J. Anim Sci. 2008. 86:2377-2384. doi:10.2527/jas.2007-0811
© 2008 American Society of Animal Science
Comparative sulfur analysis using thermal combustion or inductively coupled plasma methodology and mineral composition of common livestock feedstuffs1
B. J. Kerr*,
C. J. Ziemer*,
T. E. Weber*,
S. L. Trabue*,
B. L. Bearson*,
G. C. Shurson
and
M. H. Whitney
* USDA-ARS, Ames, IA 50011; and
University of Minnesota, St. Paul, MN 55108
 |
Abstract
|
|---|
The objective of this study was to compare the use of thermal combustion (CNS) and inductively coupled plasma (ICP) to measure the total S content in plant-, animal-, and mineral-based feedstuffs, and to provide concentrations of other macro- and micro-minerals contained in these feedstuffs. Forty-five feedstuffs (464 total samples) were obtained from suppliers as well as swine feed and pet food manufacturers throughout the United States. Mineral data from IPC analysis were summarized on a DM basis using sample mean and SD, whereas the comparison of total S content between CNS and ICP was examined by bivariate plot and correspondence correlation. Analyses of a wide range of feedstuffs by CNS and ICP for total S were comparable for all but a few feedstuffs. For potassium iodide and tribasic copper chloride, ICP estimated total S to be lower than when analyzed by CNS (bias = 2.51 ± 0.15 SE, P < 0.01). In contrast, for defluorinated phosphate and limestone, ICP estimated total S to be greater than when analyzed by CNS (bias = –1.46 ± 0.51 SE, P < 0.01). All other samples had similar estimates of total S, whether analyzed by CNS or ICP. As expected, S composition varied greatly among feedstuffs. For total S, plant-based feedstuffs generally had lower total S compared with animal-based feedstuffs, whereas minerals supplied in sulfate form had the greatest concentration of total S. In addition to total S, mineral composition data are provided for all feedstuffs as obtained by ICP analysis. Within specific feedstuffs, mineral composition was quite variable, potentially due to low concentrations in the feed-stuff causing high mathematical variation or due to the source of feedstock obtained. In general, analyzed values of P were similar to previous tabular values. These data provide feed formulators a database from which modifications in dietary minerals can be accomplished and from which mineral requirements can be met more precisely to reduce losses of minerals into the environment.
Key Words: analysis • dietary sulfur • feedstuff • inductively coupled plasma • thermal combustion
|
INTRODUCTION
|
|---|
Literature pertaining to S chemistry and metabolism is both dated (Lewis, 1924
; Du Vigneaud, 1952; Dziewiatkowski, 1962) and recent (Stipanuk, 2004; Shoveller et al., 2005; Baker, 2006), but little attention has been given to inorganic S (Baker, 1977). Gastrointestinal ecology is a growing area of science (Anderson, 2003), and it is well established that sulfate reduction does not occur in mammalian cells, but does in sulfate reducing bacteria (Gibson, 1990; Florin et al., 1991). Because evidence suggests organic and inorganic S in gastrointestinal tissues may be linked to chronic intestinal disease (Burrin and Stoll, 2007), excess dietary S reaching the intestinal tract is of great importance. In addition, high concentrations of total dissolved solids in drinking water are known to affect scour scores (Maenz et al., 1994), with sulfate being suggested as the major compound of concern (McLeese et al., 1991). Research studies specifically evaluating increased inorganic sulfate intake showed an increase in diarrhea, but did not show any adverse effect on pig BW gain or feed intake (Veenhuizen et al., 1992; Gomez et al., 1995). Limited research results have shown that reducing total dietary S reduces odor and hydrogen sulfide emissions (Apgar et al., 2002). With S known to affect gastrointestinal health, occurrence of diarrhea, nutrient utilization, and odor emission, data are needed to quantify the total concentration of S in feedstuffs commonly fed to livestock. Because standard manure values reported by the ASAE (2003) do not consider changes in dietary mineral concentrations, updated mineral content of ingredients are needed for nutritionists to modify dietary mineral intakes to ultimately reduce mineral excretion. Therefore, the objective of this research was to analyze common feed ingredients for mineral concentration and to determine the difference in total S of ingredients as measured by thermal combustion (CNS) and inductively coupled plasma (ICP).
|
MATERIALS AND METHODS
|
|---|
Animal Care and Use Committee approval was not obtained for this manuscript because no animals were used in the research.
