Molecular Characterization of High Mobility Group Box 1a (HMGB1a) Gene in Red-Bellied Pacu, Piaractus brachypomus

High mobility group box 1 (HMGB1) is a chromosomal protein in the nucleus and a potent extracellular proinflammatory cytokine, widely described in mammals, nevertheless, with scarce reports in fish. In this study, full open reading frame of HMGB1a gene from Piaractus brachypomus is reported as well as its molecular characterization, including tissue gene expression. At predicted protein level, HMGB1a showed similarities with its orthologs in teleosts and higher vertebrates. The relative gene expression of HMGB1a mRNA was measured in several tissues including the brain, where a differential expression appeared in brain regions, i.e., higher expression in the cerebellum and telencephalon. In addition, in an assay of sublethal exposure to chlorpyrifos, upregulation of HMGB1a was detected in optic chiasm. Furthermore, in a traumatic brain injury model, upregulation of HMGB1a expression was evident 24 hours after lesion and remained higher up to 14 days. These findings suggest a role for HMGB1a in brain damage and its candidature as biomarker of brain injury; however, more studies are required to elucidate the functions of HMGB1a and its regulation in P. brachypomus.


Introduction
High mobility group (HMG) proteins are a class of nonhistone proteins, normally found in the nucleus, that play a role in the regulation of DNA-dependent processes [1]. HMGs are classifed into three families: HMGA (HMGAT-hook family), HMGN (HMG-nucleosomebinding family), and HMGB (HMG-box family) [2], the latter being the one with the highest expression [3]. High mobility group box 1 (HMGB1) is highly conserved among vertebrates and can be found in the nucleus, cytoplasm, or in the extracellular milieu [1,4]. Tis protein acts as a chromosomal protein in the nucleus, and in the cytoplasm prevents mitochondrial abnormalities, and promotes autophagy, and from cytoplasm, it gets released into the extracellular medium as a danger-associated molecular pattern (DAMP) or secreted by immunocompetent cells in response to diferent stimuli, such as the recognition by pathogen-associated molecular patterns (PAMPs) or DAMPs. HMGB1 outside the cell, regulates the immune response by stimulating the release of cytokines and promoting the diferentiation, proliferation, and maturation of immunocompetent cells [5][6][7][8], through interaction with Toll-like receptors (TLR) (TLR2, TLR4, and TLR9) and the endogenous receptor for advanced glycation end-products (RAGE) [4][5][6][7][8][9][10][11]. In addition, HMGB1 is involved in neurogenesis during early development of mammals and fsh [8,9] and neuroinfammation [10], and may contribute to neuropathology and inhibition of neurogenesis through secondary tissue damage.
Recently, HMGB1 has been reported in diferent aquatic species with similar functions to its mammalian orthologs [3], including its upregulation under pathological conditions [7]. In teleost fsh and lampreys, HMGB1 recombinants can interact with DNA and participate in cytokine-like immune responses in the extracellular medium [1,4,6,11]. Furthermore, Fang et al. [12] in a zebrafsh model of spinal injury showed that upregulation of HMGB1 promotes neurogenesis, angiogenesis, and recovery of motor capacity and cell survival, and then its downregulation minimizes the efects of infammation during recovery, which suggest a dual role of this protein in neuroinfammation.
Te red-bellied pacu (Piaractus brachypomus) is an endemic freshwater fsh of economic importance in Colombia, which has been used as a biomodel in pharmacological and immunotoxicological studies [13][14][15][16]; nevertheless, more studies are required to validate biomarkers associated with pathologies and disease mechanisms in these native fsh species. To the knowledge of the authors, there are no molecular characterization studies of the HMGB1a gene and its gene expression in P. brachypomus. Tus, the present study aims for the molecular characterization of HMGB1a and its transcript expression brain tissue of P. brachypomus, after exposure to chlorpyrifos and brain injury.

Ethical Approval.
All the experimental procedures followed the guidelines of the Bioethics Committee of the Central Research Directorate of the University of Tolima in the framework of the project code 310130517, based on Law 84/1989 and Resolution 8430/1993, in addition to complying with the parameters established for the care of animals and its use in research and teaching [17,18].

