Water samples were collected at six points along the P1 transect in NPO onboard the R/V Dongfanghong 3 from October 31 to November 4, 2019 (Figure 1A). Comprehensive data for sampling sites was contained in the Supplementary Appendix Table 1. A total of sixty two seawater samples from six vertical sites (>8 samples per site) were collected using a 12-liter Niskin bottle and connected to an electroplating sampler with SeaBird CTD (SBE 911 model) to measure water depth, temperature, salinity, and Dissolved Oxygen (DO). All the seawater samples were divided into three groups, i.e., the ocean surface mixed layer (UL, 5-105 m; 19 samples), the pycnocline and North Pacific Intermediate Water layer (ML, 155-700 m; 15 samples), and the bathypelagic zone (BL, >1000 m; 28 samples). Approximately 1 L of seawater was filtered through 3μm and 0.22μm polycarbonate membranes (GTTP, 47 mm, Ispore) using a vacuum pump under low, non-disruptive pressure (< 5mm Hg). All filters were immediately frozen and stored in −80 °C onboard and transferred to a −80 °C freezer in the laboratory until DNA extraction. Samples for nutrients were collected, and the nutrients in each sample were measured based on the classical colorimetric method (Grasshoff et al., 2009), including NO2−, NO3−, NH4+, Dissolved Silicon (DSi) and Dissolved Inorganic Phosphorus (DIP). Water samples (500 ml) for Chlorophyll a (Chl a) analysis were filtered through a GF/F filter using a vacuum pump (< 10mmHg). Then, the filters were wrapped in aluminum foil and stored in the dark at −20 °C. They were extracted with 90% acetone and kept in the dark at 4 °C for 24 h, after which the concentrations of Chl a were determined by a Turner Designs Trilogy fluorometer (Parsons et al., 1984).

Using 3-μm and 0.22-μm polycarbonate membranes, DNA was extracted by the protocol described by Yin et al. (2013). Each polycarbonate membrane that seawater was filtered through was cut into pieces under sterilized conditions. A sterile tube was filled with 500 μl of sodium chloride-Tris-EDTA (STE buffer. Using FastPrep-24 homogenization equipment (MP Biomedicals, Irvine, California, U.S.A.), the solution was rapidly shaken twice to encourage cell lysis, resulting in a hydrolysis rate of 6.0 m/s. The DNeasy Power Water Kit (QIAGEN, U.S.A.) was then used to extract DNA according to the manufacturer's instructions. Then, the quantity and quality of the extracted DNA were detected by Nanodrop-2000 Spectrophotometer (ND-2000; Thermo Fisher Scientific), and the DNA samples were preserved at −80 °C until used.

16S rRNA gene-targeted qPCR was used to evaluate abundance of total Vibrio spp. Each DNA specimen was measured using the QuantStudio^TM 5 System (Applied Biosystems) and QuantStudio^TM Design and Evaluation Software. Specific 16S rRNA oligonucleotide primers for the genus Vibrio, V567F and V680R, were used in qPCR with SYBR-green detection (Thompson et al., 2004a; Vezzulli et al., 2015). The reaction system and conditions were performed according to Wang et al., 2022. The 16S rRNA genes of Vibrio rotiferianus WXL191, a species found in our lab, were used to create standard curves. According to Wang et al., 2022 technique each plate had standard curves and No-Template Control (NTC), in which ddH₂O served in place of the template DNA. To ensure the reliability of the findings, each DNA sample conducted three rounds of qPCR analysis. The qPCR assay's amplification efficiency indicated an R² value higher than 0.99, with values varied between 95% - 100%.

