Partial Replacement of Soybean Meal with Canola Meal or Corn DDGS in Low-Protein Diets Supplemented with Crystalline Amino Acids—Effect on Growth Performance, Whole-Body Composition, and Litter Characteristics

Simple Summary Dietary protein reduction with amino acid supplementation is a nutritional approach to reducing environmental pollution. Soybean meal is nutritionally superior to alternative protein feedstuffs, such as corn distillers dried grain with solubles (cDDGS) and canola meal (CM), because of its relatively good balance of amino acids. This study investigated the usage of CM or cDDGS compared to SBM in low-protein broiler diets and their effects on growth performance, carcass yield, whole-body composition, and litter characteristics. The results demonstrated that replacing soybean meal with cDDGS or CM negatively affected the performance of the birds, whereas reducing dietary protein reduced the litter surface ammonia, irrespective of protein feedstuff. Therefore, complete replacement of SBM with cDDGS or CM in low-protein diets is not feasible for optimum performance in broiler chickens. Abstract A 42-day study was conducted to explore the application of supplemental amino acids (AA) in low-protein diets with soybean meal (SBM), canola meal (CM) or corn distillers dried grain with solubles (cDDGS) as the main protein feedstuffs. The responses of interest were growth performance, carcass yield, whole-body composition, litter ammonia and litter N. On d 0, a total of 540 Cobb 500 (off-sex) male broilers were allocated to 36 floor pens. All the birds received one starter diet that met nutrient requirements during the first 10d. Thereafter, six experimental diets were provided in grower and finisher phases. The diets included a positive control (PC): a corn–SBM diet with adequate protein. The protein level of the negative control (NC) was decreased by 45 g/kg relative to the PC. The next two diets had the same protein levels as the NC but with cDDGS added at 50 or 125 g/kg. The last two diets had the same CP as the NC but with CM added at 50 or 100 g/kg. All the low-protein diets had the same level of standardized ileal digestible indispensable AA according to Cobb 500 recommended level. Gly and Ser were added as sources of non-specific N. The dietary protein reduction in corn–SBM diets at both phases decreased (p < 0.05) weight gain and increased (p < 0.05) feed conversion ratio (FCR). Increasing levels of cDDGS or CM, at a constant CP level, linearly decreased (p < 0.05) the weight gain and feed intake, whereas increasing CM level linearly increased (p < 0.05) FCR in the grower and finisher phases. The eviscerated and carcass yields decreased, whereas the fat yield increased (p < 0.05) with reduced protein in corn–SBM diet. Increasing levels of cDDGS and CM at a constant CP level quadratically decreased (p < 0.05) the eviscerated weight, whereas the fat weight linearly decreased (p < 0.05) with increasing levels of cDDGS and CM. The birds receiving the PC diet had a lower (p < 0.05) lean muscle (%) and a higher fat (%) compared to birds receiving the NC diet at d 21. However, on d42, birds receiving the PC diet had decreased (p < 0.05) bone mineral density, bone mineral content and lean weight compared to those receiving the NC diet. The litter ammonia increased (p < 0.05) with the increasing levels of protein in the SBM diets. In conclusion, 50 g/kg inclusion levels of CM and cDDGS at the same low-protein levels as SBM produced a similar growth response to the NC, whereas higher levels were detrimental. Hence under the conditions of the current experiment, complete replacement of SBM with DDGS or CM in low-protein diets was not feasible.


