Black pepper and tarragon essential oils suppress the lipolytic potential and the type II secretion system of P. psychrophila KM02

Characteristics of the P. psychrophila KM02 genome and pangenome

The growth and metabolic activities of microorganisms are considered the main causes of fish-based product spoilage. Our previous work identified P. psychrophila as a specific spoilage organism in the salmon microbiome⁸. At temperatures that can occur during the processing and storage of fish, the enzymatic activity of pseudomonads can result in off-odors³⁴. However, determination of the physiological characteristics of P. psychrophila in in vitro and in situ conditions requires detailed investigation. WGS and de novo assembly provide complete information about bacterial strains isolated from various sources and increase the value of biological studies¹². These are important because foodborne microorganisms are subjected to harsh conditions in the food chain³⁵. Therefore, in this study, WGS technology was applied to understand the spoilage potential of the KM02 strain.

Initially, KM02 was isolated from salmon fillets kept in a cold environment, and the complete genome sequence was obtained by combining the Ilumina and MinION platforms. Reads were de novo assembled, resulting in one scaffold with 20 × coverage. A complete genome sequence of KM02 comprised 5,313,922 base pairs of a circular chromosome with a 57.4% G + C content (Fig. 1), coding 4,813 total genes, out of which 4,713 were protein coding DNA sequences (CDSs) and 54 were pseudogenes. No plasmids were found. Other chromosome features of KM02 are presented in Supplementary Table S2. In comparison to KM02, a previously sequenced genome of the P. psychrophila HA-4 strain also isolated from a cold environment, consisted of 5,235,696 bases with a mean G + C percentage of 56.4% and 4,721 predicted coding sequences³⁶. Similarly, the draft genome of P. psychrophila MTCC 12,324 isolated from the Arctic was composed of 5,269,174 bases, with a mean G + C content of 57.52%³⁷. Given the above, it can be concluded that P. psychrophila species genome characteristics are rather equal between strains, in contrast to other species, such as Escherichia coli³⁸. However, only small variations in P. psychrophila genome features were able to be identified by analyzing the small number of available sequenced strains; many more sequences strains are available for E. coli species, and identification of genome variations is dependent on the number of genomes available for analysis³⁸. Nevertheless, according to the work of Wessels et al.³⁹, who sequenced 35 various Pseudomonas spp. fish- isolates, the genome sizes and number of genes ranged from 4,505,98 bp to 6,279,60 bp and 4,123 to 5,874, respectively³⁹. Similar values were also obtained for genomes of Pseudomonas fragi and Pseudomonas lundensis isolated from spoiled meat and milk samples⁴⁰. Based on the summary of the available sequenced bacterial genomes, such values are considered average, characteristic for bacteria isolated from environmental sources³⁸.

To assess the essential genomic elements of P. psychrophila species, pangenome analysis was performed. There were 8 P. psychrophila genome assemblies in the GenBank NCBI database. Due to problems with gbff file validation HA-4 assembly (GCA_000282975.1) was excluded from the pangenome analysis. Details on the contribution of specific P. psychrophila genomes to the pangenome of this species are depicted in Table 1. P. psychrophila was characterized by 3914 core genes (shared by all strains), while strain KM02 had 548 accessory genes (shared by two or more strains, but not all), 2 unique and 0 absent genes. In the most distinct strain, MF6762, the number of unique genes reached 574. Interestingly, this strain has also been isolated from food (raw chicken), while other sequenced P. psychrophila strains were isolated from cold environments (rooms of food storage or cold water). As presented in the core-pangenome plot (Supplementary Fig. S2), P. psychrophila had a small variations among strains and varied only in the range of less than 2 Mb of gene families. This indicates a rather confined and homogeneous group of P. psychrophila strains in the context of gene products³⁸. According to the outcomes of the attribution of the Clusters of Orthologous Groups (COGs) functional categories to the P. psychrophila pangenome (Fig. 2), the highest percentage of unique genes was related to the ‘replication, recombination and repair’ category. Simultaneously, the fraction of pangenome core, after the ‘genes of general function prediction only’ category, was the most abundantly represented in ‘amino acid transport and metabolism’ (above 10%). Similar results were obtained for the pangenome of fish-pathogenic Aeromonas hydrophila strains, in which the core genome was represented by 9.61% of genes annotated to the ‘amino acid transport and metabolism function’ COG category⁴¹. This functional group was also well represented in the accessory genome of P. fragi, which has been recognized as a contributor to the spoilage of fresh meat and fish and pasteurized milk by secreted lipases and proteases⁴⁰. Regarding lipid transport and metabolism, the core, accessory and unique percentages of representatives in P. psychrophila pangenome were equally distributed. Furthermore, based on KEGG pangenome analysis (Fig. 3), a high metabolic activity of P. psychrophila (almost 70% of the pangenome) was also confirmed. KEGG detailed distribution revealed that the most enriched metabolic functions were ‘amino acid metabolism’, ‘lipid metabolism’ and ‘membrane transport’ and they accounted for approx. 13%, 4%, and 8% of the core genes, respectively.

