- Email: [email protected]
Relationships Between Crude Protein and Determination of Nondispersible Lignin1
Relationships Between Crude Protein and Determination of Nondispersible Lignin1 JAMES B. REEVES, III Nutrient Conservation and Metabolism Laboratory, Livestock and Poultry Sciences Institute, USDA, ARS, Beltsville, MD 20705
ABSTRACT The objective of this study was to examine the relationships between crude proetin (CP), acid detergent fiber (ADF), ADF CP, and acid detergent lignin CP and measures of lignin, neutral detergent fiber (NDF), and digestibility. Specifically, the study examined whether a relationship existed between these measures and the CP content of feed or the residual CP found in ADF or acid detergent lignin. Sixty-seven samples (15 alfalfa hay, 16 tall fescue hay, 15 orchardgrass hay, 10 vegetative corn plants, and 11 vegetative wheat plants) were collected at different maturities for study. Samples were analyzed for CP, NDF, digestibility, and lignin (acid detergent, permanganate, chlorite, and acetyl bromide). In addition, batches of ADF and acid detergent lignin were prepared for CP determination. Results of correlations between the feed CP, ADF CP, or acid detergent lignin CP and measures of lignin, NDF, or digestibility were found to be highly dependent on feedstuffs. In some instances, such as the correlations between acid detergent lignin CP and lignin, NDF, or digestibility, different feedstuffs produced equally close, but opposite (positive versus negative) correlations for the same relationship. Consideration of the type of material involved, maturity, or the assay was of little help in explaining the results. Overall, CP contamination was a significant problem for the determination of ADF and lignin concentrations. ( Key words: lignin, crude protein, forage) Abbreviation key: ABL-ADF = acetyl bromide lignins, ABL-OR = acetyl bromide lignins extracted with hot water and organic solvent fiber, ADL = acid detergent lignin; AL = alfalfa, CL-ADF = chlorite lignins, CL-OX = chlorite lignins on ammonium oxalate, CS = corn stover, FS = tall fescue, MIR = mid
Received November 8, 1994. Accepted July 23, 1996. 1Mention of a trade name, proprietary product, or specific equipment does not constitute a guarantee or warranty by the USDA and does not imply its approval to the exclusion of other products that may be suitable. 1997 J Dairy Sci 80:692–699
infrared, OG = orchardgrass, PL-ADF = permanganate lignin on ADF, WS = wheat straw. INTRODUCTION The determination of lignin is used by many researchers in diverse research areas to monitor changes in plant composition or digestibility. At present, several different procedures are used to measure lignin including permanganate oxidation ( 3 ) , acid detergent lignin ( ADL) ( 3 ) , sodium chlorite oxidation ( 2 ) , and acetyl bromide extraction (7, 8). Lignin is an aromatic-based polymer that is produced by plants and is important to their structure and composition (15). Lignin has long been known to influence the digestibility of plant tissues, although the exact nature of its role is still debated. Acid detergent lignin is defined as the residue remaining after extraction of ground plant materials with a boiling solution of acidic acetyltrimethyl-ammonium bromide, followed by extraction with 72% sulfuric acid. Other measures of lignin include the acetyl bromide procedure (7, 8), in which lignin is determined by dissolving the entire sample in a glacial acetic acid and acetyl bromide mixture to determine the UV absorption, and the permanganate ( 3 ) and chlorite ( 2 ) procedures, in which lignin is oxidized and determined as the difference before and after oxidation. The determination of lignin by these various procedures is used in many research areas (e.g., animal feed or environment) to monitor various efforts to increase or gauge the digestibility, disappearance, or quality of plant materials. In earlier work (10), efforts were made to find the best procedure to calibrate near infrared reflectance spectrometers for lignin content. Those efforts indicated that, although near infrared reflectance spectroscopy could be used in place of the various chemical assays for lignin, the results were not as accurate as might be desired. In that same study (10), the results of various lignin methods often did not match or produced very different values for lignin content, as though the methods were measuring completely different substances. Subsequent studies using chemical extraction (11), pyrolysis-gas chromatography mass spectrometry
692
693
CRUDE PROTEIN LIGNIN
(13), mid infrared ( MIR) reflectance spectroscopy (12), and 13C-labeled solid-state nuclear magnetic resonance spectroscopy ( 1 4 ) have shown that the various procedures either do not extract all of the lignin or else determine nonlignin components as lignin. For ADL determinations, studies have indicated that protein and possibly carbohydrates could be included in the determination, but all of the lignin originally present in the ADF was apparently retained. Contamination of ADL by nonlignin fractions has also been reported by others (4, 5), as has the possible loss of lignin in the production of the ADF (4, 6), a possibility that was not examined in the studies in question (11, 12, 13, 14). The MIR examination of a few ADL samples showed spectra without the details that are generally expected for plant materials and their extracted residues (12). The many MIR peaks, representing carbohydrates, lignin, and proteins, were largely blended together into nondescript, rather featureless spectra that were more reminiscent of near infrared than MIR spectra. Results (11, 12, 13, 14) for the other measures of lignin indicated that similar problems existed, and protein contamination occurred often. In addition, the permanganate procedure showed problems with carbohydrates (11), and the chlorite procedure did not remove all of the lignin (11, 12, 13). However, results indicated that carbohydrate removal was not a problem for the chlorite procedure (11). Finally, results for the acetyl bromide procedure indicated that furfural formation from carbohydrates can be an extremely serious problem (11), which was particularly true when the acetyl bromide procedure was applied to residues from previous procedures (i.e., ADF extracted with chlorite or permanganate). The results of various lignin assays that were obtained with the acetyl bromide procedure for residues indicated furfural formation (11). These results are extremely important in regard to another issue with lignin assays. All of the procedures discussed determine lignin in some form of extracted fiber and not in the original feedstuff. Thus, if lignin were lost in the initial fiber preparation, the subsequent lignin values would be incorrect. Lowry et al. ( 6 ) reported that up to 50% of the lignin in tropical grasses might be lost before the lignin determination was ever made. Unfortunately, the estimates of the dispersible lignin were largely based on UV absorption of the ADF solution resulting from the ADF step or were based on the use of acetyl bromide on recovered residues. In both cases, the possible generation of furfurals and their subsequent interference, because they also absorb in the same wavelength range as lignin, make evaluation of the results difficult. References to digestible lignin, as indicated by soluble lignin and
carbohydrate complexes, are also used to support the concept of a large fraction of acid detergent dispersible lignin. However, work by Windham et al. ( 1 7 ) showed that such lignin and carbohydrate complexes contained no lignin at all, but rather were protein and carbohydrate complexes. Finally, work by Hatfield et al. ( 4 ) on Klason lignin and ADL support the work by Lowry et al. ( 6 ) . Those researchers suggested that, for ADL, a significant portion of the lignin may be dispersed by the acid detergent solution. Hatfield et al. ( 4 ) reported significant differences in the lignin contents of lucerne, cocksfoot, and switchgrass as determined by the Klason and ADL procedures. These differences could not be accounted for by carbohydrate or protein contamination as determined by pyrolytic and wet chemical analyses and, thus, appeared to be due to dispersion of lignin by acid detergent. However, Hatfield et al. ( 4 ) also showed that a much greater percentage of the original forage protein was retained in the Klason lignin than in the ADL (approximately 2 to 6.5 times as much). Because lignin is a known, defined material (i.e., an aromatic polymer) and not a composite of protein, carbohydrates, and lignin, it is important to understand exactly what each lignin procedure does and does not measure. An accurate procedure that truly measures lignin and only lignin is important not only for developing better calibrations from near infrared reflectance spectrophotometry, but also for obtaining a better understanding of the role of forage composition in determining digestibility. Although further investigations should determine the extent to which acid detergent dispersible lignin exists and its composition, it is still important to determine accurately the nondispersible lignin if a two-step procedure for lignin determination (dispersed lignin in the acid detergent solution by UV absorption and lignin in the ADF) is to be used, as suggested by Lowry et al. ( 6 ) . Therefore, the objective of this work was to examine the relationships between CP, ADF CP, and ADL CP and measures of nondispersible lignin, NDF, and digestibility for several lignin procedures currently used for forage analysis. Specifically, the study was conducted to determine whether a relationship exists between the CP content of the starting material or the residual CP found in ADF or ADL and various measures of nondispersible lignin, NDF, or digestibility. MATERIALS AND METHODS Samples All forages were collected in the field as grab samples of fresh plant; the samples were composited and dried immediately at 60°C. After drying, samples Journal of Dairy Science Vol. 80, No. 4, 1997