Four hundred sixty-four samples of 45 different feed-stuffs were obtained from ingredient suppliers as well as swine feed and pet food manufacturers throughout the United States for total mineral analysis. To protect the confidentiality of ingredient sourcing, however, suppliers or locations of each feedstuff are not provided in this manuscript. All samples were ground through a 1-mm screen before analysis. Dry matter was determined by drying 1-g samples at 70°C for 24 h. After drying, samples were either analyzed for minerals by ICP and for total S by CNS. For ICP, 10-g samples were shipped to the University of Minnesota and analyzed for mineral content by the University of Minnesota Soils Laboratory. Feedstuffs were digested following methods described by Miller (1998) and then analyzed in 10% HCl by ICP (model 3560, Applied Research Laboratory, Sunland, CA). For thermal combustion, total S was analyzed using a thermal combustion analyzer (VarioMAX, Elementar Analysensysteme GmbH, Hanau, Germany), which uses catalytic tube combustion to volatilize the sample. The target gas is converted to SO2, separated from other gases using adsorption columns, and after heating, measured using a thermal conductivity detector. For S analysis by CNS, sulfadiazine was used as the S standard. Mineral data obtained from ICP analysis were summarized on a DM basis using sample mean and SD. Comparison of total S content between CNS and ICP was examined by bivariate plot and correspondence correlation where it is assumed that if CNS = ICP the slope would equal 1 and the intercept would equal, with deviation from the line (i.e., bias) indicating a difference between analytical methods (Lin, 1989; Meek, 2007). For the purposes of data reporting, the feedstuffs were classified into 1 of 3 groups: plant-based, animal-based, and mineral-based feedstuffs.
|
RESULTS AND DISCUSSION
|
|---|
Mineral analysis of plant-, animal-, and mineral-based feedstuffs are presented in Tables 1, 2, and 3, respectively, whereas only a subset of the ingredients could be analyzed for Cl (Table 4). With the wide array of feedstuffs analyzed, it was fully expected that total S concentrations would vary greatly, ranging from less than 500 mg/kg for feedstuffs such as calcium iodide, tribasic copper chloride, defluorinated phosphate, and potassium iodide; to concentrations over 100,000 mg/kg for cobalt-, copper-, iron-, manganese-, and zinc-sulfate. Bivariate plot and correspondence correlation revealed a correspondence correlation coefficient (r = 0.941) for S analysis between CNS and ICP (Figure 1). Only 4 feedstuffs appeared to deviate from this relationship. For potassium iodide and tribasic copper chloride (group 1, Figure 1), S values determined by ICP were lower than S determined by CNS (bias = 2.51 ± 0.15 SE, P < 0.01) In contrast, S values for defluorinated phosphate and limestone (group 2, Figure 1) determined by ICP were greater than S determined by CNS (bias = –1.46 ± 0.51 SE, P < 0.01). Feedstuffs in group 3 representing the sulfate containing minerals (cobalt-, copper-, iron-, manganese-, and zinc-sulfate), were high in total S, but had similar S concentrations whether measured by CNS or ICP.
View larger version (20K):
[in this window]
[in a new window]
|
Figure 1. Bivariate plot and correspondence correlation between the log of thermal combustion analysis (CNS) and the log of inductively coupled plasma-mass spectrometry analysis (ICP; correspondence correlation coefficient, r = 0.941). Group 1 feedstuffs (potassium iodide and tribasic copper chloride) show that S values determined by ICP were lower than S determined by CNS [bias = 2.51 ± 0.15 (SE), P < 0.01]. Group 2 feed-stuffs (defluorinated phosphate and limestone) show that S concentrations determined by ICP were greater than S determined by CNS [bias = –1.46 ± 0.51 (SE), P < 0.01]. Group 3 feedstuffs (cobalt-, copper-, iron-, manganese-, and zinc-sulfate) were high in total S but had similar S concentrations whether measured by CNS or ICP. The gray lines refer to an empirical histogram representing the distribution of the data.
