Characterization of intronic SNP located in candidate genes influencing cattle temperament

ABSTRACT The objective of this study was to evaluate the effect of intronic single nucleotide polymorphisms (SNP) on temperament traits in a Brahman cattle population. The SNP located in CACNG4, EXOC4, NRXN3, and SLC9A4 candidate genes were genotyped in 250 animals with temperament records of exit velocity, pen score, and temperament score. Rs3423464051:G>A in the CACNG4 gene was associated with exit velocity and temperament score. An in silico analysis of the five intronic SNP showed that alternative alleles of CACNG4-rs3423464051, EXOC4-rs109393235, and SLC9A4-rs109722627 SNP could alter branch point sites during splicing, while a protein–protein interaction network analysis demonstrated a GRIA2 gene-mediated interaction between CACNG4 and NRXN3. The present results support previously reported evidence regarding bovine temperament-related candidate genes, particularly CACNG4, which is a confirmed candidate gene in need of more detailed analyses to reveal its role in temperament-related traits.


Introduction
Temperament is a complex and economically relevant trait that reflects the welfare of an animal, directly affects production, and is closely related to genetic control (Haskell et al., 2014).Few studies have explored the characteristics of cattle temperament at the genomic level, due to the biological complexity and recording difficulty.Recent discoveries based on genome-wide association studies (GWAS) have revealed candidate genes implicated in cattle temperament (Lindholm-Perry et al., 2015;Valente et al., 2016;dos Santos et al., 2017;Chen et al., 2020).
Several previously identified genes that play a role in the central nervous system and excitatory synaptic transmission show differential expression during anxiety events related to cognitive impairment in humans and mice (Riley et al., 2016).These findings represent progress in defining the genetic architecture of cattle temperament and show that temperament regulation involves multiple genes.Knowledge of gene interactions influencing temperament and other traits is limited, which hinders the elucidation of the group of genes that participate in regulating the trait of interest (Alvarenga et al., 2021).
Relevant evidence has recently been published, identifying similarities between bovine and human SNP related to temperament and involved in psychiatric and personality disorders through GWAS (Costilla et al., 2020;Paredes-Sánchez et al., 2020).These disorders may include neuroticism, schizophrenia, autism spectrum disorder, and developmental delay disorders related to the brain and cognition.Brahman cattle population studies identified 14 SNP with an effect on exit velocity in bovines (Paredes-Sánchez et al., 2020).From those, five SNP (rs3423464051, rs109393235, rs135982573, rs110864071, and rs109722627) were located on the intronic regions of the neurexin-3 (NRXN3:NC_037337.1 BTA 10), calcium voltage-gated channel auxiliary subunit gamma-4 (CACNG4:NC_037346.1 BTA 19), exocyst complex component-4 (EXOC4:NC_037331.1 BTA 4), and solute carrier family 9 member A4 (SLC9A4: NC_037338.1 BTA 11) genes (Paredes-Sánchez et al., 2020).The CACNG4 and SLC9A4 candidate genes have been previously implicated in human diseases (Guan et al., 2016;Calvete et al., 2021).Studies showed that members solute carrier gene family SLC18A2, SLC9A9, and SLCO3A1 are associated with temperament traits (Garza-Brenner et al., 2017;Chen et al., 2020;Paredes-Sánchez et al., 2020).Furthermore, it was reported that SNP located in introns represented 41% of the variations associated with behavioral traits (Li et al., 2021).The relationship and concordance between these reports demonstrate the need to continue the efforts to characterize genes and their interactions to define the genetic architecture of cattle temperament (Garza-Brenner et al., 2017;Chen et al., 2020;Paredes-Sánchez et al., 2020;Li et al., 2021).Genetics contribute to differences in the expression of the phenotypes of complex traits, and population size can play an essential role in masking these differences.Focusing on individuals showing extreme expression of the trait of interest facilitated the discovery of associations (Paredes-Sánchez et al., 2020); however, the effect of variations and their distributions must be confirmed by studies in unselected populations (Petrakova et al., 2012).
Herein, we evaluated the effect of five previously reported intronic SNP (rs3423464051-CACNG4, rs109393235-EXOC4, rs135982573-NRXN3, rs110864071-SLC9A4, and rs109722627-SLC9A4) associated with temperament traits based on both in silico functional analysis and association analysis in a Brahman population.

Ethical report
All practices complied with the Guide for the Care and Use of Agricultural Animals in Research andTeaching 2010 (AUP 2002-315).

