Introduction

For beef cattle, birth weight is an important trait that represents the first phenotypic variable of a new individual in a population and is also considered an indicator of calving ease (^{Canellas et al., 2012}). Furthermore, for beef cattle producers, calves with greater weights at weaning have the advantage of being more morphologically developed and better equipped to successfully cope with the environment (^{Jahuey-Martínez et al., 2016}).

Databases with phenotypic records that describe growth usually include birth (BW), weaning (WW), and yearling (YW) weight (^{Jahuey-Martínez et al., 2016}). The expression of growth traits is influenced by multiple environmental and genetic factors, and innumerable interactions of this trait with others, such as temperament, have been described. Temperament, defined as the response of an animal to handling by humans (^{Burrow and Dillon, 1997}), has been studied in cattle using different testing approaches (^{Haskell et al., 2014}; ^{Friedrich et al., 2015}). For example, flight speed is a measure of the time it takes for an animal to traverse a certain distance after being contained in a chute (^{Lindholm-Perry et al., 2015}). Different studies have found that cattle with slower flight speeds gain weight more rapidly than those with faster flight speeds (^{Burrow and Dillon, 1997}; ^{Voisinet et al., 1997}; ^{Cafe et al., 2011}) and that temperamental cattle have lower live weights than docile cattle (^{Fordyce et al., 1985}). Exit velocity (EV) is frequently used as a measure of cattle temperament. It is defined as the rate (m/s) at which an animal traverses a specific distance after exiting a squeeze chute (^{Curley et al., 2006}). ^{Petherick et al. (2002)} found a negative correlation between flight speed and daily weight gain in 5/8 Brahman × 3/8 Shorthorn steers; they found that flight speed was a reliable predictor of the performance of the animals. Similarly, in a study with many animals, unfavorable genetic and phenotypic relationships were observed between average daily gain (ADG) estimated from weaning to yearling in Bos indicus (Nellore) cattle, demonstrating that animals with high flight speeds had lower body weights (^{Sant’Anna et al., 2012}). Similarly, a study with Bos taurus cattle reported that animals exhibiting more excitable temperaments were phenotypically more likely to have a lower body weight at entry into the feedlot than more docile cattle (^{Reinhardt et al., 2009}). Thus, investigators and producers have increasingly focused on the reaction of livestock during management using diverse methods to determine the temperament of the animals (^{Haskell et al., 2014}). Such research is based on evidence from the literature, in which docility is correlated with not only the ease of managing cattle but also with economically relevant traits (^{Haskell et al., 2014}; ^{Friedrich et al., 2015}).

Previous studies have identified growth-influencing genes in Mexican Charolais cattle (^{Jahuey-Martínez et al., 2016}). ^{Garza-Brenner et al. (2017)} identified new genes and polymorphisms associated with bovine temperament as well as specific single-nucleotide polymorphisms (SNP) that showed effects on temperament-related traits such as EV and pen score (PS). Previous reports have not determined the effect of these genes on both characteristics (growth traits and temperament); however, these genes could have this potential due to their reported biological activities (e.g., the genes proopiomelanocortin [POMC], neuropeptide Y [NPY], and certain genes belonging to the solute carrier family [SLC18A2]).

The objective of this study was to determine how a panel of molecular markers previously associated with temperament affect growth characteristics (BW, WW, and YW) in Charolais cows.

Material and Methods

The studied animal populations were included in the Animal Biotechnology Biobank and represented animals from four herds located in the northwest of México: herd 1 (n = 50), herd 2 (n = 77), herd 3 (n = 145), and herd 4 (n = 140). All cows in this study were born between 2004 and 2013, and similar management objectives in each herd are based on the sale of breeding stock and breeding purebred Charolais cattle. For correlation analysis of growth traits and temperament, data from selected animals (412 Charolais cows) were obtained from the Charolais Herd Book de México A.C.^© (consulted July 2017) online database (http://www.ronbmexico.com/charolais/BusquedaAnimales.aspx), using the registration number of each cow. Data included the registered BW, WW, and YW. As the animal age is an important factor affecting temperament (^{Curley et al., 2006}), cows were categorized into two groups according to age: mature cows ≥ four years old (n = 270) and young cows of 2-3 years old (n = 142).

Temperament was assessed using EV and PS. For EV, the velocity of a cow as it traversed a 1.83-m (6 ft) distance was recorded with a 2-infrared sensor (FarmTek Inc., North Wylie, TX, USA) following the stimulus of hair sampling before exiting a squeeze chute (^{Curley et al., 2006}). The velocity was calculated as EV = distance (m)/time (s). Pen score was assessed by three evaluators using a five-point scale, with scores ranging from 1 (calm) to 5 (moving aggressively), as described by ^{Hammond et al. (1996)}. Individual temperament score (TS) values were calculated by averaging the PS and EV [TS = (PS + EV)/2] (Garza-Brenner et al., 2017). Records of PS, EV, and TS were obtained for each animal from the available database of the Animal Biotechnology Laboratory; these data were recorded once at the time of hair sampling from each animal (^{Garza-Brenner et al., 2017}).