Introduction
The increase in the world population brings about a spike in the demand for poultry meat. This demand is driven mainly because of poultry products' accessibility, affordability and overall acceptance across a variety of traditions, cultures and religions [1]. However, to meet this increased demand and to sustain a high rate of growth of broilers, high-protein diets are fed to birds. This poses environmental (e.g., release of gases, such as NH 3 and N) and bird welfare concerns [2]. In addition, governmental policies and environmental agencies have increased pressure on producers to lower NH 3 emissions [3,4].
One leading approach to mitigating excess N excretion is a reduction in dietary protein and supplementing such diets with limiting amino acids (AA). Several broiler experiments have demonstrated that supplementing low-protein diets with appropriate supplemental AA resulted in a growth performance equal to that of an adequate-protein diet but with a reduction in litter N [5,6] and/or ammonia [7]. However, nearly all studies on reduced protein diets with supplemental AA have used SBM as protein feedstuff. This is likely because SBM has a high protein content and a nutritionally superior AA balance [8,9]. Roberts et al. [10] reported that ammonia emissions can be diminished when feeding birds a fibrous diet because: (1) AA in a highly fibrous diet is less digestible compared to that in a low fibrous diet; and (2) AA in a fibrous diet is less likely to degrade to the urea and consequently leads to NH 3 , because dietary fiber increases the metabolism and growth of beneficial bacterial in the large intestine [11].
Canola meal (CM) and corn distillers dried grain with solubles (cDDGS) are potential alternative protein feedstuffs currently used in poultry diets in many parts of the world. As with many less-conventional feedstuffs, the chemical composition of the feedstuffs is variable but they are sufficiently high in CP and AA, which make them attractive protein feedstuffs in poultry diets [12]. Roberts et al. [10] reported that the use of cDDGS and CM (which increased dietary fiber) lowered NH 3 emissions from manure over 7 days in comparison to the positive control. The use of cDDGS in low-protein broiler diets has not been studied extensively. For example, Guney et al. [13] showed that up to 200 g/kg of low-fiber cDDGS could be added to broiler starter diets without any detrimental effects on live performance.
To the authors' knowledge, although there is abundant information on the use of SBM in low-protein diets, information on the use of CM or cDDGS in low-protein diets is currently unavailable. Therefore, the objective of the current experiment was to investigate the use of increasing levels of CM or cDDGS compared to SBM in low-protein broiler diets. A diet adequate in protein was included to ascertain the influence of reduced protein in SBM diets. The response criteria of interest were growth performance, carcass yield, whole-body composition and litter characteristics.

Materials and Methods
All animal experiment procedures used in the current study were approved by the Institutional Animal Care and Use Committee of the University of Georgia.

Animals and Diets Experimental Design
A 42-day experiment was conducted to investigate the effect of supplemental AA in low-protein diets with SBM or diets in which graded levels of CM or cDDGS partly (or nearly totally) replaced SBM. Five hundred and forty Cobb 500 (off-sex) male broiler chicks were obtained on the day of hatch (d 0) and allocated on the basis of body weight into 36 floor pens in an environmentally controlled room. All the birds received the same corn-SBM starter diet during the first 10 d, and experimental diets were fed from d 10 to 42.
The model used in the current study was of diets in which nonphytate P and Ca levels were reduced and subsequently supplemented with phytase. Phytase (Quantum Blue, AB Vista) was supplemented at the rate of 0.1 g/kg to provide 500 phytase units per kg diet (one phytase unit is the activity required to release 1 µmoL of inorganic P per minute at pH 5.5 from an excess of 15M sodium phytate at 37 • C). The matrix values used for Ca and nonphytate phosphorus were 1.75 and 1.5 g/kg, respectively, according to manufacturer's specification. Because there were no specific requirements for Gly and Ser, Gly-equivalent value used in the low-CP diets was based on the Gly-equivalent level of the PC diet. The diets were fed as pellets throughout the experiment (as crumbs on d 0 to 10).
Analyzed values of feedstuffs (SBM, CM and cDDGS) used are shown in Table 1. Diet formulas and calculated nutrient values are shown in Tables 2-4. Feed and water were provided ad libitum during the study. Temperature and light schedules followed Cobb 500 management guide.

