Bacterial strains used for cloning and expression of the scr genes are displayed in Table 1. Information regarding the source and time of isolation of Vibrio strains are shown in Table 2. Most V. parahaemolyticus strains were sent to the Vibrio consultant laboratory hosted at the BfR from official German federal state office laboratories, which investigate seafood for microbiological safety. All strains used in this study were routinely grown on lysogeny broth (LB) containing yeast extract (5 g/L), peptone (10 g/L), NaCl (10 g/L) or LB agar (Merck, Darmstadt, Germany). For selective cultivation of strains carrying the shuttle vector pVv3 (Klevanskaa et al., 2014) or its derivatives, kanamycin sulfate (Roth, Karlsruhe, Germany) was supplemented to a final concentration of 100 μg/mL (w/v).

Identification of V. parahaemolyticus strains was carried out using phenotypical tests, e.g., cultivation on TCBS agar (Oxoid, Wesel, Germany) and Vibrio Chrom-Agar (MAST-Diagnostika, Reinfeld, Germany), growth tolerance against different NaCl concentrations (0, 2, 6, 8, and 10%) in peptone water (Bacto Peptone, Gibco/Thermo Fisher Scientific, Schwerte, Germany; NaCl Merck, Darmstadt, Germany), API20E profiling (biomérieux, Nürtingen, Germany) and oxidase test (Merck, Darmstadt, Germany). Furthermore, PCR targeting the toxR gene (Bauer and Roervik, 2007; Bechlars et al., 2015) and Matrix-Assisted-Laser-Desorption/Ionization Time-of-Flight Mass Spectrometry (MALDI-ToF MS) (Biotyper microflex, LT/SH, Bruker, Bremen, Germany) was conducted as previously described (Jäckel et al., 2020). The test of sucrose fermentation was conducted by using sucrose bouillon with bromothymol blue containing peptone from meat (10 g/L), NaCl (5 g/L), and sucrose (10 g/L) (Merck, Darmstadt, Germany). 5 mL of sucrose broth was inoculated with bacterial strains and incubated for 24–48 h at 37°C. Sucrose-positive strains turn the media from blueish green to yellow.

Genomic DNA (gDNA) extraction from V. parahaemolyticus strain 16-VB00198 was carried out by using the RTP Bacteria DNA Kit (Stratec Molecular, Berlin, Germany) according to the recommendations of the manufacturer and was used for multiplex PCR amplification of the scr gene cluster. PCR reactions for the detection of the scrA, scrB, and scrK genes were performed in a volume of 25 μl with 1x PCR buffer (2 mM MgCl₂), 0.2 mM of each dNTP, 0.5 μM of each primer, and 1.5 U of Dream Taq DNA Polymerase (Fermentas, St. Leon-Rot, Germany). The following PCR running conditions were used: An initial denaturation of 94°C for 2 min, 35 cycles of 94°C for 15 s, 58°C for 30 s, and 72°C for 30 min, and a final elongation step of 72°C for 5 min. The sequences of the used oligonucleotide primers are given in Table 3.

For the amplification of the complete scr gene cluster comprising scrA, scrK, and scrB and the flanking regions, the Phusion^TM High-Fidelity DNA Polymerase Kit (ThermoFisher Scientific, Schwerte, Germany) was used with the primers VparascrAB-Fo and VparascrAB-Re (Table 3). In a volume of 50 μl, 1 x PCR buffer (2 mM MgCl₂), 0.2 mM of each dNTP, 0.5 μM of primers and 1.5 U of Phusion^TM Polymerase (Thermo Fisher Scientific, Schwerte, DE) were used for amplification. PCR conditions were applied as stated below: An initial denaturation of 98°C for 2 min, 35 cycles of 98°C for 10 s, 61.6°C for 30 s and 72°C for 2 min, and a final elongation step of 72°C for 5 min. For further downstream analysis, the PCR products were purified using the RTP Kit INVITEK according to the manufacturer’s recommendations (Stratec Molecular, Berlin, Germany). A 4,656 bp PCR fragment containing the scr operon was obtained from V. parahaemolyticus 16-VB00168. The fragment was digested with the enzymes BamHI and HindIII (Thermo Fisher Scientific, Schwerte, Germany), whose recognition sequences were included in the sequences of the primers VparascrAB-Fo and VparascrAB-Re (Table 3). The digested PCR fragment was inserted into the corresponding sites of the pUC18 cloning vector (Thermo Fisher Scientific, Schwerte, Germany) using T4 ligase (Thermo Fisher Scientific, Schwerte, Germany). After transformation of electrocompetent E. coli GeneHogs^TM cells, transformants containing the recombinant plasmid consisting of pUC18 with the inserted scr operon were obtained on LB agar with ampicillin (Merck, Darmstadt, DE) (100 mg/mL). Plasmid DNA from transformants was isolated using the RTP plasmid Kit according to the manufacturer’s recommendation and subjected to Sanger sequencing (Eurofins Genomics, Ebersberg, Germany) for verification.

