The Microbial Composition of Penaeid Shrimps’ Hepatopancreas Is Modulated by Hemocyanin Free

This article is featured in Top Reads, p.

Many biological processes, such as digestion (1), immune defense (2), host behavior (3, 4), and species evolution (5), are modulated by host-associated microbiomes. Given that most symbiotic microbiota are in the gastrointestinal tract (GIT), microbial dysbiosis could result in neurologic disorders (6), obesity (7), Alzheimer disease (8), atherosclerotic cardiovascular disease (9), and inflammatory bowel disease (10). Thus, host–microbe interaction is termed an evolutionary arms race because host and microbes undergo adaptive changes to survive and maintain equilibrium (11). Most hosts employ multiple selective immune response strategies against pathogenic and symbiotic microbes. To prevent tissue inflammation and entry of pathogenic microorganisms, minimum contact is allowed between microbes and host epithelial cell surfaces (12). To achieve this, well-defined ecological niches between host and microbiota are maintained by means of physical and biological barriers in the forms of mucus secretions, Igs, antimicrobial peptides, epithelial cells, and immune cells (13). These physical and biological barriers in mammals constitute the main factors involved in mucosal immunity that help to maintain host–microbe homeostasis (14, 15).

To exert control over microbiota, most organisms express immune genes and proteins that modulate cellular and humoral immune responses. Mammals produce secretory IgA from plasmocytes that select innocuous or metabolically useful commensal microbes (16) and restrict microbial penetration to localize damaging immune responses (17). Similarly, intestinal goblet cells secrete mucin that forms a thick viscous layer on intestinal epithelial surfaces to prevent contact with bacteria (18). In addition, some lectins (e.g., RegIIIγ) promote spatial separation between the intestinal epithelial surface and bacteria (19). Most organisms recognize microbial-derived ligands, which enable them to mount cellular and humoral innate immune responses, such as phagocytosis, antimicrobial peptide production, and generation of reactive oxygen species (ROS) (20, 21). Among arthropods, some insects (22, 23) and shrimp (24, 25) use memory-like innate immune responses to recognize and respond to the same pathogenic agent upon re-exposure (26). These species use highly diverse proteins, such as fibrinogen-related proteins and Down syndrome cell adhesion molecules (27) for immune response against various pathogens. In penaeid shrimp, various biomolecules and secondary metabolites are used as effector molecules for immune response (28). For instance, the kuruma shrimp (Marsupenaeus japonicus) produces C-type lectins that promote hemocyte phagocytosis and modulate antimicrobial peptide expression via the JAK/STAT pathway to inhibit hemolymph microbiota proliferation (29–31). C-type lectin is therefore regarded as an important component of a primitive form of mucosal immunity in shrimp (32).

The arthropodan and molluscan respiratory glycoprotein hemocyanin (33, 34) exists in the extracellular and intracellular forms (35, 36) and is implicated in numerous immune and physiological functions (37–39). In shrimp, extracellular hemocyanin constitutes ∼95% of total serum proteins (40) and is involved in multiple functions, such as melanization (41), clotting (42), agglutination (43), and toxin neutralization (44, 45), because of molecular diversity (46, 47), alternative splicing (48), and posttranslational modification (e.g., glycosylation) (49). Currently, little is known about the functions of intracellular hemocyanin, especially its role in shaping the microbial composition in crustaceans’ hepatopancreas, the main site of hemocyanin synthesis (50). In this study, we reveal that intracellular hemocyanin modulates energy metabolism and ROS production to modulate the microbial composition in the hepatopancreas of penaeid shrimp.

All shrimp experiments were carried out in accordance with the guidelines and approval of the Animal Research and Ethics Committees of Shantou University.

Healthy and diseased Pacific white shrimp (Penaeus vannamei) were obtained from a local shrimp farm in Shantou (23.28°N, 116.69°E), Guangdong Province, China. The diseased shrimp (Fig. 1A) presented typical atrophied hepatopancreas and empty GIT. Samples (20 each) were processed on the farm before being transported immediately on dry ice to the laboratory for experimentation. Another set of healthy shrimp (weighing 5–8 g), purchased from a commercial farm (Huaxun Aquatic Product Corporation, 23.36°N, 116.66°E, Shantou, China), were acclimatized to laboratory conditions for 2–3 d in artificial seawater (salinity ∼10 ppm and temperature ∼26°C) and fed once daily on a commercial diet (∼35% protein). Shrimp were starved for 24 h before being used for experiments. In challenge experiments, shrimp were i.m. injected with 1 mg of N-acetylcysteine (NAC; MedChemExpress, Monmouth Junction, NJ) or by reverse gavage with 100 μl hydrogen peroxide solution (3%) (Nanjing Reagent, Nanjing, China) or 100 μl of sterile normal saline as control.

