Effects of the Combination of the C1473G Mutation in the Tph2 Gene and Lethal Yellow Mutations in the Raly-Agouti Locus on Behavior, Brain 5-HT and Melanocortin Systems in Mice

Tryptophan hydroxylase 2 (TPH2) is the key and rate-limited enzyme of serotonin (5-HT) synthesis in the brain. The C1473G mutation in the Tph2 gene results in a two-fold decrease in enzyme activity in the mouse brain. The lethal yellow (AY) mutation in the Raly-Agouti locus results in the overexpression of the Agouti gene in the brain and causes obesity and depressive-like behavior in mice. Herein, the possible influences of these mutations and their combination on body mass, behavior, brain 5-HT and melanocortin systems in mice of the B6-1473CC/aa. B6-1473CC/AYa, B6-1473GG/aa are investigated. B6-1473GG/AYa genotypes were studied. The 1473G and AY alleles increase the activity of TPH2 and the expression of the Agouti gene, respectively, but they do not alter 5-HT and 5-HIAA levels or the expression of the genes Tph2, Maoa, Slc6a4, Htr1a, Htr2a, Mc3r and Mc4r in the brain. The 1473G allele attenuates weight gain and depressive-like immobility in the forced swim test, while the AY allele increases body weight gain and depressive-like immobility. The combination of these alleles results in hind limb dystonia in the B6-1473GG/AYa mice. This is the first evidence for the interaction between the C1473G and AY mutations.


Introduction
The enzyme tryptophan hydroxylase 2 (TPH2) hydroxylates L-tryptophan to 5-hydroxy tryptophan as the first and rate-limited stage of serotonin (5-HT) synthesis in the mammalian brain [1,2]. Tph2 gene knockout [3][4][5] or the TPH2 inhibitor p-chlorophenylalanine [6] dramatically reduce the brain's 5-HT level. The C1473G mutation in the Tph2 gene, which results in the P447R substitution in the enzyme molecule, causes a two-fold decrease in TPH2 activity in the mouse brain [7][8][9]. Taking into account the association between mutations affecting TPH2 activity and the risk of psychiatric diseases (see reviews [10,11]), it seems surprising that the 1473G allele does not appear to produce any visible effect on adaptive behavior in laboratory mice [12,13]. However, a high (wild-type) TPH2 activity seems to be important for mouse adaptation, since mice carrying the mutant 1473G allele, which decreases the activity of TPH2, are absent in wild mouse populations [14]. Mutations reducing the activities of aromatic amino acid hydroxylases are associated with dystonia [15]. In our preliminary study, we observed hind limb dystonia in young, 4-week-old mice of the 1473GG genotype. However, this movement disorder disappeared in adult 12-week-old mice (see the Supplementary Table S1).
The lethal yellow (A Y ) mutation is a large deletion in the Raly-Agouti locus that deletes the Raly gene-coding domain and puts the Agouti gene under the control of the Raly The experiments were carried out on SPF-state 12-week-old males of the B6-1473CC/aa (n = 8), B6-1473CC/A Y a (n = 8), B6-1473GG/aa (n = 8) and B6-1473GG/A Y a (n = 8) genotypes. The B6-1473CC/aa (50%) and B6-1473CC/A Y a (50%) genotypes were bred by crossing B6-1473CC/aa females and B6-1473CC/A Y a males. The B6-1473GG/aa and B6-1473GG/A Y a genotypes were bred in two steps. In the first step, the B6-1473GC/A Y a (50%) and B6-1473GC/aa (50%) genotypes were bred by crossing B6-1473GG/aa females and B6-1473CC/A Y a males. In the second step, the B6-1473CG/ A Y a males were crossed with B6-1473GG/aa females, and consequently, the B6-1473GG/aa (25%), B6-1473GG/A Y a (25%), B6-1473CG/aa (25%) and B6-1473CG/A Y a (25%) genotypes were received. The B6-1473CC/aa, B6-1473CC/A Y a, B6-1473GG/aa and B6-1473GG/A Y a mice used in this experiment have similar genetic backgrounds (C57BL/6) and differ only in the 1473C, 1473G, a and A Y alleles. The aa and A Y /a genotypes were detected by the color of their fur, which is yellow in A Y /a mice and black in a/a mice [17]. The 1473C and 1473G alleles were detected via PCR with the allele-specific primers (see Section 2.6).
After weaning, males of the same genotype were kept in groups of four per cage (Optimice, Animal Care Systems, Inc., Centennial, CO, USA) at a temperature of 24 ± 2 • C, humidity of 45-50%, and an artificial 14:10 (day/night) photoperiod with daybreak and sunset at 01:00 and 15:00, respectively. The mice were fed with sterile food and water ad libitum. The mice were tested at 12 weeks of age. Two days before the first test, the animals were isolated in cages of the same type and size to reduce group effect. We used a rapid test battery with 1-day intervals between tests [25]. The test sequence in these batteries was the following: open field, elevated plus maze, forced swim and tail suspension tests [26].
Two days after the tail suspension test, the animals were euthanized via carbon dioxide asphyxiation followed by decapitation. The hypothalamus, frontal cortex, hippocampus, striatum and midbrain were rapidly dissected, frozen in liquid nitrogen, and stored at −80 • C.

