Review pubs.acs.org/JAFC
Multiple Applications of Ion Chromatography Oligosaccharide Fingerprint Profiles To Solve a Variety of Sugar and Sugar−Biofuel Industry Problems Gillian Eggleston*,† and Eduardo Borges§ †
Southern Regional Research Center, Agricultural Research Service, U.S. Department of Agriculture, 1100 Robert E. Lee Boulevard, New Orleans, Louisiana 70124, United States § Fermentec Ltda., Av. Antônia Pazzinato Sturion 1155, Piracicaba, Brazil 13420 640 ABSTRACT: Sugar crops contain a broad variety of carbohydrates used for human consumption and the production of biofuels and bioproducts. Ion chromatography with integrated pulsed amperometric detection (IC-IPAD) can be used to simultaneously detect mono-, di-, and oligosaccharides, oligosaccharide isomers, mannitol, and ethanol in complex matrices from sugar crops. By utilizing a strong NaOH/NaOAc gradient method over 45 min, oligosaccharides of at least 2−12 dp can be detected. Fingerprint IC oligosaccharide profiles are extremely selective, sensitive, and reliable and can detect deterioration product metabolites from as low as 100 colony-forming units/mL lactic acid bacteria. The IC fingerprints can also be used to (i) monitor freeze deterioration, (ii) optimize harvesting methods and cut-to-crush times, (iii) differentiate between white refined sugar from sugar cane and from sugar beets, (iv) verify the activities of carbohydrate enzymes, (v) select yeasts for ethanol fermentations, and (vi) isolate and diagnose infections and processing problems in sugar factories. KEYWORDS: ion chromatography, pulsed amperometric detection, HPAEC, oligosaccharide fingerprint profiles, sugar cane, sugar beet, sweet sorghum
■
the column based on size, composition, and linkage.6 Separated carbohydrate anions are detected via PAD by measuring the electrical current generated by their oxidation at the surface of a gold electrode over a fixed period of time. The products of this oxidation reaction are cleaned between measurements; otherwise, there would be a loss of analyte signal due to gradual poisoning of the electrode surface.12 Cleaning occurs by applying a series of potentials for fixed time periods after the detection potential. Before 1998, cleaning occurred by first raising the potential to a level sufficient to oxidize the gold surface, which causes desorption of the carbohydrate oxidation products; the electrode potential is then lowered to reduce the electrode back to gold.11 Since 1998 and with new detectors (e.g., ED40 detector by Dionex), cleaning has occurred with negative potentials.12 The detector response is measured in coulombs because current in amperes integrated over time is electrical charge. The detector sensitivity differs for different carbohydrates; thus, calibration curves need to be created for each carbohydrate. When sodium acetate (NaOAc) is added to the carbonate-free NaOH eluent, it accelerates the elution of strongly bound species without compromising selectivity or interfering with pulsed amperometric detection. This is because acetate interacts much more strongly than hydroxide with the anion-exchange sites. Moreover, larger sized oligo- and polysaccharides have greater solubility in NaOH and NaOAc eluents, which also aids their separation and detection.14 The
INTRODUCTION In the early 1980s, Hughes and Johnson1 developed pulsed amperometric detection for carbohydrates. Rocklin and Pohl2 reported that high-performance anion-exchange chromatography (HPAEC or ion chromatography IC) could be combined with integrated pulsed amperometric detection (IPAD) to separate and detect carbohydrates with high sensitivity and selectivity. There was also no cumbersome and time-consuming derivatization or sample cleanup prior to injection.3 Since then, IC-IPAD has been applied to determine many types of carbohydrates in a large variety of samples from different research fields,4 including plant-derived reducing and nonreducing monosaccharides, alditols, oligo- and polysaccharides such as amylopectins, arabinans, arabinoxylans, fructooligosaccharides, fructans, galactans, xylans, and