Nutritional and performance viability of cactus Opuntia-based diets with different concentrate levels for Girolando lactating dairy cows

Objective The aim of this research was to evaluate the effect of different concentrate levels in diets based on cactus Opuntia Stricta (Haw.) Haw cladodes on the performance of lactating Girolando cows. Methods The experiment involved 10 Girolando multiparous dairy cows at 512.6 kg of body weight (BW) and producing 13.2 kg milk/d, allocated into two 5×5 Latin squares. The experimental treatments consisted of control diet composed by cactus Nopalea cochenillifera. Salm-Dyck. cladodes (Nopalea), forage sorghum silage and concentrate at 20% on dry matter (DM) basis, and four concentrate levels diets (20%, 24%, 28%, and 32%) plus cactus Opuntia stricta (Haw.) Haw. cladodes (Opuntia) and forage sorghum silage. Results Regarding cows fed control diet, the nutrients intake were greater than for cows fed with cactus Opuntia and concentrate. Regarding concentrate levels, intakes of DM, organic matter (OM), crude protein (CP), non-fiber carbohydrates (NFC), and total digestible nutrients of cows increased linearly. Organic matter, CP, and NDF digestibilities were similar in between to control diet and cactus Opuntia-based diets. The digestibility of NFC increased linearly when the concentrate was inserted. The N balance was the same for control diet and cactus Opuntia-based diets, irrespective the concentrate levels. Conclusion For cows producing 14 kg/d with 3.5% of fat, it is recommended 32% of concentrate to be included in cactus Opuntia-based diets, and the increase in concentrate level promotes a linear increase in milk yield.


INTRODUCTION
Dairy cattle production is one of the few economically possible activities in the semiarid regions of Brazilian northeast, in which feeding the heard is characterized by the usage of cultivated forage and native vegetation [1]. The prevailing production system is characterized by smallholders system with an average of the productive area of 37 ha [2]. Nevertheless, the reduced size of the property has limited the production of forage, thus causing difficulties in the increase of milk productivity per area.
Forage cactus due to its high production of green matter per unit of area, therefore be coming fundamental for the increase of the efficiency of the land's productivity. The cactus can be used when added to a source of fiber [3] and as a source of energy [4] when its nutri tional deficiencies are corrected [1].
The superiority of cactus Nopalea cochenillifera. SalmDyck. cladodes (Nopalea) in re lation to other genotypes mainly when compared to the cactus Opuntia stricta (Haw.) Haw cladodes (Opuntia) was already observed [5]. However, from the agronomic point of view, the cactus Opuntia has present ed less demand of nutrients, it has been more tolerant to the conditions of hydric stress and it has also presented a higher production of dry matter per unit area (37 t of DM/ha/2 yr) than cactus Nopalea (21 t of DM/ha/2 yr) [6], sparking more interest in its insertion in the feeding of dairy cows in the semiarid.
A problem observed by Rocha Filho [7] for the inclusion of cactus Opuntia in the diet of lactating cows was the decreased consumption of nutrients and of milk production, associat ed to low acceptability when compared to diets with cactus Nopalea. As an alternative to solve this problem it would be increasing the proportion of concentrate in diets with cactus Opuntia to compensate the lessened consumption. Thus, it was hypothesized the existence of concentrate level associated to the cactus Opuntia which maximizes performance in lac tating dairy cows.
The aim of this research was to evaluate the effect of con centrate levels in diets based on cactus Opuntia Stricta (Haw.) Haw cladodes over the performance of lactating Girolando cows.

MATERIALS AND METHODS
The study was approved by the Ethics committee of Federal Rural University of Pernambuco (License n° 069/2016) and was conducted at Experimental Station of the Instituto Agronômico de Pernambuco (IPA), located at Arcorverde, Pernambuco, BR, presenting a semiarid climate "Bsh".
The experiment involved 10 Girolando multiparous dairy cows at 512.6±53.66 kg of body weight (BW), 14 weeks in milking and producing 13.2±1.94 kg milk/d, allocated into two 5×5 simultaneous Latin squares. The trial lasted for 126 days, with four consecutive 21day periods divided into 14 day adaptation and sevenday sampling periods.