Open Field Test
The open field test was carried out on a brightly illuminated, white, opaque plastic arena 60 cm in diameter with a 30 cm wall. We used transmitted (inverted) lighting to maximize the mouse/arena contrast [27]. The light from two 12 W halogen lamps, each placed 60 cm beneath the semitransparent floor, was transmitted through the arena to a web camera. The mouse was placed at the wall, and its movements were automatically tracked for 5 min [27]. The EthoStudio software automatically calculates two behavioral traits: the distance travelled during the test (m), and the time (%) spent in the round part that is 30 cm in diameter (the center) [27]. At the same time, the numbers of vertical postures and grooming bouts were recorded by an experienced rater [27]. The arena was cleaned with wet and dry napkins after each test.

Elevated Plus-Maze Test
The elevated plus-maze test was carried out in an apparatus made of gray plastic with two closed and two open arms (length 30 cm and width 5 cm). The closed arms were boarded with plastic walls measuring 15 cm. The apparatus was elevated 65 cm above the floor and dimly illuminated with diffuse lighting. A mouse was placed in the apparatus center, facing a closed arm, and its movement was automatically tracked for 5 min using the Microsoft Kinect 1 3D sensor. Depth data from the sensor contain the distances (m) from each pixel of the object and each pixel of the maze to the sensor. The height threshold algorithm marks pixels higher or lower than the threshold, associated with the animal (1) or background (0), respectively [24,28]. The main advantage of the 3 D sensor over standard digital video is that the former can track an animal of any color and in any part of the maze (including the closed arms). The EthoStudio software automatically calculates three traits: the distance travelled during the test (m) and the time (%) spent in the closed and open arms, based on the rates of the number of animal-associated pixels in these parts to the total number of animal-associated pixels in the maze. The apparatus was cleaned with wet and dry napkins after each test.

Forced Swim Test
The mice were placed in a clear glass tank (diameter, 18 cm; height, 30 cm) filled up to 2/3 of its volume with water at a temperature of 25 • C for 6 min. The water tank was placed on a semitransparent platform and brightly illuminated (300 lx) with two halogen lamps (35 W) placed 40 cm below the platform [29]. Behavior in the FS test was evaluated for the last four minutes of the test using the immobility time (%). The water was changed after each test.

Tail Suspension Test
In the present study, the tail suspension test was used to measure the number and accumulated time of hind limb clasping [30]. Using adhesive tape, mice were fixed by their tails and hooked onto a horizontal bar placed 30 cm above the table surface. The number and accumulated time of hind limb clasping were registered by an experienced rater.

