This study concerns the first characterization of the BDNF gene in Dicentrarchus labrax and its transcriptional, translational and post-translational regulation following acute stress. We found that in the seabass, BDNF is highly expressed in the brain and that the transcripts 1c and 1d from the second promoters cluster are down regulated after acute stress. In addition, we show that acute stress induces a significant increase in the proBDNF levels and a corresponding reduction in mature BDNF suggesting altered regulation of proBDNF proteolytic processing. Finally we show that the proBDNF/totBDNF ratio (or its counterpart matBDNF/totBDNF) is a highly reliable novel quantitative neurological biomarker capable to detect biological stress in fishes with sensitivity 100%, specificity 87%, Positive Predictive Value of 88% and Negative Predictive Value of 100%.
All known vertebrate BDNF genes share a similar multiple exons organization and encode for a pre-pro-protein that is translocated to the endoplasmic reticulum and proteolitically processed to yield the mature protein [52–54]. The D. labrax BDNF gene consists of at least five alternative 5'-exons and one 3'-coding exon. For what concern the nomenclature, we have referred to previous studies on zebrafish  because mammalian, avian and amphibians BDNF follow a different nomenclature . Analysis of D. labrax BDNF transcripts, carried out both in the early developmental stages and in adult tissues, shows that all alternative upstream exons (1β, 1a-1d) are spliced to the protein-coding exon 2. This indicates that D. labrax BDNF transcripts structure is similar to other vertebrates. In fact, multiple transcriptional initiation sites and splicing into two-part transcripts can also be found in humans, chimpanzees, dogs, pigs, cats, cows, chicken, frogs, lampreys, zebrafish, and pufferfish .
With the exception of the exon 1a (43%), we found strong sequence homology with zebrafish and pufferfish genes for most exons. Furthermore, exon 1β contains 170 nt segment that is highly similar to human exon 1 (75% identity). Three highly conserved segments were found in the seabass BDNF, HCS2 in exon 1a which is present also in exon IIc of mammalian BDNF; HCS1 in exon 1c also found in mammalian exon IV while in the 3'UTR encoded by exon 2, we have found the HCS3 which is also present in the 3'UTR in mammals. These results suggest that mammalian exons I, II, and IV (coding for the the 5' untranslated region) evolved early in the vertebrate radiation and may play a major role in BDNF action, while more recently evolved splice variants including other 5'exons may participate in more specialized functions of BDNF such as, for example, synaptic plasticity. Further research in this direction may allow to test this hypothesis.
The post-hatching developmental analysis indicates that although in different amounts, all D. labrax BDNF transcripts, except 1d/2, are well represented at all stages analyzed (day post-hatching 6, 16, 27, 33, 44). On the other hand, the distribution of various BDNF transcripts in adult seabass is tissue specific with all transcripts being most expressed in the brain. The splice form 1c/2 was also expressed, even though at low levels, in all the examined extra nervous tissues (liver, kidney, muscle), while exon 1b/2 transcript was found only in the kidney. Organ-specific expression also holds for most BDNF exons in zebrafish and mammals suggesting conserved transcriptional regulation among the vertebrates [47, 52–55].
According to this view, our bioinformatic analysis of D. labrax BDNF gene suggests that the region upstream to exon 1c contains two potential responsive elements, belonging to the CRE family. These elements function as responsive elements also in BDNF exon IV of rat cortical neurons [47, 55] and may be responsible of the higher expression of the isoform 1c/2 in adult seabass tissue. Of note, a previous study on transcriptional analysis of Zebrafish HCS1 reported that this highly conserved sequence in the 5' exon 1c (and vertebrate exon IV) has properties of a dehancer and, depending on the sequence context, as an enhancer .
In the second part of our study we have examined the expression of BDNF transcripts after acute stress caused by water deprivation for 30 minutes. Although no significant difference was found in the total BDNF mRNA levels between stressed and control groups (measured by analysis of the protein-coding exon 2, common to all transcripts), we found a significant decrease in exons 1c and 1d. This finding is consistent with data in rodents in which single immobilization stress induces down regulation of exon IV, omologous to fish exons 1c (they both contain HCS1), due to decreased histone acetylation at this promoter immediately after acute stress . Of note, in a recent study we showed that exons of the second promoter clusters (mammalian exons IV-VII, fish 1c-d), are particularly important for cell survival in response to cellular excytotoxic stress in human neuroblastoma cells . Thus, activation of promoters upstream to these exons might related to a rapid adaptative response to various types of stress.