mannans (see Rohrer,5 Cataldi et al.,6 and Corradini et al.7 for comprehensive reviews). IC-IPAD has also been used to detect and quantitate compounds derived from carbohydrates such as hydroxymethylfurfural (HMF) in honey, high-fructose corn syrups, and milk.8 IC-IPAD is now regarded as one of the major tools for carbohydrate analyses.3 Koizumi et al.9 reported that IC-IPAD could be used to separate linear glucose polymers with a degree of polymerization (dp) as high as 50, and Hanshiro et al.10 even reported the separation of glucose polymers >80 dp. Moreover, it has been reported that carbohydrates at low picomole levels can be detected.11 At pH values >12 in sodium hydroxide (NaOH) eluents, carbohydrate hydroxyl groups are at least partially ionized to produce weakly acidic oxyanions. These oxyanions are separated on a strong anion-exchange stationary phase in the form of microbeads packed in a column. Anions interact with © XXXX American Chemical Society
Received: February 9, 2015 Revised: February 23, 2015 Accepted: February 24, 2015
A
DOI: 10.1021/jf506370s J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Review
Journal of Agricultural and Food Chemistry
Figure 1. Gradient separation on a Thermo Scientific (previously Dionex) CarboPac PA1 column of oligosaccharides in fresh (green line) and deteriorated (blue line) sugar cane juice (samples were standardized to the same Brix value). DP, degree of polymerization. See the text for full ICIPAD method conditions. cross-linked with divinylbenzene) agglomerated with MicroBead quaternary ammonium fuctionalized latex (5% cross-linked). Flow rate = 1.0 mL/min. Injection volume = 25 μL. Eluent conditions were as follows: carbonate-free 100 mM NaOH (isocratic (0.0−1.1 min; inject 1.0 min), a gradient of 0−300 mM NaOAc in 100 mM NaOH (1.1−40.0 min), and return to 100 mM NaOH (40.1−45.0 min) to reequilibrate the column to the starting conditions prior to injection). The re-equilibration is needed for good retention time reproducibility. The eluents were prepared by diluting a 50% w/w NaOH solution with 18 MΩ deionized water and then degassing them with ultrahighpurity helium for 20 min. Once the eluents were prepared, they were kept blanketed under helium (20 kPa) at all times. The samples were diluted 1 g/25 mL in deionized water, unless otherwise stated, and filtered through a 0.45 μm PVDF filter to remove particulates. IPAD detection was with either a Dionex PED-2 or Dionex ED50 detector; both detectors were equipped with Au working and Ag/AgCl reference electrodes. The PED-2 detector operated with the following working electrode pulse potentials and durations: E1 = +0.05 V (t0 = 0.00 s), E2 = 0.05 V (t1 = 0.42 s), E3 = +0.75 V (t3 = 0.43 s), E4 = +0.75 V (t4 = 0.60 s), E5 = −0.60 V (t5 = 0.61 s), and E 6 = −0.60 V (t 6 = 0.96 s). The ED50 detector operated as follows: E1 = +0.10 V (t0 = 0.00 s), E2 = 0.10 V (t1 = 0.20 s), E3 = +0.10 V (t3 = 0.40 s), E4 = +2.00 V (t4 = 0.41 s), E5 = −2.00 V (t5 = 0.42 s), E6 = +0.60 V (t6 = 0.43 s), E7 = −0.10 V (t7 = 0.44 s), and E8 = −0.10 V (t8 = 0.50 s). A refrigerated Spectra-Physics SP8880 or Dionex AS refrigerated autoinjector/autosampler was used to prevent degradation of sugars in samples while waiting for injection onto the column. Dionex Peaknet or Chromeleon chromatography software was used to accumulate multiple samples and check standards.15 Oligosaccharides, mannitol, and ethanol were identified by comparing retention times with standards and by spiking with standards. A representative chromatogram with the gradient method conditions overlaid is shown in Figure 1. To maintain baseline stability, the sodium hydroxide was kept constant during the sodium acetate gradient, because acetate has no buffering capacity at high pH.11 Brix (Percent Refractometric Dry Substance). Brix value was measured using an Index Instruments (Kissimmee, FL, USA) TCR 1530 temperature-controlled refractometer accurate to ±0.01 Brix, and results were expressed as an average of triplicates.
