Effects of substitution of dietary protein with carbohydrate on lactation performance in the mink

Thirty mink dams nursing litters of 6 kits were assigned to one of 3 dietary treatments to investigate the effects of changing the protein:carbohydrate ratio on nutrient utilization, heat production, milk production and kit growth. Three diets were formulated to contain 65:3 (LC), 48:15 (MC) or 34:33 (HC) % of the metabolizable energy (ME) from protein and carbohydrate, respectively. The diets were fed ad libitum for 4 weeks from parturition. Twelve dams were held in an intensive care unit and subjected to balance and respiration experiments and the kits were injected with deuterium oxide to measure water kinetics and milk production. Eighteen dams were kept under normal farm conditions and feed intake of dams and weight gain of the kits were determined. Milk samples were collected from the dams. Metabolizable energy intake was not affected by dietary treatment. Carbohydrates were efficiently utilized with a digestibility coefficient of 84% in dams fed the HC diet. Dams fed the HC diet had a lower (PO.05) percentage weight loss, lower (PO.05) total heat production (HE), lower (P<0.05) protein oxidation (OXP), lower (P<0.05) water intake and a lower (P<0.05) nitrogen (N) excretion than dams fed the LC diet. Milk production, and thereby * Part of the Nordic Joint Committee for Agricultural Research Project No. 100 "Stable isotopes in comparative studies of milk production and energy requirement in pigs, mink and foxes", supported by The Danish Veterinary and Agricultural Research Council, Grant No. 9701275 4 Corresponding author: e-mail: Anne-Helene.Tauson@ihh.kvl.dk 648 CARBOHYDRATES FOR MINK liveweights of the kits 4 weeks post partum, was higher (P<0.05) in dams fed the HC diet than in dams fed the LC diet. In conclusion, lactating mink dams are able to utilize digestible carbohydrates with positive effects on lactation performance and reduced nitrogen excretion.


INTRODUCTION
In the Scandinavian countries, recommendations for supply of protein, fat and carbohydrates for mink dams during the lactation period are presently a minimum of 40% of metabolizable energy (ME) from protein, 40-50% of ME from fat, and a maximum of 20% of ME from carbohydrates (Hansen et al., 1991).However, dietary protein supply is usually considerably higher than the minimum recommended, and a conventional Danish lactation diet typically consists of 55:33:12% of ME derived from protein:fat:carbohydrates, respectively.
During lactation, glucose is an important intermediate in the metabolism as precursor in lactose synthesis, and an adequate supply is essential for milk secretion (Annison et al., 1968).Borsting and Damgaard (1995) found that lactating mink fed a diet with 12% of ME derived from carbohydrates had to provide about 75% of their glucose requirement by gluconeogenesis.However, although the mink has a well-developed capacity for gluconeogenesis (Sorensen et al., 1995), lactation places very high energy demands on the female mink, and usually voluntary feed intake of the dams is insufficient to meet nutrient demands for maintenance and milk production.Weight losses in the order of 25% during a 6-week lactation period are common (Tauson, 1988;Hansen and Berg, 1998), and in extreme cases mobilization of body reserves can result in emaciation and death (Clausen etal., 1992).
The digestive tract of the mink is very simple (Szymeczko and Skrede, 1990) with a low a-amylase enzyme activity (Elnif et al., 1988) and microbial activity (Williams et al., 1998).Therefore, to ensure high digestibility, the dietary carbohydrate supply has to be either monosaccharides or gelatinised starch.However, previous experiments have shown that the mink has maintained a high capacity for utilization of digestible carbohydrates, since also the glycolytic capacity is large (Sorensen et al, 1995;Fink et al., 2002a,b).
The objective of the present experiment was to test our hypothesis that replacement of dietary protein by carbohydrates has beneficial effects in reducing nutritional stress on the dam, increasing milk production during the first 4 weeks of lactation, increasing kit growth, and reducing nitrogen emission from the mink industry.

