Baixe NMR & Physicochemical Techniques for Forage Assessment Study by D.S. Himmelsbach e outras Notas de estudo em PDF para Engenharia de Alimentos, somente na Docsity!
The use of high-resolution (HR) spectroscopic techniques (e.g. high-field nuclear magnetic resonance (NMR), mid-infrared (MIR), Raman spectroscopy and pyroly- sis mass spectrometry (PyMS)) are finding increased usage in forage assessment. These techniques, as opposed to low-resolution (LR) spectroscopic techniques (e.g. near-infrared (NIR) and ultraviolet (UV) spectroscopy), generally provide more specific and detailed information of a primary nature. The trade-off for the gain in resolution is often the sacrifice of time, increased cost and the requirement to use more skilled operators. However, the employment of HR techniques is frequently necessary when the LR techniques fail to provide the desired information or require verification. UV spectroscopy can be an excellent primary source of information, if the information sought is related to a single or isolated chromophore. However, forages tend to contain a multitude of chromophores, making such analyses difficult. NIR spectroscopy is generally not a source of direct information but an indirect or sec- ondary technique, requiring independent calibration by a primary technique, using chemometric approaches (Kowalski, 1977; Martens et al. , 1991). The primary tech- nique is normally some form of gravimetric analysis for moisture, protein, lignin or acid-detergent fibre (ADF) (AOAC, 1997). Although this approach is sufficient and cost-efficient for many needs, the dependence on proximate analyses limits the effectiveness of NIR spectroscopy (Preston et al. , 1997). To date, the focus of the HR techniques has been directed towards providing a route to primary molecular structures. If the HR spectroscopic results can be linked to the LR techniques or be made more cost-effective themselves, the accessibility to more reliable results should be greatly enhanced. The use of chemometrics with any of the techniques can also increase their utility (Bro et al ., 1997). Since anatomy and chemical composition have combinatory effects on the digestibility of forages, additional benefits can be realized by the integration of any of these spec- troscopic with microscopic techniques.
16 NMR and Other Physicochemical
Techniques for Forage Assessment
D.S. HIMMELSBACH
Richard B. Russell Agricultural Research Center, Agricultural Research Service, P.O Box 5677, Athens, GA 30604-5677, USA
© CABInternational 2000.Forage Evaluation in Ruminant Nutrition (eds D.I. Givens, E. Owen, R.F.E. Axford and H.M. Omed) 321
Introduction
Both solution and solid-state NMR have been found to be powerful tools for exam- ination of forages and their residues. True solution NMR studies are limited to extractable components that are soluble in suitable deuterated NMR solvents. If the isolated material is pure, a wide array of single and multidimensional experi- ments may be employed to unequivocally assign the molecular structure and solu- tion confirmation of a molecule (Bruch, 1996). Solid-state NMR is well suited for examining the intact structure of forage materials, without requiring extraction. It is especially suited for the in situ study of forages and their residues. However, it is more limited than solution NMR in the variety of experiments that can be per- formed and the line widths of signals are broader (i.e. show lower apparent resolu- tion). New experiments, which lie somewhere in the realm between the solution and solid methods, are now in vogue. These are experiments in which all or part of the sample is made mobile by hydration. The carbon-13 (^13 C) version of this experi- ment permits separation of component spectra based on an increased separation of relaxation times, due to selective hydration (Newman, 1992). The proton (^1 H) version of this experiment is characterized by relatively narrow line widths and per- mits greater use of multidimensional experiments (Gil et al. , 1997). These experi- ments show potential application to studies on forages, but have not yet been employed for that purpose.
