Identification of a novel enzyme from E. pacifica that acts as an eicosapentaenoic 8R-LOX and docosahexaenoic 10R-LOX

North Pacific krill (Euphausia pacifica) contain 8R-hydroxy-eicosapentaenoic acid (8R-HEPE), 8R-hydroxy-eicosatetraenoic acid (8R-HETE) and 10R-hydroxy-docosahexaenoic acid (10R-HDHA). These findings indicate that E. pacifica must possess an R type lipoxygenase, although no such enzyme has been identified in krill. We analyzed E. pacifica cDNA sequence using next generation sequencing and identified two lipoxygenase genes (PK-LOX1 and 2). PK-LOX1 and PK-LOX2 encode proteins of 691 and 686 amino acids, respectively. Recombinant PK-LOX1 was generated in Sf9 cells using a baculovirus expression system. PK-LOX1 metabolizes eicosapentaenoic acid (EPA) to 8R-HEPE, arachidonic acid (ARA) to 8R-HETE and docosahexaenoic acid (DHA) to 10R-HDHA. Moreover, PK-LOX1 had higher activity for EPA than ARA and DHA. In addition, PK-LOX1 also metabolizes 17S-HDHA to 10R,17S-dihydroxy-docosahexaenoic acid (10R,17S-DiHDHA). PK-LOX1 is a novel lipoxygenase that acts as an 8R-lipoxygenase for EPA and 10R-lipoxygenase for DHA and 17S-HDHA. Our findings show PK-LOX1 facilitates the enzymatic production of hydroxy fatty acids, which are of value to the healthcare sector.

www.nature.com/scientificreports/ of human ALOX15B displays arachidonic acid 8S-lipoxgenase activity 16 . Mouse Alox8 specifically oxidizes the eighth carbon of ARA and catalyzes the conversion of ARA to 8S-hydroxy-eicosatetraenoic acid (8S-HETE). Euphausia pacifica (North Pacific krill) is the most common krill species found in the Northern Pacific Ocean and one of the few commercially harvested Euphausiids. We previously showed E. pacifica contains 8R-HEPE, 8R-HETE and 10R-HDHA 17 . We also showed that 8-HEPE could be produced enzymatically from EPA in E. pacifica 17,18 . These observations suggest that E. pacifica has a lipoxygenase that metabolizes EPA to 8R-HEPE, ARA to 8R-HETE and DHA to 10R-HDHA. The R-lipoxygenase is rarely found in mammal and plant species. Moreover, ALOX12B is the only R-lipoxygenase found in human 5 . The R forms of HETEs are metabolized by aspirin-acetylated cyclooxygenase or cytochrome P450 enzymes in mammals 19,20 . The genome of krill which is estimated to be about 50 Gb, has not been determined due to its huge size 21,22 . Therefore, to date, the LOX gene of E. pacifica had not been identified.
In this study, we identified LOX gene (PK-LOX1 and 2) of E. pacifica and qualitatively analyzed its gene product PK-LOX1.

Results
Identification of two lipoxygenase genes from E. pacifica. We constructed 42,432 contigs from E. pacifica RNA sequencing reads by assemble using trinity (Supplemental data S1). Total contig size was 52,804,389. N50 contig size was 1487. A search for candidate lipoxygenase genes from 42,432 contigs using blastx identified two potential hits (PK-LOX1 and PK-LOX2). PK-LOX1 encoded a protein of 691 amino acids, which included a PLAT domain and LOX domain. PK-LOX2 encoded a protein of 686 amino acids, which also included a PLAT domain and LOX domain (Fig. 1). The amino acid sequence homology between PK-LOX1 or PK-LOX2 and Alox8 was analyzed by blastp. The results are summarized in Supplemental Figure S1. A multiple sequence alignment of lipoxygenase proteins from North Pacific Krill, mouse, soybean and coral is shown in Supplemental Figure S6. www.nature.com/scientificreports/ Lipoxygenase activity of PK-LOX1. Recombinant baculovirus vectors were constructed for the expression of 6xHis tagged PK-LOX1, PK-LOX2, Alox8 or AcGFP. The expression level of PK-LOX1 protein was monitored every 24 h until 6 days after baculovirus infection. LOX expression peaked 48 h after infection (Supplemental Figure S2). Recombinant PK-LOX1 (80 kDa) was detected by Western blot analysis ( Fig. 2A). By contrast, PK-LOX2 was barely detectable in Sf9 cells. We examined the level of RNA expression of PK-LOX1 and PK-LOX2. The expression of PK-LOX2 RNA was detected in Sf9 cells (Supplemental Table S1). Our analysis showed PK-LOX1 could produce 8-HETE, 8-HEPE and 10-HDHA from 40 µmol/L of ARA, EPA and DHA, respectively (Fig. 2B,C). Alox8 displayed the highest activity for 8-HETE production, whereas PK-LOX1 showed the highest activity for 8-HEPE production (Fig. 2B). The metabolic efficiency of 8-HETE, 8-HEPE or 10-HDHA production from 40 µmol/L of ARA, EPA or DHA by PK-LOX1 was 0.79%, 55.62% or 0.97%, respectively.
HEPE produced by PK-LOX1 from EPA was detected as a single peak by LC/QTOF analysis. By contrast, HEPE produced by Alox8 from EPA could be separated into two peaks (Fig. S3A). Specifically, Alox8 generated two product peaks corresponding to 15-HEPE and 8-HEPE (Fig. S3B). The specificity of omega 13 carbon oxidization activity of Alox8 was lower for EPA and DHA compared with ARA (Fig. S3C).