Growth Performance, Whole-Body Composition, Litter Surface Ammonia Measurement and Carcass Composition
Birds and feed were weighed on d 0, 10, 28 and 42. Mortality was monitored daily and used to calculate mortality-adjusted weight gain (WG), feed intake (FI) and FCR. On d 21 and 42, one bird from each pen was randomly selected for whole-body composition analysis using dual-energy X-ray absorptiometry (DEXA; pDEXA®, Bone Densitometer, Norland Medical System Inc., Fort Atkinson WI, USA). The whole bird was defined as a region of interest. Whole body region scans were conducted to measure bone mineral density (BMD), bone mineral content (BMC), total area of bone, total weights of lean and fat, as well as lean and fat percentages. Scanning was performed across each bird, placing each bird at the same position and orientation during the measurement. All scans were obtained at a scan speed of 2.5 mm/s, with a voxel resolution of 0.07 × 0.07 × 0.50 mm.
Total ammonia produced at the litter surface was determined for each pen on d 42. Ammonia concentrations were determined using a Drãger®chip measurement system (CMS portable gas meter, Drägerwerk AG & Co. KGaA Moislinger Allee 53-55 23558 Lübeck, Germany). Ammonia was measured by placing 0.03 m 3 plastic box on the bedding in an area to the side of the feeder where no caked litter was present. A tube connected the box with the Drager®CMS. The box was left on the bedding for 2 min with the Drãger ®CMS pump running. Ammonia concentration (ppm) was measured after 2 min. Litter samples close to the water lines were collected on d 42 for litter N determination.
Two birds were randomly selected from each pen on d 42, after feed was withdrawn overnight (but with access to water) and slaughtered for carcass yield evaluation. The following data were collected from the carcasses: eviscerated carcass yield, breast yield, wings yield, thighs and drumsticks, and back plus ribs. Carcass yield was calculated as: Eviscerated Carcass Yield, % = Eviscerated carcass weight, g Live weight, g × 100; whereas cuts (e.g., breast) yield was calculated as: Breast yield, % = Breast weight, g Eviscerated Carcass weight, g × 100.

Chemical Analysis
All the diets and feedstuffs were analyzed for chemical profiles using AOAC [20] procedures. Samples were dried at 100 • C for 24 h to determine the dry matter (method 934.01). Nitrogen content was measured using the combustion method (method 968.06). Amino acids were analyzed AOAC procedures (Method 994.12). Briefly, the samples were hydrolyzed with 6 N HCl containing phenol for 24 h at 110 ± 2 • C in glass tubes in an oven. Amino acids were measured using AA analyzer (ion exchange) with ninhydrin post-column derivatization. The chromatograms detected at 570 and 440 nm were integrated using dedicated software (Agilent Open Lab software, Waldbronn, Baden-Württemberg, Germany). Cys and Met were analyzed as cysteic acid and methionine sulphone, respectively, by oxidation with performic acid-phenol for 16 h at 0 • C prior to hydrolysis. For the measurement of Trp, the samples were saponified under alkaline conditions with barium hydroxide solution in the absence of air at 110 • C for 20 h in an autoclave. The internal standard α-methyl Trp was added to the mixture following hydrolysis. After adjusting the hydrolysate to pH 3.0 and diluting with 30% methanol, Trp and the internal standard were separated by reverse phase chromatography (RP-18) on an HPLC column (CORTECS C18 Column; 2.7 µm, Waters Corporation, Dublin, Ireland). Finally, detection was selectively performed by means of a fluorescence detector to prevent interference by other AAs and constituents.

Statistical Analysis
Data were analyzed as appropriate for randomized complete block design using the MIXED procedure of SAS 9.4. The blocks were treated as random variables whereas treatments were the fixed effects. Although all the birds received the same diet until d 10, in order to account for inevitable variation in body weight, d 10 body weight (252 ± 11 g) was used as covariate in the statistical analysis. Orthogonal polynomial contrasts were used to determine the linear and quadratic responses to increasing levels of CM or cDDGS in low-protein diets. When both linear and quadratic responses are significant, only the quadratic response is discussed. One df pair-wise contrast was used to determine the effect of protein reduction between the PC and the NC diets. Statistical significance was set at p ≤ 0.05. Data can be found in Supplementary Materials Table S1.