Histological examination of the shrimp digestive tract was performed by a commercial company (Hubei Biossci Biotechnology, Wuhan, China). Briefly, the entire GIT (including hepatopancreas, stomach, and intestine) was removed from healthy and diseased shrimp before being fixed with 4% paraformaldehyde for 12 h. Next, tissues were sectioned and stained with H&E. Samples were then examined and photographed with a Pannoramic MIDI light microscope (3DHISTECH, Budapest, Hungary).

dsRNAs were generated and purified using the HiScribe T7 Quick High Yield RNA Synthesis Kit (New England Biolabs, Ipswich, MA). Briefly, specific primers containing the T7 promoter sequence (Supplemental Table I) were used to generate dsRNAs (double-stranded hemocyanin [dsHMC] and dsControl) targeting the P. vannamei hemocyanin gene (GenBank identifier: XM_027383261.1) and enhanced GFP gene (GenBank identifier: U55762.1), respectively. For dsRNA-mediated gene silencing, shrimp were i.m. injected with dsHMC or dsControl (10 μg/shrimp), and hepatopancreas samples were collected at 48-, 72-, and 96 h postinjection (hpi) for immediate analysis or stored at −80°C for later use.

Genomic DNA (gDNA) and total RNA were aseptically extracted from shrimp tissues (hepatopancreas, stomach, and intestine) and bacteria using the TIANamp Marine Animals DNA Kit (TIANGEN, Beijing, China) and the TRIzol Plus RNA Purification Kit (Invitrogen, Carlsbad, CA) following the manufacturers’ protocols. For the high-throughput 16S rRNA (V3-V4 amplicon) sequencing (51) and RNA sequencing, aliquots of gDNA and total RNA were snap-frozen in liquid nitrogen before being shipped on dry ice to a commercial company (Majorbio Bio-Pharm Technology, Shanghai, China) for sequencing on the Illumina MiSeq platform (Illumina, San Diego, CA). The sequencing data were analyzed using the free online Majorbio Cloud platform (https://www.majorbio.com) and deposited in the Sequence Read Archive with BioProject accession numbers PRJNA637166 (RNA sequencing), PRJNA644045, PRJNA667790, and PRJNA690594 (16S rRNA sequencing). The concentrations of total RNA and gDNA were measured on a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Wilmington, MA), and the RNA integrity was assessed by an Agilent 2100 Bioanalyzer system (Agilent Technologies, Santa Clara, CA). The quality of the extracted DNA was examined using 1% agarose gel electrophoresis.

First-strand cDNA was synthesized from total RNA with the EasyScript One-Step gDNA Removal and cDNA Synthesis SuperMix kit (TransGen Biotech, Beijing, China) using anchored oligo(dT)₁₈ primers according to the manufacturer’s instruction. The cDNA samples were used for relative quantitative PCR (qPCR) analysis with gene-specific primers (see Supplemental Table I). gDNA (50 ng/reaction) and gene-specific primers (see Supplemental Table I) were used for absolute qPCR analysis to determine total bacteria abundance and the relative abundance of Photobacterium and Vibrio. The qPCR reactions were carried out on the qTOWER³G RT-PCR system (Analytik Jena, Jena, Germany) with the following program: one cycle at 95°C for 10 min, 45 cycles of 95°C for 15 s and 60°C for 30 s. The relative qPCR data were analyzed by the 2^−ΔΔCT method with the PvEF-1ɑ gene used as the internal control. The absolute qPCR data were quantified using a standard curve prepared using different bacteria strains.

