This article presents a method for the detection and quantification of organic acids from plant material using free zonal capillary electrophoresis. An example of the potential application of this method, determining the effects of a secondary fermentation on organic acid levels in coffee seeds, is provided.
Carboxylic acids are organic acids containing one or more terminal carboxyl (COOH) functional groups. Short chain carboxylic acids (SCCAs; carboxylic acids containing three to six carbons), such as malate and citrate, are critical to the proper functioning of many biological systems, where they function in cellular respiration and can serve as indicators of cell health. In foods, organic acid content can have significant impact on taste, with increased SCCA levels resulting in a sour or "acid" taste. Because of this, methods for the rapid analysis of organic acid levels are of particular interest to the food and beverage industries. Unfortunately, however, most methods used for SCCA quantification are dependent on time-consuming protocols requiring the derivatization of samples with hazardous reagents, followed by costly chromatographic and/or mass spectrometric analyses. This method details an alternate method for the detection and quantification of organic acids from plant material and food samples using free zonal capillary electrophoresis (CZE), sometimes simply referred to as capillary electrophoresis (CE). CZE provides a cost-effective method for measuring SCCAs with a low limit of detection (0.005 mg/ml). This article details the extraction and quantification of SCCAs from plant samples. While the method provided focuses on measurement of SCCAs from coffee beans, the method provided can be applied to multiple plant-based food materials.
Carboxylic acids are organic compounds containing one or more terminal carboxyl functional groups, each attached to an R-group containing one or more carbons (R-C[O]OH). Short chain, low molecular weight carboxylic acids (short chain carboxylic acids, SCCAs) containing between one and six carbons, are essential components of cellular respiration, and function in several biochemical pathways necessary for cell growth and development. SCCAs play critical roles in cellular metabolism1, cell signaling2, and organismal responses to the environment (such as antibiosis3). Because of this, SCCAs can serve as useful indicators of disruptions to cellular metabolism, plant stress responses4,5, and fruit quality6,7. To date, SCCAs have been quantified primarily through chromatographic techniques such as high performance liquid chromatography (HPLC) or gas chromatography-mass spectroscopy (GC-MS). While these methods, are capable of achieving very low limits of detection, they can be expensive, require the derivatization of target SCCAs using caustic and/or toxic reagents, and include lengthy separation runs on the GC or HPLC. Because of this, interest in the use of free zonal capillary electrophoresis (CZE), which does not require sample derivatization, to quantify organic acids has steadily increased8.
Free zonal capillary electrophoresis (CZE) is a chromatographic separation methodology that, due to its high number of theoretical plates, speed, and relative ease-of-use, is increasingly replacing both GC-MS and high-pressure liquid chromatography as an analytical method for the quantification (particularly for quality control purposes) of anions, cations, amino acids, carbohydrates, and short chain carboxylic acids (SCCAs)8,9,10. CZE-based separation of small molecules, including SCCAs, is based two primary principles: the electrophoretic movement of charged ions in an electrical field established across the buffer filling the capillary; and the electro-osmotic movement of the entire buffer system from one end of the capillary to the other, generally towards the negative electrode. In this system, small molecules move towards the negative electrode at varying speeds, with the speed of each molecule determined by the ratio of the net charge of the molecule to the molecular mass. As the movement of each individual molecule in this system is dependent on the charge state of the molecule and the overall rate of electro-osmotic flow (which is itself based on the ion content of the buffer used to fill the capillary), the buffer pH and ionic composition heavily impact the degree to which molecules can be efficiently separated using CZE. Because of this, SCCAs, with their relatively high charge-to-mass ratios, are ideal targets for CZE-based separation. Metabolites separated using CZE can be detected using a variety of methods, including UV absorbance, spectral absorbance (which is generally performed using a photo-diode array [PDA]), and/or mass spectroscopy (CE-MS or CE-MS/MS)8. The diversity of separation and detection methods provided by CZE makes it an extremely flexible and adaptable technique. Because of this, CZE has been increasingly applied as a standard method of analysis in the areas of food safety and quality11,12, pharmaceutical research13, and environmental monitoring13,14.