Genotyping for C1473G Polymorphism
Using protease K treatment for 1.5 h, followed by a phenol-chloroform extraction, samples of DNA were isolated from the tail tips of the mice and diluted to a concentration of 10 ng/µL. The DNA samples were genotyped according to the published protocol [7][8][9] with two positive control primers (5 -TTTGACCCAAAGACGACCTGCTTGCA and 5 -TGCATGCTTACTAGCCAACCATGAGACA) providing a 523 bp PCR product and with either 1473C (5 -CAGAATTTCAATGCTCTGCGTGTGGG)-or 1473G allele (5 -CAGAATTTCAATGCTXTGCGTGTGGC)-specific primers providing a 307 bp PCR product [9,10]. The PCR products were resolved via electrophoresis in 2% agarose gel, stained with ethidium bromide and photographed with a digital camera.

Tissue Preparation
We used the same brain samples for the total RNA extraction and for the assays of serotonin (5-HT) and 5-hydroxyindoleacetic acid (5-HIAA) levels and theTPH2 activity.
For these purposes, the structure was homogenized in 250 µL (hypothalamus), 300 µL (cortex and striatum) or 400 µL (hippocampus and midbrain) of 50 mM Tris HCl, pH 7.6, and 1 mM dithiotreitol, using a motor-driven grinder (Z359971, Sigma-Aldrich, Darmstadt, Germany). One aliquot of 50 µL of the homogenate was mixed with 150 µL of 0.6 M HClO 4 for the 5-HT and 5-HIAA extractions (see Section 2.8). Another aliquot of 100 µL of the homogenate was mixed with Trizol reagent (Bio Rad, Hercules, CA, USA) for the total RNA extraction (see Section 2.10). The rest of the homogenate was spun for 15 min at 12,700 rpm (+4 • C). The clear supernatant was transferred into a clear tube and stored at −80 • C until the TPH2 activity assay (see Section 2.9).
The standard mixes containing 1, 2 and 3 ng of 5-HT and 5-HIAA were repeatedly assayed throughout the entire procedure and used to plot the calibration curves for each substance. The peak areas were estimated using LabSolution LG/GC software, version 5.54 (Shimadzu Corporation, Kyoto, Japan), and calibrated against calibrated curves for the corresponding standards. The 5-HT and 5-HIAA contents were expressed in ng/mg protein, as assayed via the Bradford method and described elsewhere [23]. All data are the means of three replications.

mRNA Level Assay by qPCR
The total mRNA was extracted from the mixture of homogenate and Trizol reagent (see Section 2.7) according to the manufacturer's protocol and treated with RNAase-free DNAase (Promega, Madison, WI, USA) according to the manufacturer's protocol. Its concentration was assayed with a Nanodrop 2000 (Waltham, MA, USA) and it was diluted to a final concentration of 125 ng/µL. The cDNA was synthesized using a random hexanu-cleotide primer and an R01 Kit (Biolabmix, Novosibirsk, Russia). The mRNA levels of the target genes were assayed via qPCR, using the set of selective primers (Table 1) and an R401 Kit (Sintol, Moscow, Russia) according to the protocol of the manufacturer (95 • C 5 min; (95 • C, 15 s; annealing temperature, 60 s; 82 • C, 2 s; fluorescence registration) × 40 cycles). The threshold cycles were calibrated with the external standards containing 25, 50, 100, 200, 400, 800, 1600, 3200 and 6400 copies of genomic DNA extracted from a C57BL/6 mouse liver. The gene expression was presented as a relative number of cDNA copies calculated on 100 copies of Polr2 cDNA as an internal standard [24,31,32].

Statistics
All data were tested using Kolmogorov's test and met the assumption of normality. Data were presented as the mean ± SEM, and they were analyzed via one-way ANOVA. If the one-way ANOVA detected an intergroup difference, the data were reanalyzed using a two-way ANOVA with "C1473G" and "A Y " as independent factors, including their interaction. Post hoc analyses were carried out using Fisher's LSD multiple comparison test when appropriate. Statistical significance was set at p < 0.05.