Results
The analyzed CP and AA contents of the experimental diets (starter, grower and finisher) were close to the expected values (Tables 5 and 6). The analyzed AA content was the total AA, but the values reflected the expected variability due to the inclusion of supplemental AA in the applicable diets. The average analyzed phytase activity was 1278 FTU/kg, which was far in excess of the expected level of 500 FTU/kg. Table 5. Analyzed crude protein and total amino acids content (g/kg, as fed) of experimental starter grower phase diets (day 10 to 28) fed to broilers receiving diets containing soybean with reduction in crude protein or low-protein diets with serial inclusion levels of canola meal or corn-DDGS supplemented with crystalline amino acids.

Growth Performance
Birds on the PC diet were heavier and had a lower (p < 0.05) FCR compared to those on the NC diet in the grower (d 10-28) and finisher (d 28-42) phases (Table 7). In diets with similar levels of protein and digestible AA, increasing the dietary levels of cDDGS at the expense of SBM in the low-protein diets linearly reduced (p < 0.05) weight gain and feed intake in the grower phase but had no effect in the finisher phase. During the grower and finisher phases, increasing the dietary level of CM at the expense of SBM linearly increased (p < 0.01) the FCR but had no effect on feed intake. In addition, increasing the CM level in low-protein diets linearly reduced (p < 0.05) the weight gain during the finisher phase only. The day 42 BW was lower in the NC compared to the PC (p < 0.05). Increasing levels of cDDGS or CM in the low-protein diets linearly reduced (p < 0.05) the d 42 BW.

Carcass Yield
Birds receiving the NC diet had a lower (p < 0.05) eviscerated weight and carcass yield, but a greater (p < 0.05) fat yield and thigh yield compared to those receiving the PC diet (Table 8). Increasing the cDDGS level in the low-protein diets quadratically decreased (p < 0.05) the eviscerated weight, abdominal fat weight and fat yield (%), as well as linearly increased (p < 0.05) the wing yield but had no effect on the carcass yield, breast meat yield, thighs yield or back and ribs yield. Increasing the CM level in low-protein diets quadratically decreased (p < 0.05) the eviscerated weight, as well as linearly reduced (p < 0.05) the abdominal fat weight and fat yield (%); but had no effect on other responses.

Whole Body Composition
On d 21, birds receiving the NC diet had a lower (p < 0.05) lean muscle (%) and a higher fat (%) compared to the birds receiving the PC diet (Table 9). Increasing the levels of cDDGS in the low-CP diet, at the expense of SBM, linearly decreased (p < 0.05) the bone mineral density, bone mineral content and total area of bone as well as the total weight of fat, lean muscle and fat (%) at day 21 (Table 9). On d 42, birds receiving the NC diet had lower (p < 0.05) bone mineral density, bone mineral content and lean weight compared to those receiving the PC diet (Table 10). However, increasing the cDDGS or CM levels in the low-protein diet, at the expense of SBM, had no significant effect on any of the responses on d 42.

Litter Surface Ammonia and Litter Nitrogen
Ammonia was not detectable in the litter surface when measured on d 0. The litter NH 3 was lower in birds receiving the NC diet (p < 0.05) compared to those receiving the PC diet. The inclusion of cDDGS or CM in the low-protein diet, at the expense of SBM, had no effect on the litter NH 3 on d 42, irrespective of the protein feedstuffs used (Table 11). The litter N and litter N per kg body weight gain were not significantly affected by the treatments.