Hepatopancreatic cells were prepared by gentle mincing of the hepatopancreas in 1 ml of 0.01 M PBS (pH 7.2) with a sterile pestle before being filtered through a 150-μm steel mesh, followed by centrifugation at 4°C (150 × g; 10 min) to obtain cell pellets. Harvested cells were further washed at least five times with PBS before being lysed with cold cell lysis buffer (25 mM Tris-HCl [pH 7.4], 150 mM NaCl, 1 mM EDTA, 1% NP-40, 5% glycerol) for 20 min. Samples were then centrifuged at 4°C (20,000 × g; 10 min) and cell lysates were boiled for 10 min with 5× SDS loading buffer (250 mM Tris, 10% SDS, 50% glycerol, 10% 2-ME, 1% bromophenol blue [pH 6.8]), before being separated on 10% SDS-PAGE denaturing gels. After protein transfer onto PVDF membranes (Millipore, Burlington, MA) with the Bio-Rad Mini-PROTEAN Tetra Cell (Bio-Rad Laboratories, Hercules, CA), membranes were blocked for 1 h at room temperature in 5% skim milk dissolved in TBST buffer (20 mM Tris, 150 mM NaCl, 0.1% Tween 2 [pH 7.6]), followed by incubation for 1.5 h at room temperature with primary Abs (rabbit anti-shrimp hemocyanin Ab (home-made) or mouse anti–α-tubulin Ab (Sigma-Aldrich, St. Louis, MO) at 1:3000 dilution in SignalBoost Immunoreaction Enhancer Ab diluent (Millipore, Billerica, MA). After being washed three times with TBST buffer, membranes were incubated with HRP-linked goat anti-rabbit or goat anti-mouse secondary Abs (Thermo Fisher Scientific, Cambridge, MA) at 1:5000 for 1 h at room temperature. Signals were developed using the Millipore Immobilon Western Chemiluminescent HRP Substrate (Millipore) and imaged on the Amersham Imager 600 system (GE Healthcare, Boston, MA). The band intensities of immunoblot were quantified by densitometry using ImageJ software version 1.46r (National Institutes of Health, Bethesda, MD) and normalized to the internal control (tubulin).

Fecal microbiota transplantation (FMT) followed a previous method (52) with some modification. Briefly, the GIT of dsHMC- or dsControl-treated shrimp were washed with 5 ml of normal saline, and the fecal suspension was centrifuged at 5,000×g for 3 min, followed by resuspension of the pellet in 100 μl of normal saline. Normal saline was used as a negative control to challenge both the dsHMC and dsControl group as well as untreated shrimp (US). Recipient shrimp were cultured in seawater treated with an antibiotic mixture (10 mg/l streptomycin sulfate and 5 mg/l ampicillin) before being irrigated with normal saline by reverse gavage and transfer to filtered sterile seawater. Next, 100 μl of the fecal microbiota from the donor shrimp was introduced into recipient shrimp via reverse gavage. After 72 h of FMT, tissue samples (hepatopancreas) and GIT contents of recipient shrimp were collected for DNA extraction and bacterial quantification. The experimental groups were set up as follows (donor to recipient): 1) FMT–normal saline to US (designated NU), 2) FMT–normal saline to dsHMC shrimp (designated NH), 3) FMT–normal saline to dsControl shrimp (designated NC), 4) FMT-dsHMC to dsControl shrimp (designated HC), and 5) FMT-dsControl to dsHMC shrimp (designated CH). Shrimp were fed once daily on a commercial diet, starting at 12 h after the FMT.

Hepatopancreatic cells from dsControl- and dsHMC-treated shrimp, were prepared (as described under Preparation of tissues, cells, cell lysates, and Western blot analyses tissues preparation) and stained with the Mitochondrial Membrane Potential Assay Kit (with JC-1) (Beyotime Biotechnology, Nanjing, China) following the manufacturer's protocol to determine changes in mitochondrial membrane potential (ΔψM). Measurements were carried out at an excitation wavelength of 488 nm, with the emission wavelengths of JC‐monomers and JC‐aggregates observed at 529 and 590 nm, respectively. Data were recorded in terms of percentage JC‐monomers.

Intracellular ROS levels in hepatopancreas cells were determined using the ROS Assay Kit (Beyotime Biotechnology, Shanghai, China) based on the 2′,7′-dichlorodihydrofluorescein diacetate fluocrescence probe. Briefly, freshly isolated hepatopancreas cells from five shrimp were incubated with 5 µM 2′,7′-dichlorodihydrofluorescein diacetate probe in PBS for 20 min at 28°C. Next, cells were harvested by centrifugation and suspended in PBS, after which 100 μl of cell suspension was seeded onto black 96-well plates, and the fluorescence was measured at 485-nm (excitation) and 527-nm (emission) wavelengths on a microplate reader (Synergy HTX; BioTek, Winooski, VT). Another 100 μl of the cell suspension was lysed in 2× lysis buffer (50 mM Tris-HCl [pH 7.4], 300 mM NaCl, 2 mM EDTA, 2% NP-40, and 10% glycerol), and the protein level was quantified using the BCA Protein Assay Kit (GenStar, Beijing, China). The relative level of ROS in the samples was calculated as fluorescence intensity per milligram of protein. All assays were carried out in triplicates and repeated at least three times.