Solution NMR spectroscopy continues to benefit from the development of super- conducting magnets with increased field strengths. Commercially available sys- tems, with field strengths of 18.5 tesla (T) (800 MHz for protons), are now being delivered. This is a significant increase in field strength over the previously avail- able highest fields of 11.5 and 14 T (500 and 600 MHz). This means the possibility of even greater signal dispersion and sensitivity and the ability to unravel even more complex polymeric structures, such as those habitually encountered in plant materials used as forages. Examples of the effective use of the currently available solution NMR technol- ogy are provided in studies directed at determination of the specific involvement of hydroxycinnamic ester moieties in restricting the availability of cell-wall polysac- charides of forage grasses to ruminant digestion. Several hydroxycinnamic oligosac- charide esters have been isolated and their structures have been subjected to detailed characterization by solution NMR. Figure 16.1 shows examples of such compounds that are currently known. These compounds have been isolated from bagasse (Kato et al. , 1983, 1987, 1990), barley straw (Mueller-Harvey et al. , 1986), bamboo (Ishii and Hiroi, 1990a, b; Ishii et al. , 1990; Ishii, 1996), coastal Bermuda grass (Hartley et al. , 1990; Himmelsbach and Hartley, 1993; Himmelsbach et al. ,
- and maize (Kato and Nevins, 1985). The characterization of compound 3b in Fig. 16.1 provides a good example of how solution NMR can be used for unequivocal structural assignment of these
322 D.S. Himmelsbach
NMR Spectroscopy
Solution NMR spectroscopy
of each residue. This generates edited subspectra of each of the monosaccharide components. Then the simple one-bond proton–proton scalar coupling system is worked through to assign the proton signal for each unit. Knowing the proton assignments, the corresponding 13 C spectrum can be assigned by employing an experiment such as the inverse detected (^1 H) one-bond { 1 H, 13 C} heteronuclear multi-quantum correlation (HMQC) experiment (Bax and Subramanian, 1986). The results of this type of experiment are shown in Fig. 16.2. After both the 1 H and 13 C spectra have been assigned for the individual sub- units, the linkages between them can be found by employing an experiment such as the inverse-detected long-range {^1 H, 13 C} heteronuclear multiple-bond connectivity (HMBC) experiment (Bax and Summers, 1986; Bax and Marion, 1988). The circled cross-peaks, shown in Fig. 16.3, establish the linkages between the subunits
324 D.S. Himmelsbach
Fig. 16.2. Results of the inverse-detected one-bond HMQC experiment conducted on 5 mg of compound 3b (Fig. 16.1) obtained at 500 MHz for 1 H and 125 MHz for 13 C in a 20% d 6 -acetone/D 2 O solution at 25°C.
of 3b (Fig. 16.1). It shows that the ferulic acid is linked through an oxygen attached to the carbonyl carbon at F9 to the carbon at the 5 position (A5S proton shown) of the arabinofuranoside. This arabinofuranoside, in turn, is linked to two different xylopyranoside units at the 1 (A1) and 2 (A2) positions. At the 1 posi- tion, it is linked to the 3 position (X 9 3) of a xylopyranoside, which is further linked at its 1 position (X 9 1) to the 4 position of the reducing terminal xylopyranose, which displays signals for both a and b anomers (Xa4, Xb4).
NMR and Other Physicochemical Techniques 325
Fig. 16.3. Results of the inverse-detected long-range HMBC experiment conducted on compound 3b (Fig. 16.1) obtained at 500 MHz for 1 H and 125 MHz for 13 C in 20% d 6 -acetone/D 2 O solution at 25°C.
In addition, the angle of rotation of the hydrocinnamic acid in relation to the arabinofuranoside can be estimated by analysing the three-bond coupling constants (^3 J HH) made between the proton attached to the A4 carbon and each of the two protons attached to the A5 carbon (Hoffmann et al. , 1992). The final conforma- tion can then be expressed as a population of rotamers. In this case, the rotamer population was trans = 0.52, gauche^2 = 0.38 and gauche+^ = 0.10. These energy- minimized structures are shown in Fig. 16.5. Although these conformations will be fixed in the solid-state matrix of the cell walls of forage and not free to rotate, as in solution, this does give some idea of what the preferred (minimum-energy) confor- mation might be. The solid-state conformation may be critical to how these com- pounds are incorporated into the rest of the cell wall. Solution NMR has also been used to provide insight as to how the hydroxycin- namic acids linked to hemicellulosic polysaccharides are incorporated into the cell wall. It has been shown that [2+2] cyclo-addition dimers can be formed in a solid- state photochemical reaction, yielding a-truxillic acids from compounds 1a and 1b (Fig. 16.1) (Hartley et al. , 1990). Proton NMR, along with mass spectrometry (MS), has been used to verify that the hydroxycinnamic acids all prefer to couple with a head-to-tail orientation towards each other (Morrison et al. , 1992). In the case of the ferulate polysaccharide esters, the 8 position of the hydroxycinnamic
NMR and Other Physicochemical Techniques 327
Fig. 16.5. The three minimum energy-staggered rotamers of 3b (Fig. 16.1) that describe its solution conformation in a 20% d 6 -acetone/D 2 O solution at room temperature.