Western blot analysis showed that in brain, but not in liver, proBDNF content is significantly increased in the stressed samples. Mammalian BDNF transcripts produce the well-known 32 kDa propeptide precursor that is cleaved either to pro28KDa or to the mature 14 kDa BDNF forms by two different proteases . Pro28Kda BDNF peptide is not further processed into the mature 14 kDa BDNF form but it represents a true final proteolytic product generated by a specific Ca2+-dependent serine proteinase known as Membrane-Bound Transcription Factor Site-1 protease (MBTFS-1; EC = 22.214.171.124, Alternative names: S1P endopeptidase, Site-1 protease), also known as Subtilisin/kexin-isozyme 1 (SKI-1) , while mature 14 KDa BDNF is generated intracellularly by furin , or extracellularly by plasmin and matrixmetalloprotease-7 . In contrast, in the seabass we only found two BDNF forms, a proBDNF form corresponding to mammalian pro32KDa precursor and a mature BDNF, while the pro28KDa peptide was absent. Comparison of D. labrax BDNF protein with that of rodents and human BDNF, revealed that the mammalian SKI-1 cleavage site at Threonine 57 (Arg-Gly-Leu-Thr↓) is absent in fishes and amphibians and has first emerged in reptilians during vertebrates evolution . Limited proteolysis of one inactive precursor to produce active peptides and proteins is a general mechanism to generate biologically diverse products from a single gene. Here, we provide the first evidence that fishes possess a simplified proteolytic regulation of BDNF and that the pro28KDa proteolytic product, whose function remains yet to be determined, is absent at this stage of vertebrates evolution.
We found that acute stress profoundly alters the relative amount of proBDNF and mature BDNF. Our data are suggestive of a lower proteolytic activity to generate mature BDNF and thus, the uncleaved product is accumulated in the seabass brain, but not in liver, immediately after an acute stress. Although, the mechanisms by which stress can prevent efficient conversion of proBDNF into mature BDNF are presently unknown, several recent studies have pointed out that pro32KDa BDNF has a biological function distinct from that of mature BDNF. Both proBDNF precursor and mature BDNF can be released from neurons [59, 60]. While proBDNF binds only to p75 receptor, mature BDNF displays high affinity to TrkB and lower affinity to p75 . Binding of proBDNF to p75 promotes cell death and attenuates synaptic transmission by inducing long term depression [62, 63], while mature BDNF sustains long term potentiation and cell survival [59, 64, 65]. It is therefore conceivable that the shift towards higher proBDNF and lower BDNF level observed after acute stress may have the biological role of attenuating pro-active behavior inducing reduced activity in stressed animals. Stress affects the hormonal response in fish in much the same way it does in higher animals. Stress stimulates the hypothalamus, one of the oldest parts of the brain (in evolutionary terms) and is responsible for controlling the most basic functions such as hunger, thirst, sex drive and, in mammals, body temperature; all functions that are mediated also by BDNF. A reduced behavioral activity may thus represent an adaptive response to dangerous situations represented here by shallow waters, to allow for an immediate energy saving and recovery in preparation for future actions. In this context, it is striking that 100% of animals in our experimental stress group showed >1SD increase in proBDNF levels (and corresponding decrease in mature BDNF). A theoretical, optimal prediction test can achieve 100% sensitivity (i.e. predict all people from the sick group as sick) and 100% specificity (i.e. not predict anyone from the healthy group). Thus, our test performances will make it feasible to screen for stress even in low prevalence populations, particularly where samples are first pooled before testing.
In conclusion, we have determined the structure of Dicentrarchus labrax BDNF gene, its expression in neuronal and non neuronal tissues, and we have demonstrated that the proBDNF/totBDNF ratio (or its counterpart matBDNF/totBDNF) is a novel quantitative neurological biomarker capable to detect biological stress in fishes with sensitivity 100%, specificity 87%, Positive Predictive Value of 88% and Negative Predictive Value of 100%.