RNA Extraction and cDNA Synthesis.
Total RNA from brain samples for all experiments was extracted using TRIzol reagent (Invitrogen, USA), following the manufacturer's instructions. RNA quality and concentration were assessed by molecular spectrophotometry using NanoDrop ™ One (microvolume UV-Vis spectrophotometer, Termo Fisher Scientifc, USA). Finally, cDNA was synthesized using the High Capacity cDNA Reverse Transcription Kit (Termo Fisher Scientifc, USA) following the manufacturer's recommendations. Te quality of the cDNA was evaluated by amplifying elongation factor 1 alpha (EF-1α) gene by RT-PCR [13].

HMGB1a cDNA Sequencing and Primer Design.
HMGB1a gene sequence was obtained by nanosequencing, using the MinION sequencer (cDNA sequencing kit, SQK-DCS109, Flow Cell R9.4.1, N-50 of 800, Oxford Nanopore Technologies, United Kingdom) from P. brachypomus brain cDNA, mapped on HMGB1a sequence from Colossoma macropomum (accession number XM_036561022) as a reference template using the Geneious Prime v2023.04 software [19]. Based on the mapped sequence, the primers were designed (Table 1)  and GoTaq ® Flexi DNA polymerase were used (Madison, USA), following the manufacturer's guidelines. Te amplifcation products were revealed by horizontal electrophoresis using a subcell electrophoresis chamber (Bio-Rad, CA, USA) on a 2% agarose gel stained with HydraGreen ™ (ACTGene, NJ, USA) and a molecular weight marker of 100 bp (New England Biolabs, USA). Te gel was visualized and documented using the ENDURO GDS Gel Documentation System (Labnet International, NJ, USA). Te amplifcation products were purifed and sequenced by the Sanger method (Macrogen Inc., Korea).

Analysis of Sequences and Construction of Model
Structures. Te sequence of the HMGB1a full ORF was confrmed by Sanger sequencing, and amino acid (aa) sequence was predicted using the Geneious Prime software v2023.04 [19], and Expasy ProtParam tool was used for calculation of molecular weight and isoelectric point (pI), among other intrinsic parameters [20]. In addition, the online servers Signa1P 6.0 [21], NetGPI-1.1 [22], DeepLoc-2.0 [23], NetNGlyc [24] and NetOGlyc [25] (DTU Health Tech, Denmark), were used for analysis of putative signal peptide, GPI anchor, most likely location and N-glycosylation and O-glycosylation sites, respectively. Protein domains were predicted using the Conserved Domains tool (NCBI, USA). Te HMGB1a structural model was constructed in SWISS-MODEL [26], using as template the Cryo-EM structure of mouse RAG1/2 HFC complex containing a partial HMGB1 linker; the model was modifed in PyMOL 2.0 [27], to improve its interpretation.

Sequence Alignment and Phylogenetic Analysis.
Multiple sequence alignment (MSA) was performed in Geneious Prime software v2023.04 [19], using HMGB1a amino acid sequences reported in the GenBank, to Tamnophis elegans (XP_032075701). HMGB1a phylogenetic tree was built using the neighbor-joining method model of the Geneious Primer v2023.04 software [19] and the Jukes-Cantor genetic distance model and considering 1000 bootstrap hits. In addition, the Drosophila melanogaster elongation factor 1 alpha sequence (accession number NP_524611) was set as an outgroup.

Experimental Design.
For the treatments in this study, 24 P. brachypomus juveniles (2.5 ± 0.3 g) from the same spawning and regardless of the sex, which came from a hatchery, were used. Te fsh were kept in a 90 L glass tank, with a thermostatically controlled temperature of 25°C, aeration without a flter, and light periods of 12 hours, and fed twice 2 Veterinary Medicine International daily with commercial food (30% protein, Solla ® ) in an amount equivalent to 2% of their body weight. Te fsh were acclimatized in a period of 15 days and were treated with NaCl to eliminate ectoparasites [28]. Animals were randomly distributed for the experiments, and brain samples were collected.

Sublethal Chlorpyrifos (CPF) Exposure Assay.
Te animals were divided into two experimental groups: a treatment group (n � 5) exposed to a sublethal concentration of CPF (0.011 μg/L) [29] for 72 hours and a control group (n � 3) without CPF exposure. To maintain CPF concentrations in the water, half of the water volume in each glass tank was replaced every 24 hours and the corresponding CPF amount was added.