strong alkaline mobile phase also catalyzes the mutarotation of reducing sugars such as glucose and fructose very efficiently, and at room temperature the α- and β-anomers elute together in one sharp peak.3 This is an added advantage of IC-IPAD over both gas and liquid cation-exchange chromatography with RI detection, with which double peaks of the reducing sugars are often obtained.3 The high-resolving power of IC-IPAD and its ability to determine oligosaccharides of higher degrees of polymerization has allowed the “fingerprinting” of numerous products manufactured from sugar crops. This includes sugar cane and sugar beet used in the global manufacture of sugar, as well as sweet sorghum, which is increasingly being used as a sugar feedstock for the manufacture of biofuels and bioproducts. The fingerprinting capability of a NaOH/NaOAc gradient IC method utilizing a Thermo Scientific (previously Dionex) CarboPac PA1 strong anion-exchange column, combined with the sensitivity of the IPAD has allowed its application to numerous industrial problems associated with these sugar crops, which are discussed herein.
■
MATERIALS AND METHODS
Chemicals and Reagents. HPLC grade sodium hydroxide and sodium acetate were obtained from Thermo Scientific Dionex (Sunnyvale, CA, USA). Deionized water (with resistivity of 18 MΩ) was used to prepare eluents and samples. Standard sugars, oligosaccharides, and mannitol were of analytical grade. 1-Kestose, 1,1-nystose, and 1,1,1-fructofuranosylnystose were from Meiji Seika Kaisha, Ltd. (Tokyo, Japan). Theanderose (6-O-α-D-glucosylsucrose) was from Wako Chemicals (Richmond, VA, USA) and kindly purified by Dr. Greg Cote of USDA-ARS-NCAUR. Dextran was T2000 (2,000,000 Da) from Amersham Biosciences AB (Uppsala, Sweden). Dextranase was obtained from BioCat, Inc. (Troy, VA, USA), and pullanase (Promozyme D2 manufactured by Novozymes (Franklinton, NC, USA)) was obtained from Brenntag Southwest (Houston, TX, USA). Distiller’s yeast in dried form (Saccharomyces cerevisiae; Crosby & Baker Ltd., Westport, MA, USA) was donated by Delta BioRenewables LLC (Memphis, TN, USA). Absolute ethanol was from Aaper (USA). IC-IPAD Oligosaccharide Chromatograms. Carbohydrates (mainly oligosaccharides up to 12 dp), including sugar alcohols and ethanol, were separated on a Dionex (now Thermo Scientific, Sunnyvale, CA, USA) BioLC or Dx500 instrument, using a strong NaOH/NaOAc gradient over 40 min.14 The compounds were separated on Thermo Scientific CarboPac PA1 analytical anionexchange (250 × 4 mm) and guard columns (50 × 4 mm) at 30 °C or, unless otherwise stated, using a Dionex LC25 Chromatography Oven. The CarboPac PA1 column was designed for the rapid analysis of mono- and oligosaccharides, in particular, linear homopolymers.13 The CarboPac PA1 column contains 10 μm substrate of polystyrene (2%
■
RESULTS AND DISCUSSION Deterioration of Juices from Sugar Crops. IC oligosaccharide fingerprint profiles have been used to investigate the deterioration of juices extracted from sugar cane (Saccharum officinarum) and sweet sorghum (Sorghum bicolor) during storage, as well as from freeze-deteriorated sugar cane.16−18 The typical progression of deterioration in sugar cane juice during storage at ∼25 °C is illustrated in Figure 2. By overloading the column with high concentrations of glucose, fructose, and sucrose present in the juice, the presence of oligosaccharides and other related deterioration compounds B
DOI: 10.1021/jf506370s J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Review
Journal of Agricultural and Food Chemistry