The individual BW was measured at the beginning and at the end of each experimental period after milking. The cows were housed in individual pens, with approximately 24 m 2 , with individual feed bunks and unrestricted access to water. The experimental treatments consisted of control diet com posed by cactus Nopalea cochenillifera. SalmDyck. cladodes (Nopalea), forage sorghum silage and concentrate at 20% on dry matter (DM) basis [7], and four concentrate levels diets (20%, 24%, 28%, and 32%) plus cactus Opuntia stricta (Haw.) Haw. cladodes (Opuntia) and forage sorghum silage. The diets were formulated to be isonitrogenous (13.3% crude protein; CP) and meet dairy cows' nutrients requirements producing 14 kg of milk with 3.5% of fat [8], considering the ingredients composition (Table 1).
Feed was supplied ad libitum as total mixed ration, twice a day at 8:00 and 16:00, allowing 5% to 10% in orts (DM basis). Tables 1 and 2 present the diets feeds chemical composition, diets composition percentages and diets chemical composi tion. The voluntary intake was evaluated from the 15th to 21th day. In this sense, the amounts of supplied diet and orts were taken into account. The diet ingredients and orts sam ples were pooled per animal and experimental period and stored in plastic bags at -20°C. At the end of the experiment, the samples were ovendried at 60°C for 72 hours and ground to pass through a 2 mm mesh for in situ ruminal incubation and through a 1 mm screen for further chemical analyses.
For estimated apparent digestibility and total digestible nutrient (TDN) concentration, the spot fecal samples were collected directly from the animals' rectums in between the 16th and 20th days of each experimental period [9], and the samples were pooled per animal, and experimental period and stored at -20°C for chemical analyzes. The total fecal excretion was estimated using the indigestible neutral de tergent fiber (iNDF) as an internal marker, and the feces, feed and orts iNDF content were obtained after 288 hours of ruminal incubation time [10]. The diets TDN content and its conversion in lactation net energy were estimated accord ing [11,8].
After sample processing to pass through a 1 mm screen sieve, we evaluated for DM (method INCTCA G003/1), or ganic matter (OM, method INCTCA M001/1), CP (method INCTCA N001/1), ether extract (EE, method INCTCA G005/1), neutral detergent fiber (NDF) corrected for ash and protein (NDFap, methods INCTCA F002/1, INCTCA M 002/1, and INCTCA N004/1), and neutral detergent insolu ble protein (NDIP), method INCTCA N004/1), according to the standard techniques of the Brazilian National Institute of Science and Technology in Animal Science [12]. The quan tification of nonfiber carbohydrates (NFC) was performed according to Detmann and Valadares Filho [13] as follows: Blood samples were taken on the 21th day of each sam pling period, four hours after the morning feeding, from the coccygeal vein of each animal. Samples were immediately centrifuged at 2,100× g/min for a period of 15 min, and the remaining plasma or blood serum was maintained at -20°C for further analysis of urea concentration.
Simultaneously to blood sampling, spot urine samples were collected of each cow [14]. The urine was filtered through gauze and an aliquot of 10 mL was diluted immediately in 40 mL of H 2 SO 3 (0.036 N). The samples were stored at -20°C for further nitrogen, urea, allantoin (AL), uric acid (UA), and creatinine analysis.
To analyze AL in milk and urine, we used the colorimetric method as described by Chen and Gomes [15]. The urea con centration in urine was analyzed via the enzymaticcolorimetric system of the urease method, using commercial kits (Labtest Diagnóstica S.A. Lagoa Santa, MG, Brazil). The UA concentra tion in urine was analyzed via the enzymatic Trinder method, using commercial kits (Labtest Diagnóstica S.A., Brazil). The creatinine concentration in urine was analyzed using "end point" markers with picrate and acidification, using commer cial kits (Labtest Diagnóstica S.A., Brazil).
Daily total urinary volume was estimated through the rela tion of daily urinary excretion of creatinine, using the observed values of creatinine concentration in urine as described by Valadares Filho and Valadares [16]. The daily urinary excretion of creatinine was based on 24.05 mg/kg of BW of creatinine [14]. Microbial protein synthesis was estimated according Chen and Gomes [15], considering a recovery of absorbed purines of 0.85 and an endogenous contribution to the ex cretion of purines as recommended by GonzálezRonquillo et al [17].
The nitrogen balance (NB) was obtained through the dif ferences between total nitrogen intake (N intake) and feces (N fecal), milk (N milk), and urine (N urine) total nitrogen. The N milk was quantified using milk total protein (MTP/6.38), and the milk urea nitrogen (MUN) was estimated using the equation N urine (g/d) = 12.54×N milk (mg/dL).