Capillary electrophoresis has been used to detect and quantify short chain carboxylic acids for nearly two decades13. The resolving power (particularly for small, charged molecules), short run time, and low per sample cost of CZE analyses make CZE an ideal technique for the separation and quantification of SCCAs13. This method presents a protocol to utilize CZE to measure the concentration of organic acids from plant tissues. Example data was generated through the successful implementation of this protocol to measure the change in organic acid levels in coffee seeds following a secondary fermentation treatment. The protocol details the critical steps and common errors of CZE-based separation of SCCAs, and discusses the means by which this protocol can be successfully applied to quantify SCCAs in additional plant tissues.
1. Sample Preparation
2. Organic Acid Standard Preparation
3. Organic Acid Extraction
4. Setting Up the SCCA Detection Run
Table 1: Conditioning method program used to prepare the capillary for short chain carboxylic acid separation via capillary electrophoresisa.
Table 2: Separation method program used to analyses short chain carboxylic acids via capillary electrophoresisa.
5. SCCA Detection Run Execution and Data Collection
Figure 1: A comparison of PDA traces highlighting an overloaded sample. As analyte concentration increases, individual peak geometry may begin to become asymmetric. At (a) 0.05 mg/ml, acetic acid presents a well-defined, bilaterally symmetrical peak. As the concentration of acetic acid increases to (b) 0.07 mg/ml and (c) 0.10 mg/ml, a peak tail forms (arrows). This peak tailing is a good indication that the sample is overloaded. Please click here to view a larger version of this figure.
6. Data Analysis
This protocol has been successfully utilized to measure the effects of seed treatments on the SCCA content of green coffee seeds. In this experiment, the six treatments were: a saturated microbial suspension of Leuconostoc pseudomesenteroides strain GCP674 in its growth medium (1), an aqueous suspension of GCP674 microbes in water (2), an aqueous solution of acetic and lactic acids (0.15 and 0.4 mg/ml respectively) (3), a spent M1 growth medium treatment (4), dH2O water (5), and an un-treated control (no substance added to seeds) (6). Treatments were applied and allowed to ferment for 24 hours. Four independent replicates were assessed for each treatment. Analysis was conducted for citric (C), malic (M), acetic (A) and lactic (L) acid levels using adipic acid as an internal standard.
For each SCCA and the internal standard, standard curves spanning the range of concentrations predicted to be in coffee samples (5, 10, 20, 40, 60, and 80 ng/µl) were generated. As expected, each acid exhibited a linear response for this concentration range with R-squared values of 0.9876 for citric acid, 0.9987 for malic acid, 0.9998 for acetic acid, and 0.9999 for lactic acid. The internal standard (adipic acid) also exhibited a linear response over this concentration range, yielding an R-squared value of 0.9984. Prior to sample analyses, limits of detection (LOD) and limits of quantitation (LOQ) were determined for each SCCA and the internal standard (adipic acid). The limit of detection of citric, malic, acetic, and adipic acid was 1 ng/µl; and the LOD for lactic acid was 2 ng/µl. The limit of quantitation of citric, adipic, acetic, and lactic acid was 4 ng/µl; and the LOQ for malic acid was 2 ng/µl. To calculate the percent recovery SCCAs and adipic acid, coffee was spiked with individual SCCAs or adipic acid and processed using the protocol described above. The percent recovery was then calculated by using the following formula ([{peak area spiked sample – peak area control sample}/theoretical peak area of calculated concentration of spiked sample {generated by analyzing buffer + SCCA/adipic acid spike}] x 100). Using this method, we calculated percent recoveries of 106.08% ± 12.66 for citric acid, 98.35% ± 1.15 for malic acid, 91.94% ± 3.07 for acetic acid, 97.42% ± 1.48 for lactic acid, and 100.19% ± 2.57 for adipic acid.