694
REEVES, III
were ground to a size of 20 mesh (Christie mill; Christie Norris Ltd., Chelmsford, England). Reproductive portions of corn and wheat samples were removed prior to drying in order to provide material that corresponded with corn stover ( CS; n = 10) and wheat straw ( WS; n = 11). Alfalfa ( AL; n = 15) and orchardgrass ( OG; n = 15) were cut several times during the season for hay or silage, and samples were taken from the regrowth. The AL was harvested on May 21 (10% bloom), June 23 (20% bloom), and August 2 (40% bloom). The OG was harvested at the stem elongation stage on May 10, July 21, and August 25. The tall fescue ( FS; n = 16, not free of endophytes) was cut only once and represented two growth periods: growth until June 3 (during late anthesis) and regrowth between June 3 and December 22. The CS and WS were taken every week or two from fields that were harvested once at the end of the growing season for grain (September 30 and July 9, respectively) ( 9 ) . Fiber and Lignin Production All 67 samples were analyzed for CP, ADF, and ADL in triplicate (1, 3). In addition, large batches of ADF and subsequently ADL were prepared by exact scaling of the procedures of Goering and Van Soest ( 3 ) for CP determination. Neutral detergent fiber and permanganate lignin on ADF ( PL) were similarly determined according to the method of Goering and Van Soest ( 3 ) , but without the addition of decahydronapthalene or sodium sulfite. Chlorite lignins were determined on both ammonium oxalate fiber ( CL-OX) by the procedure of Collings et al. ( 2 ) with modifications to the filtering procedure ( 1 0 ) and on ADF ( CL-ADF) . For CLADF, 1.0-g samples were each extracted with 200 ml of 0.5% ammonium oxalate solution for 2 h and dried at 100°C overnight. Chlorite lignin in these fibers was then determined by treating the fiber and crucible with sodium chlorite in acetic acid (1.0%) and refiltering the residue through the same crucible (10). Acetyl bromide lignins were determined according to the method of Morrison (7,8) using fibers extracted with hot water and organic solvent ( ABLOR) and by using ADF ( ABL-ADF) . For ABL-OR, hot water (30 min at 60°C), ethanol, acetone, and diethyl ether were used to extract the samples (each fiber sample was washed with 200 ml of each solvent). The fiber was dried overnight at 100°C and scraped from the crucible. A fraction was then weighed out for determination of acetyl bromide ligJournal of Dairy Science Vol. 80, No. 4, 1997
nin, and the rest was used for ash determination. Computations were based on optical density units per gram of OM, rather than the percentage of DM as for the other lignins. All chemical assay results are the means of triplicate determinations that were calculated on an OM basis. Forage Digestibility All in vitro digestibilities were carried out according to the methods of Goering and Van Soest ( 3 ) and are presented as the means of duplicate analyses. Incubations were for 48 h using rumen fluid obtained from a fistulated steer fed an OG diet. Statistics All data computations and statistical summaries were performed using SAS software (16). Correlations were based on the Pearson correlation matrix. RESULTS AND DISCUSSION Table 1 contains the minima, maxima, means, and standard deviations for the chemical results obtained for all 67 samples examined and for the same samples grouped by feedstuff. All feedstuffs contained a wide range of sample maturities as a result of sample collection from the beginning to the end of the growing season. The wide range in sample maturities is shown by the wide distribution of assay results and the magnitude of the standard deviations. The coefficients of variation ranged from ±7 to ±70%. Thus, the available samples represented a wide range of sample types and compositions upon which to draw summary conclusions. In particular, all sample groupings contained some samples with a relatively high initial CP content. Thus, even the CS and WS groups contained samples with CP contents that would be considered good for a feed (maximum CP content for CS, 27.0; maximum CP content for WS, 16.8). Although CS and WS would not normally be found with such high CP values, the early cuttings provided such samples. These types of samples might be used in studies to determine how lignin content changes as plants mature and are important in understanding how measured lignin content may be influenced by variations in CP content. Examination of the values for ADF CP indicated wide variation in the degree to which CP was incorporated into ADF. Although the highest CP values found for AL, CS, FS, and OG were quite similar (31.8, 27.0, 24.8, and 27.7%, respectively), the maxi-
695
CRUDE PROTEIN LIGNIN
mum ADF CP values varied considerably (3.3, 0.75, 1.4, and 2.8%, respectively). The maximum CP values for CS and OG were virtually identical at 27.0 and 27.7%, but the ADF CP values varied almost fourfold. Thus, the incorporation of CP into ADF was
dependent on the sample in question and was not simply dependent on the initial CP content of the starting material. A quick examination of the data for the six lignin procedures studied (ABL-ADF, ABL-OR, ADL, CL-
TABLE 1. Chemical composition (percentage of DM) for forages and by-products. Variable1 All forages and by-products ( n = 67) ABL-ADF ABL-OR ADF ADL CL-ADF PL-ADF CL-OX NDF CWD DMD CP ADF CP ADL CP Alfalfa hays only ( n = 15) ABL-ADF ABL-OR ADF ADL CL-ADF PL-ADF CL-OX NDF CWD DMD CP ADF CP ADL CP Vegetative corn plants only ( n = 10) ABL-ADF ABL-OR ADF ADL CL-ADF PL-ADF CL-OX NDF CWD DMD CP ADF CP ADL CP
Minimum Maximum X
SD
RSD2
0.8 3.0 17.3 0.84 0.7 2.8 4.0 25.8 37.9 55.4 4.4 0.5 1.5
7.1 9.6 53.0 7.8 6.0 14.0 8.2 83.0 87.8 94.1 31.8 3.3 18.0
3.0 6.6 33.1 3.8 2.7 7.5 5.9 59.1 60.8 76.9 15.4 1.2 6.9
1.4 1.3 7.4 1.5 1.0 2.5 1.0 14.5 11.9 9.2 7.7 0.8 3.3
0.47 0.20 0.22 0.40 0.37 0.33 0.17 0.25 0.20 0.12 0.50 0.67 0.48
1.2 3.0 17.3 2.4 1.5 5.6 5.3 25.8 37.9 71.9 17.7 1.9 7.3
4.8 6.5 40.4 7.3 3.4 14.0 7.0 50.5 64.3 90.8 31.8 3.3 12.2
2.9 5.1 27.3 5.1 2.5 9.4 5.9 36.7 49.7 81.3 24.3 2.4 10.7
1.0 1.2 6.1 1.3 0.6 2.2 0.5 7.2 7.1 5.5 4.2 0.4 1.5
0.34 0.24 0.22 0.25 0.24 0.23 0.09 0.20 0.14 0.07 0.17 0.17 0.14
1.2 5.4 25.4 1.8 1.0 4.3 4.5 53.0 53.9 62.3 5.1 0.5 1.5
5.2 8.1 45.4 4.5 4.8 10.0 8.2 81.8 82.8 90.9 27.0 0.7 9.3
3.8 7.2 37.8 3.3 3.2 7.0 6.4 69.0 64.4 74.7 10.4 0.6 3.4
1.4 0.8 6.3 0.9 1.2 2.0 1.1 8.7 9.5 9.1 7.3 0.1 2.3
0.37 0.11 0.17 0.27 0.38 0.29 0.17 0.13 0.15 0.12 0.70 0.17 0.68
Variable Tall fescue hays only ( n = 16) ABL-ADF ABL-OR ADF ADL CL-ADF PL-ADF CL-OX NDF CWD DMD CP ADF CP ADL CP Orchardgrass hays only ( n = 15) ABL-ADF ABL-OR ADF ADL CL-ADF PL-ADF CL-OX NDF CWD DMD CP ADF CP ADL CP Vegetative wheat plants only ( n = 11) ABL-ADF ABL-OR ADF ADL CL-ADF PL-ADF CL-OX NDF CWD DMD CP ADF CP ADL CP
Minimum Maximum X
SD
RSD
0.8 5.5 21.3 0.8 0.7 2.8 4.9 48.2 5.8 63.9 7.5 0.6 3.8
3.1 8.1 38.4 4.4 3.2 9.1 8.1 73.6 87.8 94.1 24.8 1.4 9.7
2.5 7.3 33.7 3.1 2.5 6.6 6.5 66.0 59.7 72.7 11.6 1.0 7.3
0.6 0.6 4.4 1.1 0.7 1.5 1.0 6.9 11.0 8.9 4.6 0.3 2.0
0.24 0.08 0.13 0.35 0.28 0.23 0.15 0.10 0.18 0.12 0.40 0.30 0.27
1.0 5.2 22.9 2.3 1.1 3.5 4.1 54.4 52.0 64.6 13.4 0.7 4.2
3.3 8.1 39.0 5.6 4.1 9.7 6.6 73.8 84.3 91.4 27.7 2.8 18.0
2.0 6.6 30.3 3.2 2.0 6.0 5.6 62.6 69.9 80.9 19.4 1.1 6.8
0.6 0.9 4.3 0.8 0.7 1.6 0.6 5.8 7.9 6.5 3.6 0.5 3.2
0.30 0.14 0.14 0.24 0.35 0.27 0.11 0.09 0.11 0.08 0.19 0.45 0.47
1.5 5.7 27.7 2.4 1.5 4.5 4.0 52.1 46.3 55.4 4.4 0.5 2.6
7.1 9.6 53.0 7.8 6.0 13.9 6.5 83.0 84.4 91.8 16.8 1.3 7.6
4.4 7.3 39.6 4.7 3.4 8.9 4.8 66.2 61.8 73.4 7.9 0.8 4.5
2.0 1.4 9.5 1.9 1.6 3.5 0.9 11.5 13.6 13.0 3.9 0.3 1.7
0.45 0.19 0.24 0.40 0.47 0.39 0.19 0.17 0.22 0.18 0.49 o.38 0.38
1ADF CP = CP content of ADF (percentage of ADF DM), ADL CP = CP content of ADL (percentage of ADL DM), ABL-ADF = acetyl bromide lignins on ADF (based on optical density), ABL-OR = acetyl bromide lignins extracted from fiber on hot water and organic solvent (based on optical density), ADL = acid detergent lignin (72% sulfuric acid lignin on ADF), CL-ADF = chlorite lignins on ADF (percentage of DM), PL-ADF = permanganate lignins on ADF (percentage of DM), CL-OX = chlorite lignins on ammonium oxalate (percentage of DM), CWD = cell-wall digestibility or NDF digestibility (percentage of DM), and DMD = DM digestibility (percentage of DM). 2Relative standard deviation.
