and genomic regions that are transcribed from multiple and possibly redundant promoter elements.

Bacterial strains, mutant construction, and growth conditions
The quadruple alternative σ factor mutant (ΔBCHL; FSL C3-135) of L. monocytogenes strain 10403S [29] was used as the background strain in this study. This strain was modified to overexpress sigH from a rhamnose inducible promoter. Briefly, the sigH gene was amplified from L. monocytogenes 10403S by PCR and cloned into the plasmid pLF1 [30] downstream of the rhamnose inducible promoter P rha . The plasmid construct was confirmed by PCR and sequencing. All the cloning steps were performed in E. coli DH5α (NEB). The final plasmid was transformed into E. coli strain SM10 to allow for conjugation of the plasmid into L. monocytogenes 10403S ΔBCHL, followed by chromosomal integration of the P rha sigH construct (yielding strain 10403S::ΔBCHL P rha -sigH; FSL B2-426). A control strain (ΔBCHL-P rha ; FSL B2-429) was constructed by introducing the empty plasmid pLF1 into the chromosome of L. monocytogenes 10403S ΔBCHL through conjugation and chromosomal integration. Transconjugants were selected with 200 μg/ ml Streptomycin and 7.5 μg/ml Chloramphenicol, and confirmed by PCR. For RNA isolation, strains were streaked from frozen Brain Heart Infusion (BHI) stock, stored at −80°C in 15 % glycerol, onto a BHI agar plate, followed by incubation at 37°C for 24 h. A single colony was subsequently inoculated into 5 ml of BHI broth in 16 mm tubes, followed by incubation at 37°C with shaking (230 rpm) for 18 h (Series 25 Incubator, New Brunswick Scientific, Edison, NJ). After 18 h, 50 μl BHI culture was inoculated into fresh 5 ml BHI broth and grown to OD 600 0.4-0.5 at 37°C.

Rhamnose induction
Induction of sigH transcription was performed by adding 250 μl of 1M rhamnose stock solution to 5 ml OD 600 0.4-0.5 bacterial cultures (for a final concentration of 50 mM rhamnose), followed by incubation at 37°C for an additional 30 min. Induction with rhamnose was performed for both 10403S::ΔBCHL P rha -sigH and ΔBCHL-P rha . qRT-PCR using the SYBR Green Master Mix Reagent (Life Technologies) and the ABI Prism 7000 Sequence Detection System (Applied Biosystems, Foster City, CA) determined that the optimum rhamnose concentration for sigH induction was 50 mM rhamnose. Transcript levels were determined for sigH and the housekeeping gene rpoB in strain 10403S::ΔBCHL P rha -sigH. Expression level differences were determined by ΔΔCt method [31] using the housekeeping gene rpoB as reference gene.

RNA isolation
RNA isolation was performed as previously described by our group [32]; minor modifications were made to this previous RNA isolation protocol. Briefly, for each sample, 3 ml of RNAprotect bacteria reagent (Qiagen, Valencia, CA) was added to 3 ml of bacterial culture. The mix was incubated at room temperature for 10 min to ensure that the bacterial RNA was stabilized. Cells were pelleted by centrifugation (4,637 × g, 30 min) at 4°C and suspended in nuclease free water with proteinase K (25 mg/ml) and lysozyme (50 mg/ml), followed by incubation at 37°C for 30 min. After RNA isolation with TRI reagent, total RNA was incubated with Turbo DNase (Life Technologies) to remove remaining DNA in the presence of RNasin (Promega). Subsequently, RNA was purified using phenol-chloroform/chloroform extractions, followed by precipitation and resuspension. Purity of RNA and efficiency of the DNase treatment was assessed by UV spectrophotometry (Nanodrop, Wilmington, DE) and qRT-PCR for the housekeeping gene rpoB (all samples showed Ct > 35, indicating absence of DNA contamination at relevant levels), respectively. All experiments were performed in three biological replicas.

