This experiment was performed between March and June 2018 at the Department of Animal Science of the Federal University of Lavras (UFLA), Lavras, state of Minas Gerais, Brazil. The Animal Research Ethics Committee approved all experimental procedures under protocol no. 057/2017.

The 300 male one-day-old Japanese quails (Coturnix coturnix japonica) in the experiment were housed in a masonry shed in 30 cages (50 cm wide×70 cm deep×25.5 cm high) during the growth phase (1 to 35 days). For the production phase (36 to 96 days), the three males with the live weight closest to the cage average were selected, transferred to a screened laying shed, and housed together with nine females in laying cages (32 cm wide×38 cm deep×16 cm high). The temperature and humidity in the sheds were monitored using thermohygrometers placed at bird height, which recorded the minimum and maximum temperature and humidity. The temperature was kept at 38°C during the first three days using a wood-burning stove and was decreased by 0.5°C each day until the birds were 28 days old [8]. The photoperiod was 24 hours light:0 hours dark (natural + artificial light with 60 W lamps) for the first two days of life, 23 hours light:1 hour dark until day 15, then 14 hours light:10 hours dark until the end of the growth phase. At 36 days of age, the light period was increased by 30 min each day up to 17 hours light:7 hours dark, which was retained until the breeding period ended [9].

For both phases (growth and production), a completely randomized design was used, with five treatments and six replicates (cages). The treatments consisted of five dietary CP levels (18%, 20%, 22%, 24%, and 26%) during the growth phase only; feeds were isoenergetic and isonutritious for the remaining nutrients. Diets were formulated based on the recommendations of Rostagno et al [5] by increasing or decreasing the recommended CP level (22% CP) by 2% or 4%. During the production phase, all birds received standard feed (18%) formulated per the recommendations for that phase [5]. All diets were corn and soybean meal-based and formulated using the CP values obtained from their chemical analysis. Feed samples were also collected to determine the CP concentrations (Table 1). Feed and water were supplied ad libitum throughout the entire 96-day experimental period.

On the first day of the experiment, two previously identified males per cage were individually weighed every three days until 60 days of age to determine the growth curve. The two birds’ live weights were used to calculate the Gompertz curve, which was then used to determine the body weight at the growth maturity (A), growth rate (B), and age of maximum growth (M) [7]. The curve was calculated using Statistica 13.3, using the following model:

in which Wd is the weight at day d (in g); A is the body weight at growth maturity (g); B is the growth rate (d⁻¹); and M is the age (d) of maximal growth (inflection point).

The birds were weighed at 1, 14, and 35 days old to determine weight gain. The supplied feed and leftovers were also weighed to determine feed intake. Feed conversion was calculated as the feed intake:weight gain ratio per period.

At 33, 34, and 35 days of age, the total excreta were collected once daily to evaluate the nitrogen balance at the end of the growth phase. At the beginning and end of the collection period, the feed was weighed to determine intake. Collection trays were inserted under each cage and lined with resistant plastic to avoid contamination and loss of excreta. All collections began at 08:00. Feathers and feed particles were removed, and excreta were placed in plastic bags and stored at −20°C until the end of the collection period. The excreta were then thawed, homogenized, and weighed. Aliquots of 300 g were removed and predried in a forced air circulation oven (55°C) for 72 hours. The samples were again weighed to determine the predry matter and then ground using a knife mill equipped with a 2-mm sieve. Excreta and feed samples were analyzed to determine dry matter and nitrogen concentrations [10]. The nitrogen intake and excreta were quantified based on the results, and the percent nitrogen retained was calculated.

At 35 days of age, before transfer to the production phase housing, two males from each experimental unit were selected based on their average weight, euthanized by cervical dislocation and exsanguination, and dissected for testicular harvesting. At 60 days of age, the same procedure was performed on one male from each experimental unit. The right and left testes were removed, weighed separately, and measured (width×thickness×height) using a digital pachymeter (Digimess, São Paulo, Brazil). Testicular weight was calculated as the sum of the weight of the right and left testes. Testicular volume was calculated using the equation V = 4÷3πab², where a is half the testis height and b is half the testis width [11]. The gonadosomatic index, which is the percentage of body weight contributed by the gonads, was estimated using the equation GI = ([WTl+WTr]/LW)×100, where WTl is the left testis weight, WTr is the right testis weight, and LW is the live weight. After measuring these parameters, the testes were fixed in Bouin’s solution for approximately 12 hours at ambient temperature and then washed in 70% alcohol for subsequent histological analysis [12].

The testes were histologically analyzed at the Laboratory of Histology and Immunohistochemistry of the Department of Animal Science of UFLA. First, the testes were dehydrated in a graded ethanol-xylol series and embedded in paraffin. Sections (5 μm thick) were obtained, suspended in a water bath at approximately 37°C, mounted on silanized histological slides, and dried in an oven at 37°C overnight. The samples were then deparaffinized, rehydrated in a xylol-ethanol series, and stained with hematoxylin-eosin [13].