Journal of Dairy Science Vol. 80, No. 4, 1997
696
REEVES, III
ADF, PL-ADF, and CL-OX) using data from all 67 samples showed the wide range of results obtained. For example, for both the acetyl bromide procedures (ABL-ADF and ABL-OR) and for the chlorite procedures (CL-ADF and CL-OX), the choice of fiber from which to extract the lignin greatly influenced the mean lignin values obtained. In each case, the use of ADF resulted in mean lignin values that were approximately half of those obtained using the fiber procedures outlined in the original papers [ABL-OR (7, 8 ) and CL-OX (2)]. The differences, based on the initial values of the starting fiber, might be due to the presence of acid detergent dispersible lignin as suggested by Lowry et al. ( 6 ) , to the enhanced extraction of protein or carbohydrates by the harsher acid detergent extraction, or both. Quick computation of ratios of the maximum and minimum lignin values shows the same sort of problems. For example, for CL-OX, the ratio was approximately 2:1, but for CL-ADF, the ratio was about 8.5:1. Similar results occurred with the lignins based on acetyl bromide. Even comparing lignin procedures based only on ADF (i.e., ABL-ADF, ADL, CL-ADF, and PL-ADF, the maximum to minimum lignin ratios were approximately 9:1, 9:1, 9:1, and 5:1, respectively. Similar results can be found by examining the lignin data on an individual feed basis (Table 1). The various feedstuffs show interactions among the different lignin assays; the highest value for PL-ADF was for AL samples (14.0%), followed by WS (13.9%). The highest value for CL-ADF was for WS at 6.0%, and the value for AL was only 3.4%. Similar results were found for low lignin values also. Alfalfa was the lowest only for ABL-OR. If all of these procedures measured lignin, even on some relative basis, one would expect the same feed to be high or low by all of the assays used. The incorporation of ADF CP into ADL would appear to be even more dependent on the sample than the incorporation of CP into ADF. For example, ADF of CS contained about half of the CP content that was found in the WS samples, but ADL of CS at the maximum contained more CP than did ADL of WS (9.3% vs. 7.6%). Results were similar for the other measures of lignin presented (e.g., PL-ADF). These results indicated that no simple relationship existed between sample CP content and the CP content of its ADF or lignin and that correction of lignin for CP content based on initial sample or ADF CP content was not feasible. The results also indicated that the presence of CP can complicate the correct determination of ADL because 1.5 to 18% of the lignin value could be based on CP and not on lignin. Because CP Journal of Dairy Science Vol. 80, No. 4, 1997
measurements are actually based on N content and not protein, alteration of the protein by the ADF procedure or, more likely, by the 72% sulfuric acid that was used in the ADL step, may mean that the problem is more or less serious than the results indicated, depending on the type of effects that the treatments (ADF or ADL) have on the initial protein present. Unfortunately, as stated by Hatfield et al. ( 4 ) , investigation of the ADL residue by methods such as pyrolysis-gas chromatography mass spectrometry does not provide information on the nature of the N present. Future investigations are planned using 15N-labeled nuclear magnetic resonance spectroscopy to investigate the question of the state of N in both ADL and chlorite lignin, because past efforts have led to confusing results (10, 14). In any case, this is a problem because true lignin does not contain any N, but is rather a polymer made by polymerization of propenyl-benzene constituents ( 1 5 ) that was induced by free radicals. Because the same protein would be present in the ADF used for other lignin assays, it is highly possible that the same type of relationships found for CP versus ADL might exist for other lignin assays. Only for the acetyl bromide lignins would protein be of minor concern because of the low extinction coefficient of protein compared with that for lignin. Furthermore, CP, which can survive the ADF extraction procedure, might also escape the extraction by ammonium oxalate and by hot water and organic solvent used in the two other procedures studied. Finally, the CP in ADL might represent some specific protein that was difficult to extract. Although not lignin, the protein may represent some specific fraction of feedstuffs that may relate to other measurements of lignin, fiber or digestibility. The data for NDF, cell wall, and DM digestibilities, presented in Table 1, were used for the correlations presented in Table 2. Table 2 contains the results obtained from correlation analyses of the various chemical results. Results for CP versus ADF or ADL confirm earlier expectations of closer correlations for individual feedstuffs than for all of the samples combined. The results for ADF versus ADF CP show wide variations in feed responses. Although there was a low (–0.30) but significant negative correlation overall (all 67 samples), the correlations were positive and twice the magnitude for FS, OG, and WS (0.65, 0.64, and 0.81, respectively). Although CS and WS were both byproducts that were high in lignin and low in CP and thus would be expected to show similar relationships between compositional components such as ADF and CP, only WS showed a significant correlation between ADF and ADF CP. The results for CP versus ADF CP and ADL CP also supported conclusions drawn from Table 1.
CRUDE PROTEIN LIGNIN
Although the results on an individual feedstuff basis varied considerably, there was approximately a 0.5 correlation for both relationships for all samples combined. Individually, only FS showed a significant correlation for ADF CP versus CP, and it was negative. Finally, the relationships between ADF, CP, and ADF CP and the CP content of ADL were equally varied and complex. Overall, for AL and CS, significant negative correlations existed between ADF content and ADL CP content; however, for FS, OG, and WS, positive correlations existed. Similarly, for all samples, for AL, and for CS, significant positive correlations existed between CP content of the base material and ADL CP content, but a negative relationship was found for FS; the results were not significant for OG and WS. The final and most consistent relationship was for ADF CP and ADL CP. Except for CS, for which the results were not signifi-
TABLE 2. Results of correlation analyses (correlation coefficients) of ADF, CP content of forages and by-products, and CP content of their ADF and acid detergent lignin (ADL). 1 CP All forages and by-products ( n = 67) ADF –0.81 CP ADF CP Alfalfa hays only ( n = 15) ADF –0.68 CP ADF CP Vegetative corn plants only ( n = 10) ADF –0.93 CP ADF CP Tall fescue hays only ( n = 16) ADF –0.81 CP ADF CP Orchardgrass hays only ( n = 15) ADF –0.62 CP ADF CP Vegetative wheat plants only ( n = 11) ADF –0.83 CP ADF CP 1Results
ADF CP
ADL CP
–0.30 0.54
–0.34 0.53 0.87
NS2 NS
–0.62 0.68 0.59
NS NS
–0.72 0.90 NS
0.65 –0.83
0.51 –0.75 0.86
0.64 NS
0.63 NS 0.95
0.81 NS
0.593 NS 0.91
were significant at P < 0.05 unless otherwise noted. > 0.06. 3P < 0.06. 2P
697
cant, a positive and generally close correlation existed between the ADF CP and the ADL CP, indicating that CP remaining in the ADF would, in general, be largely incorporated into, and consequently determined as, ADL. Correlation results involving lignin are found in Table 3. E xamining the results for correlations of CP with the various measures of lignin showed that, except for the determination of CL-OX in AL samples and ADL for all samples, a negative relationship existed between CP content and lignin, regardless of the method of lignin determination. Such negative correlations were expected because high quality materials are generally high in protein and low in lignin. The results did, however, indicate differences in these feed-dependent relationships. For example, the correlations between PL-ADF and CP were much more negative for single feedstuffs than for all five feedstuffs [r = –0.23 (all samples)] versus –0.51, –0.78, –0.81, –0.67, and –0.85 for AL, CS, FS, OG, and WS, respectively. Variations in the relative strength of the correlations between CP and various lignin measures, similar to the kind of variations shown in Table 1, may also be found. Thus, for ABL-ADF, the highest correlation was for CS ( r = –0.97), but for ABL-OR the highest correlation was for FS ( r = –0.83), and these are both acetyl bromide lignins, only based on different starting fibers. Overall, AL appeared to have the greatest range in correlation values between CP and lignin, from +0.02 for CL-OX to –0.71 for ABL-ADF; CS, FS, and OG were the most consistent. The highest general correlations appeared to be for CS and WS, which were both low quality by-products that were low in CP. However, the low correlation ( r = –0.33) for CL-OX for WS and the high correlation for CS ( r = –0.89) should be noted. The correlations of ADF CP with lignin found in Table 3 were extremely variable. The results for all 67 samples varied from +0.55 for ADL to –0.48 for ABL-OR. These results indicate that, overall, the CP content of ADF was incorporated into ADL but not into ABL-OR. Even grouping the data for the individual feedstuffs was difficult. Wheat straw and CS are generally considered to be of similar nature (i.e., low quality, low protein, and high fiber); however, the correlations between the various lignin procedures and ADF CP content were vastly different. The WS samples produced correlations more similar to those for FS and OG. Also, although for CS, FS, OG, and WS the correlations were all positive, indicating that the incorporation of ADF CP into lignin was a Journal of Dairy Science Vol. 80, No. 4, 1997
698
REEVES, III
problem, regardless of the procedure used, for AL, the results were both positive and negative. The correlations between the CP found in ADL and that found in other lignin procedures are difficult to explain. For FS, OG, and WS, the more CP that was present in ADL, the more lignin was determined by other procedures, as indicated by the positive correlations. However, for CS, the opposite was found. The positive correlation is easy to explain: the CP retained from the starting ADF in the ADL procedure was extracted and thus was determined as lignin by the other lignin procedures. A zero correlation can also be easily explained because CP found in ADF might have been left by the other lignin procedures and, thus, have no relationship to the lignin values found.