Animals and feeding
Thirty, two-year-old mink dams of the standard brown genotype (Nes et al., 1987) were allocated into 3 groups at parturition.When necessary, kits were cross-fostered within 48 h after birth to obtain six kits per litter.The dams had free access to drinking water and were fed ad libitum from parturition until the fourth week of lactation on diets with, on average, 34% (range 32-37) of ME from fat and ratio (% of ME) between proteinxarbohydrates ranging from [low carbohydrates (LC): 65:3, medium carbohydrates (MC): 48:15 to high carbohydrates (HC): 34:33].The diets were mixed on a single occasion and weighed out into plastic bags and immediately frozen.Feed was taken out of the freezer the day before use and thawed overnight.Feed samples were collected daily and pooled for each week, and stored at -18°C until analysis.Dietary composition and chemical analyses of the diets are given in Table 1, and analysed amino acid composition in Table 2.

Balance and respiration experiments
Twelve of the dams (n=4 per dietary treatment) were transferred from the experimental farm to the laboratory for adaptation approximately one week prior to expected delivery.In the laboratory, the dams were maintained under natural daylight conditions (May, 55°N 12°E) in individual metabolism cages designed as described by Jorgensen and Glem-Hansen (1973), equipped with nest boxes and devices for feeding and drinking water supply and quantitative collection of feed residues, faeces and urine.The experiment was divided into four consecutive one-week balance periods, which started on May 4 and continued until June 1.Recording of drinking water intake and quantitative collection of feed residues and excreta took place once daily between 8.30 and 12.00 a.m., and the pooled amount collected from each period was stored at -18°C until analysis.The collection procedures have been described in detail by Wamberg et al. (1996).Each balance period included a 22-h respiration experiment by means of indirect calorimetry in an open-air circulation system.In the respiration chamber, temperature and relative humidity were 15-18°C and 65-75%, respectively.Air from the chambers was analysed every third min for concentrations of 0 2 and C0 2 .Oxygen was analysed by a paramagnetic analyser (Magnos 4G, Hartmann and Braun, Frankfurt, Germany), and C0 2 was analysed using the infrared principle (Uras, Hartmann and Braun, Frankfurt, Germany).Measurements of milk intake of the kits at weekly intervals by the water isotope dilution technique were performed as described in detail by Fink et al. (2001).The animals were weighed at weekly intervals.

Milking procedure
The remaining 18 dams (n=6 per dietary treatment) were kept under normal farm conditions and daily feed intake, and weekly liveweights of kits and dams were recorded.Milk samples representing each group and lactation week were collected from the dams by separating the dam from her kits for approximately one hour, after which the dam was anaesthetised [0.4 ml Xylazin (Narcoxyl®), 10 mg/ml + 0.2 ml ketamin (Ketaminol®) 50 mg/ml].The anaesthetised dam received an intramuscular injection of 0.1 ml oxytocin (10 IU/ml, Leo Vet, Ballerup, Denmark) and all milk glands were milked and emptied as much as possible by intermittent suction using a small rubber tubing connected to a water vacuum pump.The milk samples were stored at -18°C until analysis.

Ethical approval
The experimental procedures complied with Danish National Legislation and the guidelines approved by the Member States of the Council of Europe for protection of vertebrate animals (Anonymous, 1986).

Analytical procedures, feed, excreta and milk
Diets, feed residues, faeces and milk were analysed for: dry matter (DM) by evaporation to constant weight at 105°C, ash by combustion at 525°C for 6 h, nitrogen (N) by the micro-Kjeldahl technique using the Tecator-Kjeltec system 1030 (Tecator AB, Hoganas, Sweden), and crude protein (CP) was calculated as N><6.25 (diets and faeces) and N*6.38 (milk).Fat in diets and faeces was determined by petroleum ether extraction after HC1 hydrolysis, and gross energy (GE) by use of an adiabatic bomb calorimeter.In milk samples, fat was analysed by a method slightly modified from Baverstock et al. (1976) as described by Tauson et al. (2004), and GE was calculated by use of the data for milk composition and the energy factors (in kJ/g): protein (23.9), fat (39.8) and carbohydrates (17.6).In all samples, contents of carbohydrates (CHO) were calculated by difference (CHO = DM -ash -CP -fat).The DM in urine was taken to be 10% (Neil, 1988), and energy in urine (UE) was estimated from the N content in urine (UN) by the following regression equation: UE, kJ = 27.1><UN,g -0.8 (Tauson, unpublished data).Amino acids in the diets were determined by oxidation with a performic acid/phenol mixture and hydrolysed with HC1 (6 N) for 23-h.The hydrolysate was adjusted to pH 2.20.The amino acids were separated by ion exchange chromatography and determined by reaction with ninhydrin using photometric detection at 570 nm (440 nm for proline) (EC, 1998).