High-throughput sequencing aids in assessing microbial diversity in Vibrio species, using next-generation technologies like Illumina and Ion Torrent for population analysis, species identification, and ecological role investigation. Vibrio-specific primers V169F and V680R (Siboni et al., 2016) were amplified in the hypervariable regions of V2—V4 of the 16S rRNA gene to ascertain the general makeup of the Vibrio community. Using agarose gel electrophoresis, positive amplicons were verified. The PCR products were purified from 2% agarose gels using the AxyPrep DNA gel extraction kit (Axygen Biosciences, Union City, CA) and further quantified by QuantiFluor-ST (Promega) following the manufacturer's protocol. And, the amplicons were then pooled in equimolar and paired-end sequenced (2 × 300) on an Illumina Miseq PE300 platform at Majorbio Bio-Pharm Technology. Following the application of FLASH to combine raw fastq files, UPARSE (Version 11) was utilized to cluster operational taxonomic units (OTUs) at a 97% sequence similarity level (Gyraite et al., 2019; Zhu et al., 2023). The UCHIME programmed was employed to determine and dislodge chimeric sequences (Edgar et al., 2011). Using a minimum confidence level of 70%, the RDP classifier (Wang et al., 2007) was used to assign the taxonomy of each representative OTU 16S rRNA gene sequence against the SILVA 138 16S rRNA database (http://www.arb-silva.de). A more precise taxonomic identification was obtained by reassigning the Vibrio sequences to the EzBioCloud database (https://www.ezbiocloud.net/). With a “single rarefaction” QIIME script, sequences were subsampled based on the bare minimum of sample sequences for each sample to remove the impact of sampling effort on the analysis (Caporaso et al., 2010).

To minimize biases associated with sequencing coverage, the number of sequences for each sample was homogenized to the lowest number (15,839 reads) by running a script in R software. Alpha diversity including Shannon and Simpson indices was calculated using the “vegan” package. The linear correlation between environmental parameters and α-diversity indices was performed using the “psych” package. For Beta diversity, the Principal Co-Ordinates Analysis (PCoA) was performed at the OTU level by using the “vegan” package. The subsequent Analysis Of Similarities (ANOSIM) was performed by using the anosim function with 999 permutations in “vegan”. The relationships between phylotypes and environmental factors were evaluated by db-RDA (distance-based Redundancy Analysis) in Canoco version 5.0. The analysis of the distance-decay pattern for the Vibrio spp. was conducted by using the function “vegdist” (“vegan” package), and Spearman's rank correlation test was used to test the significance of the correlations. The study correlated Vibrio abundance and environmental characteristics using Spearman's rank correlation analysis. To reveal the relationship between environmental factors and microbial communities, the Mantel test based on Pearson's correlations was carried out by the “ggcor” package. Additionally, a null model analysis was carried out to quantify the relative contributions of different ecological processes (Stegen et al., 2013), which was calculated using the “picante” package. The linear correlation between environmental parameters and βNTI was also performed by the “psych” package. Species with significant differences between groups were performed using STAMP (Parks et al., 2014).

The environmental parameters of seawater collected from 6 sites in the NPO were measured (Figure 1B and Supplementary Table 1). The temperature gradually decreased along water columns from the surface (14-29 °C) to deep layers, and tended to stabilize in the BL (1-4 °C). The salinity of UL among all sites ranged from 33.6 (P1-19-1) to 34.9 PSU (P1-19-13), and increased rapidly until the maximum at 105 m, and then, it decreased slowly with depth, reaching the lowest value at 1,000 m. The Dissolved Oxygen concentration (DO) rapidly declined in the ML, and then slowly rebounded under 1,000 m. Additionally, the concentrations of SiO32−, NO3− and PO44− with depth showed consistent patterns. All these concentrations were relatively low in the UL, gradually increased with depth and stabilized at depths deeper than 1,000 m. The concentration of NO2− was relatively stable along the whole depths, and fluctuated significantly within the ML. Interestingly, at site P1-19-1, the temperature and salinity of the UL and ML were significantly lower than those of other stations. Similarly, the variation patterns of the concentrations of SiO32−, NO₃⁻ and PO43− with depth in the water column at site P1-19-1 differed from other stations, and the concentrations in the UL are significantly higher than other samples.

The abundance of total Vibrio spp. in water columns among water layers was detected by qPCR (Figure 2). In general, the abundance of Vibrio at each site was ranged from 1.0 × 10⁷ ± 9.3 × 10⁶ to 3.9 × 10⁴ ± 1.5 × 10⁴ copies/L. At station P1-19-1, the abundance of Vibrio in the water column showed a decreasing trend with depth, from 3.7 × 10⁶ to 1.4 × 10⁵ copies/L. At stations P1-19-5, P1-19-9, and P1-19-13, the abundance of Vibrio fluctuated greatly in the water layers below 80 m (2.1 × 10⁴ to 2.4 × 10⁷ copies/L). As the sampling sites moved southward, i.e., at stations P1-19-21 and P1-19-25, the abundance of Vibrio spp. showed a decreasing trend with depth (from ~10⁶ to ~10⁴ copies/L) and was relatively stable in the BL.