A K-mer phylogenetic tree based on the genome assembly sequences showed the closest evolutionary relationship with P. psychrophila BS3667 (Fig. 4) with 99.9181% identity; however the sample and isolation source of this strain is not known.

To characterize the KM02 genome, a number of available bioinformatic tools were applied. Similar to pangenome analysis, the distribution of COG of the KM02 strain was also determined. As presented in Fig. 5, gene products with ‘unknown function’ and with ‘general function prediction only’ comprised approx. 40% in total. Therefore, to clarify the significance of the remaining known function assigned to the COG, the numbers of those families were subtracted from the percentage calculations. Consequently, ‘amino acid and lipid transport and metabolism protein’ orthologs accounted for 9 and 3% of all categorized proteins, respectively. COG results for KM02 are in agreement with the genome of other seafood spoilage contributors, such as Shewanella baltica isolated from spoiled shrimp⁴². These psychrotrophic bacteria, similar to Pseudomonas spp. dominate in spoilage of iced-stored fish meat^43,44 and the COG category of ‘amino acid’ and ‘lipid transport and metabolism’ represented 8.66% and 3.8% of the genome, respectively⁴². In regard to the proteins related to those activities, the most dominant annotations involved COG1028 (dehydrogenases with different specificities), COG0834 (ABC-type amino acid transport/signal transduction system), COG0665 (glycine/D-amino acid oxidases), COG1280 (putative threonine efflux protein), COG0612 (predicted Zn-dependent peptidases), COG0006 (Xaa-Pro aminopeptidase), COG1686 (D-alanyl-D-alanine carboxypeptidase), and COG0024 (methionine aminopeptidase) among others. Similarly, in the genome of Pseudomonas fluorescens SRM1 isolated from spoiled milk, the operon containing proteases, lipases and the ABC-transporter, which directs enzyme secretion, was identified⁴⁵. High enzymatic activity is required for efficient utilization of complex compounds from which bacteria produce energy. Furthermore, Gene Ontology system was used to determine the biological relevance of genes and gene products. In KM02, 870 GO terms were assigned to biological process, 77 GO terms to cellular component and 680 GO terms to molecular function (Supplementary Fig. S1). Among the ‘biological process’ GO terms, the most abundant were genes involved in ‘metabolic process’ (GO:0008152) and ‘cellular process’ (GO:0009987) as they consist of ancestor annotations of predicted genes, which are also annotated to child GO terms representing more specific entities⁴⁶. Among the child terms that confirmed a wide spoilage potential of KM02, those containing: ‘catabolic’, ‘proteolysis’, ‘protein’, ‘lipid’, and ‘fatty acid’ phrases were revised and selected. In the ‘organic substance catabolic process’ GO term, the most abundant genes were genes involved in chemical reactions and pathways of organonitrogen compounds, organic acids, organic cyclic compounds, carbohydrates, organophosphates, macromolecules, proteins and lipids (Fig. 6A). In ‘cellular catabolic process’ group, which indicates the activity of individual cells, the most abundant pathways were pathways resulting in the degradation of aromatic compounds, nitrogen compounds, drugs, neurotransmitters, macromolecules, peptides, and sulfur compounds (Fig. 6B). As reported in Liu et al.