Source of data
Data from 250 Brahman calves (126 males and 124 females) born in 2018 (n = 116; 56 males and 60 females), 2019 (n = 67; 37 males and 31 females), and 2020 (n = 66, 33 males and 33 females) were randomly selected and included in this study.The cattle management and temperament evaluation procedures were previously described (Schmidt et al., 2014).Briefly, the following phenotypic evaluations were performed at weaning.Exit velocity (EV), which is an objective test that measures the velocity of an animal traveling 1.83 m after receiving a stimulus, was measured with an infrared sensor (FarmerTek Inc., North Wylie, TX, USA) (Curley et al., 2006).Pen score (PS), which is a subjective measure based on the visual evaluation of animal behavior while confined to a pen in a group of five animals, was recorded on a scale from 1, indicating calm, to 5, indicating aggressive (Hammond et al., 1996).Finally, temperament score (TS) was calculated as the average of the EV and PS values: [TS = (EV + PS)/2] (Burdick et al., 2011).
Based on the reported EV records of the population, the animals were grouped into three temperament categories: calm, with a range of 0.16-1.82m/s and 0.4-1.56m/s; intermediate with a range of 1.83-3.12m/s and 1.57-3.04m/s; and temperamental with a range of 3.13-7.66m/s and 3.05-10.83m/s for females and males, respectively (Garza-Brenner et al., 2017;Paredes-Sánchez et al., 2020).

Genotyping
Ear notch samples were obtained from all animals at weaning.Samples were stored at −80 °C until DNA extraction.DNA extraction was performed using the commercial Genelute Mammalian Genomic DNA kit (Cat.G1N350, Sigma-Aldrich Co. LLC, St. Louis, Missouri, USA).Primers for the amplification of a gene region encompassing the rs3423464051-CACNG4, rs109393235-EXOC4, rs135982573-NRXN3, rs110864071-SLC9A4, and rs109722627-SLC9A4 were designed in AmplifX 2.0.7 software.As part of Nextera ® XT sequencing protocol (Part #15044223 Rev. B), primer design included an adapter sequence to each primer (forward: TCGTCGGCAGCGTCAGATGTGTATAAGAGACAG and reverse: GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAG).Table 1 shows the primer sequences and the expected PCR product sizes.Polymerase chain reaction (PCR) amplifications were done in 25 µL volumes on a Thermocycler DNA Engine TETRAD 2 Peltier thermal cycler (MJ Research, Inc., Waltham, MA, USA).The reaction mixtures contained 50 ng of genomic DNA, 1.5 mM MgCl2, 0.1 µM of each primer, 0.4 mM dNTPs, and 2.5 U of GoTaq ® DNA Polymerase (Promega, Madison, WI, USA).A touchdown method was used, and the amplification profile included an initial denaturation step of 95 °C for 10 min, five three-step cycles of 45 s at 95 °C, an annealing step for 45 s that started at 65 °C but decreased by 2 °C during each cycle and 45 s at 72 °C, and 25 three-step cycles of 45 s each at 95 °C, 60 °C and 72 °C.The PCR products were verified in 2% agarose gel electrophoresis.Allelic discrimination of each SNP was achieved either by sequencing using the Nextera ® XT DNA protocol in MINIseq Illumina equipment (Part #15044223 Rev. B) and by Restriction Fragment Length Polymorphism (PCR-RFLP).For the PCR-RFLP, amplicons were generated as in the previous description and were digested with the specific enzymes selected for allelic discrimination with Watcut software (Palmer, 2007), with digestion patterns verified in silico by NEBcutter (Vincze et al., 2003) (Table 2).The PCR products were digested using standard protocols under the conditions Gene ID includes gene abbreviation, reference sequence number, and Bos taurus autosome (BTA); gene structure -gene length and number of exons and introns; SNP -Ensembl database identifier; POS -intron locations after colon number of genomic position; primers: F -forward and R -reverse; PCR size -amplicon size in pair base (pb).
suggested by each enzyme provided and incubated at 37 °C during 5 h.The digested fragments were electrophoresed in 2.5% agarose gels stained with Syber Gold 1X and visualized on a transilluminator (Kodak Gel-Logic 112, Burlington, USA).Determination of genotype of each sample was achieved by visual analysis of the gels (Figure 1).A database with the genotypes was generated in the Excel tool of the Microsoft Office package.We used the database for computed allelic frequencies and Hardy-Weinberg equilibrium through the software Cervus 3.0.7 (Kalinowski et al., 2007).