|
|
For plant-based feedstuffs (Table 1
), concentrations of total S were relatively low compared with other feed-stuffs, except for feedstuffs known to contain moderate concentrations of total sulfur amino acids (NRC, 1998) or by-products derived from corn or soybean processing. Although various nutrients typically increase 3-fold due to the removal of starch during dry-grinding and fermentation of corn to produce ethanol (Spiehs et al., 2002; Stein et al., 2006; Pedersen et al., 2007), the use of sulfuric acid in the dry grind process appears to increase the total S content of dried distillers grains with solubles by 5-fold compared with corn (Table 1). Slightly greater concentrations of S relative to corn were also noted for corn gluten feed and corn gluten meal, which are by-products of the wet-milling industry, where sulfuric acid is used in the steeping process. In contrast, this same magnitude of S increase was not noted when comparing values from the NRC (1998) for corn relative to its by-products, suggesting that either corn processing methods have changed by-product composition or the values reported in NRC (1998) may underestimate the true S content. Because whole soybeans were not analyzed, comparisons between soybean milling byproducts relative to whole soybeans could not be made. However, S content would be expected to increase due to removal of the oil during soybean processing, but we are not aware of any S containing ingredients used during processing that would increase S content.
Animal-based feedstuffs contained greater concentrations of total S relative to most plant-based ingredients. Because sulfur amino acids are prevalent in various body components (condroitin, glutathionine, taurine, etc.) are high in S (21% in Met and 26% in Cys), this was expected. Two examples of animal-based feed ingredients with high S content are feather meal and plasma protein. Both of these ingredients are known to have high concentrations of sulfur-containing amino acids, 4.1 and 2.6% Cys, respectively (NRC, 1998); thus they had high concentrations of total S (Table 2). For comparison purposes, the value for feather meal is close to that listed in the most recent swine NRC (1998), but no such value exists for plasma protein.
The range in S concentration in mineral-based feed-stuffs was the greatest, depending on whether the mineral source was sulfate- or nonsulfate-based. In addition, the method used to process specific minerals also affected its total S concentration. For example, sulfuric acid is used in the production of dicalcium- and mono-calcium-phosphate, whereas phosphoric acid is used in the production of defluorinated phosphate. Consequently, total S in defluorinated phosphate is approximately 5% of that found in dicalcium- and monocalcium-phosphate (Table 3).
One key aspect of our research results is the large variation that can be noted in feedstuff mineral concentrations. Overall, the variability of S content in whole grains (corn, oats, rice, and wheat) was relatively low, usually less than 10%. Depending upon the processing technology, variation in S content may or may not be affected. For example, the S concentration in corn from the ICP analysis was 1,132 mg/kg with a CV of 9.4%, whereas the S concentration for dried distillers grains was 6,866 mg/kg with a CV of 33%. In contrast, no such increase in variation was noted in the S content of corn gluten feed and corn gluten meal (CV of 6.3 and 10.8%, respectively). There was no consistent effect of wheat milling on the variability of S analysis where the CV for S was 14.8, 25.7, and 11.9% for wheat, wheat flour, and wheat middlings, respectively (Table 1). Although our data are limited, this suggests that technology that uses S containing compounds in their process may add additional variability to S analysis. Mineral sources were not immune to variability. For example, the analyzed S value for cobalt carbonate was 12,669 mg/kg with a CV of 108%. However, among the 6 samples, 3 samples from one supplier averaged 25,082 mg/kg of S, whereas 3 samples from a different supplier averaged 256 mg/kg of S. Likewise, calcium concentrations of these samples was also variable, with the 3 high S samples having a high Ca content (32,821 mg/kg), whereas the 3 low S samples had a low Ca concentration (876 mg/kg).
With potential linkages of dietary S to chronic intestinal disease (Deplancke et al., 2000; Attene-Ramos et al., 2006; Burrin and Stoll, 2007), scour scores (Anderson and Stothers, 1978; Paterson et al., 1979; Maenz et al., 1994), protein digestibility (Anderson et al., 1994), and hydrogen sulfide emissions (Sutton et al., 1998; Whitney et al., 1999; Apgar et al., 2002), data on total dietary S is vital. In evaluation of the data in Tables 1, 2, and 3, a 3-pronged approach to reduce total dietary S would be to 1) lower dietary protein and still maintain Met and Cys requirements; 2) use defluorinated as the P source; and 3) use oxide-based trace minerals.