Calculations, balance, respiration and milk energy data
Metabolizable energy (ME) was calculated as GE -energy in faeces (FE) -UE.The N balance was calculated as the difference between daily N intake in feed and N output in faeces, urine and milk.Heat production (HE) was calculated from 0 2 consumption, C0 2 production and UN according to the formula by Brouwer (1965): HE, kJ -(16.18x 1, 0 2 ) + (5.02 x 1, C0 2 ) -(5.99 x g, UN) Retained energy (RE) was calculated as ME -HE.If RQnp (non-protein respiratory quotient) was higher than 0.707, net oxidation of fat (OXF), carbohydrates (OXCHO) and protein (OXP) was calculated from measurements of gas exchange and excretion of UN in accordance with the methods described and validated for pigs by Chwalibog et al. (1992).However, in cases where RQnp<0.707there was no net oxidation of carbohydrates (OXCHO=0), whereas the amount of oxidized fat included both net fat oxidation and energy cost of fat mobilization (Chwalibog, personal communication).The respiration experiments were performed with dam and litter, and since dam and kits are huddling, results relating to metabolic weight (kg 0 75 ) consist of the total metabolic weight of dam and litter.Nutrient and energy output in milk (LE) was calculated based on the estimated milk production, the chemical composition and the DM content of the milk, e.g., energy output in milk (LE).

Statistical analyses
Statistical analyses on animal liveweights, feed and drinking water consumption, energy metabolism data, milk intake, milk production and chemical composition of milk were carried out by means of the MIXED procedure in SAS (Littell et al., 1996) by the following model: where Y.. k is the Y jjk th observation, \i the general mean, a. the fixed effect of treatment (diet: LC, MC or HC), p k the fixed effect of time (weeks one to four of lactation), aP jk the interaction effects between treatment and week, and e the residual error.
Lactation week was used as a repeated measure and the autoregressive order 1 (AR (1)) covariance structure was fitted (Littell et al., 1996).Animal liveweights showed no differences between animals housed on the farm or in the laboratory, therefore results on liveweights were pooled (females n=30, kits n=180).Results are presented as least squares means (LS-means), and effects were considered significant at PO.05.

RESULTS
The fat contents in % of ME were very similar and thus the protein was replaced by carbohydrates as planned (Table 1).However, the ME content increased from 15.7 kJ/kg DM in the LC diet to 17.2 kJ/kg DM in the HC diet (Table 1).The content of amino acids (g/kg DM) increased with increased protein content of the diets (Table 2).However, when calculated as g amino acids per 16 g N, the total amount of amino acids decreased with increasing dietary protein, being 76.5, 67.0 and 65.9 g per 16 g N in the HC, MC and LC diet, respectively (Table 2).

Feed and drinking water consumption
Feed intake increased (PO.001) as lactation progressed in all dams.Measured in grams per day, dams fed the LC diet consumed approximately 20% more (P<0.01)per day than dams fed the HC diet, the difference increasing for each lactation week (Table 3).The average carbohydrates intakes were 5, 14 and 32 g per day for dams fed the LC, MC and HC diets, respectively.However, the metabolizable energy intake was not significantly affected by dietary treatment (Table 3).The average digestibility coefficients were 82% (range 81-83%) and 97%) (range 97-98%>) for protein and fat, respectively.For carbohydrates the digestibility coefficients were 81 and 84%> in dams fed the MC and HC diet, respectively, and 57%> in dams fed the LC diet.Feed intake provided the main water supply in all dams.However, drinking water consumption was affected (P<0.01) by lactation week and dietary treatment, dams fed the LC diet consuming approximately 43% more drinking water per day during weeks one to four of lactation than dams fed the HC diet (Table 3).