But negative correlations, such as those found for all the CS and most of the AL lignin determinations (except CL-OX), are not easy to explain. Because the various samples were collected at different stages of growth, the negative correlations might have been due to the low lignin and high CP contents of early maturity samples. The same maturity differences could be true for the other feedstuffs, especially for the WS samples, which were cut periodically, as was the CS, from a field of continuously maturing plants (as opposed to the regrowth found in fields of AL, OG, and FS, which were cut periodically for hay). Overall, these results indicate that, although relationships existed between ADL CP content and lignin, the relationships were very complex
TABLE 3. Correlation results by forage group for CP, ADF CP, and acid detergent lignin (ADL) CP with various measures of lignin, NDF, and digestibility. Variable1 All forages and by-products ( n = 67) ABL-ADF ABL-OR ADL CL-ADF PL-ADF CL-OX NDF CWD DMD Alfalfa hays only ( n = 15) ABL-ADF ABL-OR ADL CL-ADF PL-ADF CL-OX NDF CWD DMD Vegetative corn plants only ( n = 10) ABL-ADF ABL-OR ADL CL-ADF PL-ADF CL-OX NDF CWD DMD
CP
ADF CP
ADL CP
–0.63* –0.77* NS –0.61* –0.23 –0.25*
–0.06 –0.48* 0.55* 0.07 0.45* 0.21† –0.65* –0.59* 0.05
–0.23 –0.37* 0.35* –0.02 0.26* 0.24* –0.50* –0.45* 0.02
–0.71* –0.60* –0.78* –0.23 –0.51* 0.02
–0.22 –0.18 NS 0.067 –0.084 0.64* –0.16 –0.0045 0.13
–0.70* –0.56* –0.54* –0.24 –0.62* 0.47† –0.66* 0.29 0.58*
–0.97* –0.76* –0.92* –0.93* –0.78* –0.89*
0.08 0.25 NS 0.095 0.39 0.23 0.31 –0.08 –0.22
–0.78* –0.71* –0.69* –0.72* –0.48 –0.66* –0.64* 0.79* 0.69*
Variable Tall fescue hays only ( n = 16) ABL-ADF ABL-OR ADL CL-ADF PL-ADF CL-OX NDF CWD DMD Orchardgrass hays only ( n = 15) ABL-ADF ABL-OR ADL CL-ADF PL-ADF CL-OX NDF CWD DMD Vegetative wheat plants only ( n = 11) ABL-ADF ABL-OR ADL CL-ADF PL-ADF CL-OX NDF CWD DMD
CP
ADF CP
ADL CP
–0.79** –0.83** –0.84** –0.92** –0.81** –0.70**
0.54 0.59** 0.93** 0.86** 0.80** 0.83** 0.71** –0.89** –0.88**
0.42† 0.56 0.74** 0.75** 0.60** 0.71** 0.66** –0.76** –0.77**
–0.66* –0.47 –0.62* –0.57* –0.67* –0.47
0.55* 0.48 0.95* 0.91* 0.75* 0.41† 0.66* –0.71* –0.77*
0.53* 0.50 0.89* 0.86* 0.66* 0.37 0.62* –0.67* –0.74*
–0.90* –0.78* –0.82* –0.81* –0.85* –0.33
0.75* 0.76* 0.81* 0.86* 0.80* 0.90* 0.80* –0.60* –0.72*
0.49 0.59* 0.61* 0.65* 0.56 0.93* 0.59* –0.33 –0.49
1ADF CP = CP content of ADF (percentage of ADF DM), ADL CP = CP content of ADL (percentage of ADL DM), ABL-ADF = acetyl bromide lignins on ADF (based on optical density), ABL-OR = acetyl bromide lignins extracted from fiber on hot water and organic solvent (based on optical density), ADL = acid detergent lignin (72% sulfuric acid lignin on ADF), CL-ADF = chlorite lignins on ADF (percentage of DM), PL-ADF = permanganate lignins on ADF (percentage of DM), CL-OX = chlorite lignins on ammonium oxalate (percentage of DM), CWD = cell-wall digestibility or NDF digestibility (percentage of DM), and DMD = DM digestibility (percentage of DM). †P < 0.10. *P < 0.05. **P < 0.01.
Journal of Dairy Science Vol. 80, No. 4, 1997
699
CRUDE PROTEIN LIGNIN
and specific to the material in question, thus making generalizations extremely risky. Finally, consideration of the results for correlations of ADF CP and ADL CP with NDF and digestibility showed that, in general, the higher the ADF CP, the lower the digestibility of the samples (particularly cell wall digestibility) and the higher the NDF content. However, as for the lignin results, the findings were dependent on the sample. Thus, CS behaved similar to AL and opposite to the other samples. CONCLUSIONS Results of correlations among the CP content of feedstuffs, their ADF or ADL, and measures of lignin, NDF, or digestibility were found to depend greatly on the samples in question. In some instances, such as the correlation of ADL CP with lignin, NDF, or digestibility, different feedstuffs produced equally close, but opposite (positive vs. negative) correlations for the same relationships. Consideration of the type of material involved (high quality hays vs. low quality by-products), stage of maturity (regrowth vs. continual growth), or the particular assay involved was of no help in producing generalizations or in explaining the results. Because of the wide variation in the relationships between these variables, it would appear to be impossible to develop simple equations to correct lignin measures or ADF for CP content without the need for more chemical determinations (i.e, actual measurement of ADF CP or ADL CP). Because of the small amount of residues obtained from lignin determinations, such determinations are impractical on a routine basis; therefore, it would be best to modify the fiber procedure (ADF or other) in order to eliminate the incorporation of CP into ADF and subsequently into lignin, as well as the possible dispersion of lignin into the extracting solvent (4, 6). Efforts using various protease enzymes as a pretreatment have not proved to be successful (data not shown). In conclusion, although often relationships appear to exist between the values obtained by various measures of lignin and CP in feedstuffs, the exact relationships are extremely dependent on samples,
and more research is needed to determine their exact nature and causes. REFERENCES 1 Association of Official Analytical Chemists. 1984. Official Methods of Analysis. 14th ed. AOAC, Arlington, VA. 2 Collings, G. F., M. T. Yokoyama, and W. G. Bergen. 1978. Lignin as determined by oxidation with sodium chlorite and a comparison with permanganate lignin. J. Dairy Sci. 61:1156. 3 Goering, H. K., and P. J. Van Soest. 1970. Forage Fiber Analyses (Apparatus, Reagents, Procedures, and Some Applications). Agric. Handbook No. 379. ARS-USDA, Washington, DC. 4 Hatfield, R. D., H. G. Jung, J. Ralph, D. R. Buxton, and P. J. Weimer. 1994. A comparison of the insoluble residues produced by the Klason lignin and acid detergent lignin procedures. J. Sci. Food Agric. 65:51. 5 Kirk, T. K., and J. R. Obst. 1988. Lignin determination. Page 96 in Methods in Enzymology. Vol 161. W. A. Wood and S. T. Kellogg, ed. Academic Press, Inc., New York, NY. 6 Lowry, J. B., L. L. Conlan, A. C. Schlink, and C. S. McSweeney. 1994. Acid detergent dispersible lignin in tropical grasses. J. Sci. Food Agric. 65:41. 7 Morrison, I. M. 1972. Improvements in the acetyl bromide technique to determine lignin and digestibility and its application to legumes. J. Sci. Food Agric. 23:1463. 8 Morrison, I. M. 1972. A semi-micro method for the determination of lignin and its use in predicting the digestibility of forage crops. J. Sci. Food Agric. 23:455. 9 Reeves, J. B., III. 1987. Lignin and fiber compositional changes in forages over a growing season and their effects on in vitro digestibility. J. Dairy Sci. 70:1583. 10 Reeves, J. B., III. 1988. Chemical assays for fiber, lignin, and lignin components: interrelationships and near infrared reflectance spectroscopic analysis results. J. Dairy Sci. 71:2976. 11 Reeves, J. B., III. 1993. Chemical studies on the composition of fiber fractions and lignin determination residues. J. Dairy Sci. 76:120. 12 Reeves, J. B., III. 1993. Infrared spectroscopic studies on forage and by-product fibre fractions and lignin determination residues. J. Vib. Spectrosc. 5:303. 13 Reeves, J. B., III, and G. C. Galletti. 1993. Use of pyrolysis-gas chromatography/mass spectrometry in the study of lignin assays. J. Anal. Appl. Pyrolysis 24:243. 14 Reeves, J. B., III, and W. F. Schmidt. 1994. Solid-state 13C NMR analysis of forage and by-product-derived fiber and lignin residues. Resolution of some discrepancies among chemical, infrared, and pyrolysis-gas chromatography-mass spectroscopic analyses. J. Agric. Food Chem. 42:1462. 15 Sarkanen, K. V., and C. H. Ludig. 1971. Lignins: Occurrence, Formation, Structure, and Reactions. John Wiley & Sons, New York, NY. 16 SAS User’s Guide: Statistics, Version 5 Edition. 1985. SAS Inst., Inc., Cary, NC. 17 Windham, W. R., D. S. Himmelsbach, H. E. Amos, and J. J. Evans. 1989. Partial characterization of a protein-carbohydrate complex from the rumen of steers fed high-quality forages. J. Agric. Food Chem. 37:912.