acid has been shown (by the HMBC experiment) to couple to the b (8) position of both guaiacyl and syringyl alcohol monomers (Ralph et al. , 1995). The ferulate esters are also known to form dehydrodiferulates (Harris et al. , 1980) and possibly link to core lignin by ether bonds (Jung and Ralph, 1990), although this has not been verified by NMR spectroscopy. Solution 13 C NMR has additionally been used to compare isolated forage lignins and show that warm-season grass forages differ from cool-season in that they incorporate p -coumaryl units (Himmelsbach and Barton, 1980). The solution 13 C NMR spectra of two representative spectra of these two classes of forages are shown in Fig. 16.6. The incorporation of p -coumaryl units in the warm-season forage (Fig. 16.6a) is indicated by the signals labelled ‘2,6H’ and ‘3,5H’. It has been shown, from solution 13 C NMR spectra of bagasse, that treatment of lignin isolated from warm-season grasses with alkali releases most of the p -coumaryl units (Fernandez, 1990). This indicates that most of the p -coumaryl units exist as acids esterified to lignin and the remainder are esterified to arabinoxylans. This esterification to lignin has been confirmed, using the HMBC experiment, in isolated maize lignin, as compared with synthetic materials. Further, it has been revealed by this experi- ment that p -coumaric acid is extensively incorporated into the a and g positions of
328 D.S. Himmelsbach
Fig. 16.6. Solution 13 C NMR spectra at 25 MHz of (a) Coastal Bermuda grass (warm- season grass) and (b) Kentucky-31 tall fescue (cool-season grass) lignins in d 6 -DMSO and referenced to internal TMS. H, 4-hydroxycinnamyl (p-coumaryl); G, guaiacyl; S, syringyl; a, b and g, 7, 8, and 9 positions, respectively, of aliphatic three-carbon side- chains off aromatic rings of lignin precursors.
a change in lignin structure can be directly detected in the intact forage. Thus, the isolation of the lignin is not required in order to obtain the desired information. A comparison of the results of using the basic CP/MAS and interrupted decoupling experiments is shown for pearl millet in Fig. 16.8. The loss of the signal intensity at 154 p.p.m., due to decrease of syringyl units, can barely be detected in going from the normal to the mutant (Fig. 16.8a and b, respectively) using the standard CP/MAS experiment. However, using the interrupted decoupling experiment, this difference is readily apparent (Fig. 16.8c and d). It also shows that the experiment is not effective when molecular motion averages 1 H-^13 C interactions of proton- bearing carbons. Thus, signals also appear for some protonated carbons, the aliphatic CH 3 in lipids (~22 p.p.m.) and the aromatic OCH 3 (~56 p.p.m.) in lignin. Fortunately, these signals do not interfere with others and the aromatic OCH 3 actually helps in providing confirming information for the decrease in syringyl units.
330 D.S. Himmelsbach
Fig. 16.7. Solid-state 13 C CP/MAS NMR spectra at 50 MHz of: (a) Coastal Bermuda grass (warm-season grass) and (b) Kentucky-31 tall fescue (cool-season grass). Spectra were externally referenced to the aromatic signal in hexamethylbenzene, set at 132.3 p.p.m.
Solid-state 13 C CP/MAS NMR spectroscopy was used, in the same study (Morrison et al. , 1993), to show that treatment with 1 mol l^21 sodium hydroxide (NaOH) removed almost all of the phenolics from the bmr mutant rind tissue but was not effective in doing so in the normal tissue. It required treatment with 4 mol l^21 NaOH to remove phenolics from the normal rind tissue of pearl millet. This indi- cated that the phenolics in the mutant were ester-linked, rather than ether-linked as in the normal line. This suggests that lignin polymerization had not occurred to a large extent in the mutant and this could account for the greater digestibility of bmr mutants by ruminants. One of the most effective uses of solid-state 13 C CP/MAS NMR spectroscopy for the study of forages is the investigation of the nature of the indigestible residues that remain after rumen microbial digestion. It has been used to characterize the residues from both in vitro (Akin et al. , 1993b) and in sacco digestion (Himmelsbach et al. , 1988).
NMR and Other Physicochemical Techniques 331
Fig. 16.8. Solid-state 13 C NMR spectra at 75 MHz of pearl millet rind tissue: (a) CP/MAS of normal line; (b) CP/MAS of mutant line; (c) CP/MAS with interrupted decoupling of normal line; (d) CP/MAS with interrupted decoupling of mutant line.