Nitrogen metabolism
The N intake increased as lactation progressed (P<0.001) in all dams, and with increased protein content of the diets (P<0.05), the N intake being higher (P<0.01) in dams fed the LC diet than in dams fed the MC and HC diets (Table 4).Consequently, the total N excretion increased (P<0.001) as lactation progressed, but with the N excretion via faeces and urine being higher (P<0.01) in dams fed the LC diet than in dams fed the MC and HC diets.The N output in milk was not affected (P>0.05) by dietary treatment, but there was a tendency for lower excretion in HC dams, this difference being significant (P<0.01) in the fourth week of lactation.The resulting N balance was not affected by dietary treatment, and was not significantly different from 0 in any group.However, it decreased significantly (P<0.001) from week 1 post partum, reaching a negative value in week 4 (Table 4).Because the N balance was 0, the efficiency of utilization of digested N (DN) for milk production could be calculated: it increased significantly (P<0.001) with decreasing dietary protein supply from 31%) for LC dams to 44%) for HC dams.Also the effect of lactation week was significant (P<0.05) and the efficiency of utilization DN was moderately increased from 32%o in lactation week 1 to 38%o in the following weeks (Table 4).

Respiration experiments
Heat production, calculated in relation to metabolic liveweight of dam with litter (kJ/kg 075 ), was relatively constant during the lactation period, but was significantly (P<0.05)higher in LC dams (676 kJ/kg 0 75 ) than in dams fed the HC diet (546 kJ/kg 0 75 ), the difference being approximately 19% (Table 5).The oxidation of protein (OXP) in relation to total HE was about twice as high (P<0.001) in dams fed the LC diet than in dams fed the HC diet, whereas the opposite was the case for the oxidation of fat (OXF) (P<0.05),OXF contributing more than 50% to the total HE in all diets.Oxidation of carbohydrates (OXCHO) in relation to total HE was not affected (P>0.05) by dietary treatment, or lactation week, and values were generally low (Table 5).The energy output in milk (LE) increased (P<0.001) during the lactation period, but the LE in relation to metabolic liveweight of the dams was not significantly different between dietary treatments (Table 5).Average values for retained energy (RE) for all 4 measurement weeks were negative, and most negative for MC dams (Table 5).During lactation week 4, LE was close to (LC and HC) or even slightly higher (MC) than the ME intake, which was reflected in clearly negative RE values, suggesting mobilization in the order of 10 g fat per day.

Dam liveweights
The absolute liveweights of the dams in weeks 1, 2, 3 and 4 post partum did not differ.However, the individual weight loss from week one to week four post partum, was significantly (PO.001) higher in dams fed the LC diet, losing 126 g or 10% compared with 34 (3%) and 13 g (1%) in dams fed the MC and HC diets, respectively (Table 3).

Milk production
Daily milk intake of the kits, and thereby the daily milk production, increased significantly in all treatment groups as lactation progressed (P<0.001).The daily milk intake was highest in kits nursed by dams fed the HC diet, and during weeks three and four post partum the milk production was significantly higher (PO.05) in dams fed the HC diet than in dams fed the MC and LC diets (Table 6).

Chemical composition of milk
The DM, fat and gross energy (GE) contents of the milk increased (P<0.05) as lactation progressed, but were not affected (P>0.05) by dietary treatment (Table 6).There was a tendency for decreasing protein content with decreasing dietary protein supply, and in the fourth week of lactation the difference between dams fed the HC and LC diets was significant (P<0.05).The carbohydrates concentration in milk was higher (P<0.05) in dams fed the HC diet compared with dams fed the LC diet (Table 6).

Kit liveweights
Growth of the kits was affected (P<0.05) by dietary treatment of the dams, HC kits being the heaviest (Table 6).At four weeks post partum, individual body weights of kits nursed by dams fed the HC diet were 6 and 10% higher (PO.001) than those of kits nursed by dams fed the MC and LC diets, respectively.