cDNA libraries and RNA-seq
Preparation of directional cDNA fragment libraries was performed using the ScriptSeq Complete Kit (Bacteria)-Low Input (Epicentre, Madison, WI). 16S and 23S rRNA was removed from total RNA with Ribo-Zero rRNA Removal Reagents (Bacteria)-Low Input and Magnetic Core Kit-Low Input. rRNA-depleted samples were run on the 2100 Bioanalyzer (Agilent Technology, Santa Clara, CA) to confirm reduction of 16S and 23S rRNA and followed by purification using Agencourt RNA-Clean XP Kit (Beckman Coulter Inc, Brea, CA). Indexed RNA-seq libraries were quantified by digital PCR and sequencing was carried out on a Hiseq 2500 (single-end, 150-bp per read) at the Cornell Core Facility for RNA-sequencing.

RNA-Seq alignment, coverage and differential expression analysis
Sequence reads were aligned to a 10403S genome using the BWA mem algorithm in BWA version 0.7.3a [33] and the data for coverage per base on sense and antisense strands were analyzed separately using samtools [34].

Promoter search using a Sliding Window approach
In order to identify further σ H -dependent promoters that may not result in differential expression of the actual gene due to the coexistence of σ A -dependent promoters regulating a given gene, a sliding window approach was implemented. Three different window sizes and window sliding values were used (window sizes = 26, 51 and 102 nt; window sliding values = 13, 25 and 50 nt, respectively), which all resulted in the identification of the same promoters. However, the combination of window size = 51 nt and window sliding = 25 nt resulted in a lower number of fragments that were not preceded by upstream promoters (which was considered a "false positive" finding); a sliding window size of 51 yielded 3 false positive fragments as compared to 18 and 10 with sliding windows sizes of 25 and 100, respectively. Therefore, results obtained with window size = 51 nt and window sliding = 25 nt are presented here. Briefly, the 10403S genome was divided into windows of 51 nt (window size) with 25 nt overlap (window sliding) and the RNA-Seq coverage was obtained for each of the 116,123 resulting windows. RNA-Seq coverage was obtained as described above and the coverage per nt was used to obtain the total coverage of each window (e.g., the coverage of the window ranging from nt 1 to nt 51 is the sum of the coverage of each nt between nt 1 and nt 51). Bayseq version 2.2.0 was then used, as described above, to identify windows with significant differential expression between the ΔBCHL::P rha-sigH strain and the ΔBCHL::P rha strain. Fold change was also calculated as described above and windows presenting an FDR < 0.05 and FC > 2.0 were considered for further analysis. Windows matching the required thresholds were then sorted and overlapping windows were considered part of the same fragment. Fragments that were mapped to the same gene or operon were considered as being part of the same transcriptional unit (TU). The 5' end of significant TUs were then manually scanned for promoter sequences located 5 to 30 nt upstream the identified transcriptional start sites using the genome browser Artemis to identify  Table S3).
Identification and characterization of σ H -dependent promoters in additional L. monocytogenes and Listeria spp. genomes A total of 28 finished genome sequences representing 10 L. monocytogenes lineage I strains, 10 L. monocytogenes lineage II strains (including 10403S), 4 L. monocytogenes lineage III strains, 1 L. innocua strain, 1 L. ivanovii strain, 1 L. welshimeri strain and 1 L. seeligeri strain were retrieved from NCBI GenBank database (Additional file 2: Table S2). Standalone Blast searches using the σ H -dependent promoter sequences or their respective regulated genes as queries were carried out against two databases containing (i) all 24 L. monocytogenes genomes (L. monocytogenes database) and (ii) all 4 non-L. monocytogenes genomes (Listeria spp. database). Blast searches against both databases were carried out with the following parameters: "Expectation Value (E)" set to 0.2, "Word Size" set to 8, and "Filter Query Sequence (DUST)" set to FALSE. For searches against the L. monocytogenes database, the "Penalty for Nucleotide