Journal of Dairy Science Vol. 80, No. 4, 1997
ABSTRACT The objective of this study was to examine the relationships between crude proetin (CP), acid detergent fiber (ADF), ADF CP, and acid detergent lignin CP and measures of lignin, neutral detergent fiber (NDF), and digestibility. Specifically, the study examined whether a relationship existed between these measures and the CP content of feed or the residual CP found in ADF or acid detergent lignin. Sixty-seven samples (15 alfalfa hay, 16 tall fescue hay, 15 orchardgrass hay, 10 vegetative corn plants, and 11 vegetative wheat plants) were collected at different maturities for study. Samples were analyzed for CP, NDF, digestibility, and lignin (acid detergent, permanganate, chlorite, and acetyl bromide). In addition, batches of ADF and acid detergent lignin were prepared for CP determination. Results of correlations between the feed CP, ADF CP, or acid detergent lignin CP and measures of lignin, NDF, or digestibility were found to be highly dependent on feedstuffs. In some instances, such as the correlations between acid detergent lignin CP and lignin, NDF, or digestibility, different feedstuffs produced equally close, but opposite (positive versus negative) correlations for the same relationship. Consideration of the type of material involved, maturity, or the assay was of little help in explaining the results. Overall, CP contamination was a significant problem for the determination of ADF and lignin concentrations. ( Key words: lignin, crude protein, forage) Abbreviation key: ABL-ADF = acetyl bromide lignins, ABL-OR = acetyl bromide lignins extracted with hot water and organic solvent fiber, ADL = acid detergent lignin; AL = alfalfa, CL-ADF = chlorite lignins, CL-OX = chlorite lignins on ammonium oxalate, CS = corn stover, FS = tall fescue, MIR = mid
Received November 8, 1994. Accepted July 23, 1996. 1Mention of a trade name, proprietary product, or specific equipment does not constitute a guarantee or warranty by the USDA and does not imply its approval to the exclusion of other products that may be suitable. 1997 J Dairy Sci 80:692–699
infrared, OG = orchardgrass, PL-ADF = permanganate lignin on ADF, WS = wheat straw. INTRODUCTION The determination of lignin is used by many researchers in diverse research areas to monitor changes in plant composition or digestibility. At present, several different procedures are used to measure lignin including permanganate oxidation ( 3 ) , acid detergent lignin ( ADL) ( 3 ) , sodium chlorite oxidation ( 2 ) , and acetyl bromide extraction (7, 8). Lignin is an aromatic-based polymer that is produced by plants and is important to their structure and composition (15). Lignin has long been known to influence the digestibility of plant tissues, although the exact nature of its role is still debated. Acid detergent lignin is defined as the residue remaining after extraction of ground plant materials with a boiling solution of acidic acetyltrimethyl-ammonium bromide, followed by extraction with 72% sulfuric acid. Other measures of lignin include the acetyl bromide procedure (7, 8), in which lignin is determined by dissolving the entire sample in a glacial acetic acid and acetyl bromide mixture to determine the UV absorption, and the permanganate ( 3 ) and chlorite ( 2 ) procedures, in which lignin is oxidized and determined as the difference before and after oxidation. The determination of lignin by these various procedures is used in many research areas (e.g., animal feed or environment) to monitor various efforts to increase or gauge the digestibility, disappearance, or quality of plant materials. In earlier work (10), efforts were made to find the best procedure to calibrate near infrared reflectance spectrometers for lignin content. Those efforts indicated that, although near infrared reflectance spectroscopy could be used in place of the various chemical assays for lignin, the results were not as accurate as might be desired. In that same study (10), the results of various lignin methods often did not match or produced very different values for lignin content, as though the methods were measuring completely different substances. Subsequent studies using chemical extraction (11), pyrolysis-gas chromatography mass spectrometry
692
693
CRUDE PROTEIN LIGNIN
(13), mid infrared ( MIR) reflectance spectroscopy (12), and 13C-labeled solid-state nuclear magnetic resonance spectroscopy ( 1 4 ) have shown that the various procedures either do not extract all of the lignin or else determine nonlignin components as lignin. For ADL determinations, studies have indicated that protein and possibly carbohydrates could be included in the determination, but all of the lignin originally present in the ADF was apparently retained. Contamination of ADL by nonlignin fractions has also been reported by others (4, 5), as has the possible loss of lignin in the production of the ADF (4, 6), a possibility that was not examined in the studies in question (11, 12, 13, 14). The MIR examination of a few ADL samples showed spectra without the details that are generally expected for plant materials and their extracted residues (12). The many MIR peaks, representing carbohydrates, lignin, and proteins, were largely blended together into nondescript, rather featureless spectra that were more reminiscent of near infrared than MIR spectra. Results (11, 12, 13, 14) for the other measures of lignin indicated that similar problems existed, and protein contamination occurred often. In addition, the permanganate procedure showed problems with carbohydrates (11), and the chlorite procedure did not remove all of the lignin (11, 12, 13). However, results indicated that carbohydrate removal was not a problem for the chlorite procedure (11). Finally, results for the acetyl bromide procedure indicated that furfural formation from carbohydrates can be an extremely serious problem (11), which was particularly true when the acetyl bromide procedure was applied to residues from previous procedures (i.e., ADF extracted with chlorite or permanganate). The results of various lignin assays that were obtained with the acetyl bromide procedure for residues indicated furfural formation (11). These results are extremely important in regard to another issue with lignin assays. All of the procedures discussed determine lignin in some form of extracted fiber and not in the original feedstuff. Thus, if lignin were lost in the initial fiber preparation, the subsequent lignin values would be incorrect. Lowry et al. ( 6 ) reported that up to 50% of the lignin in tropical grasses might be lost before the lignin determination was ever made. Unfortunately, the estimates of the dispersible lignin were largely based on UV absorption of the ADF solution resulting from the ADF step or were based on the use of acetyl bromide on recovered residues. In both cases, the possible generation of furfurals and their subsequent interference, because they also absorb in the same wavelength range as lignin, make evaluation of the results difficult. References to digestible lignin, as indicated by soluble lignin and
carbohydrate complexes, are also used to support the concept of a large fraction of acid detergent dispersible lignin. However, work by Windham et al. ( 1 7 ) showed that such lignin and carbohydrate complexes contained no lignin at all, but rather were protein and carbohydrate complexes. Finally, work by Hatfield et al. ( 4 ) on Klason lignin and ADL support the work by Lowry et al. ( 6 ) . Those researchers suggested that, for ADL, a significant portion of the lignin may be dispersed by the acid detergent solution. Hatfield et al. ( 4 ) reported significant differences in the lignin contents of lucerne, cocksfoot, and switchgrass as determined by the Klason and ADL procedures. These differences could not be accounted for by carbohydrate or protein contamination as determined by pyrolytic and wet chemical analyses and, thus, appeared to be due to dispersion of lignin by acid detergent. However, Hatfield et al. ( 4 ) also showed that a much greater percentage of the original forage protein was retained in the Klason lignin than in the ADL (approximately 2 to 6.5 times as much). Because lignin is a known, defined material (i.e., an aromatic polymer) and not a composite of protein, carbohydrates, and lignin, it is important to understand exactly what each lignin procedure does and does not measure. An accurate procedure that truly measures lignin and only lignin is important not only for developing better calibrations from near infrared reflectance spectrophotometry, but also for obtaining a better understanding of the role of forage composition in determining digestibility. Although further investigations should determine the extent to which acid detergent dispersible lignin exists and its composition, it is still important to determine accurately the nondispersible lignin if a two-step procedure for lignin determination (dispersed lignin in the acid detergent solution by UV absorption and lignin in the ADF) is to be used, as suggested by Lowry et al. ( 6 ) . Therefore, the objective of this work was to examine the relationships between CP, ADF CP, and ADL CP and measures of nondispersible lignin, NDF, and digestibility for several lignin procedures currently used for forage analysis. Specifically, the study was conducted to determine whether a relationship exists between the CP content of the starting material or the residual CP found in ADF or ADL and various measures of nondispersible lignin, NDF, or digestibility. MATERIALS AND METHODS Samples All forages were collected in the field as grab samples of fresh plant; the samples were composited and dried immediately at 60°C. After drying, samples Journal of Dairy Science Vol. 80, No. 4, 1997