Other spectroscopic techniques have been used in attempts to obtain similar infor- mation to that obtained by NMR spectroscopy. Diffuse-reflectance infrared Fourier- transform spectroscopy (DRIFTS), in the MIR region, performed on the same samples as were used in the study on in sacco residues (Himmelsbach et al. , 1988) are shown in Fig. 16.11. Comparison of the MIR spectra reveal, even more clearly, the digestion of carbohydrate. This is shown by the decrease in O-H stretch bands (~3000–3600 cm^21 ) and C-O mixed-mode heavy-atom vibrations (~900–1200 cm^21 ). Like the NMR spectra, the MIR spectra show that there is no digestion of lignin or lipids. Again, the indications for lignin or phenolics ( and 1505 cm^21 bands) actually show a slight increase as digestion proceeds. One indicator for lipids, the C-H stretch at 2850 cm^21 , is overshadowed by the loss of similar bands due to carbohydrate, but a sharpening of the shoulder at this position is evident. The other indicator for lipids is the carbonyl (CO) band at 1750 cm^21. It shows the most dramatic change, indicating that lipids are not digested.
NMR and Other Physicochemical Techniques 333
Fig. 16.10. Solid-state 13 C CP/MAS NMR spectra at 75 MHz of neutral detergent residues (NDRs) fromin sacco digestion of a 28% crude-fibre mixed silage after: –––– 0 h (dipped in digesta), – – – – 48 h and -------- 336 h digestions.
Other Spectroscopic Techniques
The Fourier-transform (FT) Raman spectra of these same samples, taken using an NIR laser source, are shown in Fig. 16.12. The baselines in these spectra have been corrected to remove fluorescence effects. The loss of carbohydrate is not as evident in the Raman spectra as in the MIR or NMR spectra. However, the bands in the 900–1200 cm^21 region do decrease in intensity as digestion proceeds. The biggest difference in the Raman spectra is the response to aromatic components. This is to be expected, since the Raman active vibrations must be accompanied by a change in polarizability (Colthup et al. , 1990) and aromatic compounds contain highly polarizable bonds. Small differences in the phenolic (lignin or hydroxycin- namic acid) content of the residue, barely detectable in the solid-state NMR and MIR spectra, are revealed as large differences in the Raman spectra at ~1600 cm^21. Lipids are only detectable by a sharpening of the shoulder for C-H stretch bands at 2850 cm^21 , as in the MIR. The carbonyl band gives only a small response in the Raman, compared with the MIR. The bands in the ~1200–1500 cm^21 region can- not be unequivocally assigned to any single component, but could be associated with an initial loss in protein with digestion at 48 h and a reassociation of micro- bial protein by 336 h of digestion.
334 D.S. Himmelsbach
Fig. 16.11. DRIFTS MIR spectra of NDRs fromin sacco digestion of a 28% crude-fibre mixed silage after: ––––– 0 h (dipped in digesta), – – – – 48 h and --------- 336 h digestions. All spectra are baseline-corrected.
carbohydrates can be emphasized. An example of the use of PyMS is provided in the examination of Coastal Bermuda grass and Kentucky-31 tall fescue, using EI PyMS (Fig. 16.14; Morrison et al. , 1991). Specific markers were found for p -coumaric and ferulic acid esters at 120 and 150 mass-to-charge ratio ( m / z ), respectively. The marker for p -coumaric acid showed greater abundance in Coastal Bermuda grass. This is in agreement with the NMR spectroscopic results on lignin already dis- cussed. In addition, PyMS results revealed that (after sequential ozone and base treatment) all traces of phenolics could be removed from the Kentucky-31 tall fes- cue residue but not from the Coastal Bermuda-grass residue. This suggested that all cell-wall carbohydrates were more available in Kentucky-31 tall fescue and could explain the greater rumen digestibility of this forage after treatment with ozone.
Morphological, anatomical and chemical factors all potentially contribute to limit- ing ruminant digestibility of plants (Akin and Chesson, 1989). Thus, being able to study chemical factors within anatomical compartments at specific morphological stages is a highly desirable goal.
336 D.S. Himmelsbach
Fig. 16.13. NIR spectra of NDRs fromin sacco digestion of a 28% crude-fibre mixed silage after: –––––– 0 h (dipped in digesta), – – – – 48 h and --------- 336 h digestions. All spectra baseline-corrected and normalized around the water band.