oligosaccharides, permitting their quantitation. Moreover, extensive column washings between runs are not required.15 Oligosaccharides form on the deterioration of sugar crops, with the main oligosaccharides produced being glycosyl acceptor (secondary) products produced from the action of dextransucrase by Leuconostoc mesenteroides bacteria.16,19 Examples of Leuconostoc oligosaccharides are leucrose, isomaltose, and isomaltotriose (Figure 2). Other oligosaccharides formed on sugar cane deterioration include kestoses (GF2 trisaccharides), which can form from invertase activity. However, kestoses are also primary oligosaccharides that occur naturally in the sugar cane plant, particularly in the leaves and top part of the stalk.14 Early in our sugar crop deterioration research, we discovered that mannitol was a major degradation marker of sugar cane deterioration16 and later found it in sweet sorghum deterioration.20 Mannitol is also formed by Leuconostoc lactic acid bacteria via a metabolic pathway separate from that for dextran biosynthesis. Fructose is reduced to mannitol by the enzyme mannitol dehydrogenase in that pathway.21 As illustrated in Figure 2, mannitol is a very sensitive indicator of sugar cane deterioration and is even more sensitive than dextran as a deterioration marker.22 Steinmetz et al.23 had also reported that mannitol was even a better marker of sugar beet deterioration than dextran. Mannitol (a sugar alcohol) is a reduced carbohydrate that is a weaker acid than its nonreduced counterpart mannose and, therefore, it is poorly retained on the CarboPac PA1 column, which explains its early elution (Figure 2). Ethanol is also formed during the
Figure 2. Changes in IC-IPAD profiles and Brix value on the deterioration of sugar cane juice during storage time. S, sucrose; G, glucose; F, fructose. The chromatograms were obtained on a Dionex BioLC system with a PED-2 detector at room temperature (∼25 °C). See Materials and Methods for full details. Reprinted with permission from ref 16. Copyright 2002 Elsevier.
was detectable (Figure 2). This highlights another marked advantage of this IC-IPAD method in that the anion-exchange CarboPac PA1 column can tolerate high concentrations of samples (up to 20 °Brix15), yet still detect low levels of oligosaccharides and other compounds of interest. The very large sucrose peak is not quantitated at this high concentration but is completely separated from these other low-level
Figure 3. Dextran concentration correlations with (A) mannitol, (B) isomaltose, (C) 1-kestose, and (D) nystose determined in sugar cane juice. C
DOI: 10.1021/jf506370s J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Review
Journal of Agricultural and Food Chemistry
Figure 4. IC-IPAD fingerprint oligosaccharide chromatograms of stored juice extracted from KN Morris sweet sorghum cultivar (November 2013): (A) raw juice; (B) heated juice; (C) clarified juice. Times of storage and Brix values of the juices are denoted on the overlaid chromatograms. Reprinted with permission from ref 20. Copyright 2014 Elsevier.
Leuconostoc deterioration of sugar cane but is only a minor metabolite of Leuconostoc compared to dextran and mannitol.16 Similarly to mannitol, ethanol elutes very early due to its weak interaction with the stationary phase in the column. Detection of ethanol by PAD using a gold working electrode is also less sensitive than for mannitol.16 In sugar cane juice collected from a Brazilian factory, there was a strong correlation between the mannitol concentration
and the dextran concentration when the latter was measured accurately by the Roberts copper method24 (Figure 3A). Other microorganism and plant metabolites, including isomaltose, 1kestose, and nystose, were also shown to strongly correlate with the dextran concentration (Figure 3B−D). However, as concluded in previous research,16,22 mannitol was shown to be the most sensitive deterioration indicator (Figure 3A). D
DOI: 10.1021/jf506370s J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Review
Journal of Agricultural and Food Chemistry
Figure 5. Example IC-IPAD oligosaccharide fingerprint profiles to illustrate the effect of sugar cane harvest method and cut-to-crush delays on deterioration (Brix value was standardized in the samples). The chromatograms were obtained on a Dionex BioLC system with a PED-2 detector at ∼25 °C. Adapted from ref 14.