⁴⁷, the transcriptome of the P. fluorescens strain strongly associated with food spoilage differed from the RpoS-mutant strain in regard biological processes; the greatest differences were seen in GO biological processes such as signaling, protein catabolic process and secretion. Because RpoS contributes to the spoilage activities of P. fluorescens⁴⁸, we can conclude that the abovementioned significantly downregulated genes are strongly involved in the spoilage of foods. Regarding the secretion system, according to the cellular component GO categories, the presence of the T2SS complex and transmembrane transporter activity molecular function was also noted. Among the other molecular function terms, the most abundant were ‘catalytic activity’, and those important for protein and lipid degradation were ‘hydrolase activity’, ‘catalytic activity, acting on a protein’, and ‘hydrolase activity, acting on ester bonds’ (Table 2). The functions of the assigned genes were also deduced on the basis of the sequence similarity of their presumptive protein products to the protein motifs in the Pfam database⁴⁹. From among a wide range of annotated Pfam protein domains, only those related to hydrolase activity, protein secretion, lipid degradation and other fish spoilage aspects were selected, and their genome abundances are presented in Table 3. These data will significantly complement the current knowledge on the lipolytic activity of pseudomonads. Highly enriched Pfam domains were involved in the hydrolase activity represented by Aminohydro_1 and Abhydrolase_1 with relative abundances of 11,828 and 7768, respectively. Furthermore, most of the Pfam annotations assigned to the KM02 genome were related to peptidase activity, e.g., Peptidase_M20, Peptidase_M24 or Peptidase_M23. Most of these enzymes are classified as metallopeptidases, whose catalytic activity involves metals⁵⁰. Proteins and lipids degraded to smaller molecules such as oligopeptides or single fatty acids are further metabolized by bacteria to form derivatives with undesirable odors. For example, ELFV_dehydrog represents the family of dehydrogenases of amino acids that catalyze the oxidative deamination of an amino acid to its keto acid analogs, known from spoiled fish⁵¹. In the context of lipid degradation, Pfam results revealed the presence of proteins representing Lipase_3 and Lipase_GDSL domains with abundance values of 2131 and 1246, respectively. Our results are in agreement with the work of Lo et al.⁴⁵, who sequenced the P. fluorescens SRM1 strain and found that its genome contains heat-stable lipases encoded by lipA and lipB genes, which are responsible for spoilage of raw milk⁴⁵.