Comparison of our values to table values is a daunting task depending upon which ingredient and mineral to compare. In light of this, we chose to compare 3 plant- and 3 animal-based feedstuffs in terms of their P content and variability. On a DM basis, analyzed P concentration compared with NRC (1998) values were corn, 0.30 vs. 0.31; distillers dried grains with solubles, 0.89 vs. 0.83; soybean meal, 0.80 vs. 0.77; dried whey, 0.82 vs. 0.75; meat and bone meal, 4.78 vs. 5.35; and plasma protein, 1.53 vs. 1.88% P, respectively. Despite differences of when the samples were obtained and location differences among samples, the average values reported herein are similar to tabular values. Similar to that in discussing the S concentrations previously, there can be a large range of concentrations within an ingredient. Coefficients of variations in corn, distillers dried grains with solubles, soybean meal, dried whey, meat and bone meal, and plasma protein were 14.1, 12.7, 28.7, 25.8, 25.1, and 10.9%, respectively. Consequently, fine tuning of dietary levels of minerals, in this case P, requires continual evaluation of each specific ingredient.
For mineral-based feedstuffs, the NRC (1998) does not report DM for minerals, consequently values we reported in Table 3 were converted to an as-is basis for comparison. In general, our values were relatively close to those reported in the swine NRC (1998). For macro-minerals, we reported 34.2% Ca and 20.3% P for defluorineated phosphate vs. 32.0% Ca and 18.0% P in the NRC (1998), 21.8% Ca and 17.8% P for dicalcium phosphate vs. 22% Ca and 18.5% P in the NRC (1998), 16.6% Ca and 20.0% P for monocalcium phosphate vs. 17.0% Ca and 21.1% P in the NRC (1998), and 39.9% Ca and 0.01% P for limestone vs. 38.5% Ca and 0.02% P in the NRC (1998). Comparative evaluation of a several micro-minerals were: copper sulfate, 27.2 vs. 25.2% Cu; iron sulfate, 32.1 vs. 30.0% Fe; manganese sulfate, 26.4 vs. 29.5% Mn; tribasic copper chloride, 62.8 vs. 58.0% Cu; zinc oxide, 79.8 vs. 72.0% Zn; zinc sulfate, 38.1 vs. 35.5% Zn; for the current data vs. NRC (1998), respectively. Similar to plant- and animal-based feedstuffs, mineral content of mineral-based feedstuffs was variable.
We were only able to analyze a few feedstuffs for chloride (Table 4). We selected ingredients that might be used in swine nursery rations because past research on electrolyte balance has generally focused in this size of pig (Haydon and West, 1990; Patience, 1990; Patience and Chaplin, 1997) but may also affect sow and litter performance (Dove and Haydon, 1994) and water consumption (Shaw et al., 2006). Our values were comparable to the NRC (1998) for corn (0.08 vs. 0.05% Cl, respectively), distiller grains with solubles (0.19 vs. 0.20% Cl, respectively), dried whey (1.77 vs. 1.40% Cl, respectively), meat and bone meal (0.90 vs. 0.69% Cl, respectively), soybean meal (0.03 vs. 0.05% Cl, respectively), and wheat middlings (0.08 vs. 0.04% Cl, respectively). Our values were greater for fish meal (0.90 vs. 0.55% Cl, respectively) but lower for plasma protein (0.72 vs. 1.40% Cl, respectively). We did not detect chloride in our dicalcium or monocalcium phosphate sources, and like the NRC (1998), a very low level was found in limestone. Sodium chloride was analyzed to contain 35.9% Na and 60.8% Cl, similar to the 39.5% Na and 59.0% Cl reported in the NRC (1998).
Data from this study show the relative abundance of minerals in a wide array of ingredients and that dietary manipulation of mineral intake can be achieved by altering feedstuff inclusion rates. In addition, data from this feed ingredient survey indicate that care is needed to determine the actual concentration of minerals, due to potential wide variation in mineral content due to ingredient source or processing. These results provide researchers and nutritionists relative concentrations of the mineral composition of feedstuffs commonly used in the livestock diets.
|
Footnotes
|
|---|
1 Mention of a trade name, proprietary product, or specific equipment does not constitute a guarantee or warranty by the USDA or the University of Minnesota and does not imply approval to the exclusion of other products that may be suitable. The authors gratefully acknowledge the assistance of J. Cook (USDA-ARS, Ames, IA) for laboratory assistance and D. Meek (USDA-ARS, Ames, IA) for statistical assistance. 