Microspectroscopy and Imaging
UV spectroscopy has been employed in conjunction with microscopy, called UV microspectroscopy or microspectrophotometry, in an attempt to attain this goal. By restricting the spectra to specific cell types, the interference between com- pounds that give similar spectral responses can also be reduced. One of the better examples of the use of UV microspectroscopy is its use in the detection of phenolic components in specific cells of forages that show different degradabilities (Akin and Rigsby, 1992). Figure 16.15 shows the UV spectra of the mestome sheath of leaf blades, which is mostly degraded in legumes, partially degraded in cool-season grasses and not degraded in warm-season grasses by rumen microorganisms after 7 days. The important bands in the UV spectra are the absorptions for aromatics, which occur at ~280 nm for condensed lignin or lignin-like compounds and at ~320 nm for those with extended conjugation (e.g. p -coumaryl and feruloyl esters). This suggests that, when only condensed lignin or lignin-like structures are present, digestibility is not particularly hindered in this tissue, as in the case of lucerne. When esters are present, digestion is restricted, as with the cool-season grasses (cocksfoot and tall fescue). When both are present to a large extent, then digestion is blocked, as with the warm-season grass Coastal Bermuda grass.
NMR and Other Physicochemical Techniques 337
Fig. 16.14. PyMS, electron impact (16 eV) fingerprint of Coastal Bermuda grass and Kentucky- 31: (a) and (b) cell walls, respectively, and (c) and (d) chemically treated residues, respectively.
the vascular bundle were imaged at 1509, 1550 and 1469 cm^21 , respectively. The greatest concentration of phenolics was found in the sclerenchyma and parenchyma- bundle sheath (which included the mestome sheath). Lipids appeared to be located in the epidermal tissue and protein nearby. It is apparent that MIR microspec- troscopy is capable of locating more chemical constituents than UV micro- spectroscopy, but it is limited by its spatial resolution in discriminating between anatomical structures. Raman microspectroscopy has been around as long as MIR microspectroscopy but has been more difficult to routinely access. Raman imaging was introduced as the Raman microprobe in 1975 (Delhaye and Dhamelincourt, 1975). Like MIR imaging, it has seen a resurgence for many of the same reasons (Turrell and Corset, 1996). However, the development of the FT-Raman confocal systems (Brenan and Hunter, 1995) or systems based on acousto-optic tunable filter or liquid-crystal tun- able filter sources (Morris et al. , 1994) and charge-coupled device (CCD) detectors have made this a more viable method. Spatial resolution can be achieved on the order of 1 mm and spectral resolution is nearly equivalent to that in MIR. This technique is not very applicable to the study of forage materials, which contain highly fluorescing compounds, such as chlorophyll. However, it is applicable to the study of residues. It would be especially useful in mapping the location of lignin in digested residues, although it has not yet been used for this purpose.
NMR and Other Physicochemical Techniques 339
Fig. 16.16. UV absorption spectra of leaf-blade parenchyma bundle sheath of: –––– Coastal Bermuda grass, – – – – – cocksfoot grass and ......... Kentucky-31 tall fescue.
NIR imaging is just starting to be explored extensively and most MIR imaging systems can be converted to work in the NIR spectral region. Some commercial NIR imaging systems, which are optimized for the NIR, are just becoming available (Treado, 1995) and will make this technique more accessible. Its main advantage is that it can be used to analyse samples that are totally absorbing, which is a situa- tion that can be expected to be encountered in studies on forages. NIR imaging represents a compromise between MIR and Raman imaging with regard to spatial and spectral resolution. It generally has higher spatial resolution but presents diffi- culties in assigning spectral features, due to severe spectral bands overlapping. Although most of the recent applications of NIR imaging have been in the medical arena, its use is beginning to become more widespread in other biological areas (Dempsey et al. , 1996). No applications of NIR imaging to forages could be found, but this technique certainly lends itself to their study, as NIR spectroscopy has already done. NMR imaging or magnetic resonance imaging (MRI) can also be used to obtain chemical imaging information. This technique has been restricted to solu- tion studies in plants and much of the work is at low spatial resolution, but it can yield information that is inaccessible by other microscopic techniques (Blümich and Kuhn, 1992). The combination of solution 2-D {^1 H, 1 H} correlation spec- troscopy with chemical-shift imaging NMR in an experiment called correlation- peak imaging (CPI) is an example of the state of the art. It has been performed on a plant seedling to reveal the locations of soluble carbohydrates and amino acids (Metzler et al. , 1995). The 1 H image obtained at 500 MHz clearly shows the vascu-
340 D.S. Himmelsbach
Fig. 16.17. MIR microspectroscopy of cell walls in pearl millet stem: (a) normal rind (sclerenchyma and vascular tissue); (b) vascular bundle within the pith ofbmr mutant; (c) normal pith.