Brain Injury (BI) Assay. Brain injury was performed
following the stab wound model [30,31]. For this, the animals were divided into 4 groups: Group 1 (control group, n = 4, 0 h) without puncture, Group 2 (n = 4) 24 hours postinjury; Group 3 (n = 4) at 7 days post-injury, and Group 4 (n = 4) 14 days post-injury. Before brain puncture, the fsh were anesthetized by immersion in a glass tank containing 50 mg/L of dissolved eugenol (eugenol, clove oil, Proquident S.A., Colombia) [13] until reaching the state of general anesthesia. Te lesion was performed with a 000-caliber sterile entomological needle, with a puncture depth of 5 mm on the left side of the skull until reaching the telencephalon and the optic bulb. Te control group was treated with the same manipulation and anesthesia procedure, without puncture, and sacrifced after sedation.

Brain Region Basal Diferential Expression.
Basal relative expression of the HMGB1a in brain regions was measured from brain samples of healthy P. brachypomus belonging to a tissue bank collection from our laboratory, where dissected olfactory bulb, telencephalon, optic bulb, hypothalamus, cerebellum, medulla oblongata, and optic chiasm were used. RNA extraction and cDNA synthesis were performed as mentioned above.
2.9. Sampling. From each experiment before sampling, fsh were anesthetized by immersion in a glass tank with dissolved anesthetic until they reached the stage of general anesthesia [32]. Ten, individuals were euthanized by cervical dislocation [33], and brain samples were collected, and the olfactory bulb, optic chiasm, and telencephalon were dissected and collected separately. All tissue samples were snap frozen in liquid nitrogen and stored at −20°C until analysis.

Assessment of HMGB1a Transcript Relative Expression by qPCR.
For qPCR, primers were designed based on the confrmed ORF sequence of HMGB1a from red-bellied pacu ( Table 1). qPCR assays were run by duplicate on a Quant-Studio 3 real-time PCR system (Termo Fisher Scientifc, USA) following the fast ramp program and using Luna ® Universal qPCR Master Mix (New England Biolabs, USA). Each amplicon was validated by melting curve analysis and gel electrophoresis imaging to ensure that there was no amplifcation of genomic DNA or possible misannealing. Relative gene expression was calculated using the 2 −ΔΔCt method [34], and EF1α mRNA level was used for normalization.

Statistical
Analysis. Data were analyzed by descriptive statistics, and normality was determined by the Shapiro-Wilk test. Diferences in gene expression between treatments with CPF were assessed using the Mann-Whitney U test, while for the injury treatment and expression in the brain, Kruskal-Wallis test and the two-step augmentation method of Benjamini, Krieger, and Yekutieli as post hoc were used, using GraphPad Prism v 9.0 for MacOS (La Jolla, CA, USA).

Tissue Expression of HMGB1a by RT-PCR Assay.
cDNA showed optimal values of quality by biomolecule spectrophotometry, and the reference gene EF-1α was detected in all tissues (data not shown). Full ORF of HMGB1a from Piaractus brachypomus (PbHMGB1a) was amplifed from the brain, liver, gill, and head kidney, showing a band of the expected size ( Figure 1).  A model of the protein was predicted using the Cryo-EM structure of mouse RAG1/2 HFC complex containing partial HMGB1 linker as a template, with an identity of 88.89% and a Global Model Quality Estimate of 0.53 ( Figure 2). Te secondary structure of the protein reveals the presence of 7 α-helices, of which 4 are in the HMG-box A domain and 3 in the HMG-box B domain.

Sequence Alignment and Phylogenetic Analysis.
Te MSA of the sequences showed an identity between PbHMGB1a and 16 orthologous sequences from diferent orders of fsh in a range from 74.27 to 98.52%, with the higher identity with Colossoma macropomum. On the other hand, PbHMGB1a showed a range from 73.61 to 77.03% identity with its higher vertebrate orthologs ( Figure 3). In addition, the MSA demonstrates few variations in the sequences of the domains, as well as in the C-terminal acid tail, the latter presenting more residues in higher vertebrates ( Figure 3).
From the phylogenetic analysis, two clearly defned clades were obtained, in which the HMGB1a sequences of fsh are grouped, separated from those of higher vertebrates ( Figure 4). On the other hand, subgroups of fsh belonging to the same taxonomic order are formed, which demonstrates the evolutionary diference of the protein between the different orders of teleost fsh.