Seminiferous tubule images were analyzed at 400× magnification using an Olympus CX31 microscope (Olympus, Tokyo, Japan) coupled to an Altra SC30 digital camera (Olympus, Japan) and Axio Vision software (Carl Zeiss, Oberkochen, Germany). In each section, the largest and smallest diameters of ten random round-shaped tubules were measured to calculate the seminiferous tubular area using the equation A = πr², where r is the tubule radius calculated from the average of the largest and smallest diameters. The seminiferous epithelial height was obtained by performing seven measurements per tubule and analyzed to calculate the area.

Immunohistochemical analyses were performed in the Laboratory of Histology and Immunohistochemistry of the Department of Animal Science of UFLA. Slides were prepared as described for the histological analyses per Suvarna et al [13] with some adaptations. All slides were incubated in a humidified chamber, and all washes consisted of three consecutive immersions for five minutes in 0.1 M phosphate-buffered saline (PBS) at pH 7.2. Endogenous peroxidase activity was blocked by incubation in Peroxidase Block (K4011, DakoCytomation, USA) for 30 minutes and washed in PBS. Nonspecific antibody binding was blocked by incubating the sections with Block Serum (X0909, DakoCytomation, USA) for 15 minutes. The histological sections were then washed again in PBS, incubated for 2 hours with mouse anti-PCNA monoclonal primary antibody (M0879, DakoCytomation, USA) for staining the Spermatogonia, or incubated for 2 hours with rabbit anti-Androgen receptor polyclonal antibody (AB168834, Abcam, Waltham, MA, USA) for detection of Leydig cells [14]. After incubation the histological sections were washed in PBS, incubated for 15 minutes in Biotinylated Link Universal (K0690 DakoCytomation, USA), washed again in PBS for 5 minutes, and incubated in Streptavidin-HRP (K0690 DakoCytomation, USA) for 15 minutes per the recommendations from the LSAB+System/HRP anti-rabbit and anti-mouse secondary antibody kit (K0690, DakoCytomation, USA). After washing, the histological sections were developed enzymatically using 3,3-diaminobenzidine tetrahydrochloride (DakoCytomation, USA) and immersed in distilled water to stop the reaction after 30 seconds (samples from birds aged 35 days) or after one minute (samples from birds aged 60 days). The slides were then counterstained with hematoxylin and mounted under a coverslip.

Sertoli cells were identified in counterstained histological section slides of the PCNA test through their morphology. They are tall simple columnar cells, which span from the basement membrane to the lumen surrounded by the proliferating and differentiating germ cells. After counting the number of Sertoli cells per tubule the percentage of proliferating spermatogonia were determined in ten tubule sections per testis and observed at 400× magnification using a light microscope (Olympus CX31; Olympus, Japan). The Sertoli cell number per area (10,000 μm²) was calculated using the tubular area. The number and percentage of proliferating spermatogonia were determined by counting the stained cells relative to the total number of cells located closer to the basal membrane (Figure 1).

The number of Leydig cells and the tubular and intertubular areas were evaluated at 1,000× (35 days of age) 400× (60 days of age) using an Olympus CX31 microscope (Olympus, Japan) coupled to an Altra SC30 digital camera (Olympus, Japan) in Axio Vision software (Carl Zeiss, Germany). Ten fields randomly distributed in the testicular parenchyma were evaluated on a projected grid in ImageJ version 1.50i (NIH, Bethesda, MD, USA), and 102 intersection points were considered. The Leydig cell number per area (10,000 μm2) was calculated using the proportion of Leydig cells per intertubular area and the intertubular area.

On days 59 and 60 of the experiment, all eggs were collected from each cage. Eggs that were cracked, broken, dirty, too large, or too small were discarded, and the remainder (n = 100/treatment) were labeled and stored for 24 hours at 20°C. The eggs were then weighed, and ten eggs were selected per cage based on average weight. The selected eggs were sterilized with a 2:1 formaldehyde (37%):potassium permanganate (99%) solution [15] and incubated at 37.5°C and 60% humidity [16] in an automatic incubator (Luna 480; ChocMaster, Piraquara, Brazil). The automatic egg-turning tray turned the eggs every two hours until day 15 of incubation [17]. After 21 days of incubation, the numbers of hatched and fertilized but unhatched eggs were counted, and percent fertility was calculated.

Semen quality was evaluated from one male per cage. From day 50 of the experiment, semen was collected every two days via dorsal-abdominal massage.

Semen was evaluated with or without the presence of foam [18] from the two remaining quails in each experimental unit. Foam was collected from each male’s cloacal gland at days 80, 81, and 82 of the experiment at 08:00 and 16:00 each day by gently squeezing the cloacal gland on each side. After collection, the foam was stored in Falcon tubes at −20°C. At the end of the collection period, the samples obtained on the previous days were thawed at ambient temperature and diluted to 1:4 (foam:saline) in saline solution (0.9%). The samples were then centrifuged at 3,000 g for 45 min (Sorvall ST 16 Centrifuge; Thermo Fisher Scientific, Waltham, MA, USA). The supernatant was collected and stored at −80°C until semen quality analysis.