The uptake of glucose by shrimp hepatopancreas cells was determined using the fluorescent glucose analogue 2-NBDG (Cayman Chemical, Ann Arbor, MI). Briefly, shrimp were injected i.m. with 2 μmol 2-NBDG followed by collection of hepatopancreas samples at 2 hpi for preparation of cell suspensions. Glucose uptake (in terms of the proportion of FITC⁺ cells) of samples was measured by flow cytometry on the BD Accuri C6 Plus (BD Biosciences, San Jose, CA) at the FITC channel (excitation 488-nm, emission 525-nm optical filter). Eight shrimp were analyzed per group and repeated at least three independent times.

The ATP levels in hepatopancreas cells were measured using an Enhanced ATP Assay Kit (Beyotime Biotechnology, Shanghai, China) according to the manufacturer’s instructions. Briefly, isolated hepatopancreas cells from five shrimp were lysed with ATP lysis buffer before being centrifuged at 12,000 × g for 5 min at 4°C. Supernatants were collected and mixed with 100 µl of ATP working solution (preincubated for 5 min at room temperature) in 1.5-ml tubes. Next, luminescent signals (ATP levels) were measured by a luminometer (Promega, Madison, WI), and the level of ATP was presented as nmol/mg protein. All assays were carried out in triplicates and repeated three independent times.

All statistical analyses used GraphPad Prism Version 8.0.1 (GraphPad Software, San Diego, CA). Data are mainly presented as mean ± SD, unless otherwise stated. An unpaired two-tailed Student t test was used to determine the significant difference (considered at p < 0.05) between two groups for which data were distributed normally with the equal variance between conditions.

Routine field examination of the Pacific white shrimp (P. vannamei) revealed that the GIT of some diseased shrimp present with atrophic hepatopancreas and empty stomachs and intestines (Fig. 1A). Subsequent histological examination of the diseased shrimp revealed that their hepatopancreases were atrophied with loose hepatic tubules and hemocytic infiltration (blue arrows), whereas the stomach walls were thinner (red arrows) and display some injury to the intestinal epithelial brush border (green arrows) (Fig. 1B). In crustaceans, the GIT plays a crucial role in many physiological and pathophysiological processes, especially in the synthesis of multifunctional hemocyanin by the hepatopancreas (50, 53, 54). Thus, we examined whether atrophy of the hepatopancreas in diseased shrimp had any consequence on hemocyanin expression and GIT microbiota. In diseased shrimp, we found significantly low levels of hemocyanin transcripts in the hepatopancreas (Fig. 1C), stomach (Fig. 1D), and intestine (Fig. 1E) compared with healthy shrimp. Total bacterial abundance was also significantly higher in diseased shrimp, especially in the hepatopancreas (Fig. 1F) and intestine (Fig. 1H), but not the stomach (Fig. 1G). In both healthy and diseased shrimp, hemocyanin expression and total bacterial abundance correlated negatively in the hepatopancreas (Fig. 1O, 1P) but not in the stomach and intestine (Fig. 1Q–T). A significant decrease in bacterial diversity (Shannon index) was found in the hepatopancreas, stomach, and intestine of diseased shrimp compared with healthy shrimp (Supplemental Fig. 1A–C). Most bacteria in the hepatopancreas and stomach of diseased shrimp were opportunistic pathogens, mainly Vibrio (Fig. 1I–K, Supplemental Fig. 1D, 1E). Vibrio constituted 51.36% of the total bacterial abundance in the hepatopancreas of diseased shrimp compared with 21.91% in healthy shrimp (Fig. 1L). Similarly, in the stomach of diseased shrimp, Vibrio represented 26.98% of the total bacterial abundance compared with 6.24% in healthy shrimp (Fig. 1M). However, no significant difference in Vibrio levels was found in the intestines of diseased and healthy shrimp (Fig. 1N). Moreover, significant negative correlation was observed between hemocyanin expression and Vibrio abundance in the hepatopancreas and stomach but not the intestine of both diseased and healthy shrimp (Fig. 1U–Z). This set of data suggests that low hemocyanin levels in the GIT of diseased penaeid shrimp could be the reason for the high relative abundance of Vibrio.