Due to financial restrictions, we implemented a selective genotyping strategy for the association analysis. From the 412 analyzed cows, a group of 80 individuals identified as the most docile (n = 41) and temperamental (n = 39) were genotyped with a 151 SNP panel using the Sequenom MassARRAY^® platform (GeneSeek, Inc., Lincoln, NE, USA). All SNP were selected from a previously reported SNV (single nucleotide variations) panel described by ^{Garza-Brenner et al. (2017)}. The ID and gene location of each SNP are described in Table 1. The genotypic and allelic frequencies were estimated using Genepop^® web version 4.0.10 software (^{Rousset et al., 2008}). Before the association analysis, the quality of the genotypic data was verified. The SNP that were monomorphic (n = 14) or presented minor allele frequencies <0.01 (n = 19) (Table 1) were eliminated. The allelic frequencies of each tested SNP are shown in Table 1.

Pearson's correlation coefficients among traits evaluated were determined. A general linear model was fitted for BW, WW, and YW as follows:

in which Y_ijk = BW, WW, and YW, which represent the dependent traits of this study; µ = the overall mean value; HD_i = the i-th herd effect (herd 1, herd 2, … herd 4); GE_j = the j-th age group effect (young and mature cows); G_k = the effect of the k-th genotype in each individual SNP; and ε_ijk = the random error. The genetic diversity information of those SNP that resulted with effect on live growth traits was calculated using the GenAlex Software (^{Peakall and Smouse, 2012}).

The least squares mean of the genotypes was estimated for the SNP that demonstrated significant effects (P<0.05), and means were compared using the PDIFF statement. All statistical procedures were performed in SAS software (Statistical Analysis System, version 9.4).

Results

The least square mean values for live growth traits are described in Table 2.

In the young cow group, low and moderate but significant correlations with variables BW (P<0.0001) and WW (P<0.03) were found only for two temperament measurements (EV and TS). However, the variable YW was not correlated with any temperament measurements (PS, EV, and TS) (Table 3). In the mature cow group, no significant correlations were observed between temperament and weight traits (Table 3).

Ten SNP located on six genes (Table 4) resulted in an association with one or two of the growth measures (nine for BW and three for WW). No significant associations were found for YW. Genetic diversity parameters of the ten SNP are shown in Table 5, while in Table 6, the least square means of each growth trait are reported, also including the least square means data of temperament traits (^{Garza-Brenner et al., 2017}). Seven markers (rs41749779 [DRD2]; rs134256715 [MAOA; monoamine oxidase A]; rs134604486, rs136809285, rs137756569, rs41257366 [POMC; proopiomelanocortin]; and rs43696138 [HTR2A; serotonin 5-hydroxytryptamine receptor 2A]) were associated with BW, in which the most significant associations were observed for markers rs41749779 and rs134256715 (P = 0.008 and P = 0.002, respectively). Marker rs17871681, located on the POMC gene, was significantly associated (P = 0.0085) with WW.

Markers rs109576799 and rs385054562, located on the dopamine D3 receptor (DRD3) and tryptophan 2,3-dioxygenase (TDO2) genes, respectively, were significantly associated (P = 0.05) with BW and WW. Cows with genotype CC of marker rs385054562 had a higher BW than cows with the TT genotype (39.36 vs 36.74 kg; P = 0.042); similarly, for WW, cows with CC genotype had higher WW than cows with the TT genotype (240.19 vs 224.49 kg; P = 0.043). Cows with the CC genotype of marker rs109576799, located on the DRD3 gene, had higher BW and were 5.32 kg (P = 0.004) and 4.59 kg (P = 0.014) heavier than cows with the AA and CA genotypes, respectively.

Discussion

From a molecular perspective, the use of candidate genes that affect economically important traits has been shown to be a direct method of understanding biological functions that focuses on the trait of interest. This approach consists of exploring genes that are implicated in known biological pathways and define whether the genetic variation present in populations is associated with phenotypic differences (^{Mormède et al., 2005}). By using this approach, investigations have been performed to elucidate the molecular basis of bovine temperament (^{Glenske et al., 2011}; ^{Lühken et al., 2010}). Recently, ^{Garza-Brenner et al. (2017)} conducted a study of genes and markers associated with bovine temperament in Mexican Charolais cows. The selected markers were located on genes of the dopamine and serotonergic pathways as well as five protein-protein interacting genes. Among complex and productive traits of interest, growth has been extensively studied, and its genetic architecture is currently under investigation (^{Lühken et al., 2010}). Because cattle temperament affects other important traits such as cattle weight, we analyzed whether SNP known to be associated with temperament traits are also associated with live weight traits.