The verified scr region was excised from the recombinant pUC18 derivative by BamHI/HindIII cleavage and inserted into the corresponding sites of the pVv3 shuttle cloning vector (accession number HG326273) (Klevanskaa et al., 2014). To avoid re-ligation of the vector DNA, BamHI/HindIII cleaved pVv3 was incubated with alkaline phosphatase (Thermo Fisher Scientific, Schwerte, Germany) before T4 ligase treatment. Ligation mixtures were subjected to transformation to electrocompetent E. coli GeneHogs cells. Transformed bacteria were selected on LB agar supplemented with kanamycin (100 μg/mL) and the presence of the src gene cluster was confirmed by primer walking (Supplementary Table S1) and Sanger sequencing (Eurofins Genomics, Ebersberg, Germany). pVv3 plasmid or derivative isolation was conducted using Qiagen Plasmid Plus Kits (Qiagen GmbH, Hilden, Germany).

Preparation of competent cells of E. coli GeneHogs^TM was done following a standard protocol (Sambrook and Russel, 2001). The preparation of electrocompetent Vibrio cells was conducted as described previously (Klevanskaa et al., 2014). For electroporation, 200–400 ng gDNA and 40–50 μl freshly prepared competent cells were mixed, and incubated on ice for 10 min. Electroporation was performed at settings of 7.5 kV cm^–1, 25 μF, and 200 Ω. Thereafter, 950 μl SOC medium (Sambrook and Russel, 2001) was added, and the bacteria were incubated for 37°C for 1 h. Afterward the bacteria were plated on LB agar under selective conditions, but potential Vibrio transformants were investigated by MALDI-ToF MS before further processing.

Whole-genome sequencing was performed for 17 V. parahaemolyticus isolates. Therefore, isolates were grown in LB and gDNA was extracted using the PureLink Genomic DNA Kit (Invitrogen, Karlsruhe, Germany). Paired-end, short-read WGS (MiSeq, Illumina, San Diego, CA, United States) was performed as previously described (Schwartz et al., 2019). SPAdes de novo assemblies using raw reads and genome annotation were performed using the Pathosystems Resource Integration Center (PATRIC) (release 3.5.39) (Wattam et al., 2017) and the automated Prokaryotic Genome Annotation Pipeline (PGAP) of the National Center for Biotechnology Information (NCBI), respectively.

Average nucleotide identity (ANI) prediction was used for assigning genomes at species level (Rodriguez and Konstantinidis, 2014). Genome sequences were compared pairwise to the genome of the V. parahaemolyticus type strain ATCC 17802 (accession NZ_LATW01000001-51). The ANI online calculation tool¹ was used with default settings (alignment options: minimum length 700 bp, minimum identity 70%, minimum alignments 50; fragment options: window size 1,000 bp, step size 200 bp).

Multilocus sequence typing (MLST) was conducted using MLST Finder (Larsen et al., 2012), which is based on the V. parahaemolyticus scheme of the pubMLST database² (Jolley et al., 2018). Prediction of putative prophage sequences was performed using the PHAge Search Tool Enhanced Release (PHASTER) with default settings (Arndt et al., 2016). Initial plasmid replicon prediction was performed using PlasmidFinder (v 2.0³) (Carattoli et al., 2014). In addition, genomic contigs showing significantly higher sequence coverages than the rest of the contigs were screened for similarities to known plasmids using the BlastN algorithm of the NCBI database All BlastN searches were carried out using the NCBI database (https://blast.ncbi.nlm.nih.gov/Blast.cgi?PROGRAM=blastn&PAGE_TYPE=BlastSearch&LINK_LOC=blasthome) with default settings.