To determine the precision of the method, the intra- and inter-day variability of measurements made using CZE was also calculated. To determine intra-day variability, samples of each SCCA and adipic acid were measured from a single coffee sample five times across a 24-hour period, and coefficients of variability (COV; calculated by dividing the standard deviation in peak area [or concentration] for each SCCA by the average peak area [or concentration] for that SCCA, and then multiplying the result by 100) for each compound were calculated using the peak area. To assess the importance of correcting samples using the internal standard, the coefficient of variance for each SCCA was also calculated using the concentrations of each SCCA calculated using adipic acid as an internal standard. As expected, the CZE method was able to precisely measure SCCAs and adipic acid, with COVs of 3.02% for citric acid, 2.64% for malic acid, 3.74% for acetic acid, and 2.01% for adipic acid (lactic acid was below LOD/LOQ for all coffee samples measured). These measurements were repeated across a five-day period, and the CVs ranged from 2.11-5.25% for citric acid, 2.01-5.32% for malic acid, 1.72-3.74% for acetic acid, and 1.26-3.82% for adipic acid. When peak areas were corrected using the internal standard and calculated concentrations of SCCA were used to calculate COVs, the intra-day COVs ranged from 1.08-4.17% for citric acid, 1.47-3.40% for malic acid, and 2.68-5.10% for acetic acid. To assess inter-day variability, SCCAs in a single coffee sample were measured once per day across a five-day period. This experiment was then repeated five times for each SCCA and adipic acid. The COV for each SCCA was then calculated by dividing the standard deviation in peak area (or concentration) for each SCCA by the average peak area (or concentration) for that SCCA and multiplying by 100. To assess the importance of correcting samples using the internal standard, COVs were also calculated using the concentrations of each SCCA calculated using adipic acid as an internal standard. The CZE method was able to reliably reproduce SCCA measurements, with COVs ranging from 5.24-10.02% for citric acid, 6.55-9.47% for malic acid, 7.67-8.63% for acetic acid, and 3.08-6.57% for adipic acid. When peak areas were corrected using the internal standard and calculated concentrations of SCCA were used to calculated COVs, the inter-day COVs ranged from 4.56-6.23% for citric acid, 3.39-4.99% for malic acid, and 9.5-10.94% for acetic acid.
Figure 2: Example PDA traces with peaks indicated. Green coffee samples were diluted 1:10 before loading into CE vials. Acids of interest are indicated: acetic acid (A), citric acid (C), lactic acid (L), malic acid (M) and the internal standard, adipic acid (IS). Please click here to view a larger version of this figure.
Statistical analysis was conducted on corrected acid concentrations using a general linear model (GLM) to determine whether or not treatments altered organic acid levels. A model incorporating "treatment" x "run" (coded as a random factor) was implemented for the GLM. Tukey pairwise comparisons at 95% confidence were used to determine the directionality of changes.
Treatment altered SCCA concentrations in the green coffee. Acetic acid was significantly enriched in all treatments over the no treatment control (GLM, p <0.0001).
Figure 3: Impact of treatment on organic acid levels in green coffee. Treatment had no impact on (a) citric or malic acid levels in green coffee. However, all treatments significantly increased (b) acetic acid levels compared to no treatment controls (GLM, p <0.001). Increases in acetic acid levels were greatest with the microbe + medium treatment. Letters indicate significant difference by Tukey pairwise comparisons at 95%. Please click here to view a larger version of this figure.
The greatest increase was observed in the microbe + media treatment, which increased 0.25-fold (0.5 vs 0.4 mg/g coffee in GCP674 vs. NT). A similar trend was observed in lactic acid levels, though they were not significantly different. There were no significant changes or trends in the other acids tracked (citric and malic acid).
As with any analytical technique, there are several critical factors that can significantly impact the quality and reliability of the data generated. First, it is important to process samples efficiently, with a minimum of freeze/thaw cycles. Repeated freezing and thawing can compromise the chemical composition of the sample before processing or analysis. Second, it is critical to apply the steps of this protocol to all samples consistently and evenly. Technical errors arising from inconsistent sample preparation and handling can significantly impact the quality of the data generated, and result in increased "noise" in the SCCA measurements. For example, the particle size of the sample post-grinding will impact the speed and efficiency of SCCA extraction. It is therefore critical to ensure that the samples are ground to a uniform particle size, as this will maintain extraction efficiency across samples and help to reduce sample-to-sample variability. Similarly, accurate measurement and pipetting during the preparation of standard solutions, accurate measurement of sample masses, and careful monitoring (and normalization) of extraction times will result in the generation of more uniform data. Taking time to carefully normalize, measure, and record each of these parameters will ensure the reliability of the data and allow for accurate post-run correction of data for any processing errors which may occur. Third, it is critical to select and use an appropriate internal standard. Calculations of the final analyte concentration will rely heavily on accurate correction based on the peak area of the internal standard observed in each sample. Because of this, it is critical that the internal standard be both accurately and precisely measured; and should ideally not occur naturally in the sample or co-elute with other compounds. Not only does the proper selection and use of an internal standard in all samples enable the researcher to control for changes in metabolite levels observed due to differences in extraction efficiency; it also greatly simplifies downstream processing and peak identification. Finally, it is necessary to monitor the peak identification and integration process. While automated processing algorithms are useful, they are not perfect. The data generated by this technique rely on the accuracy of integration, and the traces must be checked manually to ensure the quality of integration events, particularly the definition of peak boundaries and baselines.