Given that crustaceans’ hepatopancreas express the highest levels of hemocyanin transcripts (53, 54), we used RNA interference to deplete levels of intracellular hemocyanin and explored the effect on microbial composition and abundance. Upon obtaining sustained (48–96 h) knockdown of hemocyanin in the hepatopancreas (Fig. 2A, 2B) and other shrimp tissues (Supplemental Fig. 2A–F), we analyzed the effect on bacterial composition and abundance. As shown by the Shannon index (Fig. 2C), Simpson index, and other microbial diversity indexes (Supplemental Table II), a significant decrease in bacterial diversity was observed after hemocyanin knockdown. At the genus level, distinct bacteria communities (Fig. 2D) were observed in the two groups (i.e., dsControl and dsHMC). The most dominant bacteria at the genus level in the hemocyanin depleted (dsHMC) group were Vibrio, constituting 77.42% of the total average reads, compared with only 23.11% in the control group (dsControl). Conversely, Photobacterium were the dominant bacteria in the control (dsControl) group compared with the dsHMC group (Fig. 2E, 2F). The relative abundance of other bacteria genera (i.e., Halomonas, Chromohalobacter, Salinishaera, etc.) also decreased significantly in shrimp hepatopancreas upon hemocyanin knockdown (Fig. 2F).

When the dynamic changes in relative abundance of eubacteria was examined at different time points (after dsControl and dsHMC injection), total bacterial abundance decreased significantly at 24 h and 48 h, recovering to pretreatment levels at 72 h, followed by a significant increase at 96 h (Fig. 2G). A notable significant change was in the relative abundance of Vibrio and Photobacterium; the relative abundance of Vibrio increased significantly at 7–96 h (Fig. 2H), whereas the relative abundance of Photobacterium in the dsHMC samples was significantly lower at 48–96 h compared with dsControl (Fig. 2I). These results indicate that hepatopancreas microbiota composition is modulated by hemocyanin levels, with Vibrio constituting the abundant species at low hemocyanin levels.

After observing that hemocyanin knockdown changed the bacterial composition in the hepatopancreas, we wondered whether this phenomenon could be reversed by FMT. At 48 hpi of shrimp with dsRNA (dsHMC or dsControl), FMT was performed, followed by determination of total bacterial abundance and relative abundances of Vibrio and Photobacterium in the hepatopancreas and GIT contents (Fig. 3A). With sustained hemocyanin knockdown and FMT (Fig. 3B), a significant decrease in total bacterial abundance in the hepatopancreas (Fig. 3C) and GIT contents (Fig. 3D) of the HC FMT group (dsHMC to dsControl) was observed compared with the CH FMT group (dsControl to dsHMC). Similarly, in the HC FMT group, Vibrio abundance was significantly decreased in the hepatopancreas (Fig. 3E) and GIT contents (Fig. 3F) compared with the CH FMT group, whereas Photobacterium abundance in the hepatopancreas of the HC FMT group increased significantly compared with the CH FMT group (Fig. 3G). For the CH FMT group, the relative abundance of Vibrio increased significantly, whereas that of Photobacterium decreased significantly in the hepatopancreas (Fig. 3E, 3G) and GIT contents (Fig. 3F, 3H), respectively, compared with the NH FMT group (normal saline to dsHMC). These results indicate that, at low hemocyanin levels, changes in hepatopancreas microbiota diversity cannot be reversed by FMT, which shows that intracellular hemocyanin could be a key host factor that modulates crustaceans’ hepatopancreas microbiota.

Crustaceans’ hepatopancreas integrates immune and metabolic processes (55) and is also the tissue that expresses the highest levels of intracellular hemocyanin transcripts (54). To explore the role of hepatopancreas intracellular hemocyanin on biological processes, we performed transcriptome analysis after depleting hemocyanin in penaeid shrimp hepatopancreas. The data showed 410 differentially expressed genes (DEGs) after hemocyanin knockdown; 137 upregulated and 273 downregulated (Supplemental Fig. 2G). Most of the 410 DEGs were enriched in energy metabolism, such as carbohydrate, lipid, and amino acid metabolism (Fig. 4A). Notably, 46 of the energy metabolism-related genes were classified into 10 functional annotations, including lipid regulation, lipid transport, lipid degradation, lipid synthesis, glucose transport, glycogen degradation, glycolysis, amino acid transport, amino acid synthesis, and oxidative phosphorylation (Supplemental Table III). Most of the metabolism-related genes were downregulated, except for a few genes involved in lipid and amino acid transport (Fig. 4B).