Nine lakh eighty two thousand hundred and eighteen high-quality reads were acquired after merging and filtering raw data for the 62 water samples. The total sequences yielded 611 Operational Taxonomic Units (OTUs) at a 97% sequence similarity. The sequencing coverages of all water samples were above 0.99, indicating that the retrieved sequences could represent most of the Vibrio community in the studied sites. The Shannon, Simpson, Chao 1, and Pielou's evenness indices were calculated to estimate α-diversity (Supplementary Table 2). Though no significant linear correlation was observed between the α-diversity indices and depth (P > 0.05), the community diversity (Shannon and Simpson) and evenness (Pielou's evenness) decreased from surface to ~3,000m layers and increased near the bottom (Supplementary Figure 1). And, the community richness (Chao 1 and Sobs) fluctuated along with depths (Supplementary Table 2). Regression analysis revealed that both Shannon and Simpson indices exhibit a positive correlation with temperature and SiO32−, and Shannon index also positively correlated to PO43−, and NO3− (Figure 3A). As to the β-diversity, the results of Principal Co-ordinates Analysis (PCoA) indicated that total samples were divided into the UL, ML and BL and samples at similar depths were clustered together (Figure 3B). There were significant differences among the UL, ML and BL (ANOSIM, P < 0.05, Figure 3C). The relationship between the similarity of the Vibrio community and the vertical distance of the samples conformed to the distance-decay model (P < 0.001), and the community similarity decreased with the increase of depth (Figure 3E).

To identify specific taxa that contributed to the observed vertical dynamics of Vibrio communities, representative sequences of each OTU were compared against the EzBioCloud database to do the accurate identification. Almost all sequences (96.17%) belonged to the Vibrionaceae family, and 66.25% were assigned to the genus Vibrio. Forty-five abundant species (relative abundance >0.1%) were found in total samples across all sites, and accounted for 97.14% of all sequences (Figure 4A). V. pomeryi occupied the highest relative abundance across all samples (32.45%), followed by V. sagamiensis (16.42%), P. marinum (16.27%), P. phosphoreum (6.95%) and V. caribbeanicus (6.50%; Figure 4A). V. pomeroyi, P. marinum, and V. caribbeanicus were detected along the whole water column, whereas their relative abundances significantly varied among the different depth groups (Tukey, P < 0.001, Figure 5). V. sagamiensis and P. marinum were the dominant species in the UL, V. caribbeanicus, V. campbellii and P. marinum were concentrated in the UL and ML, and V. pomeroyi exhibited a high relative abundance in the BL (Figure 4A). In addition, different from other sites, V. hangzhouensis and Vibrio sp. OTU380 showed high relative abundance in the UL at station P1-19-1.

Distance-based redundancy analysis (db-RDA) was performed to assess the impacts of environmental parameters on the composition of Vibrio communities. Temperature, SiO32−, NO3− and PO43− drove the composition of Vibrio communities (Figure 3D). Further, the Mantel test showed that temperature, salinity, DO, NO3−, PO43− and SiO32− have significant effects on shaping the community diversities of vibrios across the vertical profile (P < 0.05, Figure 4B). The correlation between the relative abundance of Vibrio (>0.1%) and environmental parameters was calculated by Spearman's rank correlation coefficients (Figure 4C). Most of the abundant species showed significant positive correlations with temperature and DO, whereas significantly negatively correlated with depth and concentrations of NO3−, PO43− and SiO32− (Figure 4C). In detail, V. pomeryi, the most abundant species, had significant correlations with SiO32− (P < 0.05). V. caribbeanicus, V. campbellii, and P. marinum, the dominant species in the UL and ML, were positively related to temperature and DO, and negatively to depth, PO43−, SiO32− and NO3− (P < 0.05); whereas Photobacterium phosphoreum, the dominant species in the BL, positively correlated to temperature, and negatively to depth and SiO32− (P < 0.05). The abundant species in the UL at site P1-19-1, Vibrio sp. OTU380 showed negative correlations with depth and salinity, and positive correlations with DO (P < 0.05; Figure 4C).