Effect of subMICs of BPEO, TEO, and their major compounds on P. psychrophila KM02 lipolytic activity

In light of the increasing use of EOs as modern fish biopreservatives, the current study assessed the anti-lipolytic potentials of BPEO and TEO and their major compounds toward KM02. Although the spoilage of fishery products is mainly caused by gram-negative microbes⁵¹ it is advisable to inhibit pseudomonad metabolic activity in seafoods. For this study, the subMIC concentrations of all agents were used; the concentrations of EOs above subMIC levels in foods can be sensorily unacceptable for consumers⁵². Many studies have demonstrated that plant EOs (e.g., oils of cinnamon and glove) can suppress bacterial metabolic activities/production of virulence factors when used at subMIC concentrations⁵³. However, to date the inhibition of lipase production by plant- derived antimicrobials has only been shown in Serratia marcescens and P. fluorescens cultures^54,55.

To investigate whether the analyzed compounds changed the lipolytic activity in KM02 cells, first, the spectrophotometric method with p-NPP reagent was used. As presented in Fig. 7, all treatments resulted in a considerable lipolysis decrease depending on the compound and medium, and it ranged from 11 to 46%. The highest inhibition potential was observed for the bulk of BPEO and TEO applied in modified TSB medium, and it was approximately twice that of major compounds used alone. The anti-lipolytic action of EO obtained from juniper toward fish-related P. fluorescens was also investigated, where the whole oil inhibited lipase production by 45%, while its major compounds, i.e., α-pinene and sabinene were significantly less effective⁵⁵. A similar outcome was observed in the context of proteolytic enzyme inhibition in KM02, which was reduced to less than 20% by PHE and 28% by CAR, while TEO and BPEO resulted in significantly higher effects^8,14. The explanation of such findings may be related to the lower antimicrobial effect of single terpenes, which was also seen in in vitro tests⁵². Notably, a combination of two different EO constituents or the presence of minor components in the entire EO volume can cause additive or synergistic antimicrobial effects⁵⁶. For example, the inhibitory effect of LIM on P. aeruginosa was enhanced by the addition of on equal volume of eucalyptol⁵⁷.

In fish juice medium, the overall anti-lipolytic activity was significantly (p < 0.05) lower than that under in vitro conditions, with the exception of ME. However, based on the results, this compound inhibited lipolytic activity by only 17%. The decreased antimicrobial effectiveness of EOs and their major constituents is probably due to the more complex composition and physiochemical characteristics of the medium extracted from fish fillets. According to our previous study⁵⁵, fish juice medium constituted 1.8 mg/g protein and 0.0635 mg/g lipids, which ideally mimics fish muscle conditions. These food components usually considerably reduce EOs bioactivity and protect bacteria by absorbing some volume of added EOs⁵⁸. Therefore, to maintain equal antimicrobial efficacy in real food matrices where high molecular compounds are present, higher concentrations of antimicrobials are needed⁵⁹. Furthermore, P. psychrophila was the least sensitive analyzed fish isolate for rosemary extract applied in a food model of common carp fillets, and its lipolytic potential was arrested by only one day in relative to the control culture⁷.

Similarly to Actinobacter baumannii and Vibrio cholerae strains, the genome of KM02 harbors genes encoding lipases (lipA, lipB)⁶⁰. In this work, the anti-lipolytic activity of subMICs among the compounds was verified by RT-qPCR experiments and evaluation the expression of genes encoding lipases (lipA, lipB) in KM02 was evaluated. Based on the Pffafl calculations, the ratio of the expression of the lipA gene encoding the major lipase synthase ranged from 0.1 to 0.9 regardless of the culture medium used (Fig. 8). The highest decrease in lipA gene transcription was observed in cells treated with BPEO and TEO, which confirmed the phenotypic observations. Because the lipA and lipB genes are linked in a single operon, a disruption of even one of them results in a lipase-negative phenotype⁶¹. In the previous work, the LipB gene was also downregulated to the highest extent by whole EOs. According to the work of Christensen et al.⁶², the absence of the LipB protein in Serratia proteamaculans resulted in no spoilage of milk-based products. The aforementioned works indicate that changes in the expression of the lipA and lipB genes may result in a reduced rate of food biodeterioration⁶.

Effect of subMICs of BPEO, TEO, and their major compounds on P. psychrophila KM02 T2SS

LipA and LipB are known type 2 substrates that degrade lipids⁶⁰. Ogierman et al.⁶⁰ noticed that a lipase-deficient lipA mutant of Vibrio cholerae was not able to grow on olive oil, and complementing the T2SS mutant with a plasmid expressing LipAB did not reverse this defect suggesting that secretion of lipases is T2SS-dependent. T2SS apparatus proteins are considered antimicrobial targets; thus in this work changes in the mRNA levels of T2 secretome genes in KM02 cells were evaluated. For RT–qPCR experiments, KM02 cells were treated with subMICs of BPEO, TEO and major compounds of the oils both in vitro and in fish juice medium mimicking the seafood ecosystem.