Effects of C1473G and A Y Mutations on Mouse Behavior in the Forced Swim Test
The males of the four genotypes differed in immobility time (F(3,28) = 4.6, p = 0.01) in the forced swim test. Marked effects of the "C1473G" (F(1,28) = 4.68, p = 0.04) and "A Y " (F(1,28) = 8.19, p = 0.008) factors, but not of their interaction (F(1,28) < 1), on this trait were revealed. The C1473G mutation decreases the immobility time, while the A Y mutation increases the immobility time ( Figure 4).

Effects of C1473G and A Y Mutations on Mouse Behavior in the Forced Swim Test
The males of the four genotypes differed in immobility time (F(3,28) = 4.6, p = 0.01 the forced swim test. Marked effects of the "C1473G" (F(1,28) = 4.68, p = 0.04) and " (F(1,28) = 8.19, p = 0.008) factors, but not of their interaction (F(1,28) < 1), on this trait w revealed. The C1473G mutation decreases the immobility time, while the A Y mutation creases the immobility time ( Figure 4).

Effects of C1473G and A Y Mutations on 5-HT, 5-HIAA Levels and the 5-HIAA/5-HT Turnover Rate in the Brain
The males of the four genotypes did not differ in 5-HT, 5-HIAA levels and the over rate of 5-HIAA/5-HT in the cortex   (Figure 6).

Discussion
The A Y mutation increases the body mass [17,23,24] and depressive-like immobility in the forced swim test [24,33,34]. However, this mutation did not affect the brain 5-HT system [24]. The C1473G mutation decreases brain TPH2 activity [8,9], but it does not seem to affect the behavior of mice in the widely used laboratory tests such as the open field, elevated plus-maze, and forced swim tests [12,13,35]. Recently this mutation was associated with an increase in hind limb dystonia in young, but not adult, B6-1473GG mice (Supplementary Table S1). It is hypothesized that the combination of the C1473G and A Y mutations will increase the severity of behavioral disorders caused by each of these mutations separately.
The main aim of the present study was to compare the effects of the combination of the C1473G and A Y mutations on body mass, behavior and the 5-HT and melanocortin systems in the brain with the effects of each of the two mutations. For this purpose, the four congenic mouse lines of the B6-1473CC/aa, B6-1473CC/A Y a, B6-1473GG/aa and B6-1473GG/A Y a genotypes were bred. It was confirmed that the A Y mutation caused the ectopic overexpression of the Agouti gene in the brain [20] and increased body mass and depressive-like behavior in the forced swim test [24].
A new result of this study is that the 1473G allele decreases body mass and depressivelike immobility in the forced swim test. In other words, the 1473G and A Y alleles have opposite effects on these traits. Moreover, no statistically significant effect of the interaction of these alleles on body mass and depressive-like immobility was shown. Therefore, the 1473G and A Y alleles regulate these traits by means of different and independent molecular mechanisms.
The α-, βand γmelanocyte-stimulating hormones receptors stimulate energy expenditure and reduce eating via the hypothalamic melanocortin 3 and 4 (MC3R and MC4R) [36][37][38][39][40]. The agouti protein is a natural inhibitor of the MC3R and MC4R receptors [20,21]. Usually, this protein is not expressed in the brain; however, the A Y mutation puts the Agouti gene under the promoter of the ubiquitously expressing Rally gene [17] and causes its ectopic overexpression in the brain [18][19][20] and this study. The agouti protein inhibits the hypothalamic MC3R and MC4R receptors and reduces energy expenditure and increases eating [17].
5-HT increases eating via the 5-HT1A and 5-HT2B receptors on arcuate neurons [41]. The Tph2 gene knockout that dramatically decreases brain 5-HT level reduces body mass in Tph2Ko mice compared to wild-type mice [4,5,42,43]. Moreover, the Tph2 gene knockout increases the expression of the Mc4r gene in the hypothalamus [41], which can also reduce eating and body weight gain. Here, it was found that the 1473G allele also reduces body mass. Although the reduction in TPH2 activity caused by the 1473G allele is more moderate and is not accompanied with any alterations to the Mc3r and Mc4r receptors in the hypothalamus, it seems to be sufficient to reduce body weight gain.