Statistical analysis
To determine the effect of loci on the studied temperament characteristics (EV, PS, and TS), a mixed model was fitted as follows:   in which Y ijklm was the random variable (EV, PS, TS), μ was the overall mean, Si i was the random effect of the sire (i = 11), Sx j was the fixed effect of animals' sex (j = male, female), Y k was the fixed effects of animals' year of birth (k = 2018, 2019, 2020), β 1 X l + β 2 X l 2 was the linear and quadratic effect of age of dam (l = 2022-2019), G m was the fixed effect of genotypes in assessed loci (m = 0, 1, 2), and ε ijklm was de random residual error.Least square means were estimated of each trait and compared using a Bonferroni adjustment.All analyses were performed using SAS software (Statistical Analysis System, SAS OnDemand for Academics).
2.5.Functional analysis of intronic SNP 2.5.1.Prediction of the effect of intronic SNP on splicing sites To perform an in silico analysis of the effects of previously associated intronic SNP (Paredes-Sánchez et al., 2020), a search for splicing sites was carried out in the region where the SNP were located.Briefly, for the generation of mature mRNA, exons need be identified and joined together in a precise process that requires the coordinated activity of small nuclear RNA.The ESEfinder 3.0 (Cartegni et al., 2003;Smith et al., 2006) tool enabled the identification of splicing site motifs.This tool analyzes 5ʹ,3ʹ splice sites and branch sites to reveal motifs needed to correct splicing.A pair of input sequences (500 bp flanking the position of the SNP) that contained the alleles to be evaluated were identified, and the existence of a splicing site was determined based on the default threshold values (5 'donor = 6.67, 3 'acceptor = 6.632 and branch site = 0) suggested for the platform.

Interaction networks
To evaluate the relationships between the studied genes and thus verify whether they act together in the control of bovine temperament, the protein-protein interactions of the CACNG4, EXOC4, NRXN3, and SLC9A4 genes were considered based on the STRING database (Szklarczyk et al., 2015), displaying a reference of the interactions reported for homologous proteins in humans.The references used by the STRING database were known interactions based on curated databases and experimental determination; predicted interactions based on the gene neighborhood, gene fusions, and gene co-occurrence; and other interactions based on text mining, co-expression, and protein homology.Information that was in accord with the STRING interaction database was then employed to produce an individual score and a combined score.

Temperament assessment
Regarding the temperament distribution of the population, the frequency of intermediate animals was highest, followed by temperamental and calm animals.The distribution of animals in the three groups differed depending on the year of birth (Figure 2A).Animals born in 2018 presented higher frequencies of the intermediate and temperamental phenotypes than the calm phenotypes, while in the 2019 population, the calm phenotype showed the highest frequency.The temperamental phenotype was greater in those animals born in 2020.
The results of our mixed model-based population analysis allowed us to identify differences in the mean results of the temperament tests according to year of birth.Calves born in 2019 presented lower mean EV and TS values (1.6741 and 1.8021) than those born in 2018 and 2020, which presented higher mean values (Figure 2B), whereas no differences by year were found between PS means.

Effect of SNP on temperament traits
Four of the five evaluated SNP were determined to be polymorphic (CACNG4-rs3423464051, NRXN3-rs135982573, SLC9A4-rs110864071, and rs109722627) in the population (minor allele frequency > 10%).Based on the allelic frequency (Table 3), the EXOC4 SNP was excluded from further evaluation.Only rs3423464051:G>A of the CACNG4 gene was significantly associated with EV and TS (P<0.05;Table 4).For EV and TS, both homozygous genotypes were different (P = 0.0223 and P = 0.0211) from the heterozygote genotype and had a higher mean EV than heterozygote, ranging from 0.51-0.54m/s for AA and GG; in the TS test, both genotypes also presented values higher than the heterozygous range of 0.36-0.46for AA and GG.

Prediction of the effect of SNP on splicing sites
Because of the locations of the SNP in the introns of genes, the SNP were analyzed to determine the effect of each alternative allele in the splicing process (Table 5).The allelic variants in CACNG4, EXOC4, and SLC9A4 caused a change in the branch point (BS) recognition site, although the SNP in the NRXN3 gene did not.According to the prediction matrix, the CACNG4-rs3423464051, EXOC4-rs109393235, and SLC9A4-rs109722627 SNP had an effect on BS due to a change in the alternative allele.

Protein-protein interaction analysis
The protein-protein interaction analysis showed that the tested proteins did not interact directly (Figure 3).However, CACNG4 and NRXN3 were connected through an interaction with the GRIA2 gene (glutamate receptor 2).This interaction was supported by functional evidence from co-expression, experimental or biochemical data, and data mining in the STRING database, where the combined scores representing evidence of interactions with GRIA2 were 0.862 for CACNG4 and 0.578 for NRXN3.