To determine the phylogenetic relationship of the isolates, a CSI Phylogeny (v 1.4⁴) based single nucleotide polymorphism (SNP) tree was prepared. The tool was used under default settings and the exclusion of heterozygous SNPs. As reference for comparison of the sucrose-positive strains the WGS dataset of 16-VB00498 was used. Nucleotide variations were predicted according to the specifications provided (Kaas et al., 2014).

The relation between the sucrose region (scrA-scrK-scrB) was determined by using MEGA X (Kumar et al., 2018) with the Maximum Likelihood method and Tamura–Nei model (Tamura and Nei, 1993). Initial tree(s) were obtained automatically by applying Neighbor-Join and BioNJ algorithms. Bootstrapping was performed with 500 repetitions using all sites.

Genome sequences of the investigated V. parahaemolyticus isolates have been deposited in GenBank at the National Center for Biotechnology Information (NCBI) under the accession numbers given in Supplementary Table S2.

16-VB00198 was assigned to the species V. parahaemolyticus by MALDI-ToF MS and PCR targeting toxR (data not shown). The strain was positive for sucrose utilization when tested with API20E strips. Furthermore, it changed the color from blue to yellow when grown in sucrose bouillon indicating that it is able to utilize sucrose as sole carbon source.

Whole-genome sequencing and ANI calculation showed a result of 98.2% identity between 16-VB00198 and the V. parahaemolyticus type strain ATCC17805 (WDCM00037). Through BlastN searches using published genes involved in sucrose metabolism, a CDS cluster consisting of three scr genes was identified. Nucleotide sequence (accession M30194) containing the sucrose uptake region with scrA and scrB genes of V. alginolyticus (Blatch et al., 1990) showed a nucleotide identity of 97% to a DNA region (contig_16) of 16-VB00198. Further bioinformatic analysis revealed that the strain contains a sucrose utilization gene cluster containing three genes (scrA, scrK, and scrB) arranged head to tail in the order 5′-3′. The scrA gene encodes a sucrose uptake protein of the phosphoenolpyruvate dependent phosphotransferase system (PTS), scrK encodes a fructokinase and scrB a sucrose-6-phosphate hydrolase (Scholle et al., 1987, 1989; Blatch and Woods, 1991; Postma et al., 1993). Upstream of scrA a gene coding for a transcriptional regulator is located whose coding sequence is counter orientated to scrA.

To examine the functionality of the scrA-scrK-scrB cluster, the region was amplified by PCR with a High Fidelity^TM DNA-polymerase using forward and reverse primers containing the restriction sites HindIII and BamHI, respectively. The 4,656 bp PCR fragment was digested with both restriction enzymes and ligated with BamHI/HindIII cleaved DNA of the high copy number E. coli cloning vector pUC18 (Amp^R) (Norrander et al., 1983). After transformation of E. coli K12, a recombinant plasmid with the scr region was isolated from ampicillin-resistant transformants (data not shown).

The verified pUC18 (scrAKB) derivative was used for re-cloning of the 16-VB00198 scrA-scrK-scrB cluster into the Vibrio/E. coli shuttle vector pVv3 (Km^R) (Klevanskaa et al., 2014). Finally, through selection of kanamycin-resistant E. coli K12 two recombinant plasmids derived from pVv3 were obtained. Both plasmids differed in the orientation of the inserted scr fragment with respect to the transcription direction of the lacZα-fragment (Figure 1) and were further verified by commercial Sanger sequencing. In the recombinant plasmid pC50, the 30 bp BamHI-HindIII fragment from the multiple cloning site of pVv3 is inserted on both sides of the scr fragment resulting in an opposite transcription direction of the scr genes and the lacZα-fragment. In contrast, pC98 exhibited the same transcription direction of scr genes and lacZα-fragment. Out of the literature (Blatch et al., 1990), two promoter and three transcription terminator sequences were deduced within the cloned fragment (Figure 1B), which are up to now not experimentally verified.

Escherichia coli K12 strains are unable to utilize sucrose (Schmid et al., 1988). However, it was shown that after introduction of the V. alginolyticus scr gene cluster into E. coli K12, the clones are able to metabolize sucrose (Scholle et al., 1987). The expression of the scr genes of V. parahaemolyticus 16-VB00198 in E. coli K12 was tested in sucrose bouillon. While the original strain (GeneHogs^TM ) and its derivative carrying the vector pVv3 were negative for sucrose utilization, transformants harboring the recombinant plasmids pC50 or pC98 were capable of metabolizing the disaccharide (Figure 2A). To confirm the E. coli K12 transformants, a multiplex PCR targeting scrA, scrB, and scrK was applied using specific primers (Table 3). PCR amplification yielded PCR products of the expected sizes (Figure 2B).