DISCUSSION
The experimental diets, in this study, were not designed with the purpose of determining the requirements of individual amino acids, but the performance results indicate that the amino acid requirements for lactation and growth of the suckling kits were sustained even by the HC diet.However, the HC diet had a relatively better protein quality than the other diets measured in terms of contents of essential amino acids and total amino acid content per 16 g N.This was caused by the higher content of chicken and lower content of cod offal.
The digestibility of the nutrients was high and almost the same in all 3 experimental diets, except for a lower carbohydrates digestibility in dams fed the LC diet, probably being an artefact caused by the low carbohydrates intake of only 5 g/day.However, the high carbohydrates digestibility coefficients in dams fed the MC and HC diet, confirm that mink is able to metabolize high amounts of heat-treated carbohydrates, as previously found by Sorensen et al. (1995) and Fink et al. (2002a,b).The ME consumption was not significantly affected by dietary treatment.Monogastric animals are, within certain limits, able to regulate their feed intake independent of dietary energy concentration (pigs: Cole et al., 1967Cole et al., , 1969;;hens: de Groot, 1972).Previous experiments in mink have shown that lactating dams are able to compensate for decreased dietary energy concentration by an increased feed intake when dietary energy concentration ranged from 15.4 to 19.2 MJ ME/kg DM (Tauson, 1988).The diets used in this experiment were in the range where compensation could be expected, and the animals adjusted their food intake to the different dietary energy densities.
Diets with high protein content impose great demands on the animals' water supply, because excess N is excreted in urine after deamination.In the present experiment, dams fed the LC diet had a significantly higher drinking water consumption and excretion of urinary water than dams fed the HC diet, corresponding to results by Neil (1988).Although mink have a great capacity to concentrate their urine (Valtonen et al., 1982), this ability is reduced during lactation, and lactating animals do not regulate the amount of water secreted in milk (Olsson, 1986).Therefore, lactating dams fed high protein diets may be vulnerable to shortages of water supply.
Nitrogen balance studies can be used to indicate if the protein supply meets the requirement, but in carnivores they often overestimate N retention (mink: Skrede, 1978;ferret: Jarosz and Barabasz, 1988;cat: Zentek et al., 1998), mainly because of incomplete recovery of urinary N as demonstrated by Tauson et al. (1997b).However, the presented N balance data are based on very careful collection routines, and thus are as reliable as possible without direct determination of N recovery (Tauson et al., 1997a).The negative N balances recorded during the fourth week of lactation might indicate that some mobilization of body protein occurred, but this was not supported by the protein oxidation data, since they were not affected by stage of lactation.
The HE was calculated from the 0 2 consumption and C0 2 production from the dam with litter, because separation of mother and young during the 22-h measurement period is impossible.However, age of the kits and litter size were equal, and hence differences in HE and substrate oxidation could be equally related to the dam in all treatment groups.The significantly higher HE per kg metabolic weight found in dams fed the LC diet compared with dams fed the HC diet, demonstrated clearly that when the protein supply is above the requirement, excess protein is deaminated and used as an energy source and excess N is excreted in urine by an energetically costly process.Furthermore, among animals fed the LC diet, protein oxidation contributed with 42% to the total HE, whereas it declined to 21%> of HE when the protein supply was reduced to 33% of ME (HC).However, this level is higher than the about 13%> of HE reported for lactating sows (Theil, 2002) and indicates, most likely, that also the HC diet sustained the protein requirement of the lactating dams.
It is suggested that evolutionary pressures have resulted in deletion or modulation of the activities of different key enzymes involved in protein and carbohydrates metabolism, which may explain the high protein requirement of mink.In another strict carnivore, the cat, the activities of some key transaminating and gluconeogenic enzymes did not change noticeably in response to changes in dietary protein supply (Rogers et al., 1977), but recent results suggest that the cat is capable of metabolic flexibility: hence net protein oxidation of cats fed diets with 35.3 vs 51.9%o protein energy was significantly higher among cats on the highest protein supply (Russell et al., 2002).Furthermore, data from whole-body protein turnover studies on cats given more extreme dietary protein supplies (20 vs 70%o protein energy) demonstrated that flux, protein synthesis and protein breakdown were all significantly lower in cats given the lowest protein supply (Russell et al., 2003).Corresponding data are almost lacking for the mink but the few available results indicate that the mink has a high activity of hepatic gluconeogenic enzymes (Sorensen et al., 1995) as well as a high glycolytic capacity (Fink et al., 2002a,b), and the present results concur with the concept of ability to regulate protein oxidation rate.
Individual kit body weights four weeks post partum were 27% higher in kits nursed by dams fed the HC diet than in kits nursed by dams fed the LC diet, irrespective of housing conditions.This further confirms that dams fed the HC diet had a higher milk production than dams fed the LC diet.The higher body weights of kits nursed by dams fed the low fat:carbohydrates ratio (33:33) compared with dams fed the high ratio (32:3), were in contrast to results by Skrede (1981), who found improved kit growth performance already twenty-one days post partum, indicating positive effects on milk production by increased fat:carbohydrates ratio from 32:24 to 50:6.The diverging results can probably be explained by a relatively constant fat content and substitution of dietary protein with carbohydrates in our experiment, hence a substitution with limited effect on dietary energy density, whereas Skrede (1981), by increasing the dietary fat content, simultaneously increased the energy density of the diet.However, variation in growth rate may not depend solely on the volume of milk consumed by the kits.Alterations in milk composition may also modulate the differences in volume intake, and differences may occur in the efficiency of nutrient utilization (Fiorotto et al., 1991).
During the four weeks of lactation the DM and fat content of the milk increased from approximately 18 to 27%, and 5 to 11%>, respectively, unaffected by dietary treatment, corresponding to previous results in mink (Olesen et al., 1992;Fink et al., 2001).Mink is a rapidly growing species, and it is believed that the suckling mink has a higher protein requirement than that of most mammalian species.In the present experiment, the protein contents of milk in the fourth week of lactation were approximately 11.0, 9.4 and 6.5% in dams fed the LC, MC and HC diets, respectively, and thereby significantly lower in dams fed the HC diet compared with dams fed the LC diet.Olesen et al. (1992) found protein contents in mink milk of 8-10%), which are in accordance with most other carnivores (Widdowson, 1965;Oftedal, 1984).This indicates that 6.5% milk protein may be low for a carnivore, however, kits from these dams had the highest body weight four weeks post partum.The fact that the carbohydrates content of milk was highest in dams fed the HC diet and that the carbohydrates oxidation was low in all dams, may indicate that the dietary carbohydrates fraction was used mainly in the synthesis of milk.However, in all dams the carbohydrates fraction calculated by difference was several times higher than the lactose content of 1% found in mink milk by Olesen et al. (1992), indicating as Conant (1962) claimed, that lactose is not the dominant carbohydrates in mink milk.However, a detailed profile of the carbohydrates fraction of mink milk has yet to be established.