Based on bioinformatic analysis of the KM02 genome, the following 7 potential genes responsible for T2SS function were identified: pulG and gspG genes encoding pseudopilins PulG and GspG respectively; tadB1 and tadC1 genes encoding integral proteins of the inner membrane involved in the general secretion pathway (GSP); gspH2 and gspH1 genes encoding proteins required for energy-related secretion from the periplasm; and pulF gene involved in lipase export (Supplementary Table S3). As presented in Fig. 9, regardless of the agents used, the expression levels of most T2SS genes were inhibited to relative levels of between 0.9 and 0.02 that of control. Under in vitro conditions the most efficient treatments for downregulation of the expression of T2SS genes were BPEO and LIM, with relative transcript levels ranging from 0.2 to 0.02 that of control. In fish juice medium the highest reductions in the mRNA levels of T2SS genes were recorded for TEO and its singular components (i.e., ME and PHE). The most considerable inhibition concerned gspH1 (0.05) and gspH2 (0.02). These results are in line with the work of Jain, Nale and Dabur⁶³, in which the response of pseudomonads to natural antimicrobials was evaluated at a proteomic level. Downregulation of proteins involved in secretion systems (e.g. xcp, PilS) in P. aeruginosa was caused by water extracts of the active fraction of catechins from Saraca asoca flowers⁶³. Additionally, in the work of Singh et al.⁶⁴, the authors noted the role of thyme EO in targeting the virulence arsenal regulated by the T2SS of Xanthomonas oryzae pv. oryzae strains. The downregulation of virulence gene expression in Xanthomonas strains remained insignificant when the bacteria were treated with thymol alone⁶⁴.

Changes in P. psychrophila KM02 growth in salmon-model products with subMICs of BPEO/TEO

The observed effect of BPEO and TEO on the T2SS-dependent lipolytic activity of the KM02 strain, triggered the need to verify the biopreservative properties of the examined agents in salmon-model products. The application of EOs in food requires additional techniques to mask their strong odor and simultaneously ensure their effectiveness⁶⁵. In this work, an oil-vinegar marinade supplemented with subMICs of BPEO and TEO was formulated to improve the quality of fresh salmon fillets. The model product inoculated with standard amount of KM02 was next packed under vacuum conditions to prevent the degradation of EO components by oxygen.

The KM02 count after 1, 3, and 5 days of storage were evaluated (Fig. 10). After 1 day from an initial value of 4 log CFU/g, the KM02 control culture reached 4.8 log CFU/g, while BPEO and TEO impeded cell proliferation to 4.2 and 4.3 log CFU/g, respectively. At 3 and 5 days of storage, both treatments resulted in a significant (p < 0.05) reduction in KM02 counts in relation to the control product, where only marinade and vacuum packaging were applied. KM02 cells in the control sample reached a critical spoilage value of 6 log CFU/g between 3 and 5 days of refrigerated storage. BPEO- and TEO-treated samples did not exceed the value of 5.5 log CFU, which indicates the antimicrobial effect of marinade supplemented with EOs. Interestingly, even as strictly aerobic bacteria, KM02, was still able to proliferate under vacuum conditions. A substantial number of pseudomonad cells, despite vacuum packaging, were also observed in refrigerated trout fillet⁶⁶. This is probably due to an inadequate barrier material used for packaging or not completely evacuating the gas from samples and the ability of cells to thrive in microaerophilic conditions. However, in comparison with aerobic storage, reducing the amount of oxygen results in a considerable decrease in microbiological counts and is an effective method for fish preservation based on hurdle technology⁶⁷. Inhibitory effects of marinades enriched with oregano, rosemary and juniper EOs on the growth kinetics of psychotropic bacteria- contaminated foods were also observed by Siroli et al.⁶⁸. In that work, marination showed the highest inhibition against Pseudomonas spp. and total coliforms. The molecular studies of Wu et al.⁶⁹ revealed that selective compounds of EOs may competitively interact with the ATP binding site of the DNA gyrase B subunit of bacteria. Thus, natural antimicrobials combine with DNA to form a complex that eventually leads to DNA degradation, blocking cell transcription and replication⁷⁰. Some bioactive agents may also cause the rearrangements of the nucleic acid double chain²¹. Interference with nucleic acids by bioactive compounds of EOs regulates bacterial metabolism and proliferation²¹. Moreover, aside from the ability to improve the safety and shelf-life of marinated fish, the utilization of EOs may also enhance consumers’ willingness to buy, in light of the recent increasing consumption of clean-label products⁷¹.