694
REEVES, III
were ground to a size of 20 mesh (Christie mill; Christie Norris Ltd., Chelmsford, England). Reproductive portions of corn and wheat samples were removed prior to drying in order to provide material that corresponded with corn stover ( CS; n = 10) and wheat straw ( WS; n = 11). Alfalfa ( AL; n = 15) and orchardgrass ( OG; n = 15) were cut several times during the season for hay or silage, and samples were taken from the regrowth. The AL was harvested on May 21 (10% bloom), June 23 (20% bloom), and August 2 (40% bloom). The OG was harvested at the stem elongation stage on May 10, July 21, and August 25. The tall fescue ( FS; n = 16, not free of endophytes) was cut only once and represented two growth periods: growth until June 3 (during late anthesis) and regrowth between June 3 and December 22. The CS and WS were taken every week or two from fields that were harvested once at the end of the growing season for grain (September 30 and July 9, respectively) ( 9 ) . Fiber and Lignin Production All 67 samples were analyzed for CP, ADF, and ADL in triplicate (1, 3). In addition, large batches of ADF and subsequently ADL were prepared by exact scaling of the procedures of Goering and Van Soest ( 3 ) for CP determination. Neutral detergent fiber and permanganate lignin on ADF ( PL) were similarly determined according to the method of Goering and Van Soest ( 3 ) , but without the addition of decahydronapthalene or sodium sulfite. Chlorite lignins were determined on both ammonium oxalate fiber ( CL-OX) by the procedure of Collings et al. ( 2 ) with modifications to the filtering procedure ( 1 0 ) and on ADF ( CL-ADF) . For CLADF, 1.0-g samples were each extracted with 200 ml of 0.5% ammonium oxalate solution for 2 h and dried at 100°C overnight. Chlorite lignin in these fibers was then determined by treating the fiber and crucible with sodium chlorite in acetic acid (1.0%) and refiltering the residue through the same crucible (10). Acetyl bromide lignins were determined according to the method of Morrison (7,8) using fibers extracted with hot water and organic solvent ( ABLOR) and by using ADF ( ABL-ADF) . For ABL-OR, hot water (30 min at 60°C), ethanol, acetone, and diethyl ether were used to extract the samples (each fiber sample was washed with 200 ml of each solvent). The fiber was dried overnight at 100°C and scraped from the crucible. A fraction was then weighed out for determination of acetyl bromide ligJournal of Dairy Science Vol. 80, No. 4, 1997
nin, and the rest was used for ash determination. Computations were based on optical density units per gram of OM, rather than the percentage of DM as for the other lignins. All chemical assay results are the means of triplicate determinations that were calculated on an OM basis. Forage Digestibility All in vitro digestibilities were carried out according to the methods of Goering and Van Soest ( 3 ) and are presented as the means of duplicate analyses. Incubations were for 48 h using rumen fluid obtained from a fistulated steer fed an OG diet. Statistics All data computations and statistical summaries were performed using SAS software (16). Correlations were based on the Pearson correlation matrix. RESULTS AND DISCUSSION Table 1 contains the minima, maxima, means, and standard deviations for the chemical results obtained for all 67 samples examined and for the same samples grouped by feedstuff. All feedstuffs contained a wide range of sample maturities as a result of sample collection from the beginning to the end of the growing season. The wide range in sample maturities is shown by the wide distribution of assay results and the magnitude of the standard deviations. The coefficients of variation ranged from ±7 to ±70%. Thus, the available samples represented a wide range of sample types and compositions upon which to draw summary conclusions. In particular, all sample groupings contained some samples with a relatively high initial CP content. Thus, even the CS and WS groups contained samples with CP contents that would be considered good for a feed (maximum CP content for CS, 27.0; maximum CP content for WS, 16.8). Although CS and WS would not normally be found with such high CP values, the early cuttings provided such samples. These types of samples might be used in studies to determine how lignin content changes as plants mature and are important in understanding how measured lignin content may be influenced by variations in CP content. Examination of the values for ADF CP indicated wide variation in the degree to which CP was incorporated into ADF. Although the highest CP values found for AL, CS, FS, and OG were quite similar (31.8, 27.0, 24.8, and 27.7%, respectively), the maxi-
695
CRUDE PROTEIN LIGNIN
mum ADF CP values varied considerably (3.3, 0.75, 1.4, and 2.8%, respectively). The maximum CP values for CS and OG were virtually identical at 27.0 and 27.7%, but the ADF CP values varied almost fourfold. Thus, the incorporation of CP into ADF was
dependent on the sample in question and was not simply dependent on the initial CP content of the starting material. A quick examination of the data for the six lignin procedures studied (ABL-ADF, ABL-OR, ADL, CL-
TABLE 1. Chemical composition (percentage of DM) for forages and by-products. Variable1 All forages and by-products ( n = 67) ABL-ADF ABL-OR ADF ADL CL-ADF PL-ADF CL-OX NDF CWD DMD CP ADF CP ADL CP Alfalfa hays only ( n = 15) ABL-ADF ABL-OR ADF ADL CL-ADF PL-ADF CL-OX NDF CWD DMD CP ADF CP ADL CP Vegetative corn plants only ( n = 10) ABL-ADF ABL-OR ADF ADL CL-ADF PL-ADF CL-OX NDF CWD DMD CP ADF CP ADL CP
Minimum Maximum X
SD
RSD2
0.8 3.0 17.3 0.84 0.7 2.8 4.0 25.8 37.9 55.4 4.4 0.5 1.5
7.1 9.6 53.0 7.8 6.0 14.0 8.2 83.0 87.8 94.1 31.8 3.3 18.0
3.0 6.6 33.1 3.8 2.7 7.5 5.9 59.1 60.8 76.9 15.4 1.2 6.9
1.4 1.3 7.4 1.5 1.0 2.5 1.0 14.5 11.9 9.2 7.7 0.8 3.3
0.47 0.20 0.22 0.40 0.37 0.33 0.17 0.25 0.20 0.12 0.50 0.67 0.48
1.2 3.0 17.3 2.4 1.5 5.6 5.3 25.8 37.9 71.9 17.7 1.9 7.3
4.8 6.5 40.4 7.3 3.4 14.0 7.0 50.5 64.3 90.8 31.8 3.3 12.2
2.9 5.1 27.3 5.1 2.5 9.4 5.9 36.7 49.7 81.3 24.3 2.4 10.7
1.0 1.2 6.1 1.3 0.6 2.2 0.5 7.2 7.1 5.5 4.2 0.4 1.5
0.34 0.24 0.22 0.25 0.24 0.23 0.09 0.20 0.14 0.07 0.17 0.17 0.14
1.2 5.4 25.4 1.8 1.0 4.3 4.5 53.0 53.9 62.3 5.1 0.5 1.5
5.2 8.1 45.4 4.5 4.8 10.0 8.2 81.8 82.8 90.9 27.0 0.7 9.3
3.8 7.2 37.8 3.3 3.2 7.0 6.4 69.0 64.4 74.7 10.4 0.6 3.4
1.4 0.8 6.3 0.9 1.2 2.0 1.1 8.7 9.5 9.1 7.3 0.1 2.3
0.37 0.11 0.17 0.27 0.38 0.29 0.17 0.13 0.15 0.12 0.70 0.17 0.68
Variable Tall fescue hays only ( n = 16) ABL-ADF ABL-OR ADF ADL CL-ADF PL-ADF CL-OX NDF CWD DMD CP ADF CP ADL CP Orchardgrass hays only ( n = 15) ABL-ADF ABL-OR ADF ADL CL-ADF PL-ADF CL-OX NDF CWD DMD CP ADF CP ADL CP Vegetative wheat plants only ( n = 11) ABL-ADF ABL-OR ADF ADL CL-ADF PL-ADF CL-OX NDF CWD DMD CP ADF CP ADL CP
Minimum Maximum X
SD
RSD
0.8 5.5 21.3 0.8 0.7 2.8 4.9 48.2 5.8 63.9 7.5 0.6 3.8
3.1 8.1 38.4 4.4 3.2 9.1 8.1 73.6 87.8 94.1 24.8 1.4 9.7
2.5 7.3 33.7 3.1 2.5 6.6 6.5 66.0 59.7 72.7 11.6 1.0 7.3
0.6 0.6 4.4 1.1 0.7 1.5 1.0 6.9 11.0 8.9 4.6 0.3 2.0
0.24 0.08 0.13 0.35 0.28 0.23 0.15 0.10 0.18 0.12 0.40 0.30 0.27
1.0 5.2 22.9 2.3 1.1 3.5 4.1 54.4 52.0 64.6 13.4 0.7 4.2
3.3 8.1 39.0 5.6 4.1 9.7 6.6 73.8 84.3 91.4 27.7 2.8 18.0
2.0 6.6 30.3 3.2 2.0 6.0 5.6 62.6 69.9 80.9 19.4 1.1 6.8
0.6 0.9 4.3 0.8 0.7 1.6 0.6 5.8 7.9 6.5 3.6 0.5 3.2
0.30 0.14 0.14 0.24 0.35 0.27 0.11 0.09 0.11 0.08 0.19 0.45 0.47
1.5 5.7 27.7 2.4 1.5 4.5 4.0 52.1 46.3 55.4 4.4 0.5 2.6
7.1 9.6 53.0 7.8 6.0 13.9 6.5 83.0 84.4 91.8 16.8 1.3 7.6
4.4 7.3 39.6 4.7 3.4 8.9 4.8 66.2 61.8 73.4 7.9 0.8 4.5
2.0 1.4 9.5 1.9 1.6 3.5 0.9 11.5 13.6 13.0 3.9 0.3 1.7
0.45 0.19 0.24 0.40 0.47 0.39 0.19 0.17 0.22 0.18 0.49 o.38 0.38
1ADF CP = CP content of ADF (percentage of ADF DM), ADL CP = CP content of ADL (percentage of ADL DM), ABL-ADF = acetyl bromide lignins on ADF (based on optical density), ABL-OR = acetyl bromide lignins extracted from fiber on hot water and organic solvent (based on optical density), ADL = acid detergent lignin (72% sulfuric acid lignin on ADF), CL-ADF = chlorite lignins on ADF (percentage of DM), PL-ADF = permanganate lignins on ADF (percentage of DM), CL-OX = chlorite lignins on ammonium oxalate (percentage of DM), CWD = cell-wall digestibility or NDF digestibility (percentage of DM), and DMD = DM digestibility (percentage of DM). 2Relative standard deviation.