We used the βNTI metric to quantify the relative importance of deterministic (|βNTI|>2) or stochastic (|βNTI| < 2) factors to community structure. The results revealed that the value of βNTI were mainly distributed between −2 and 2, indicating that stochastic processes dominated the formation of Vibrio communities (Figure 6A). And, there was a significant difference in βNTI between the UL and ML (Kruskal-wallis, P < 0.05), the ML and BL (Kruskal-wallis, P < 0.01; Figure 6A). Meanwhile, the null model was used to explore the compositional processes of Vibrio communities among the UL, ML, and BL, including deterministic processes (i.e., heterogeneous and homogeneous selection) and stochastic processes (i.e., homogenizing dispersal, ecological drift and dispersal limitation). Throughout the water column, drift was the most significant process, governing the process of community assembly (Figure 6B). Heterogeneous selection had a greater impact on the UL and BL compared to the ML (Figure 6B). The relationships between βNTI and differences in temperature, DO and salinity were further analyzed, and the significant correlations were found (P < 0.01; Figure 6C). Increases in temperature, DO and salinity led to the elevated stochasticity of Vibrio community assembly, weakening environmental selection (Figure 6C).

qPCR analysis for Vibrio spp. abundance across different depths and locations has provided valuable insights into the intricate connections between microbial distribution and environmental gradients (Diner et al., 2021). In the vertical distribution of marine microbes, most studies have reported a general decrease in abundance with depth (Treusch et al., 2009). In this study, without any surprise, significantly higher abundances were observed in the UL with the values ranging from 5.66 × 10⁴ to 4.53 × 10⁶ copies/L, and showed a decreasing trend to the bottom in sites P1-19-1, P1-19-21 and P1-19-25. Vertical declines in temperature, salinity, and Dissolved Oxygen (DO) are key factors shaping microbial community composition in the ocean (Brown et al., 2009; Eloe et al., 2011; Siboni et al., 2016; Zhu et al., 2023). As previously reported (Zhu et al., 2023) and shown in Supplementary Table 1, surface waters are characterized by warmer temperatures, moderate salinity, and higher oxygen levels, all of which decrease with depth. Certainly, Vibrio spp. distribution is also shaped by seasonal and interannual changes (Asplund et al., 2011; Liang et al., 2019), and our study is based on a single-time-point sampling which limits a comprehensive analysis of temporal dynamics and highlights the need for further investigation in the future.

Vibrio abundance varies across sampling sites due to unique environmental conditions. In stations P1-19-5, P1-19-9 and P1-19-13, the values below 155m were much higher than those in surface seawater (Figure 2). This phenomenon has been reported in the EITO, where the Vibrio abundance obviously reduced with the increasing depth of water until 2,000 m and slightly raised from 2,000m to the bottom (Zhu et al., 2023). Specific environments like water mass characteristics, biological interactions, lower temperatures, higher hydrostatic pressure, and reduced predation may increase the proliferation of vibrios (Thompson and Polz, 2006; Takemura et al., 2014; Gregg et al., 2018; Sampaio et al., 2022). Indeed, the inorganic nutrients like NO3−, PO43− and SiO33− showed accumulated trend from UL to BL in the NPO (Supplementary Table 1). Additionally, the growth of Vibrio is also positively correlated with high concentrations of organic carbon and total suspended solids (Wong et al., 2019), which may contribute to observed fluctuations (Comeau and Suttle, 2007). Due to the limitations of cruise timing and data sharing, we were unable to obtain data such as Dissolved Organic Carbon (DOC) and Particulate Organic Carbon (POC), and thus could not conduct directly correlated analyses. In future studies, more organic factors would be detected to find the main influence parameters of vibrios.

Different environmental conditions give rise to diverse microbial assemblages, and the dominant species usually show regional distribution characteristics (Wong et al., 2019). A diverse community of Vibrio species has been recorded in the marginal seas worldwide. In the Sydney Harbor estuary, Vibrio sp. OTU13800 and V. mimicus are the dominant groups (Siboni et al., 2016), and V. fluvialis in the Maowei Sea (Liang et al., 2019; Chen et al., 2020; Wang et al., 2020b, 2022). In the Indian Ocean, P. marinum and V. rotiferianus are the most abundant species (Zhu et al., 2023). Differently, V. pomeryi, V. sagamiensis, P. marinum and P. phosphoreum become the dominant species in the NPO (Figure 4A), likely due to the local environmental conditions which were selected for specific species to survive (Zhang et al., 2018; Wang et al., 2019). It has been reported that V. pomeroyi can utilize a lot of recalcitrant organic matters like cellobiose, and it even can grow at 4 °C (Thompson et al., 2003). P. marinum has been considered as specific bacteria to the pelagic environment and has existed from the surface to the deep extreme hydrothermal regions (Huang et al., 2016), and Photobacterium species can produce polyunsaturated fatty acid, cold-adapted lipase, esterase, and antimicrobial compounds (Nogi et al., 1998; Moi et al., 2017) to survive in deep seawater and sediment. Furthermore, V. caribbeanicus exhibited high relative abundance in both the ETIO and NPO, which may be attributed to its broad environmental tolerance (Liang et al., 2019; Zhu et al., 2023). It is worth noting that seasonal and interannual sampling would give more reliable results.