Discussion
In this study, we identified two novel lipoxygenase genes (PK-LOX1 and 2) from E. pacifica (Fig. 1). PK-LOX1 encodes a protein of about 80 kDa ( Fig. 2A) that oxidizes the omega 13 carbon of ARA, EPA, DHA, GLA, SDA and DPA (Figs. 2, 5). The lipoxygenase activity of PK-LOX1 was higher for EPA than ARA and DHA (Fig. 2B). Stereospecific oxidization of ARA, EPA and DHA mediated by PK-LOX1 generated 8R-HETE, 8R-HEPE and 10R-HDHA (Fig. 3). Lipoxygenase activity of PK-LOX1 was observed at neutral pH and temperatures between 4 to 40 °C (Fig. 4). PK-LOX1 also oxidizes the R position of the omega 13 carbon of 17S-HDHA to produce 10R,17S-DiHDHA (Fig. 5). PK-LOX1 is a novel lipoxygenase that acts as an 8R-lipoxygenase for EPA and 10R-lipoxygenase for DHA. Arachidonate 8-lipoxygenase has been identified from mouse (Mus musculus) 16 and coral (Plexaura homomalla) 36 , however eicosapentaenoic 8-lipoxygenase and docosahexaenoic 10-lipoxygenase had not been identified. We compared the main differences between PK-LOX1 and Alox8. Alox8 displayed only about 12% of the carbon specific oxidation activity for ARA compared with PK-LOX1. However, Alox8 acted as an 8/15-lipoxygenase for EPA and 10/17-lipoxygenase for DHA (Fig. S3). PK-LOX1 displayed an eighth and a tenth of the carbon specific oxidation activity for EPA and DHA, respectively. Moreover, Alox8 is an S type lipoxygenase, whereas PK-LOX1 is an R type lipoxygenase (Fig. 3). S type lipoxygenases are common in mammals and plants, but R type lipoxygenases tend to be confined to marine organisms. For example, arachidonate 8R-lipoxygenase was identified from coral 36 , and 11R and 12R-lipoxygenase activity were found in the eggs of sea urchins 37 . Here, we identified 8R-lipoxygenase from North Pacific krill, which is consistent with the suggestion that R type lipoxygenases are more common in marine organisms.
In this study, we show that PK-LOX1 can generate 10R,17S-DiHDHA from 17S-HDHA (Fig. 6). The enzymatic production of 17S-HDHA and 10S,17S-DiHDHA has previously been demonstrated using soybean lipoxygenase 29,30 . However, this study is the first to report the enzymatic production of 10R,17S-DiHDHA. Previous work showed 10,17-DiHDHA displays an anti-inflammatory effect [23][24][25][26][27] , but the structure-activity relationship of this compound was not determined. The method outlined in this study for the enzymatic production of 10R,17S-DiHDHA using PK-LOX1 will facilitate investigations into the structure-activity relationship. www.nature.com/scientificreports/ Although krill is a key species in marine ecosystems, there is a paucity of genome data for these organisms. Currently, the only available gene database of krill is the transcriptome analysis of Antarctic krill (Euphausia superba) 22 . Thus, the contig data of E. pacifica constructed in this study is a new source of gene information for krill. Nonetheless, the biological function of 8-HEPE in E. pacifica remains unclear. The lipoxygenase gene sequence from E. pacifica will be useful in studying the physiological role of 8-HEPE in krill.
In summary, we have identified two novel lipoxygenases from E. pacifica. One of these lipoxygenases mediates the enzymatic production of hydroxy fatty acids that are beneficial for human health. We believe the identification and characterization of lipoxygenases from various species will form the basis of a novel method for the stereochemical specific synthesis of oxidized fatty acids. As such, marine organisms are a potential source of novel types of lipoxygenase activity.
Determination of de novo RNA sequence. Total RNA was purified from E. pacifica using an RNeasy Lipid tissue mini kit (Qiagen, Tokyo, Japan). The library of E. pacifica for next generation sequencing was made using a TruSeq RNA library prep kit v2 (Illumina, Tokyo, Japan). RNA purification and library preparation were performed according to the manufacturers' instructions. The library was analyzed by Miseq using a Miseq reagent kit v3 (600 cycle) (Illumina). The fastaq data was assembled by Trinity 38 . Contigs encoding similar amino acid sequences to human and mouse lipoxygenases were identified using blastx 39 . Our analysis identified two such contigs, which were named PK-LOX1 and PK-LOX2.