Journal of Dairy Science Vol. 80, No. 4, 1997
696
REEVES, III
ADF, PL-ADF, and CL-OX) using data from all 67 samples showed the wide range of results obtained. For example, for both the acetyl bromide procedures (ABL-ADF and ABL-OR) and for the chlorite procedures (CL-ADF and CL-OX), the choice of fiber from which to extract the lignin greatly influenced the mean lignin values obtained. In each case, the use of ADF resulted in mean lignin values that were approximately half of those obtained using the fiber procedures outlined in the original papers [ABL-OR (7, 8 ) and CL-OX (2)]. The differences, based on the initial values of the starting fiber, might be due to the presence of acid detergent dispersible lignin as suggested by Lowry et al. ( 6 ) , to the enhanced extraction of protein or carbohydrates by the harsher acid detergent extraction, or both. Quick computation of ratios of the maximum and minimum lignin values shows the same sort of problems. For example, for CL-OX, the ratio was approximately 2:1, but for CL-ADF, the ratio was about 8.5:1. Similar results occurred with the lignins based on acetyl bromide. Even comparing lignin procedures based only on ADF (i.e., ABL-ADF, ADL, CL-ADF, and PL-ADF, the maximum to minimum lignin ratios were approximately 9:1, 9:1, 9:1, and 5:1, respectively. Similar results can be found by examining the lignin data on an individual feed basis (Table 1). The various feedstuffs show interactions among the different lignin assays; the highest value for PL-ADF was for AL samples (14.0%), followed by WS (13.9%). The highest value for CL-ADF was for WS at 6.0%, and the value for AL was only 3.4%. Similar results were found for low lignin values also. Alfalfa was the lowest only for ABL-OR. If all of these procedures measured lignin, even on some relative basis, one would expect the same feed to be high or low by all of the assays used. The incorporation of ADF CP into ADL would appear to be even more dependent on the sample than the incorporation of CP into ADF. For example, ADF of CS contained about half of the CP content that was found in the WS samples, but ADL of CS at the maximum contained more CP than did ADL of WS (9.3% vs. 7.6%). Results were similar for the other measures of lignin presented (e.g., PL-ADF). These results indicated that no simple relationship existed between sample CP content and the CP content of its ADF or lignin and that correction of lignin for CP content based on initial sample or ADF CP content was not feasible. The results also indicated that the presence of CP can complicate the correct determination of ADL because 1.5 to 18% of the lignin value could be based on CP and not on lignin. Because CP Journal of Dairy Science Vol. 80, No. 4, 1997
measurements are actually based on N content and not protein, alteration of the protein by the ADF procedure or, more likely, by the 72% sulfuric acid that was used in the ADL step, may mean that the problem is more or less serious than the results indicated, depending on the type of effects that the treatments (ADF or ADL) have on the initial protein present. Unfortunately, as stated by Hatfield et al. ( 4 ) , investigation of the ADL residue by methods such as pyrolysis-gas chromatography mass spectrometry does not provide information on the nature of the N present. Future investigations are planned using 15N-labeled nuclear magnetic resonance spectroscopy to investigate the question of the state of N in both ADL and chlorite lignin, because past efforts have led to confusing results (10, 14). In any case, this is a problem because true lignin does not contain any N, but is rather a polymer made by polymerization of propenyl-benzene constituents ( 1 5 ) that was induced by free radicals. Because the same protein would be present in the ADF used for other lignin assays, it is highly possible that the same type of relationships found for CP versus ADL might exist for other lignin assays. Only for the acetyl bromide lignins would protein be of minor concern because of the low extinction coefficient of protein compared with that for lignin. Furthermore, CP, which can survive the ADF extraction procedure, might also escape the extraction by ammonium oxalate and by hot water and organic solvent used in the two other procedures studied. Finally, the CP in ADL might represent some specific protein that was difficult to extract. Although not lignin, the protein may represent some specific fraction of feedstuffs that may relate to other measurements of lignin, fiber or digestibility. The data for NDF, cell wall, and DM digestibilities, presented in Table 1, were used for the correlations presented in Table 2. Table 2 contains the results obtained from correlation analyses of the various chemical results. Results for CP versus ADF or ADL confirm earlier expectations of closer correlations for individual feedstuffs than for all of the samples combined. The results for ADF versus ADF CP show wide variations in feed responses. Although there was a low (–0.30) but significant negative correlation overall (all 67 samples), the correlations were positive and twice the magnitude for FS, OG, and WS (0.65, 0.64, and 0.81, respectively). Although CS and WS were both byproducts that were high in lignin and low in CP and thus would be expected to show similar relationships between compositional components such as ADF and CP, only WS showed a significant correlation between ADF and ADF CP. The results for CP versus ADF CP and ADL CP also supported conclusions drawn from Table 1.
CRUDE PROTEIN LIGNIN
Although the results on an individual feedstuff basis varied considerably, there was approximately a 0.5 correlation for both relationships for all samples combined. Individually, only FS showed a significant correlation for ADF CP versus CP, and it was negative. Finally, the relationships between ADF, CP, and ADF CP and the CP content of ADL were equally varied and complex. Overall, for AL and CS, significant negative correlations existed between ADF content and ADL CP content; however, for FS, OG, and WS, positive correlations existed. Similarly, for all samples, for AL, and for CS, significant positive correlations existed between CP content of the base material and ADL CP content, but a negative relationship was found for FS; the results were not significant for OG and WS. The final and most consistent relationship was for ADF CP and ADL CP. Except for CS, for which the results were not signifi-
TABLE 2. Results of correlation analyses (correlation coefficients) of ADF, CP content of forages and by-products, and CP content of their ADF and acid detergent lignin (ADL). 1 CP All forages and by-products ( n = 67) ADF –0.81 CP ADF CP Alfalfa hays only ( n = 15) ADF –0.68 CP ADF CP Vegetative corn plants only ( n = 10) ADF –0.93 CP ADF CP Tall fescue hays only ( n = 16) ADF –0.81 CP ADF CP Orchardgrass hays only ( n = 15) ADF –0.62 CP ADF CP Vegetative wheat plants only ( n = 11) ADF –0.83 CP ADF CP 1Results
ADF CP
ADL CP
–0.30 0.54
–0.34 0.53 0.87
NS2 NS
–0.62 0.68 0.59
NS NS
–0.72 0.90 NS
0.65 –0.83
0.51 –0.75 0.86
0.64 NS
0.63 NS 0.95
0.81 NS
0.593 NS 0.91
were significant at P < 0.05 unless otherwise noted. > 0.06. 3P < 0.06. 2P
697
cant, a positive and generally close correlation existed between the ADF CP and the ADL CP, indicating that CP remaining in the ADF would, in general, be largely incorporated into, and consequently determined as, ADL. Correlation results involving lignin are found in Table 3. E xamining the results for correlations of CP with the various measures of lignin showed that, except for the determination of CL-OX in AL samples and ADL for all samples, a negative relationship existed between CP content and lignin, regardless of the method of lignin determination. Such negative correlations were expected because high quality materials are generally high in protein and low in lignin. The results did, however, indicate differences in these feed-dependent relationships. For example, the correlations between PL-ADF and CP were much more negative for single feedstuffs than for all five feedstuffs [r = –0.23 (all samples)] versus –0.51, –0.78, –0.81, –0.67, and –0.85 for AL, CS, FS, OG, and WS, respectively. Variations in the relative strength of the correlations between CP and various lignin measures, similar to the kind of variations shown in Table 1, may also be found. Thus, for ABL-ADF, the highest correlation was for CS ( r = –0.97), but for ABL-OR the highest correlation was for FS ( r = –0.83), and these are both acetyl bromide lignins, only based on different starting fibers. Overall, AL appeared to have the greatest range in correlation values between CP and lignin, from +0.02 for CL-OX to –0.71 for ABL-ADF; CS, FS, and OG were the most consistent. The highest general correlations appeared to be for CS and WS, which were both low quality by-products that were low in CP. However, the low correlation ( r = –0.33) for CL-OX for WS and the high correlation for CS ( r = –0.89) should be noted. The correlations of ADF CP with lignin found in Table 3 were extremely variable. The results for all 67 samples varied from +0.55 for ADL to –0.48 for ABL-OR. These results indicate that, overall, the CP content of ADF was incorporated into ADL but not into ABL-OR. Even grouping the data for the individual feedstuffs was difficult. Wheat straw and CS are generally considered to be of similar nature (i.e., low quality, low protein, and high fiber); however, the correlations between the various lignin procedures and ADF CP content were vastly different. The WS samples produced correlations more similar to those for FS and OG. Also, although for CS, FS, OG, and WS the correlations were all positive, indicating that the incorporation of ADF CP into lignin was a Journal of Dairy Science Vol. 80, No. 4, 1997
698
REEVES, III
problem, regardless of the procedure used, for AL, the results were both positive and negative. The correlations between the CP found in ADL and that found in other lignin procedures are difficult to explain. For FS, OG, and WS, the more CP that was present in ADL, the more lignin was determined by other procedures, as indicated by the positive correlations. However, for CS, the opposite was found. The positive correlation is easy to explain: the CP retained from the starting ADF in the ADL procedure was extracted and thus was determined as lignin by the other lignin procedures. A zero correlation can also be easily explained because CP found in ADF might have been left by the other lignin procedures and, thus, have no relationship to the lignin values found.