The cows were milked twice a day (6:00 and 15:00), and the milk yield (MY) was registered from the 15th to the 21th day of each experimental period. Milk samples were collected on days 18th and 19th of both milking periods, after last col lection, and composed samples were made for each cow. A milk aliquot of 50 mL was conditioned in plastic bottles with preservative (Bronopol, D & F Control Systems, Inc., New York, NY, USA), maintained between 2°C and 6°C, and sent to the PROGENE Laboratory for evaluation of lactose, fat, protein, total solids, casein, and urea, following the methods of ISO [18]. Another 10 mL aliquot of milk was deproteinated with 5 mL of trichloroacetic acid (25%), filtered, and stored at -15°C for further AL analysis. The 3.5% fatcorrected milk yield (FCMY) was estimated as the equation FCMY (3.5%) = ([0.432+0.1625×% milk fat]×MY kg/d) [19].
The data were submitted for analysis of variance and regres sion using the MIXED procedure of the statistical program SAS (version 9.4, SAS Institute Inc., Cary, NC, USA), adopt ing 5% as significance level for the type I error, according to the following model: Y ijkl = μ+T i +S j +(P/S) jk +(A/S) jl +ε ijkl Where: Y ijkl = observation ijkl; μ = over mean; T i = fixed effect of treatment i; S j = random effect of square j; (P/S) jk = random effect of period k within the square j; (A/S) il = random effect of animal l within the square j; ε ijkl = random residual error associated with each observation, assuming the NID (0; σ2).
Dunnett test was used to compare each treatment group mean (concentrate levels), with the average of control diet. Comparisons between concentrate levels in the diets were conducted by the decomposition of sum of squares in ortho gonal linear contrasts, and quadratic effects at 5% probability, with subsequent adjustments of the regression equations.

RESULTS AND DISCUSSION
The control diet provided 14 kg of milk with 3.5% of fat as expected. The cows fed with the control diet presented lower intake for most nutrients (DM, OM, CP, NFC, and TDN), except for NDF, than those which received diets with cactus Opuntia, regardless of the concentrate levels. The concen trate inclusion of cactus Opuntiabased diets allowed for a linear increase in the intakes of DM, OM, CP, NFC, and TDN. Nevertheless, the NDF intake remained unaltered ( Table 3).
The digestibility of DM was greater for the control diet than for the diet with cactus Opuntia and 20% of concentrate. The increase of concentrate in cactus Opuntiabased diets did not alter the digestibility of nutrients (Table 3).
It was observed a similarity of data collected in the present study with those obtained by Rocha Filho [7] which served as a base for the control diet; the intake of DM (14.96 and 12.11 kg/d) and the MY (14.0 and 13.0 kg/d) for cactus Nopalea and Opuntia, respectively. Regarding the proportion of leftovers observed (10.46%) the levels of the different ingredients remained unaltered. Based on this assumption and considering that the cows did not se lected the feeds, it was clear the inhibitory effect of DM intake for the genotype of cactus Opuntia. [5,7] also observed this effect. Nevertheless, in data presented by Monteiro et al [20], this effect was not observed. The difference lies in the pro portion of concentrate which was significantly higher (30%), thus there was a lesser proportion of cactus (39%).
Even though it was not quantified other causes that could be aroused would be the interference of the malic acid. The increase in the diet malic acid concentration diminishes the palatability intake of DM, which can have potentially nega tive effects on the animal performance [21]. Also, the cactus Opuntia presents more organic acids content than other geno types implying in acidic smell after chopped [22], a fact that may negatively interfere in the feed intake.
The fixation of CO 2 in the plants Crassulacean acid me tabolism (CAM) only occurs at night, in the dark, when the stomata are open. In this moment, the fixation of CO 2 in phos phoenolpyruvate to form oxaloacetate. This last substance is rapidly transformed in malate and stored all night long in the vacuoles in the form of malic acid. When the day breaks the stomata close themselves and the malic acid is removed from the vacuole, transported to the chloroplast of the cell and de carboxylated, thus, producing pyruvate and CO 2 . The fixed Therefore, it is possible to infer that in the CAM plants the formation of malic acid occurs at night and its consump tion during the day. This causes a change in the taste of the plant during the day because at night an acid taste is observed. During the day the plant becomes sweeter. It is important to highlight that the first feed was offered at 8 h. This fact seems to be more evident in the cactus Opuntia. Silva et al [22] measured both the pH of the cactus Nopalea and Opuntia, harvested in the morning and a significantly lesser pH was observed in the cactus Opuntia (4.59) in relation to cactus Nopalea (5.01).