Beginning at 90 days of age, the bird semen was collected three times at 3-day intervals per Burrows and Quinn [19]. Before collection, the feathers in the pericloacal region were removed, and the cloacal gland was measured (width×thickness×height) using a digital pachymeter (Digimess, São Paulo, Brazil) to calculate the cloacal gland area (GA) using the equation GA = W×H, where W is the lateral width and H is the dorsoventral height [20].

Foam was removed from the cloacal gland during semen collection. The birds were massaged dorsally, beginning near the wing base and ending near the cloaca, with the standard six movements per animal. Slight pressure was then applied with the fingers on the base of the phallus and on the ampoules of the vas deferens to discharge the semen. The seminal content was collected in graduated capillary tubes, where its volume was measured, and then immediately diluted to 1:1 (semen:saline) in saline solution (0.9%). A 1 mL aliquot of solution was diluted with 499 ml of formalin for subsequent sperm concentration analysis, which was performed using a Neubauer chamber. Another 1 mL aliquot was diluted with 98 mL of 0.9% saline solution or 0.9% saline solution + 5% foam [21].

Immediately after dilution in saline solution or saline solution + foam, three trained evaluators blinded to the sample identity evaluated the sperm motility and movement intensity in three subsamples mounted between glass slides and coverslips at 37°C, observing them at 200× magnification using a light microscope (Olympus CX31; Olympus, Japan). Sperm motility was expressed as the percentage of motile spermatozoa, and movement intensity was classified from 0 to 5, with 0 being the lowest and 5 the highest.

Sperm viability was evaluated by mixing one drop of semen with one drop of eosin-nigrosin on glass slides and observing the mixture at 400× magnification using a light microscope (Olympus CX31; Olympus, Japan). The numbers of living (no color) and dead (pink) cells were counted, and sperm viability was calculated as living cells/total cells ×100.

The Gompertz curve parameters (A, B, and M) obtained for each experimental unit and the male reproductive performance and quality data were checked for normality (Anderson-Darling), homoscedasticity (Breusch-Pagan), and independence of error (Durbin-Watson). When the assumptions of normality were met, an analysis of variance (one-way analysis of variance [ANOVA]) was performed, and a regression analysis was performed for the protein levels. Only for live weight a two-way ANOVA (CP levels × age of birds) was performed. When a linear regression could not be fit (R²<0.70), a broken-line analysis was performed to determine the best protein level [22]. When a curve could not be fit by broken-line analysis, the averages were compared using the Student-Newman-Keuls (SNK) test at p≤ 0.05. When the ANOVA assumptions were unmet and the Box-Cox and Johnson data transformations could not be used to normalize the data, the data were subjected to a nonparametric analysis, and the averages were compared using the Kruskal-Wallis test. The fetility test was analyzed by a binomial score (1 or 0) based on the presence or absence of fertilization. All statistical analyses were performed using Statistica 13.3 and Action 3.5 software.

The different dietary CP levels in the growth diets resulted in different growth patterns in male quails (Table 2; Figure 2). The growth rate increased linearly (p<0.01), and the age of maximum growth (curve inflection point) decreased (p< 0.05) with increasing dietary CP up to 22% CP (Figure 3). No effect was observed on body weight at maturity (p>0.05).

Live weight from the second to the seventh week of the birds’ lives was affected by the inclusion of CP (Table 3); 23% CP or higher resulted in greater live weights up to 36 days of age (Figure 4). To obtain greater weights at 42 and 48 days of age, 23.5% CP is recommended.

Diets with 24% and 26% CP resulted in greater weight gain and improved feed conversion up to 14 days of age (p<0.01) (Table 4). Feed intake was lower in the 18% CP group (p< 0.01). Considering the whole growth phase (1 to 35 days), diets with 22% CP or more resulted in less weight gain (p< 0.01). No effect on feed conversion was observed in this phase (p>0.05). At the end of the growth phase (33 to 35 days of age), both nitrogen intake and nitrogen excretion increased linearly with increasing dietary CP (p<0.01). No differences in nitrogen retention were observed (p>0.05).

At 35 days of age, anatomical and histological evaluations of the quail testes were affected by the dietary CP level (p<0.01) (Table 5). The sizes (height, width, and thickness), as well as weight and volume, of the right and left testes and seminiferous tubular areas increased linearly (p<0.01) with increasing dietary CP levels (Figure 5). Germinal epithelial height increased (p<0.01) only up to 22% CP (Figure 6). A higher intertubule: tubule ratio was observed (p<0.01) in the 24% and 26% CP groups. There was a linear effect (p<0.01) of dietary CP on the number of spermatogonia and Leydig cells in the testis.

At 60 days of age, a linear increase (p<0.01) in Sertoli cell numbers was observed with increasing CP levels in the diet. Higher Leydig cell numbers were observed (p<0.01) with 18% CP and more spermatogonia/area were observed with 24% and 26% CP (p<0.05).

At 90 days of age, the CP levels in the growth diet did not affect cloacal gland size, foam weight and protein concentration, semen volume and quality, or flock fertility (p>0.05) (Table 6).