Discussion
The objective of the current study was to assess the growth performance response, carcass yield, whole-body composition, litter ammonia and N levels for broilers fed lowprotein diets with supplemental AA using SBM, cDDGS or CM. Due to the fact that many studies investigating low-protein diets have used diets that are limiting only in protein or AA, in the current study, we utilized diets corrected for Ca and non-phytate P reduction with phytase supplementation in order to mimic standard practice with phytase supplementation. Because all the diets had the same level of phytase, Ca and non-phytate P, the responses obtained can be attributed solely to treatment differences in the level and type of protein feedstuff used. In addition, in order to partition the growth performance responses to phasespecific effects, the body weight at the end of starter phase (d10) was used as a covariate in the analysis of growth performance in both the grower and finisher phases.
Soybean meal is well suited to be the plant protein feedstuff of choice in low-protein AA-supplemented diets. The promotion of alternative protein feedstuffs, such as CM and cDDGS, in low-protein diets compared to SBM is popular because these feedstuffs have a relatively well-balanced amino acid profile [21].
The growth performance of broiler chickens receiving the NC diet in the current study was lower than those receiving the PC diet both in the grower and finisher phases, even though both diets were supplemented with essential AA, including glycine-equivalent to meet the Cobb 500 requirement. This is similar to the observations of several other studies [22][23][24]. However, other authors have reported a comparable performance for broiler chickens fed adequate or low-protein diets [5][6][7]. The poorer performance of birds receiving the NC diet in the current study could be partly explained by the drastic reduction in dietary protein (185 vs. 140 g/kg for grower diet) as a consequence of the drop in SBM inclusion from 265 to 75 g/kg in the PC vs. the NC diets, respectively. This 72% reduction in the dietary SBM level implied a reduction in intact protein in the low-protein diets. On one hand, reducing the amount of protein reaching the hindgut can be beneficial because it translates to less substrates from which putrefactive bacteria in the gastrointestinal tract would generate harmful substances, such as amines or phenols, that may impair growth [25][26][27]. Nevertheless, a comparatively poorer performance can result from such diets due to the inherently different digestion location and dynamics of 'intact' protein and non-bound amino acids. The implications may include an insufficient N pool for the synthesis of non-essential amino acids, dietary imbalances between assumed essential and non-essential amino acids and the possibility that the amino acid requirements have not been identified with sufficient accuracy in the context of reduced-protein diets [28][29][30]. In addition, the lower feed intake in the NC diet (more markedly different in the finisher phase) may account for the reduced weight gain. Protein, and consequently SBM, reduction in the NC diet of the current study was accompanied by the addition of corn starch and cellulose. These dietary modifications will result in increased dietary starch:protein ratio, with consequences on foregut digestion and distal digestive tract fermentation pattern and implications on the weight gain response [31].
By far, the majority of studies utilizing low-protein diets have focused on SBM [32][33][34][35]. However, other plant protein feedstuffs (such as cDDGS and CM) are widely used in broiler diets in many parts of the world. Consequently, one main aim in the current study was to compare the effect of the partial or near complete replacement of SBM with cDDGS or CM while maintaining the same level of standardized digestible AA. Corn DDGS is characterized by a comparatively high protein content and has been studied in many broiler experiments [12,15,36]. However, these studies have incorporated cDDGS for broilers in diets with adequate protein, not in the context of low-protein diets as used in the current study.
In the current study, birds receiving 125 g/kg of cDDGS in the grower phase had a markedly lower weight gain and feed intake compared with the corn-SBM diet at the same protein level. The high cDDGS inclusion level resulted in near total replacement of SBM and higher supplementation levels of supplemental AA. However, no significant weight differences as a result of the cDDGS addition to the diets were observed during the finisher phase. This may indicate that the weight gain effect of the cDDGS replacement of SBM is age-or phase-dependent, or that an adaptation of the birds to the diet had taken place [37,38]. In studies utilizing high dietary levels of cDDGS (although not in low-protein diets), Campasino et al. [39] reported a lower body weight gain in broiler chickens fed 150 g/kg cDDGS compared to those fed diets containing 0 or 50 g/kg DDGS. The highest cDDGS level used in the current study is 25 g/kg lower than Campasino et al. [39]; the effect of high cDDGS inclusion may have been exacerbated by low dietary protein.