Despite the DM intake of Opuntia plus concentrate diets had been lower than DM intake of control diet, MY was guar anteed, and the BW gain was 22, 36, 40, and 70 g/d for diets with 20%, 24%, 28%, and 32% of concentrate inclusion, re spectively. Also, the balance of nutrients presented explains the similarity in the MY observed between the control diet and the diet with cactus Opuntia and 32% of concentrate. However, the highest DM intake for control diet can be beneficial after lactating peak for recover the body reserves and preparing the body condition for the next calving, considering that cows fed control diet gained 200 g/d during the experiment.
Regarding NDF there was a compensation, that is to say with the increase in the proportion of the concentrate there was a decrease in the levels of fiber in the diets (Table 2) which probably was one of the factors that stimulated the rise in the consumption of DM which is normally justified by a higher concentration of fastdigesting ruminal carbohydrates (i.e. NFC) [23], as observed by Inácio et al [24] who tested differ ent concentrate levels for heifers fed sugarcane bagasse as an exclusive roughage source. On the other hand, Chung et al [25] relate the increase in nutrients intake to the higher physical density of the concentrate by the decrease in the size of the particles in relation to the roughage. The main effects are the augmentation of the passage rate of the digesta through the gastrointestinal tract, making the increase in the consumption possible.
It was not found any difference between the control diet and the different levels of concentrate in the diets based on cactus Opuntia for the different sanguine parameters. The concen trations of blood urea nitrogen (BUN), urea and glucose were not altered with the inclusion of concentrate in the diets based on cactus Opuntia. On the other hand, the concentrations of nonesterified fatty acids (NEFA) and βhydroxybutyrate (BHBA) decreased linearly ( Table 4).
The NEFA and BHBA are important metabolic parameters to measure the nutritional status and of the adaptation to negative energetic balance of dairy cows during the body tissue mobilization [26]. The levels of NEFA observed (Table  4) were lower than the levels which are considered as normal (0.60 mmol/L) by Enjalbert et al [27], however, above this value it indicates an augmentation of the risk of occurrence of metabolic diseases as abomasum displacement, clinic keto sis, metritis and placenta retention.
The BHBA is important in dairy cattle as an indicator of subclinic ketosis, caused by the mobilization of body fat to meet the energetic deficit [28]. The concentration of BHBA observed in the present study was below the level considered as an indicator of subclinic ketosis less than 10 mg/dL or 0.97 mmol/L by Enjalbert et al [27], which indicates other meta bolic conditions indicating there was not fat mobilization. Regarding the BUN of the control diet (22.9 mg/dL) and of the diets with cactus Opuntia (22.3, 23.2, 21.4, and 23.7 mg/dL) ( Table 4), they are above the 19 mg/dL limit, which indicates loss of dietetic nitrogen in the cows [29] which shows the in efficiency of the usage of dietetic protein by these animals.
It was not observed difference between the control diet and the different levels of concentrate in the diets with cactus Opuntia for urinary volume, urea urine, BUN, MUN, micro bial nitrogen, microbial CP and efficiency of microbial protein synthesis which also remained unaltered due to the inclusion of concentrate in the diets with cactus Opuntia ( Table 5).
The absence of variation in estimated values of the syn thesis and synthesis efficiency of the microbial CP (Table 5) with the offer of diets containing cactus cladodes can be ex plained by the amount of diet carbohydrates (Table 2), which were sufficient for supplying the necessary energy for the fer mentation of fiber and the microbial synthesis. The N urea in milk (12.53, 13.39, 12.08, 11.54, and 12.16 mg/dL), are within the variation of 12 to 17 mg/dL, which, according to Abrahamse et al [30], values within this variation would in dicate adequate balance of degraded protein and fermented energy in the rumen. The NB observed in cows fed with the control diet was simi lar to those which received the diet based on cactus Opuntia, regardless of concentrate level. The N excretion in the milk was greater for control diet in relation to those with cactus Opuntia with 20% and 24% of concentrate. The N intake and the N fecal in milk increased linearly with the inclusion of con centrate in the diets with cactus Opuntia, thus not affecting the N fecal and N urine ( Table 6).