SNP effects on temperament: Role of the CACNG4 gene in bovine temperament
Polymorphisms in four genes previously associated with temperament traits in beef cattle were assessed, and rs3423464051 of the CACNG4 gene was confirmed to be associated with EV and TS in a Brahman cattle population.Previous studies of this SNP showed that genotype AA was significantly different from genotypes GG and AG (Paredes-Sánchez et al., 2020), and the authors ascertained a genotypic effect reflected by the AA genotype with a higher mean EV value than the heterozygous AG and homozygous GG genotypes.Here, we were unable to differentiate the effect of each homozygous genotype.It is important to increase the investigated sample number to confirm whether rs3423464051 may be useful for identifying temperamental animals.
The functions of most genes harboring markers associated with cattle temperament and their roles in determining this trait are still unclear.To elucidate these functions, the study of temperament relies on gene orthology, since neural disorders and addictions present similarities with temperament regarding the genes involved in these phenotypes.Bioinformatic approaches have been used as an essential tool facilitating the analysis of complex traits in different organisms.These efforts focus on the identification of a candidate gene and the characterization of potential mutations affecting the trait (Zhang et al., 2017).Here, we applied two bioinformatic approaches to obtain evidence about the roles of the studied intronic SNP.Specifically, the confirmed candidate gene CACNG4 (encoding Candidate genes linked by a line have a prediction based on physical or functional interaction.The GRIA2 gene mediates the interaction between the NRXN3 and CACNG4 genes.However, the EXOC4 and SLC9A4 genes do not exhibit direct interactions with each other or with other genes.the gamma-4 subunit) identified in cattle has been reported to be involved in the MAP kinase and oxytocin signaling pathways and in contraction function and heart muscle problems, as reported in the KEGG pathway database (Kanehisa and Sato, 2020).
A specific function of this gene in representative cattle temperament pathways is still unknown.However, CACNG4 plays an integral role in calcium channels, allowing the entry of Ca 2+ into the cell, which in turn uses it as a secondary messenger in the functions and differentiation of nerve cells.The associations of CACNG4, thus, indicate that it is related to information transmission activities and the formation of nerve cells (Yin et al., 2016).
Evidence indicates that gamma 2-8 subunits are not found in neuronal or cardiac calcium complexes but act as modifying proteins of the AMPA glutamate receptor (Heyes et al., 2015).The modifying activity of CACNG4 toward the AMPA receptor is an important finding in temperament research, and it is essential to investigate whether CACNG4 modification acts in a general way or is specific to AMPA receptor subunits.
Our in silico protein-protein analysis included modifying proteins of AMPA receptors.The interaction between CACNG4 and NRXN3 is mediated by glutamate receptor 2 (GRIA2, AMPA subunit), which acts as a ligand in channels of the central nervous system and plays an essential role in excitatory synaptic transmission according to the STRING database.GRIA2 has been previously associated with temperament (Lindholm-Perry et al., 2015), and subsequently identified an interaction between NRXN3 and GRIA2 (Paredes-Sánchez et al., 2020).Reports in humans have shown that the GRIA2 gene works in conjunction with AMPA-type glutamate receptors.One of its primary activities is to connect excitatory transmissions underlying perception, cognition, and movement.Its function requires auxiliary proteins such as calcium channel subunits (CACNG4).It participates in multiple activities in the brain; its expression has been identified in subcortical regions (Shen and Limon, 2021) and its dysregulation has been associated with abnormal behavior.Some gamma calcium channel subunits, such as CACNG2 and CACNG4, show similarities in the binding sites between AMPA receptor subunits and PDZ proteins (proteins with AMPA subunit interaction domains), which suggests a physical association with AMPA receptors and PDZ proteins during synapse formation (Kious et al., 2002).The AMPA receptor must contain at least one GRIA2 subunit, which strengthens the interaction domain (Tomita et al., 2001).In contrast to CACNG4, acting in the subcortex of the brain, both GRIA2 and NRXN3 act in the hippocampus.Specifically, the NRXN3 protein is necessary for the control of postsynaptic AMPA expression and the formation of the postsynaptic density protein complex (Bayés et al., 2011;Aoto et al., 2015).The NRXN3 gene has been associated with human neurodevelopmental disorders and with a locus related to characteristics of empathy or fear in mice (Di Gregorio et al., 2017;Keum et al., 2018).In general, the pathways involving the CACNG4, GRIA2, and NRXN3 genes include postsynaptic activities, protein aggregation, and functions within complexes formed at synapses, and variants in these genes can disturb these complexes, with proportionate changes in signal transmission to the environment.
An interesting characteristic of intronic SNP is their potential to exert regulatory effects.Intronic SNP have been associated with the splicing and alteration of recognition sites for spliceosome binding (Kol et al., 2005).Due to the gene locations of all intronic SNP analyzed in this report, their possible roles in splicing were evaluated.Three of these five SNP exhibited one allelic variant with the potential to alter branch point recognition.Although the splicing machinery is highly conserved in mammals, branch point sequences are highly variable, and this variability may be a mechanism allowing the alternative splicing of a gene (Ast, 2004).It has been reported that the presence of a SNP near a branch point site that does not interrupt splicing recognition can lead to a change in splicing efficiency and, consequently, increased or decreased expression at the mRNA and protein levels (Zhang et al., 2018;Mucaki et al., 2020).
Based on the characterization and prediction of intronic branch points in the bovine genome, it has been suggested that branch points in bovine introns occur within degenerate heptamers with a consensus sequence of "nnyTrAy", indicating conserved thymine and adenine residues in the branch point sequence (Kadri et al., 2021).Therefore, the conserved sequence can change in the presence of an SNP allele that interferes with the recognition of the sequence by the spliceosome.Branch points analysis in human model indicate a strong preference toward adenine (Kadri et al., 2021).Simultaneously, the EXOC4 and CACNG4 markers harbor the A allele, where we found that allelic change in the preferential residues of the heptamer causes the loss of the recognition site.
In human diseases, it is common to find structural variations associated with a pathological condition or that contribute to it (Kozyrev et al., 2008).The SNP located at a branch point accompanied by a nonsynonymous change with strong linkage disequilibrium caused variation in gene expression and protein localization by inducing splicing and modifying the length of the protein (Martínez-Bueno et al., 2018).This was experimentally confirmed through the construction of a minigene with allelic variants at intronic and exonic positions corresponding to the associated SNP (Kozyrev et al., 2012).