PbHMGB1a mRNA Expression. Te levels of
PbHMBG1a mRNA were measured in brain regions of P. brachypomus, showing a higher expression in the telencephalon and cerebellum ( Figure 5(a)). Te expression in the telencephalon and cerebellum was higher compared to the hypothalamus, optic bulb, and, and medulla oblongata.
On the other hand, PbHMGB1a mRNA level was higher in individuals exposed to CPF compared to the control group in the optic chiasm (p < 0.05) ( Figure 6). In the case of brain injury, PbHMGB1a mRNA was upregulated at 24 hours and 7 and 14 dpi compared to the control group (p < 0.05) in the three tissues evaluated. However, in the telencephalon at 7 days post-injury, PbHMGB1a mRNA levels decreased compared to levels observed at 24 hours and then increased slightly at 14 days post-injury (Figure 7).

Discussion
Te high mobility group box (HMGB) family of proteins was originally discovered in the bovine thymus [5] and is divided into four subgroups, HMGB1, HMGB2, HMGB3, which share high identity, and HMGB4 lacking in identity acidic Cterminus [3,7,35]. HMGB1 was identifed as a ubiquitous bifunctional protein [1], which in the nucleus acts as a nonhistone chromatin-associated factor [11], binds to DNA in a structure-specifc manner, stabilizes nucleosome formation, aids in error repair, and plays important roles in gene transcription in both higher vertebrates and teleosts [1,6,7]. Pathogenic stimulation and oxidative stress induce the nucleocytoplasmic translocation of HMGB1 [4], which in grass carp induces heat shock protein 70 (HSP70) to move to the nucleus where it interacts with HMGB1b generating the nucleocytoplasmic translocation of the protein and activation of the HMGB1b-Beclin 1-mediated autophagy [4]. However, outside the cell, HMGB1 is a linker of innate and acquired immunity, as it acts as a cytokine [1,5], through interaction with pattern recognition receptors such as TLR and RAGE [4,11], leading to activation of the primary myeloid diferentiation response 88 protein (MyD88) dependent on the NF-kB pathway [36]. In addition, it plays a role in apoptosis, cell migration, and cytoskeleton reorganization [1].
In teleosts, HMGB presents a similar activity, demonstrating the capacity to stimulate the respiratory burst, the production of nitric oxide, and the proliferation of immunocompetent cells, as well as increasing the release of proinfammatory cytokines (TNF-a and IL-1b), by mediating the activity transcription of NF-kB [7,11]. On the other hand, HMGB1 expression increases after exposure to pathogens and prevents replication of the pathogen [6,7]. Tey suggest that HMGB1 activates the immune response after injury or infection by stimulating the release of cytokines and the activation of immune cells. Te early increase in PbHMGB1a expression 24 hours after brain injury indicates that it may be released due to cell injury and could be involved in the initiation of infammation; however, whether PbHMGB1a can modulate the immune response remains to be clarifed like their teleost counterparts.