Vibrio communities are primarily structured through vertical stratification of environmental factors (Thompson and Polz, 2006; Li et al., 2020a; Zhu et al., 2023). In this study, a distinct depth-dependent stratification of Vibrio communities was found in the NPO (Figure 4A). While environmental variables including temperature, SiO32−, NO3− and PO43− contributed to shaping Vibrio communities (Figure 3D), stochastic processes (drift and heterogeneous selection) exerted an even greater influence on their overall assembly (Figure 6A). Drift suggests that chance significantly influences species frequencies, particularly in dynamic environments (Martiny et al., 2006), whereas heterogeneous selection may reflect more distinct environmental gradients or greater habitat heterogeneity in these zones (She et al., 2021). Among all species, V. pomeroyi showed the highest abundance in the BL. It can thrive in colder and deeper waters, which is more common in lower depths below 200m, and especially 1,000m (Martin-Cuadrado et al., 2007; Sutton et al., 2008; Beleneva and Kukhlevskii, 2010). The reason might be that V. pomeroyi positively correlated with SiO32− concentration (Figure 1B, Figure 4C), and it could thrive in deeper layers where was the silicate rich environments. In contrast, V. sagamiensis and P. marinum persist in the UL and ML with varying abundances. P. marinum and V. sagamiensis are found in various varieties of marine habitats (Wang et al., 2022; Zhao et al., 2023). For example, P. marinum can adapt its physiology and metabolism to cope with the changing conditions encountered across different depths (Zhu et al., 2023). In this study, P. marinum showed positive correlations with temperature and DO, and negatively correlated to PO43−, SiO32− and NO3− (Fig. 4C). Additionally, V. hangzhouensis and Vibrio sp. OTU380 showed high relative abundance in the UL at site P1-19-1. The reason may be that Vibrio sp. OTU380 negatively correlated to depth and salinity (Figure 4C), and V. hangzhouensis exhibits adaptation to cold environments where the temperature at site P1-19-1 were significantly lower than those of other sites (Figure 1B) (Xu et al., 2009).

In this study, Shannon, Simpson and Pielou's indices increased near the bottom layers (Supplementary Figure 1), and showed positive correlations with PO43−, SiO32− and NO3− (Figure 3A). The vertical motions accompanied from the frontal waves along the complex water masses transferred nutrients from BL to ML (Kouketsu et al., 2007). In the study area, the Kuroshio and the Kuroshio Extension transported warm and saline water, whereas the Oyashio transported cold and low-salinity water mass with low potential vorticity characteristics (Hiroe et al., 2002; Masujima et al., 2003). Due to the different potential temperature and salinity meet along the Kuroshio Extension, there is formed a strong front of water properties (Hiroe et al., 2002; Masujima et al., 2003). These frontal waves may accompany vertical motions, for example, the upwelling associated with frontal waves in the Gulf Stream lifts nutrient-rich subsurface water to shallower depths and enhances biological productivity (Kouketsu et al., 2007). Meanwhile, nutrient enrichment may increase microbial diversity by promoting niche variations within the Vibrio population (Zhao et al., 2023). Depends on the physico-chemical conditions at the sediment-water interface, the resuspension may induce the benthic dynamics of inorganic nutrients, e.g., Dissolved Inorganic Phosphorus (DIP) and Dissolved Silicate (DSi) (Couceiro et al., 2013; Niemistö and Lund-Hansen, 2019). Interestingly, there are significant positive relationships between Vibrio diversity and SiO32−. Silicate is not readily bioavailable to most organisms, and usually plays a crucial role in the growth of diatom (Wear et al., 2015). The observed correlation may reflect an indirect effect of silicate (Figure 3A), where higher silicate supports diatom growth, leading to conditions that favor a more diverse Vibrio community (Wear et al., 2015). However, the mechanisms for such interactions are not well-established (Von Moos and Slaveykova, 2014), and should be taken into consideration in future.