To test if sucrose-negative V. parahaemolyticus and V. vulnificus can be converted to utilize the sugar, plasmids pC50 or pC98 were introduced into selected strains by electroporation using a protocol developed for pVv3 (Klevanskaa et al., 2014). Following transformation of three Vibrio strains (Table 1), kanamycin-resistant transformants were tested for sucrose utilization by growth in sucrose bouillon (Figure 2A). The transformants of the two V. parahaemolyticus strains (BfR-VB00009 and BfR-VB00016) and the V. vulnificus strain BfR-VB00034 were shown to be able to metabolize the sugar. The presence of the scrABK genes was also confirmed by scr multiplex PCR (Figure 2B).

To determine the occurrence of the scr gene cluster among Vibrio strains of our culture collection, a number of sucrose-positive V. parahaemolyticus strains was subjected to PCR application using the above described scr multiplex PCR. In total, 43 sucrose-positive strains, which had been found through biochemical testing were investigated for the presence of the scr genes. From all sucrose-positive strains the three expected PCR products were obtained confirming the presence of the scr DNA region (data not shown). Additionally, 10 sucrose-negative V. parahaemolyticus strains were tested by PCR, but did not yield any visible PCR products confirming that the PCR is specific for the detection of the scr gene cluster.

Whole-genome sequencing was performed to find out if the sucrose cluster is restricted to a small group of genetically related V. parahaemolyticus strains. In total, 17 sucrose-positive strains from our collection originating from different geographical origin, host animals and time point of isolation were selected for further dissection of their genetic basis. The resulting genomes of the strains were analyzed with respect to their phylogenetic relationship and their scr regions. WGS revealed that the genome sizes of the strains varied between 4.9 and 5.9 Mb (Supplementary Table S2). A MLST analysis of all 17 strains revealed that some strains belong to the same MLST sequence type (ST), however, several diverse alleles of housekeeping genes are found (see below).

The genomic region encompassing the coding sequences of the scrA, scrK, and scrB genes including the intergenic sequences (in total 4,027 bp) were analyzed using the Maximum Likelihood method. The sequences were assigned to three clusters, which are indicated by different colors (Figure 3, left half). The nucleotide identity of the scr regions of the strains ranged between 95.8 and 100%.

For determination of the phylogenetic relationship of the strains, the genome of 16-VB00198 was chosen as a reference for a SNP analysis. CSI phylogeny was used for SNP tree generation out of 17 sucrose-positive strains (see Figure 3, right half). In Supplementary Table S3, the SNP counts are displayed numerically. The percentage of the reference genome covered by all isolates is 73.59%, so 4,359,330 nucleotides represent the basis of the SNP comparison.

The SNP tree shows a distinct diversity between the sucrose-positive strains, as the SNP counts range between 7 (minimum) and 40,506 (maximum). The strains 16-VB00184, 16-VB00204, 17-VB00364, and 17-VB00065 differ between each other and all other strains by more than 35,000 SNPs. WGS also revealed that a number of strains with low differences in SNP counts between each other are closely related isolates. Such strains probably belong to the same yet unassigned multilocus sequence type (ST) according to the V. parahaemolyticus pubMLST scheme. In Supplementary Table S2, the alleles of housekeeping genes and their resulting STs are given. For unknown STs, the nearest alleles were given (Jolley et al., 2018).

As expected, the strains belonging to the same ST, are highly identical and exhibit scrAKB region only differing in a few nucleotides. In conclusion, the comparison of all genome sequences based on SNP differences reveals that the presence of sucrose utilizing genes is found in genetically diverse strains and is not restricted to a subset of closely related strains.

Additionally, the ANI values of all strains were calculated using the genome of 16-VB00198 as reference. The ANI values differ between 98.19 and 98.42% when compared to the reference. The genome of strain 17-VB00250 is 100% identical to the reference genome, corresponding to the low the difference in SNPs (only seven SNPs). The ANI values are displayed in Figure 3, however, their distinctiveness is less than that of the SNP analysis.