In spite of the change in the N intake, regarding the NB, the similarity of the results observed not only in animals receiving the control diet but also those supplemented with concentrate and even within the levels, they can be explained by a higher excretion via milk or by the amount of milk produced as well as by the higher proportion of protein ( Table 6).
The FCMY of cows fed with 32% of concentrate in the diet based on cactus Opuntia was similar to those which re ceived the control diet ( Table 7). The FCMY increased linearly with the insertion of concentrate in the diets based on cactus Opuntia ( Table 7). The levels of fat, protein, total solids (TS) were higher for the control diet when compared to diets based on cactus Opuntia. When concentrate was included in the diets with cactus Opuntia it was not affected the content of fat, protein, lactases, TS, casein, urea nitrogen and feed effi ciency ( Table 7). The efficiency of concentrate usage in the control diet was less than in the diet with 20% of concentrate and higher than the levels of 24%, 28%, and 32% of concen trate ( Table 7). As the levels of concentrate in the diets were increased, the efficiency of usage of the concentrate dimin ished linearly ( Table 7).
The concentrate inclusion caused an increase in the nutri  ent's intake, thus in the MY, without altering the level of fat.
Only the diet with the highest level of concentrate (32%) pro vided a production of milk identical to that verified for control diet (Table 7), which can be explained by the similarity in the nutrient's intake, notably TDN ( Table 3). The higher amount of protein (3.5 g/100 g) of milk for con trol when compared to the others experimental diets (3.25, 3.26, 3.27, 3.32 g/100 g), is related to a higher quantity of NFC of the control diet, supplying easily available energy for the microorganisms of the rumen to synthesize microbial protein.
According to Abrahamse et al [30], diets with a higher pro portion of NFC, even with a similar intake of digestible energy provide a higher supply of fermentable carbohydrates and this can result in a higher level of protein in the milk.
As palatability is defined by the physical and chemical char acteristics which "tease appetite", besides, ruminant select their feed based on flavor and color, these observations prove the necessity of more studies with cactus Opuntia to unfold the reasons why the animals do not ingest the same amount of DM than when they are offered the cactus Nopalea.
The efficiency of concentrate usage becomes better when diets with 20% of concentrate, 50% of cactus Opuntia and 30% of forage sorghum silage (4.91 kg of milk/kg of concentrate) are used when compared to the other levels of concentrate in diets (4.0, 3.5, and 3.07 kg of milk/kg of concentrate). This fact can be explained by the small difference in the MY (Table  7) and the increase in the concentrate intake with the higher levels of concentrate in the diets. Also, for each US$ spent to feeding the cows it will return 0.49, 0.57, 0.53, 0.52, 0.50 US$/d of milk, considering the control diet and, 20%, 24%, 28%, and 32% of concentrate inclusion in Opuntia diets, respectively.
Despite the lowest return calculated for 32% concentrate inclusion, it should be considered the daily gain of 70 g pro moted by this diet, which implies in better reproductive herd efficiency and important energy source to replace the body reserve at the beginning of the lactation. Increase in the total volume of produced milk per day (1.5 kg/d more than the volume promoted by the 20% concentrate inclusion). In the other hand, a lower concentrate level could be used in the pro duction system to feed dairy cows in late lactation stage. Thus, Opuntia is a viable option in the dairy system and the main advantages are the agronomic characteristics, once it demands fewer nutrients and is more tolerant of hydric stress. Also, Opuntia presents a higher production of dry matter per unit area (37 t of DM/ha/2 yr) than Nopalea (21 t of DM/ha/2 yr) [6], which may offset the concentrate feeding costs.
In conclusion, to Girolando cows producing 14 kg/d with 3.5% of fat, it is recommended the inclusion of 32% of con centrate in cactus Opuntiabased diets to achieve similar performance to those fed cactus Nopaleabased diet with 20% of concentrate. In addition, the concentrate inclusion on cactus Opuntiabased diets promotes a linear increase in MY.

IMPLICATIONS
The study provides applicable animal production and nutrition information relating to the new genotype of cactus cladodes for lactating cow's diets raising in semiarid regions. Although cactus cladodes have been used around the world in arid lands, there is a necessity to study other genotypes viability to inclu sion on the herd diet promoting the production sustainability and agricultural diversity. Despite the use of the new genotype requires more concentrate inclusion on diets to guarantee ad equate milk production, the agronomic characteristics, as high green matter production could compensate for this challenge.

CONFLICT OF INTEREST
We certify that there is no conflict of interest with any financial organization regarding the material discussed in the manu script.