The birds receiving 100 g/kg CM had a lower weight gain during the finisher phase, and a higher FCR at both the grower and finisher phases, compared to the broilers receiving SBM. Previous studies [40,41] also reported a decrease in performance when more than 150 g/kg CM was added in the diet of broiler chickens. These observations could be attributed to the high starch content relative to protein in CM, which in turn changes the digestive dynamics in the birds as a result of increasing the dietary starch:protein ratio, leading to poor digestibility as a result of the negative impact on the intestinal structure and function [21,31]. Liu et al. [42] observed that starch is more rapidly taken up than protein in the intestine of broilers fed a sorghum-based diet. Selle et al. [43] suggested that the important role that starch plays in the intestinal passage rate and digestive dynamics is the reason for its rapid digestibility when compared to protein in the gut. Moss et al. [44] observed that birds fed low protein diets with a higher starch content will flood their intestine with glucose which will compete with amino acids for absorption through their shared sodium dependent pathways, thereby causing poor digestibility. In addition, Mc Neill et al. [45] attributed the poorer performance in CM to the presence of a trypsin inhibitor, which would have consequences on the feed intake as well as nutrient utilization.
The eviscerated weight and carcass yield in the current study were lower for birds receiving the NC compared to the PC diet. Others have similarly observed that the carcass yield becomes inferior in broilers fed low-protein diets with more than a 30 g/kg protein reduction, even when all known nutrient requirements are met [22,28,46]. Others have observed that dietary protein reduction has no effect on the carcass yield of broiler chickens even when the reduction is 30 g/kg, or higher [7,47]. It is reasonable to expect that lower live body weight as a consequence of the dietary protein reduction will produce a lower eviscerated carcass weight unless the nutritional intervention produced a differential accretion of the economic and non-economic parts of the carcass. The result of the current experiment suggests otherwise, in that the carcass yield responses followed the same pattern observed for the weight gain responses to both reduced protein as well as the replacement of SBM with cDDGS or CM.
In the current study, broilers fed low-CP diets accreted 10% more abdominal fat compared to those fed an adequate protein diet. It is known that the protein and AA levels of the diets influence the carcass composition of broiler chickens, and that decrease in dietary protein usually precipitates a decrease in the carcass protein and an increase in the carcass fat content [48]. The explanation for the increase in abdominal fat with reduced protein has been attributed to the increased ME:protein ratio leading to poor digestibility as a result of the negative impact on the intestinal structure and function [49,50]. Because all the diets in the current study were isocaloric, in the content of reducing the protein level, the increased ME:protein ratio is an inevitable consequence of the nutritional modification.
The abdominal fat weight and fat yield were lower in birds receiving cDDGS and CM compared to those receiving the NC diet, reflecting the disparities in body weight. This could be attributed to the presence of relatively high fiber content in cDDGS when compared to SBM, thereby leading to lower nutrient digestibility as a result of the negative impact on the intestinal structure and function [37,38,40,51]. On the other hand, the abdominal fat weight and fat yield were greater (or comparable to the NC diet) in diets with lower inclusion levels of cDDGS and CM. Similar observation have been made in CM diets [40,51]. In the case of CM, it has been suggested that the presence of high polyunsaturated fatty acid in CM compared to SBM may decrease the fat deposit in animals [52].