results as well as the fact that the clarification process destroyed or removed all lactic acid bacteria, although other bacteria remained at ∼103 cfu/mL levels. Overall, it was shown that the IC-IPAD fingerprint profiles are extremely sensitive, being able to detect as low as 102 cfu/ mL lactic acid bacterial growth.20 Additionally, the effect of heat pasteurization was very similar to the whole clarification process, which confirmed that heat was the major contributor to juice stabilization. Optimization of Sugar Cane Harvesting Methods and Cut-to-Crush Times. The IC-IPAD profile method has also been used to determine the formation of oligosaccharides and other deterioration products caused by various sugar cane harvest and storage treatments, and example chromatograms are illustrated in Figure 5. The changing oligosaccharide profiles allowed for a much more sensitive elucidation of the different deterioration reactions contributing to overall sugar cane deterioration than typical deterioration analyses, those being pH and titratable acidity.14 Using the IC-IPAD method it was possible to show that at 0 h there were no marked differences in oligosaccharide profiles for harvest treatments, which reflected the freshness of all of the samples and indicated that, initially, when the field cane is cut, freshness is more important than harvest method.25 This result had considerable consequences for the factory in that as long as billeted cane is transported to the factory quickly on the day it is harvested, deterioration problems should be no worse than encountered in fresh-cut whole stalks (soldier harvested). There were, however, marked
More recently, the IC-IPAD has been used to successfully confirm bacterial count results in a study to stabilize and preserve sweet sorghum juices.20 Sweet sorghum is similar to sugar cane in that it has a juicy stalk that is rich in sugars. The effects of heat pasteurization (80 °C; 30 min) and clarification (80 °C; limed to pH 6.5; addition of 5 ppm of polyanionic flocculant; settling for 30 min) to juice from sweet sorghum cultivar KN Morris, which was then stored for up to 48 h at 25 °C, are illustrated in Figure 4 oligosaccharide fingerprint profiles. Treated juices were compared to the susceptible raw juice (not treated) (Figure 4). No formation of deterioration products was detected in KN Morris clarified juices stored from 0 to 48 h, although a slight amount of oligosaccharides after the sucrose peak was detected in the 24 and 48 h heated juice samples (Figure 4). This agreed with the lactic acid microbial data, with no detectable growth in the amount of total and lactic acid bacteria in the clarified juice across the 48 h storage period.20 In comparison, deterioration products were found in the 24 h stored raw juice, which was worse after 48 h of storage (Figure 4). This was no surprise, as before storage the raw juice contained 2.8 × 108 total bacteria/mL, which increased ∼5-fold during the 48 h storage period.20 In contrast, total bacteria in the 0 h heated and clarified juices were approximately 4-log less than in the 0 h raw juice, and no lactic acid bacteria were detected in the clarified juice at 0 h.20 Lactic acid formation was slightly worse in KN Morris heated than in clarified juice, but this was still considerably lower than in the raw juice.20 Thus, the IC fingerprint profiles confirmed the bacterial growth E
DOI: 10.1021/jf506370s J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Review
Journal of Agricultural and Food Chemistry
create a destabilized market and disrupt the regional economy. The detection of adulteration of refined sugars is also of interest to sugar and food manufacturers, consumers, and regulators because of issues of legal compliance. Manufacturers need to ensure that correctly labeled refined sugar is available to the consumer and/or buyer. Consumers also have an interest in ensuring the food they purchase is authentic. Although raffinose (O-α-D-Galp-(1→6)-O-α-D-Glup-(1→2)β-D-Fruf) and theanderose (O-α-D-Glcp-(1→6)-O-α-D-Glup(1→2)-β-D-Fruf) had both been advocated as differential markers, raffinose is present in both BWS and CWS (although to a much lesser extent in CWS) and pure theanderose was not commercially available; also, small IC peaks had