Results of Pearson's correlation coefficient in this study were inconsistent with those described in other studies focused on correlating temperament with growth, because all the correlations found here were positive. This finding could be related to the time at which temperament was recorded (^{Curley et al., 2006}; ^{Schmidt et al., 2014}). Temperament of animals is usually recorded around weaning time, either pre- or post-weaning. The mature cows likely did not show a correlation with the temperament traits because of their age; therefore, age could have had an influence, because older cows are acclimated to the location and environment where they have lived for years (^{Curley et al., 2006}; ^{Schmidt et al., 2014}).

Using the 151 SNP panel previously reported by ^{Garza-Brenner et al. (2017)}, we identified associations of ten SNP located on six genes with two of the three recorded growth measures (BW and WW). Interestingly, four of the SNP associated in this study have been previously described as having associations for temperament traits (Table 4) (^{Garza-Brenner et al., 2017}).

Research focusing on determining pleiotropic interactions has recently increased because of the availability of dense SNP arrays and the development of statistical methodologies capable of probing multitrait-marker interactions (^{Gianola et al., 2015}). In cattle, for instance, ^{Lindholm-Perry et al. (2015)} analyzed the role of SNP located on BTA6 previously associated with phenotypic and efficiency traits, including frame size, average daily gain, and average daily feed intake. The authors found that they were also associated with flight speed. The markers identified are located within a coding region of non-SMC condensin I complex subunit G (NCAPG) and in the 30-UTR of the ligand-dependent nuclear receptor corepressor-like (LCORL) gene.

The present study revealed that the markers rs109576799 (DRD3), rs134604468, rs137756569 (POMC), and rs43696138 (HTR2A), previously associated with bovine temperament traits (^{Garza-Brenner et al., 2017}), were also associated with live weight traits (BW and WW). Further studies analyzing our results could determine whether the four markers had a pleiotropic effect on temperament and live weight traits. The results of these studies require the conformation and analysis of a larger cattle population to confirm the effect of the identified markers on the traits of interest.

Except for the POMC gene, there are no previous reports describing any associations between the genes found in the present study and body weight. For the POMC gene, an association was reported between polymorphisms located at the 3’ flanking region with body weight and weight gain of Nanyang cattle (^{Zhang et al., 2009}).

The specific association of marker rs134256715, which is located on the MAOA gene, with BW found in our study is important because it is a non-synonymous SNP (Cys/Trp) and, therefore, has the potential to modify the protein structure. The MAOA gene has been termed as “warrior gene” because of its association with the response to aggression in several behavioral studies, mainly in humans (^{Kim-Cohen et al., 2006}; ^{McDermott et al., 2009}). The MAOA gene plays an important role in inter-individual variability in aggressiveness, impulsive response, and serotoninergic response capacity of the central nervous system (CNS) as well as in complex behavior regulation. ^{Lühken et al. (2010)} studied the genetic variation of the gene in two cattle breeds (Angus and German Simmental), which are known to differ in their behavior during handling (German Simmental was reported as more difficult to handle than Angus). The authors evaluated five SNP different from those included in our study and did not observe a significant association between the polymorphism and the recorded behavioral scores.

An interesting result of our study was the novel associations found for two body weight traits (BW and WW) and marker rs385054562 located on TDO2 and marker rs109576799 located on DRD3. The TDO2 gene does not have any reported associations in cattle. In studies conducted with humans and mice, TDO2 has been identified as a potential candidate gene for autism and as a modulator of behavioral diseases associated with inflammatory states, psychiatric disorders, behavioral modulation, and cognitive function (^{Nabi et al., 2004}; ^{Too et al., 2016}). Marker rs109576799 is located in an intron of the DRD3 gene and has been associated with EV and TS (^{Garza-Brenner et al., 2017}). It is important to determine the role of this marker in gene function, because variations in non-coding regions may be involved in splicing (and alternative splicing) and in splicing efficiency, thus affecting gene expression and regulation (^{Ramírez-Bello and Jiménez-Morales, 2017}).

Further attention should be given to polymorphisms and genes that have shown important associations with the studied traits to validate their effects on other cattle populations, especially because four of the same markers previously associated with temperament have also been associated with growth traits.

Conclusions

This report presents the first findings in which single-nucleotide polymorphisms located on candidate genes for temperament traits also had an effect on birth weight and weaning weight in Charolais cows, which indicates that both traits could be influenced by the same genes. Once validated, this information would assist in the selection of appropriate genotypes for the economically relevant traits temperament and live weight.