But negative correlations, such as those found for all the CS and most of the AL lignin determinations (except CL-OX), are not easy to explain. Because the various samples were collected at different stages of growth, the negative correlations might have been due to the low lignin and high CP contents of early maturity samples. The same maturity differences could be true for the other feedstuffs, especially for the WS samples, which were cut periodically, as was the CS, from a field of continuously maturing plants (as opposed to the regrowth found in fields of AL, OG, and FS, which were cut periodically for hay). Overall, these results indicate that, although relationships existed between ADL CP content and lignin, the relationships were very complex
TABLE 3. Correlation results by forage group for CP, ADF CP, and acid detergent lignin (ADL) CP with various measures of lignin, NDF, and digestibility. Variable1 All forages and by-products ( n = 67) ABL-ADF ABL-OR ADL CL-ADF PL-ADF CL-OX NDF CWD DMD Alfalfa hays only ( n = 15) ABL-ADF ABL-OR ADL CL-ADF PL-ADF CL-OX NDF CWD DMD Vegetative corn plants only ( n = 10) ABL-ADF ABL-OR ADL CL-ADF PL-ADF CL-OX NDF CWD DMD
CP
ADF CP
ADL CP
–0.63* –0.77* NS –0.61* –0.23 –0.25*
–0.06 –0.48* 0.55* 0.07 0.45* 0.21† –0.65* –0.59* 0.05
–0.23 –0.37* 0.35* –0.02 0.26* 0.24* –0.50* –0.45* 0.02
–0.71* –0.60* –0.78* –0.23 –0.51* 0.02
–0.22 –0.18 NS 0.067 –0.084 0.64* –0.16 –0.0045 0.13
–0.70* –0.56* –0.54* –0.24 –0.62* 0.47† –0.66* 0.29 0.58*
–0.97* –0.76* –0.92* –0.93* –0.78* –0.89*
0.08 0.25 NS 0.095 0.39 0.23 0.31 –0.08 –0.22
–0.78* –0.71* –0.69* –0.72* –0.48 –0.66* –0.64* 0.79* 0.69*
Variable Tall fescue hays only ( n = 16) ABL-ADF ABL-OR ADL CL-ADF PL-ADF CL-OX NDF CWD DMD Orchardgrass hays only ( n = 15) ABL-ADF ABL-OR ADL CL-ADF PL-ADF CL-OX NDF CWD DMD Vegetative wheat plants only ( n = 11) ABL-ADF ABL-OR ADL CL-ADF PL-ADF CL-OX NDF CWD DMD
CP
ADF CP
ADL CP
–0.79** –0.83** –0.84** –0.92** –0.81** –0.70**
0.54 0.59** 0.93** 0.86** 0.80** 0.83** 0.71** –0.89** –0.88**
0.42† 0.56 0.74** 0.75** 0.60** 0.71** 0.66** –0.76** –0.77**
–0.66* –0.47 –0.62* –0.57* –0.67* –0.47
0.55* 0.48 0.95* 0.91* 0.75* 0.41† 0.66* –0.71* –0.77*
0.53* 0.50 0.89* 0.86* 0.66* 0.37 0.62* –0.67* –0.74*
–0.90* –0.78* –0.82* –0.81* –0.85* –0.33
0.75* 0.76* 0.81* 0.86* 0.80* 0.90* 0.80* –0.60* –0.72*
0.49 0.59* 0.61* 0.65* 0.56 0.93* 0.59* –0.33 –0.49
1ADF CP = CP content of ADF (percentage of ADF DM), ADL CP = CP content of ADL (percentage of ADL DM), ABL-ADF = acetyl bromide lignins on ADF (based on optical density), ABL-OR = acetyl bromide lignins extracted from fiber on hot water and organic solvent (based on optical density), ADL = acid detergent lignin (72% sulfuric acid lignin on ADF), CL-ADF = chlorite lignins on ADF (percentage of DM), PL-ADF = permanganate lignins on ADF (percentage of DM), CL-OX = chlorite lignins on ammonium oxalate (percentage of DM), CWD = cell-wall digestibility or NDF digestibility (percentage of DM), and DMD = DM digestibility (percentage of DM). †P < 0.10. *P < 0.05. **P < 0.01.
Journal of Dairy Science Vol. 80, No. 4, 1997
699
CRUDE PROTEIN LIGNIN
and specific to the material in question, thus making generalizations extremely risky. Finally, consideration of the results for correlations of ADF CP and ADL CP with NDF and digestibility showed that, in general, the higher the ADF CP, the lower the digestibility of the samples (particularly cell wall digestibility) and the higher the NDF content. However, as for the lignin results, the findings were dependent on the sample. Thus, CS behaved similar to AL and opposite to the other samples. CONCLUSIONS Results of correlations among the CP content of feedstuffs, their ADF or ADL, and measures of lignin, NDF, or digestibility were found to depend greatly on the samples in question. In some instances, such as the correlation of ADL CP with lignin, NDF, or digestibility, different feedstuffs produced equally close, but opposite (positive vs. negative) correlations for the same relationships. Consideration of the type of material involved (high quality hays vs. low quality by-products), stage of maturity (regrowth vs. continual growth), or the particular assay involved was of no help in producing generalizations or in explaining the results. Because of the wide variation in the relationships between these variables, it would appear to be impossible to develop simple equations to correct lignin measures or ADF for CP content without the need for more chemical determinations (i.e, actual measurement of ADF CP or ADL CP). Because of the small amount of residues obtained from lignin determinations, such determinations are impractical on a routine basis; therefore, it would be best to modify the fiber procedure (ADF or other) in order to eliminate the incorporation of CP into ADF and subsequently into lignin, as well as the possible dispersion of lignin into the extracting solvent (4, 6). Efforts using various protease enzymes as a pretreatment have not proved to be successful (data not shown). In conclusion, although often relationships appear to exist between the values obtained by various measures of lignin and CP in feedstuffs, the exact relationships are extremely dependent on samples,
and more research is needed to determine their exact nature and causes. REFERENCES 1 Association of Official Analytical Chemists. 1984. Official Methods of Analysis. 14th ed. AOAC, Arlington, VA. 2 Collings, G. F., M. T. Yokoyama, and W. G. Bergen. 1978. Lignin as determined by oxidation with sodium chlorite and a comparison with permanganate lignin. J. Dairy Sci. 61:1156. 3 Goering, H. K., and P. J. Van Soest. 1970. Forage Fiber Analyses (Apparatus, Reagents, Procedures, and Some Applications). Agric. Handbook No. 379. ARS-USDA, Washington, DC. 4 Hatfield, R. D., H. G. Jung, J. Ralph, D. R. Buxton, and P. J. Weimer. 1994. A comparison of the insoluble residues produced by the Klason lignin and acid detergent lignin procedures. J. Sci. Food Agric. 65:51. 5 Kirk, T. K., and J. R. Obst. 1988. Lignin determination. Page 96 in Methods in Enzymology. Vol 161. W. A. Wood and S. T. Kellogg, ed. Academic Press, Inc., New York, NY. 6 Lowry, J. B., L. L. Conlan, A. C. Schlink, and C. S. McSweeney. 1994. Acid detergent dispersible lignin in tropical grasses. J. Sci. Food Agric. 65:41. 7 Morrison, I. M. 1972. Improvements in the acetyl bromide technique to determine lignin and digestibility and its application to legumes. J. Sci. Food Agric. 23:1463. 8 Morrison, I. M. 1972. A semi-micro method for the determination of lignin and its use in predicting the digestibility of forage crops. J. Sci. Food Agric. 23:455. 9 Reeves, J. B., III. 1987. Lignin and fiber compositional changes in forages over a growing season and their effects on in vitro digestibility. J. Dairy Sci. 70:1583. 10 Reeves, J. B., III. 1988. Chemical assays for fiber, lignin, and lignin components: interrelationships and near infrared reflectance spectroscopic analysis results. J. Dairy Sci. 71:2976. 11 Reeves, J. B., III. 1993. Chemical studies on the composition of fiber fractions and lignin determination residues. J. Dairy Sci. 76:120. 12 Reeves, J. B., III. 1993. Infrared spectroscopic studies on forage and by-product fibre fractions and lignin determination residues. J. Vib. Spectrosc. 5:303. 13 Reeves, J. B., III, and G. C. Galletti. 1993. Use of pyrolysis-gas chromatography/mass spectrometry in the study of lignin assays. J. Anal. Appl. Pyrolysis 24:243. 14 Reeves, J. B., III, and W. F. Schmidt. 1994. Solid-state 13C NMR analysis of forage and by-product-derived fiber and lignin residues. Resolution of some discrepancies among chemical, infrared, and pyrolysis-gas chromatography-mass spectroscopic analyses. J. Agric. Food Chem. 42:1462. 15 Sarkanen, K. V., and C. H. Ludig. 1971. Lignins: Occurrence, Formation, Structure, and Reactions. John Wiley & Sons, New York, NY. 16 SAS User’s Guide: Statistics, Version 5 Edition. 1985. SAS Inst., Inc., Cary, NC. 17 Windham, W. R., D. S. Himmelsbach, H. E. Amos, and J. J. Evans. 1989. Partial characterization of a protein-carbohydrate complex from the rumen of steers fed high-quality forages. J. Agric. Food Chem. 37:912.
Journal of Dairy Science Vol. 80, No. 4, 1997