Sequence
Analysis. HMGB1 is bipolar [7], has two positively charged folded helical DNA-binding domains (HMG-box A and B), and a negatively charged C-terminal acid tail rich in glutamic and aspartic acids [11], which regulates the interaction between domains, as well as other nuclear proteins with DNA [7,35]. PbHMGB1a presents these domains which are highly conserved in vertebrates; the length, molecular weight, and isoelectric point are similar to those previously described in other teleosts and lampreys [1,[5][6][7]11]. On the other hand, the C-terminus of higher vertebrates has acidic residues that are not present in any of the teleosts, including P. brachypomus; however, it is believed that the length of this region may not be critical for the biological function of HMGB1 [7]. Box A has been shown to contain receptor-binding sites [37], its reduced form is anti-infammatory, and binding to the C-terminus of the protein increases this activity [35]; however, it presents two cysteines (Cys23, Cys45), which mediate autophagy by interacting with Beclin 1 and form a disulfde bond under mild oxidative conditions [38], which gave it infammatory activities [5]. Box B is a proinfammatory domain [37] and presents a cysteine (Cys106) that promotes cytoplasmic localization and is related to TLR4 activation and TNF-α release [1]. Likewise, these components are involved in the nucleocytoplasmic translocation of HMGB [4]. Te presence of cysteines (Cys22, Cys44, and Cys105) in PbHMGB1a supports the possible role of this protein in immunity, as has been described in other teleosts and lampreys [1,5,39], due to the presence of these cysteines in positions corresponding to those described in mammals.
Te absence of a signal peptide in PbHMGB1a indicates that it may be secreted via an alternative pathway that bypasses the endoplasmic reticulum and the Golgi apparatus [7] via acetylation of lysine residues [2]; however, the presence of HMGB1 in plasma membranes has been reported, which is associated with a non-classical secretory signal peptide, the 18 N-terminal amino acids of hydrophilic acylated surface protein B (HASPB) [35].
Two N-glycosylation (Asn 36 -Phe 37 -Ser 38 and Asn 133 -Lys 134 -Tr 135 ) and O-glycosylation (Ser 14 and Tr 175 ) sites were predicted in PbHMGB1a. Glycosylation alters proteolytic resistance, protein solubility, stability, local structure, circulating lifetime, and immunogenicity; specifcally, N-glycosylation and O-glycosylation modify secreted and membrane proteins [24,25]. On the other hand, posttranslational modifcations such as acetylation, phosphorylation, methylation, ADP-ribosylation, and oxidation have been shown to be involved in nucleocytoplasmic translocation [35]. Tese fndings suggest that the interaction between diferent post-transcriptional modifcations ultimately determines the location of HMGB1, which remains to be demonstrated for glycosylation.
Te predicted PbHMGB1a structure revealed 7 α-helices distributed in the domains, like those of Scophthalmus maximus L., whose secondary structure was composed of 9 α-helices, 7 β-turns, 2 c-turns, and other structures [3]. In humans and other mammals, the presence of α-helices generates two loops in both domains, conferring an Lshape, presenting a higher content of α-helices in HMGbox A [35], like what was observed in the present model. Te aromatic amino acids Phe 10 , Phe 13 , Trp 41 , Lys 49 , and Tyr 52 in HMG-box B and their distribution within the secondary structure of PbHMGB1a (Figure 2(b)) demonstrate that they Veterinary Medicine International could interact to give the L-shape to this domain, since they are found in same positions to those reported for humans, where Phe 14 , Phe 17 , Trp 45 , Lys 53 , and Tyr 56 are found at the junction between the two helical arms [40].
On the other hand, the HMG-box A domain presents structural changes in its oxidized form, related to disulfde bonds in Cys 23 and Cys 45 that displace the N-terminus of helix II, and this allows hydrophobic interaction between the phenyl rings of Phe 38 [38]; given the presence of these amino acids in similar positions (Figure 2), it is likely that these interactions occur in PbHMGB1a.
Tese similarities in sequence and structure suggest that PbHMGB1a may exhibit behavior and function like those described in humans and other mammals [35]. Tis is

Phylogenetic Analysis.
MSA reveals high percentages of identity between vertebrate and teleost HMGB sequences, which could be the result of low selective pressure for evolutionary maintenance of amino acid sequences [14]. Analysis of the constructed phylogenetic tree indicates that PbHMGB1a is highly related to its teleost orthologs, grouping in a diferent clade from higher vertebrates, which is consistent with other phylogenetic analyses of teleost HMGB [3,5]. Te phylogenetic relationships also demonstrate the evolutionary divergence between the diferent orders of teleosts, where each one of these is grouped in a diferent clade, and thus PbHMGB1 is grouped with the species of Characiformes, Colossoma macropomum and Pygocentrus nattereri.