2 Corresponding author: brian.kerr{at}ars.usda.gov
Received for publication December 18, 2007.
Accepted for publication April 17, 2008.
|
LITERATURE CITED
|
|---|
Anderson, D. M., and S. C. Stothers. 1978. Effects of saline water high in sulfates, chlorides and nitrates on the performance of young weanling pigs. J. Anim. Sci. 47:900–907.[Abstract/Free Full Text]
Anderson, J. S., D. M. Anderson, and J. M. Murphy. 1994. The effect of water quality on nutrient availability for grower/finisher pigs. Can. J. Anim. Sci. 74:141–148.
Anderson, K. L. 2003. The complex world of gastrointestinal bacteria. Can. J. Anim. Sci. 83:409–427.
Apgar, G., K. Griswold, B. Jacobson, and J. Salzar. 2002. Effects of elevated and reduced dietary N and S concentration upon growth and concentration of odor causing components in waste of finishing pigs. J. Anim. Sci. 80(Suppl. 1):395. (Abstr.)
ASAE. 2003. Manure Production and Characteristics. ASAE Data D384.1. Am. Soc. Agric. Eng., St. Joseph, MI.
Attene-Ramos, M. S., E. D. Wagner, M. J. Plewa, and H. R. Gaskins. 2006. Evidence that hydrogen sulfide is a genotoxic agent. Mol. Cancer Res. 4:9–14.[Abstract/Free Full Text]
Baker, D. H. 1977. Sulfur in Nonruminant Nutrition. Natl. Feed Ingredients Assoc., West Des Moines, IA.
Baker, D. H. 2006. Comparative species utilization and toxicity of sulfur amino acids. J. Nutr. 136:1670S–1675S.[Abstract/Free Full Text]
Burrin, D. G., and B. Stoll. 2007. Emerging aspects of gut sulfur amino acid metabolism. Curr. Opin. Clin. Nutr. Metab. Care 10:63–68.[Medline]
Deplancke, B., K. R. Hristova, H. A. Oakley, V. J. McCracken, R. Aminov, R. I. Mackie, and H. R. Gaskins. 2000. Molecular ecological analysis of the succession and diversity of sulfate-reducing bacteria in the mouse gastrointestinal tract. Appl. Environ. Microbiol. 66:2166–2174.[Abstract/Free Full Text]
Dove, C. R., and K. D. Haydon. 1994. The effect of various diet nutrient densities and electrolyte balances on sow and litter performance during two seasons of the year. J. Anim. Sci. 72:1101–1106.[Abstract]
Du Vigneaud, V. 1952. A Trail of Research in Sulfur Chemistry and Metabolism and Related Fields. Cornell Univ. Press, Ithaca, NY.
Dziewiatkowski, D. D. 1962. Sulfur. Pages 175–220 in Mineral Metabolism: An Advance Treatise, Volume II, The Elements Part B. C. L. Comar and F. Bronner, ed. Academic Press, New York, NY.
Florin, T., G. Neale, G. R. Gibson, J. J. Christl, and J. H. Cummings. 1991. Metabolism of dietary sulphate: Absorption and excretion in humans. Gut 32:766–773.[Abstract/Free Full Text]
Gibson, G. R. 1990. Physiology and ecology of the sulphate-reducing bacteria. J. Appl. Bacteriol. 69:769–797.[Medline]
Gomez, G. G., R. S. Sandler, and D. Seal Jr. 1995. High levels of inorganic sulfate cause diarrhea in neonatal piglets. J. Nutr. 125:2325–2332.[Abstract/Free Full Text]
Haydon, K. D., and J. W. West. 1990. Effect of dietary electrolyte balance on nutrient digestibility determined at the end of the small intestine and over the total digestive tract in growing pigs. J. Anim. Sci. 68:3687–3693.[Abstract]
Lewis, H. B. 1924. Sulfur metabolism. Physiol. Rev. 4:394–423.[Free Full Text]
Lin, L. I. 1989. A concordance correlation coefficient to evaluate reproducibility. Biometrics 45:255–268.[CrossRef][Medline]
Maenz, D. D., J. F. Patience, and M. S. Wolynetz. 1994. The influence of the mineral level in drinking water and the thermal environment on the performance and intestinal fluid flux of newly-weaned pigs. J. Anim. Sci. 72:300–308.[Abstract]
McLeese, J. M., J. F. Patience, M. S. Wolynetz, and G. I. Christison. 1991. Evaluation of the quality of ground water supplies used on Saskatchewan swine farms. Can. J. Anim. Sci. 71:191–203.