Conclusions
The genotype-phenotype association analysis showed the effect of CACNG4 gene polymorphism on exit velocity and temperament score.In silico analysis allowed us to infer that CACNG4 and NRXN3 are regulated by the GRIA2 gene, previously associated with bovine temperament.Likewise, intronic allelic variants can alter branch points, which are essential sites for spliceosome recognition.
Gene -gene abbreviation; SNP ID -Ensembl database identifier; CACNG4 -calcium voltage-gated channel auxiliary subunit gamma-4; NRXN3neurexin-3; SLC9A4 -solute carrier family 9 member A4; digestion pattern (bp) -digest pattern size in base pairs (bp).MP -DNA marker ladder.The figures show the PCR-RFLP results for each marker-genotype assignations.A: For CACNG4-rs3423464051, the GG genotype is characterized by 104, 77, and 61 pb bands; AA genotype is distinguished for the 120 and 77-pb bands.B: In the SLC9A4-rs110864071, the CC genotype displays two bands of 190 and 138-bp, while TT genotype has a differential 235 and 138-pb pattern.C: The SLC9A4-rs109722627 GG genotype has two bands of 189 and 140 pb, while the TT genotype shows a main band of 329.D: The GG genotype of NRXN3-rs135982573 shows a differential band of 316-pb and AA genotype, and has three bands of 182, 134, and 52-pb.In all markers, the heterozygote carriers are characterized by a band pattern combining both homozygotes.

A:
Distribution of animals segregated into three temperament classifications and year of birth (black: calm; dark gray: intermediate; light gray: temperamental).B: Mean values of exit velocity (EV), pen score (PS), and temperament score (TS) by year of birth (black: 2018, dark gray: 2019, and light gray: 2020).

Figure 2 -
Figure 2 -Population and temperament values by year of birth.

Table 1 -
Characteristics of candidate genes, SNP, and primers

Table 2 -
Allelic discrimination of intronic SNP via RFLP analysis

Table 3 -
Allelic frequencies of candidate genes for temperamental traits in the Brahman cattle population Gene -gene abbreviation; SNP ID -Ensembl database identifier; N -number of samples; HW -Hardy-Weinberg equilibrium; NS -non-significant.*** Significant at the 0.1% level.

Table 4 -
Least-square means of the effect of assessed SNP on the temperament traits of Brahman cattle, exit velocity (EV), pen score (PS), and temperament score (TS) in the confirmation population Gene-gene abbreviation; SNP ID -Ensembl database identifier; N -number of samples.P-values are in parentheses.a,b -Means with different letters for each trait and SNP are significantly different (P<0.05).

Table 5 -
Sites and prediction values of the effect of intronic SNP on the splicing mechanism -gene abbreviation; SNP ID -Ensembl database identifier; motif -splicing motif site; site: sequence heptamer of recognition. Gene