Given that the citric acid cycle (TCA cycle) is central in energy metabolism in aerobic organisms and also connects most metabolic pathways (56), we teased out all the relevant energy metabolism-related DEGs to explore their relationship using an integrated schematic diagram. Hemocyanin knockdown resulted in significant decrease in the expression of key enzymes and metabolic intermediates involved in lipid metabolism (e.g., long-chain fatty acid-CoA ligase [LC-FACL], 3-ketoacyl-CoA thiolase [KAT], and acetyl-CoA carboxylase [ACC]), glycolysis (e.g., pyruvate carboxylase [PC]), TCA cycle/electron transport chain (e.g., succinate dehydrogenase [SDH]) and amino acids metabolism (e.g., argininosuccinate lyase [ASL], cysteine sulfinic acid decarboxylase (CSAD), and 4-aminobutyrate aminotransferase [ABAT]) (Fig. 4C, Supplemental Fig. 2H). To ascertain the effect of hemocyanin depletion on energy metabolic pathways, we examined glucose uptake and ATP levels in hepatopancreas cells. Both glucose uptake (2-NBDG glucose homologue) (Fig. 4D) and ATP production (Fig. 4E) by hepatopancreas cells decreased significantly (p < 0.01) after hemocyanin knockdown. These results indicate the role of penaeid shrimp hemocyanin in energy metabolism.

To further explore the role of hepatopancreas intracellular hemocyanin on energy metabolism via any potential effects on the mitochondria, which plays a central role in energy metabolism (57), we examined ΔψM after hemocyanin knockdown. The ΔψM of hepatopancreas cells decreased significantly (p < 0.01) after hemocyanin depletion, as indicated by an increased in the percentage of JC-1 monomers (Fig. 4F). These results indicate that hepatopancreas cells mitochondrial integrity is affected by low hemocyanin levels, which consequently dysregulates energy metabolism-related pathways.

Loss of mitochondrial membrane potential increases ROS production, resulting in oxidative stress (58). Given that mitochondrial ROS production is determined by a balance between the electron transport chain and oxidative phosphorylation (59), we further explored the effect of hemocyanin depletion on ROS production and microbiota diversity in the hepatopancreas. After hemocyanin knockdown, ROS levels in the hepatopancreas increased significantly, especially at 48 - 96 h (Fig. 5A). Next, we examined the consequence of increased ROS levels (reverse gavage treatment of shrimp with 3% H₂O₂) on hepatopancreas microbiota. A significant decrease in total bacterial abundance was observed at 24 h and 48 h, recovering steadily to pretreatment levels at 72 h and increased thereafter (Fig. 5B). Increased ROS levels (H₂O₂ treatment) had a similar effect on the relative abundance of Vibrio, where levels decreased significantly at 24 h, recovering steadily to pretreatment levels at 48 and 72 h, before peaking at 96 h (Fig. 5C). The relative abundance of Photobacterium decreased sharply at 24 h post-H₂O₂ treatment, remaining significantly low throughout the time points (Fig. 5D). These results indicate that both Vibrio and Photobacterium are sensitive to ROS, but Vibrio recovers quickly to become the dominant species in the hepatopancreas of penaeid shrimp.

To further explore the effect of ROS on hepatopancreas microbiota diversity and composition, we treated shrimp with the ROS scavenger NAC after hemocyanin knockdown. ROS levels increased significantly after hemocyanin knockdown (dsHMC) compared with control (dsControl), whereas ROS levels decreased significantly after hemocyanin depletion followed by NAC treatment (dsHMC plus NAC) compared with hemocyanin knockdown without NAC treatment (dsHMC) (Fig. 5E). Hemocyanin knockdown followed by NAC treatment (dsHMC plus NAC) also resulted in a significant increase in bacterial diversity in the hepatopancreas compared with untreated (dsHMC), as indicated by the Shannon index and other diversity indexes (Fig. 5F, Supplemental Table IV). In contrast, a significant decrease in bacterial diversity was observed in the NAC-treated control group (dsControl plus NAC) compared with untreated control group (dsControl) (Fig. 5F). Moreover, principal component analysis (PCA) analysis (genus level) revealed distinct separation among the bacterial communities in the four treatment groups (Fig. 5G).