Gene
Expression. HMGB1 is ubiquitously distributed in teleosts, being found in the kidneys, gills, brain, blood, heart, liver, spleen, intestine, muscle, and skin [5][6][7]; however, the site of highest expression varies according to the species [3,[5][6][7]11]. Tis wide distribution, under normal conditions, suggests a fundamental role of HMGB1 in teleosts [7]. In this study, the absence of HMGB1a in the blood and spleen samples of P. brachypomus may indicate lower expression in the tissue below the limit of detection of conventional PCR.
On the other hand, diferential expression of PbHMGB1a in brain regions evidenced highest expression in the cerebellum and the telencephalon. Tis can be explained by neurogenesis in the teleosts which occurs in the telencephalon and cerebellum [8,9,41], where the HMGB1 plays a pivotal role. In contrast with mammals, the expression of HMGB1 is almost null in the brain of adult mammals and the neurogenesisis limited to certain regions [35],such as the granular cells of the gyrus dentatus of the hippocampus, in the olfactory bulb, in certain areas of the telencephalic ventricles, and in almost all nuclei of the granular cells of the cerebellar cortex [8,9]. HMGB1 expression in the mammal cerebellum is relevant in granular cells, since it can interact with RAGE and the human natural killer cell glycan (HNK-1) to stimulate migration and growth of neurites in these developing animal cells, regulating downwards after the migration [8]. In teleosts, granular cells are the main population of constantly proliferation in the cerebellum [41], which may explain the high expression of PbHMGB1a in this region, which is related to the migration of granular cells, as well as in the remodeling of neuronal circuits in the red-bellied pacu.
HMGB1 may play an important role in the mammalian and teleost nervous system as it is involved in early brain development, cell migration, and neurite formation [2,8]. Similarly, it is important in neurogenesis and angiogenesis after injury; however, its release produces multiple infammatory and neurotoxic factors, contributing to secondary tissue damage from injury [12].
In our study, PbHMGB1a mRNA levels increase in the telencephalon 24 hours after the lesion and then decrease at 7 dpi, maintaining similar levels to 14 dpi. However, PbHMGB1a mRNA levels in the optic chiasm and olfactory bulb remain high during the experiment, possibly due to factors related to the site of the lesion and its severity. In zebrafsh, Danio rerio, HMGB1 mRNA increases shortly after a lesion in the central nervous system, associated with the stimulation of neurogenesis and angiogenesis; however, it undergoes a rapid decrease, preventing the efects of infammation, resulting in increased cell survival and less tissue damage during recovery [12], which is similar to the results in the present study. Likewise, in an extra-neuronal tissue, the valve intestine of the Siberian sturgeon, Acipenser baerii and the challenge with Streptococcus iniae, induced an upregulation of HMGB1 expression in the acute phase followed by a downregulation [5].
On the other hand, CPF, an organophosphate, is stable in water and can be rapidly absorbed through the gills, skin, and intestinal tract, causing neurotoxicity, hepatotoxicity, immunotoxicity, oxidative stress, and alterations in blood biochemical parameters [42], along with efects on neurobehavioral and locomotor behavior [43,44]. In P. brachypomus, subchronic exposure to CPF increased the density of astrocytes, modifed their morphology, and increased the expression of biomarkers related to reactivity [45]. In Clarias gariepinus, multifocal vacuolization was generated in the brain after exposure to CPF [46]. In the present study, the PbHMGB1a mRNA levels increased in the optic chiasm after sublethal exposure to CPF, possibly due to its high blood supply [15]; it could have been exposed to higher concentrations of the pesticide, generating greater injury and expression of PbHMGB1a. Te efects of CPF could also be increased given the tank temperature (25°C) despite the low dose and short exposure time, since exacerbation of the efect has been reported at high temperatures [44].
Because fsh can be exposed to diferent concentrations of pesticides in aquatic ecosystems close to places of agricultural activity [44], the study of molecular biomarkers of neuronal and tissue damage after exposure to pollutants may help to assess the environmental impact of the contamination and establish proper and accurate measurements.

Conclusion
Full ORF of PbHMGB1a was detected which allowed its bioinformatic and phylogenetic analysis. Expression analysis showed a wide distribution of the gene in red-bellied pacu tissues as well as its response in toxicity and brain injury models, which denote it as a candidate brain biomarker in the central nervous system, and due to its diferential expression distribution in brain regions, further studies are needed to describe its role in pathophysiology of specifc brain areas.

Data Availability
Te data used to support the fndings of this study are available from the corresponding author upon request.

Conflicts of Interest
Te authors declare that there are no conficts of interest regarding the publication of this article.

Authors' Contributions
ISRB was responsible for conceptualization, supervision, project administration, funding acquisition, and resources. NCG was responsible for original draft preparation and investigation. NCG and ISRB were responsible for methodology, validation, formal analysis, data curation, review and editing, and visualization. All authors have read and agreed to the published version of the manuscript. Veterinary Medicine International 9