Meek, D. W. 2007. Two macros for producing graphs to assess agreement between two variables, paper D6. Pages 28–30 in Data Visualization and Graphics Section. A. Katschke, ed. 2007 Midwest SAS Users Group Annu. Conf. Proc., Des Moines, IA. SAS Inst., Cary, NC.
Miller, R. O. 1998. Chapter 6. Nitricperchloric acid wet digestion in an open vessel. Pages 57–61 in Handbook of Reference Methods for Plant Analysis. Y. P. Kalra, ed. CRC Press, Boca Raton, LA.
NRC. 1998. Nutrient Requirements of Swine. 10th rev. ed. Natl. Acad. Press, Washington, DC.
Paterson, D. W., R. C. Wahlstrom, G. W. Libal, and O. E. Olson. 1979. Effects of sulfate in water on swine reproduction and young pig performance. J. Anim. Sci. 49:664–667.[Abstract/Free Full Text]
Patience, J. F. 1990. A review of the role of acid-base balance in amino acid nutrition. J. Anim. Sci. 68:398–408.[Abstract]
Patience, J. F., and R. K. Chaplin. 1997. The relationship among dietary undertermined anion, acid-base balance, and nutrient metabolism in swine. J. Anim. Sci. 75:2445–2452.[Abstract/Free Full Text]
Pedersen, C., M. G. Boersma, and H. H. Stein. 2007. Digestibility of energy and phosphorus in ten samples of distillers dried grains with solubles fed to growing pigs. J. Anim. Sci. 85:1168–1176.[Abstract/Free Full Text]
Shaw, M. I., A. D. Beaulieu, and J. F. Patience. 2006. Effect of diet composition on water consumption in growing pigs. J. Anim. Sci. 84:3123–3132.[Abstract/Free Full Text]
Shoveller, A. K., B. Stoll, R. O. Ball, and D. G. Burrin. 2005. Nutritional and functional importance of intestinal sulfur amino acid metabolism. J. Nutr. 135:1609–1612.[Abstract/Free Full Text]
Spiehs, M. J., M. H. Whitney, and G. C. Shurson. 2002. Nutrient database for distiller’s dried grains with solubles produced from new ethanol plants in Minnesota and South Dakota. J. Anim. Sci. 80:2639–2645.[Abstract/Free Full Text]
Stein, H. H., M. L. Gibson, C. Pedersen, and M. G. Boersma. 2006. Amino acid and energy digestibility in ten samples of distillers dried grain with solubles fed to growing pigs. J. Anim. Sci. 84:853–860.[Abstract/Free Full Text]
Stipanuk, M. H. 2004. Supfur amino acid metabolism: Pathways for production and removal of homocysteine and cysteine. Annu. Rev. Nutr. 24:539–577.[CrossRef][Medline]
Sutton, A. L., J. A. Paterson, O. L. Adeola, B. A. Richert, D. T. Kelly, A. J. Heber, K. B. Kephart, R. Mumma, and E. Bodus. 1998. Reducing sulfur-containing odors through diet manipulation. Pages 125–130 in Proceedings Volume 1: Oral Presentations, Animal Production Systems and the Environment, Iowa State University. Iowa State University, Ames.
Veenhuizen, M. F., G. C. Shurson, and E. M. Kohler. 1992. Effect of concentration and source of sulfate on nursery pig performance and health. J. Am. Vet. Med. Assoc. 201:1203–1208.[Medline]
Whitney, M. H., R. Nicolai, and G. C. Shruson. 1999. Effects of feeding low sulfur starter diets on growth performance of early weaned pigs and odor, hydrogen sulfide, and ammonia emissions in nursery rooms. J. Anim. Sci. 77(Suppl. 1):70. (Abstr.)[Abstract/Free Full Text]
Top
Abstract
INTRODUCTION