There is an association between obesity and depressive disorders [44][45][46][47]. It can be hypothesized that obesity is the cause of the increase in depressive-like immobility observed in the forced swim test in A Y /a mice [24]. Indeed, some authors report that the obesity caused by the leptin receptor gene (Lepr) knockout increases depressive-like immobility in Lepr db /Lepr db mice [48,49]. However, a relationship between obesity and depression seems to be more complex, and other authors do not show any increase in depressive-like behavior in these mice [50].
The relationship between TPH2 activity and depressive-like behavior is obscure. There are two commonly used indexes positively correlating to depressive-like behavior: the duration of immobility in the forced swim and tail suspension tests [51]. The effects of the TPH2 deficiency on these indexes are rather contradictory [12]: the Tph2 gene knockout can both increase [52] and decrease [4] the immobility time, and the 1473G allele can both decrease [9] and have no effect on [13] the immobility time in the forced swim test. The G1449A mutation in the Tph2 gene, which reduces the TPH2 activity by 80%, increases immobility time in the tail suspension test [53][54][55]. However, other authors did not report this trait alteration in Tph2 gene knockout mice [52] and in mice homozygous for the 1473G allele [35]. In the present study, the 1473G allele not only decreases immobility time in the forced swim test but also attenuates the increase in this trait caused by the A Y mutation.
Dystonia is a symptom of the tyrosine hydroxylase deficiency. Hind limb clasping in mice suspended by the tail is a model of dystonia [30]. Hind limb clasping was recently also found in 3-4-week-old B6-1473GG mice, while adult mice of this genotype do not show hind limb dystonia (Supplement Table S1). The most intriguing finding in the present study is the statistically significant effect of the "C1473G" and "A Y " factors' interaction on the hind limb clasping rate and duration. Although the 1473G and A Y alleles separately did not show hind limb dystonia, the combination of these combination resulted in a high level of expression of this trait in adult B6-1473GG/A Y a mice. Therefore, the A Y mutation can aggravate the negative effect (dystonia) of the C1473G mutation. At the present time, the molecular mechanism of the interaction of these two mutations unknown: no alterations in the brain 5-HT system distinguishing the B6-1473GG/A Y a and B6-1473GG/aa mice have been found. As previously mentioned, possible causes of hind limb dystonia could be deficiencies in tyrosine hydroxylase or its natural cofactor, 5,6,7,8-tetrahydrobiopterin [30]. Therefore, alterations in the tyrosine hydroxylase activity and/or 5,6,7,8-tetrahydrobiopterin availability may be expected in mice heterozygous for the A Y allele. However, this hypothesis requires additional experimental verification.
The 1473G allele is eliminated from wild mouse populations by the forces of natural selection [14]. The present study provides one hypothetical mechanism for this elimination. Although the 1473G allele in the SPF conditions does not alter physiological functions and behavior, in extremal conditions or in combinations with other adverse mutations (for example A Y ), this allele causes hind limb dystonia that can decrease mouse adaptation in wild populations.

Conclusions
The present work is a logical development of studies concerning the effects of the A Y [24] and C1473G [11] mutations in the Raly-Agouti locus and the Tph2 gene, respectively, on the mouse brain and behavior. Here, the effects of the combination of these mutations on body mass, behavior and the brain 5-HT and melanocortin systems in B6-1473GG/A Y a mice were compared to the effects of each of these mutations separately in B6-1473CC/aa, B6-1473CC/A Y a and B6-1473GG/aa mice. The following results were obtained:

1.
As expected, the A Y allele increased body mass and depressive-like immobility in the forced swim test.

2.
The 1473G allele decreased body mass and depressive-like immobility and increased locomotor activity in the open field test.

3.
No effect of the A Y allele on the 5-HT system and no effect of the 1473G allele on the melanocortin system in the brain were observed.

4.
A marked effect of the combination of the A Y and 1473G alleles on hind limb dystonia was revealed. Although these alleles do not cause hind limb dystonia separately, their combination results in a high level of expression of this trait in adult B6-1473GG/A Y a mice.
This result is the first evidence for the interaction between the C1473G and A Y mutations.