Although there were treatment effects on the carcass and fat yields, there were only marginal treatment effects on the yields of breast meat, thigh, and back and ribs. This is similar to the observations of Kobayashi et al. [53] and suggests that even though dietary protein level modification may influence weight gain or abdominal fat accretion, its influence on carcass cuts is less pronounced. The effects of reduced protein on growth performance may not be mirrored by the effect on carcass cuts. Possible explanations are the possible differences in the effect of dietary modifications on total weight gain or the growth of specific parts and organs. For example, the effect of dietary manipulation may be manifested to a greater degree in the growth and development of the digestive organs but not the breast muscle. Under such circumstances, the effect is seen in the total weight of the animal but not in the carcass cuts. This is why, depending on objective of the experiment, it is important to complement data on weight gain with those on carcass cuts; the latter representing the parts of the birds that are of greater commercial importance [54,55].
A DEXA scan provides an accurate, in-depth body composition analysis of a bird for its fat, muscle, bone and water components. This analysis is relevant, as it can be used to identify health risks and other problems, such as metabolic bone disorder, before they occur [56,57]. In the current study, the scan on d 21 showed that the whole-body fat (%) increased in the NC compared to PC, whereas the lean muscle (%) followed the opposite pattern. This could be attributed to the increased energy to protein ratio in the low CP diets. However, bone characteristics were not affected by the reduction in dietary protein. There were no differences among the treatments during the first 3 weeks of age; this observation could be explained by the proposition that their early life is a period when the birds channel most of their nutrients into organ development and overall intestinal growth [58][59][60].
On the other hand, BMC, BMD, total bone area, fat weight, lean weight, lean muscle % and fat % linearly decreased with the increasing level of cDDGS (lean muscle % increased linearly with the increasing level of cDDGS). The decrease in these responses was more pronounced in the cDDGS50 diet (compared to the NC), and the further addition of cDDGS in the diet did not produce further significant responses. These responses might be attributed to limited ability of broiler chickens in the grower phase (d 10 to 28 in the current study) to utilize a relatively more fibrous cDDGS diet [37,38].
For the scans performed on d 42 in the current study, the only significant observations were of the effect of protein reduction on BMD, BMC and lean weight. Bone mineral density (BMD) is important for diagnosing osteoporosis [61,62]. In broilers, bone strength is relevant both for carcass quality and bird welfare [63]. By d 42 in the current study, both BMD and BMC were lower in all the birds receiving low-protein diets compared to the PC diets. Yang et al. [64] observed no significant effects of different protein dilution levels of broiler starter diets on bone responses. A likely explanation for the observation in the current study regarding the effect of protein reduction is the overall reduction in the feed intake of birds receiving the low-protein diet. The implication of this is a reduction in digestible nutrient intake, including minerals. The reduced mineral intake may be responsible for a reduced bone mineralization, which is observable as bone mineral density.
The lean tissue weight from the DEXA scanning is synonymous with the muscle weight [65]. Therefore, it is no surprise that result observed in lean tissue weight at d 42 in the current experiment followed a similar trend to the carcass weight of broilers at d 42. This result is similar to that of others [66,67], who have observed that feeding broiler chickens low-protein diets impairs their lean muscle weight.
The litter NH 3 in the current study decreased with the 45 g/kg CP reduction. Our observations of the litter surface ammonia response to the protein reduction in the current experiment are similar to those of others [68,69], who have reported reduced litter NH 3 in birds receiving low-CP diets, which can be explained by the reduced water intake in birds receiving low-protein diets due to a decreased need for water to excrete excess N [70]. Others [69,71] also reported significant reductions in ammonia emissions following relatively modest reductions in dietary protein levels. However, litter N (%) and litter N per kg of body weight were not affected by the dietary CP reduction in the current study. The implication, therefore, is that the dietary protein reduction, per se, and not necessarily the feedstuff, is the primary factor for the reduced litter N and NH 3 .

Conclusions
In conclusion, under the conditions of the current experiment, reducing dietary protein by 45 g/kg in corn-SBM diets produced an inferior growth performance, carcass yield and altered whole-body composition (on d 21) but also beneficially reduced the litter surface ammonia. In addition, the partial or near-total replacement of SBM with cDDGS or CM in the low-protein diets had a further negative effect on growth performance. Therefore, complete replacement of SBM with cDDGS or CM in low-protein diets as used in the current experiment is not feasible.