Similarly, hemocyanin knockdown followed by NAC treatment (dsHMC plus NAC) or control samples treated with NAC (dsControl plus NAC) resulted in significant increase in the relative abundance of Photobacterium compared with untreated knockdown (dsHMC) or control (dsControl) samples (Fig. 5H, 5I). However, the relative abundance of Vibrio decreased significantly in the dsHMC plus NAC and dsControl plus NAC samples compared with the untreated samples (dsHMC or dsControl) (Fig. 5H, 5J). The dynamic changes in bacterial communities after hemocyanin knockdown followed by NAC treatment (dsHMC plus NAC) was also examined. Total bacteria abundance at all time points decreased significantly in the dsHMC plus NAC samples (Fig. 5K). In contrast, significant increase in Vibrio levels were observed at 24 h, followed by significant decrease at 72 and 96 h in the dsHMC plus NAC samples (Fig. 5L), whereas at 24 h, Photobacterium levels were at baseline, followed by significant increase at 48–96 h (Fig. 5M) compared with untreated knockdown samples (dsHMC). These results indicate that hemocyanin in the penaeid shrimp hepatopancreas modulates ROS levels to affect microbiota composition and diversity.

Aquatic ecosystems abound with various bacteria species, especially in coastal and estuarine environments, where many Gram-negative bacteria are found, including pathogenic and zoonotic species (60, 61). Many Vibrio species have been isolated from aquaculture ponds (62); however, among aquaculture animals, most Vibrio infections are found in shrimp (63). In this study, we were curious why some diseased penaeid shrimp (P. vannamei) presented with atrophied hepatopancreas and empty GIT, expressed low hemocyanin levels, and had high total bacterial abundance, especially Vibrio spp. In crustaceans, the hepatopancreas integrates immune-metabolic processes (55) and expresses the highest hemocyanin transcripts (53, 54). Thus, the nature and position of the hepatopancreas within the GIT of shrimp (64) makes it easily susceptible to microbial pathogens (65), especially Vibrio (62). As the tissue with the highest hemocyanin transcripts, we were therefore intrigued that the hepatopancreas (including other GIT tissues) in diseased shrimp expressed low hemocyanin levels, which correlated negatively with total bacterial abundance. This suggests that hemocyanin modulates microbial composition in the hepatopancreas of shrimp.

The bulk (∼95%) of total hemolymph proteins in crustaceans is made of extracellular hemocyanin (40), which plays a vital role in antimicrobial immune response via direct binding to bacteria (47, 66). However, little is known about the functions of intracellular hemocyanin. Using in vivo knockdown of hemocyanin in the hepatopancreas, we observed an increase in both total bacterial abundance and relative abundance of Vibrio but not Photobacterium (Fig. 2E–I), which is consistent with the low hemocyanin expressed in the hepatopancreas (also in other GIT tissues) of diseased shrimp. This indicates that microbial composition in the GIT of penaeid shrimp is affected by intracellular hemocyanin levels. Interestingly, our results reveal the importance of intracellular hemocyanin as the key host factor that modulates microbial composition in the hepatopancreas because, when hemocyanin levels are attenuated, opportunistic pathogens (Vibrio spp.) become dominant, whereas FMT could not reverse hepatopancreas microbial composition (Fig. 3C–H) or change microbial diversity when hemocyanin levels are depleted (Fig. 2C, Supplemental Fig. 1A). The reason for the selective dominance of Vibrio at low hemocyanin levels is unknown, but competition among different bacterial species in communities (67, 68) for space and resources changes the living environment and their metabolic state, which mostly favors bacteria with competitive advantage (69) such as Vibrio (70–72). Besides, the dominance of Vibrio at low hemocyanin levels (diseased shrimp and knockdown) could be due to the ability of some Vibrio strains to produce bacteriocins that are lethal to other related species (73, 74) or produce effector molecules that enable them counteract host antimicrobial response (75). Most importantly, hemocyanin can neutralize bacterial toxins (44); hence, low hemocyanin levels would favor bacteria with competitive advantage. In addition, reduced microbial diversity has been shown to correlate with increased pathogenic infections (76). The competitive dominance of Vibrio over Photobacterium at low hemocyanin levels could also be due to interbacterial antagonism (77, 78) between these two species in the hepatopancreas, which together constitute the bulk of the microbiota (Fig. 2E, 2F; Supplemental Fig. 1D, 1E).