TABLE 1
Dietary composition and chemical analysis (g/kg) of the three experimental diets [low carbohydrate (LC), medium carbohydrate (MC) and high carbohydrate (HC)], and calculated metabolizable energy(ME) 1 chicken prepared for human consumption, i.e. without head, feet, skin and entrails

TABLE 2
Amino acid composition of the experimental diets [low carbohydrate (LC), medium carbohydrate (MC) and high carbohydrate (HC)]

TABLE 3
Effect of dietary treatment [low carbohydrate (LC), medium carbohydrate (MC) and high carbohydrate (HC)] on feed and drinking water consumption (n=4 per treatment group) and liveweight change (n=10 per treatment group) of mink dams in weeks one to four post partum(LS means) * RR: square root of residuals ab values that share no common superscript differ significantly (P<0.05);effect of diet

TABLE 4
Effect of dietary treatment [low carbohydrate (LC), medium carbohydrate (MC) and high carbohydrate (HC)] on nitrogen (N) intake, N excretion and N balance, g/day and utilization of digested nitrogen (DN) for milk production and retention, % of mink dams (n=4 per treatment group) during weeks one to four of lactation * RR: square root of residuals ab ' c values that share no common superscript differ significantly (P<0.05);effect of diet

TABLE 5 Effect
abc values that share no common superscript differ significantly (P<0.05);effect of diet

TABLE 6
Effect of dietary treatment [low carbohydrate (LC), medium carbohydrate (MC) and high carbohydrate (HC)] on milk production and chemical composition of the milk of mink dams weeks one to four post partum, and liveweights of kits (n=180) nursed by these dams abc values that share no common superscript differ significantly (P<0.05);effect of diet