been found in BWS samples with the same retention time as theanderose. Although some highly sophisticated methods exist to differentiate the source of white sugar, such as isotope ratio mass spectrometry (IRMS) and deuterium NMR, these were considered only to be reference techniques28 because the equipment is expensive and not readily available. A general analytical screening method was needed to detect a wide range of compounds in the same run. Suspect samples could then be detected and evaluated by more sophisticated techniques as just described. Because IC-IPAD had (i) become a preferred technique for the routine monitoring of sugars in sugar industry products, (ii) been used to detect simple types of adulteration such as the addition of sucrose to honey,29 and (iii) a very high sensitivity that offered the great advantage of detecting very low levels of adulterant compounds, we investigated its use to differentiate between beet (BWS) and cane white (CWS) refined sugar.30 By overloading the CarboPac PA1 column with ∼7.0 Brix blind BWS/CWS samples and running the IC-IPAD NaOH/NaOAc 45 min gradient method, low raffinose combined with numerous characteristic cane marker peaks was successfully used to detect 20% CWS adulteration, which is illustrated in Figure 7. In general, IC profiles of CWS samples had more peaks of 2−12 dp than BWS samples. Increasing the Brix levels to 10.0 allowed detection of 10% CWS adulteration.30 Chromatography libraries of CWS, BWS, and BWS/CWS samples for direct comparisons would aid adulterant detection, as would pattern recognition chemometric techniques.31 The use of an internal standard would greatly help in the standardization of chromatographic patterns and be necessary for chemometric analysis. The use of column ovens and refrigerated autosamplers also aids in the stabilization of retention times. Overall, at the least, the use of IC profiles can be used as a screening method before the further verification and quantitation with more sophisticated techniques. Verification of the Activities of Carbohydrate Enzymes. Dextran (α-1→6-α-D-glucan) is a detrimental glucose polysaccharide that is found in deteriorated sugar cane, sugar beet, and sweet sorghum products. In the sugar industry, dextranase (EC 3.2.1.11; α-D-1,6-glucan-6-glucanohydrolase) is applied to hydrolyze high molecular weight (HMW) dextran into more manageable lower MW poly- and oligosaccharides. The goal is to improve viscosity and throughput rates in sugar cane and sugar beet factories (in sugar cane factories crystal elongation is also reduced). Most dextranases currently available are manufactured from Chaetomium gracile or erraticum fungi32 and are generally recognized as safe (GRAS). One of the greatest sources of confusion for factory or refinery personnel is that it is not possible to compare the activities of commercial dextranases because suppliers/vendors
differences in the formation of oligosaccharides and other deterioration products on cut-to-crush/delay time (Figure 5). For the whole-stalk treatments shown in Figure 5, no significant formation of either Leuconostoc oligosaccharides and mannitol or kestose oligosaccharides occurred during storage. 1-Kestose (1F-O-β-fructofuranosylsucrose) was the most abundant isomer, followed by 6-kestose (6F-O-β-fructofuranosylsucrose) and neo-kestose (1G-O-β-fructofuranosylsucrose), and separate peaks for 1- and 6-kestose isomers are shown in Figure 5. In strong contrast, deterioration products formed within 24 h during the storage of both green and burnt billeted cane, with greater deterioration occurring in burnt than green billets (Figure 5). Figure 6 illustrates a correlation between the sum of oligosaccharide concentrations related to sugar cane deterio-
Figure 6. Correlation between sum of oligosaccharides concentration (isomaltose + palatinose + isomaltotriose + 1-kestose + nystose + 1Ffructofuranosylnystose) with sugar cane cut-to-crush time.
ration (isomaltose, palatinose, isomaltotriose, 1-kestose, nystose, and 1F-fructofuranosylnystose) and cut-to-crush time in a Brazilian sugar factory that had problems with delivered sugar cane from two different fields. The results were compared with relatively fresh sugar cane delivered to the factory with a cut-to-crush time of