The respiratory protein hemoglobin exists in the intracellular and extracellular forms (79) akin to hemocyanin. Intracellular hemoglobin is concentrated in the mitochondria, where it associates closely with mitochondrial complex I to help protect mitochondria against oxidative stress or regulate mitochondrial electron transport (80). Thus, hemoglobin helps protect cells from the cytotoxicity of ROS by inhibiting oxidative stress-induced mitochondrial dysfunctions (80–82). Intriguingly, we observed that intracellular hemocyanin confers protection on mitochondria because hemocyanin knockdown resulted in mitochondrial depolarization and increased ROS production (Figs. 4F, 5A). Moreover, the expression of several energy metabolism-related (glucose and fatty acids) genes was attenuated (Fig. 4B, 4C), coupled with decrease in ATP production (Fig. 4E) through the TCA cycle. This indicates that hemocyanin depletion results in metabolic perturbation and mitochondria damage (83). Despite the antimicrobial role of ROS (79), increased ROS levels (i.e., after hemocyanin knockdown or upon treatment with H₂O₂ via reverse gavage) could only transiently decrease the relative abundance of Vibrio, which quickly became dominant (Figs. 2H, 5C), whereas Photobacterium levels remained low (Figs. 2I, 5D). At low hemocyanin levels (depleted), treatment with the ROS scavenger NAC to decrease ROS levels resulted in increased microbial diversity but decreased Vibrio abundance, which could probably be due to increased competition from other fast-growing bacteria or opportunistic pathogens such as Photobacterium (Fig. 5F, 5H–J).

Host-generated ROS is a vital part of the innate immune response against pathogens because ROS can oxidize microbial components such as nucleic acids, proteins and lipids (84). However, excess ROS could also cause irreversible damage to host cellular macromolecules (85). We observed that in diseased shrimp, besides their atrophied hepatopancreas and empty GIT, they expressed low hemocyanin and had high bacterial abundance (Fig. 1C–1H), which could be due to excess ROS production. Given that hemocyanin has antioxidant effects (86, 87), the low levels of hemocyanin in diseased shrimp could not counteract the excess ROS, hence the cellular damage to the hepatopancreas and GIT tissues. The extracellular hemocyanin of some arthropods promotes the production of several oxygen radicals, including reactive cytotoxic quinones (88) and ROS (89). Similarly, extracellular hemoglobin (from hemolyzed RBCs) binds to oxygen to generate free oxygen radicals (90, 91), whereas intracellular hemoglobin (expressed in nonhematopoietic organs) protects cells from oxidative stress (80, 92). Our data in this study also confirm the importance and antioxidant activity of intracellular hemocyanin, given that in vivo knockdown of hemocyanin increased ROS production in the hepatopancreas (Fig. 5A). In any case, the contrasting roles of intracellular and extracellular hemoglobin or hemocyanin in the regulation of ROS production seems to be an evolutionarily conserved function of respiratory proteins. In crustaceans, the high levels of hemocyanin in hemolymph and the hepatopancreas (40, 50), coupled with our current data, indicate the importance of hemocyanin in modulating ROS production. Thus, decreased intracellular hemocyanin levels result in mitochondria depolarization, metabolic perturbation, and increased ROS production, which consequently affects microbial composition and diversity in the hepatopancreas. Although our current study provides convincing evidence that indicates the importance of hemocyanin in modulating microbiota in the hepatopancreas of penaeid shrimp, we are still left with the following question: what or which is the initiating factor? Given that the aquatic ecosystem is inundated with various microbes, the metabolic perturbation and increased ROS levels in the hepatopancreas could be a host response to pathogenic microbes. It is equally possible that these microbes induced these changes, which consequently damaged the hepatopancreas and attenuated synthesis of intracellular hemocyanin, hence changing the microbial composition.

In conclusion, we revealed that hemocyanin modulates microbial composition and diversity in the hepatopancreas of penaeid shrimp. The proposed mechanism that supports the current findings are that decreased hemocyanin induces mitochondria depolarization, which consequently affects energy metabolism and ROS production (Fig. 6). Thus, the antioxidant activity of hemocyanin meant that, when hemocyanin levels are attenuated, the oxidant effects of ROS cannot be completely counteracted, hence the damaging effect of ROS on the hepatopancreas and other GIT tissues that further increases susceptibility to colonization by opportunistic pathogens such as Vibrio.

The authors have no financial conflicts of interest.