We compared the scrAKB region of strain V. parahaemolyticus 16-VB00198 to the genome of the V. alginolyticus reference strain ATCC 33787 (accession CP013485). In general, these two Vibrio species are related belonging to the Vibrio harveyi clade (Ke et al., 2018) and sucrose utilization has been used as an important phenotypic feature for biochemical differentiation of these bacteria. The comparison of the scr region of the two Vibrio strain genomes is depicted in Figure 4A. The cloned region is part of a genomic island, which may have been acquired through horizontal gene transfer from V. alginolyticus. In this species all strains are sucrose-positive (Farmer et al., 2015), whereas in V. parahaemolyticus 99% of the strains are unable to utilize sucrose. In V. parahaemolyticus 16-VB00198, the sucrose genes scrA, scrK, and scrB are part of a genomic island of approximately 8.0 kb. The island has an overall nucleotide similarity of >95% to the region of V. alginolyticus. Interestingly, the location of the scr gene cluster was determined to be at the same position in chromosome II of V. alginolyticus and V. parahaemolyticus. In the flanking sequences of the scr genomic island the similarity between the chromosomal DNA sequences of the two species drops to approximately 70%. The identified scr island contains additional coding sequences (CDS) in both species: scrR, a gene coding for a putative transcriptional regulator upstream of scrA, a gene encoding a porin protein possibly involved in disaccharide uptake and a CDS of a hypothetical protein. The direction of transcription of all island genes is depicted in Figure 4A. Downstream of the scrB gene, the genomic island ends in a short pseudogene sequence (110 bp), which is derived from a gene encoding a dihydrofolate reductase. On the 5′ end of the scr genomic island the same genes encoding metabolic enzymes are present in both strains (immediately upstream of the island is a gene for a ß-ketoacyl-ACP reductase). On the 3′ end of the V. parahaemolyticus island the first common gene with a chromosomal gene of V. alginolyticus is a gene coding for a S-glutathione transferase. Two genes present in V. alginolyticus ATCC 33787 between the scrB gene of the genomic island and the S-glutathione transferase gene are absent in V. parahaemolyticus 16-VB00198. In the 17 genomes of the sucrose-positive V. parahaemolyticus strains, the same arrangement of genes is conserved. The overall nucleotide identity of the scr genomic island and its flanking chromosomal region among the V. parahaemolyticus strains is higher than 97% (14 kb of contig_16 of 16-VB00198 from nucleotide position 25,001–40,000). The GC contents of the 8 kb region of the scr island of each V. parahaemolyticus genome (start codon of the hypothetical gene to the stop codon of scrB gene) was calculated and compared to the GC contents of the corresponding whole genome. In all 17 cases the GC contents of the scr island is 1.5% lower than that of the whole genome (see Supplementary Table S2). This result supports the hypothesis of acquisition of the scr island through horizontal gene transfer.

We also performed BlastN searches with the chromosomal scr region of V. alginolyticus ATCC 33878 (accession CP013485; nucleotide position 1,012,100 to 1,027,000) against other published V. alginolyticus genomes. In some V. alginolyticus strains, the region contains the same number and identical arrangement of genes as depicted in Figure 4A. However, in various other V. alginolyticus strains the two genes between scrB and the S-glutathione transferase gene are missing (just as with the sucrose-positive V. parahaemolyticus strains, data not shown).

We were furthermore interested which genes are present in the corresponding chromosome II region of sucrose-negative V. parahaemolyticus strains (Figure 4B). This genomic region displays a greater genetic variability between the strains. Downstream of the gene encoding ß-ketoacyl-ACP reductase two conserved genes encoding proteins of unknow function are common in the analyzed strains (indicated by dark gray boxes, Figure 4B). Downstream of these two hypothetical protein CDS, more genetic variability is found. Four examples of genomes of sucrose-negative strains are shown in Figure 4B: In strains CDC-K4557, FORC-006, and ATCC 17802 different gene clusters follow (see accession for details), whereas in strain CDC_K5579 nothing more than the two common hypothetical proteins CDS are present. In all genomes of sucrose-negative strains two genes for the dihydrofolate reductase gene and the transcription regulator are present upstream of the S-glutathione transferase gene as found in the genome of V. alginolyticus ATCC 33878 (Figure 4B).