Construction of baculovirus.
PK-LOX1 cDNA on pUC19 vector was amplified using the forward primer (5′-TGT ATT TTC AGG GCG CCA TGG CGC CAA TTA AGG AAA AGAA-3′) that had homologous DNA sequence to ligate to 6xHis-TEV DNA and the reverse primer (5′-AGT GAG CTC GTC GAC GTA GGC TAT ACA CTG ATG GCA TTT GGA A-3′) that had homologous DNA sequence to ligate to pFastBac vector. PK-LOX2 cDNA on pUC19 vector was amplified using the forward primer (5′-TGT ATT TTC AGG GCG CCA TGG TAG CGC TGC GCT GCT TCAA-3′) that had homologous DNA sequence to ligate to 6xHis-TEV DNA and the reverse primer (5′-AGT GAG CTC GTC GAC GTA GGC TAA ATA CTT ATT GCA TTT GGA A-3′) that had homologous DNA sequence to ligate to pFastBac vector. The PK-LOX DNA sequences were cloned into pFastBac1 vector at the Stu1 restriction enzyme site together with the DNA encoded 6xHis-TEV (ATG TCG TAC TAC CAT CAC CAT CAC CAT CAC GAT TAC GAT ATC CCA ACG ACC GAA AAC CTG TAT TTT CAG GGC GCC ATG ) using NEBuilder (NEB, Ipswich, MA, USA). We constructed bacmid including 6xHis-TEV-PK-LOX1 or 6xHis-TEV-PK-LOX2 using the Bac-to-Bac Baculovirus Expression System (Thermo Fisher Scientific).

Cell culture and baculovirus infection. Sf9 cells were cultured with Grace's medium (Thermo Fisher
Scientific) containing 10% heat-inactivated fetal bovine serum (Thermo Fisher Scientific) and antibiotic-antimycotic solution at 27 °C. We transfected 1 µg of Bacmid DNA into 8 × 10 5 Sf9 cells using 6 µL of Cellfectin II Reagent (Thermo Fisher Scientific). After 7 days culture, media was collected as the first passage (P1) of recombinant baculovirus stock. The titer of P1 virus stock was measured using BacPAK qPCR Titration Kit (Takara). The baculovirus stock was amplified to infect Sf9 cells with P1 virus (MOI = 1). We infected baculovirus including 6xHis-AcGFP, 6xHis-PK-LOX1, 6xHis-PK-LOX2 or 6xHis-Alox8 genes into Sf9 cells. After 2 days of culture, cells were collected and stored at − 80 °C to use for analysis of lipoxygenase activity.
Western blot analysis. The details of Western Blot analysis was described previously 40 . Briefly, total protein was extracted from baculovirus infected Sf9 cells using RIPA lysis buffer containing protease and phosphatase inhibitor cocktail (Thermo Fisher Scientific). The protein concentration of supernatant was determined by Bradford assay. One µg of protein was subjected to electrophoresis using a NuPAGE 4-12% Bis-Tris Protein gel (Thermo Fisher Scientific). Anti-6xHis antibody was used for His-tagged protein detection.
Enzyme reaction of lipoxygenase. For enzymatic analysis of lipoxygenase (Figs. 2, 3, 4), we infected baculovirus with 2.5 × 10^7 of Sf9 cells (MOI = 1) and cultured them for 2 days. Cells were collected, divided into 60 microtubes and stored at − 80 °C until required. The stored cell pellet was re-ssuspended in 25  Analysis of the lipoxygenase products. We used LC/QTOFMS (Triple TOF 5600, SCIEX) for qualitative and quantitative analyses of the lipoxygenase products. The conditions used for LC/QTOFMS were described previously 17  Statistical analysis. Statistically significant differences between the experimental groups were identified using one-way ANOVA and Tukey's post-hoc tests.

Data availability
The www.nature.com/scientificreports/