As with any analytical procedure, it is important to establish that the CZE method presented here is: a. accurate: able to determine the amount of a given SCCA present; b. precise: capable of reproducibly quantifying SCCAs within the same day and in measurements made across several days of analysis; and c. robust: capable of recovering most of the SCCAs present in the samples being analyzed. As shown above in the representative results, the method described here is capable of accurately determining the amounts of SCCAs present in samples. All the SCCAs measured (citric, malic, acetic, and lactic acid) and the internal standard (adipic acid) exhibited a linear response in the 0-80 ng/µl concentration range. Additionally, the LOQs and LODs for these acids ranged from 2-4 ng/µl and 1-2 ng/µl, respectively. The method was also precise, as all of the SCCAs measured and the adipic acid internal standard exhibited low intra-day variability, falling between 2.0-5.5% of the peak area (or between 1.0-5.2% of the calculated concentration) for all SCCAs measured. The inter-day variability of SCCA measurements was also relatively low, falling between 3.05-10.10% of the peak area (or between 3.30-10.95% of the calculated concentration) for all SCCAs measured. Interestingly, with the exception of acetic acid, which was consistently at or close to the limit of detection/quantification in all coffee samples analyzed, correcting samples using adipic acid as an internal standard resulted in more reproducible measurements and a decrease in the coefficient of variability (see Representative Results, above). These data are consistent with previously published results indicating that correction with an internal standard is essential in maintaining precision in CE analyses over time16. While some methods have employed inorganic ions as an internal standard, we felt that adipic acid was a more appropriate standard for the CZE method presented here, as it is an organic acid and therefore more likely to experience loss in the sample preparation steps similar to that experienced by the SCCAs of interest in the sample. Adipic acid had the additional advantages of not being present in the coffee samples being examined, exhibiting a linear response across the same concentrations used for SCCAs in this study, exhibiting relatively low LOD/LOQ values (1 ng/µl and 4 ng/µl, respectively), and a high percent recoverability from the coffee matrix (100.19% ± 2.57), making it a useful standard for quantifying SCCAs in coffee samples. Interestingly, the high percent recovery observed for adipic acid was achievable for all organic acids measured in this study, which exhibited percent recovery values ranging from 91.94-106.08%. These data indicate that the CZE method presented here is capable of recovering most of the SCCAs present in coffee samples.
Adapting this protocol to new sample types requires the collection of preliminary data, based on literature or empirical evidence, concerning the expected amounts of targeted analytes. This information will help guide the experimental design, sample preparation, and standard curve construction. Determining the correct sample dilution to use can be easily accomplished empirically by running a dilution series of a test extract to find a proper working dilution. This protocol was designed to detect the SCCAs outlined in the example results, but it could easily be adapted to detect other SCCAs by altering the separation parameters, such as separation voltage, pressure, or time. It is important to remember, however, that altering the separation parameters may alter the number of runs that can be accomplished on a single set of separation buffers as these buffers degrade with use. Baseline degradation or changes in current are often manifestations of over-used/depleted buffers. Replacing the running buffer and reconditioning the capillary can often resolve these problems. After extended use, the ability of the capillary to resolve analytes will decrease as well. Decreased baseline stability, drastic shifts in retention time, and reduced peak resolution are all signs that the capillary may need to be replaced. It is also possible to receive a pressure error resulting from improper sealing between the sample vial and the capillary due to an ill-fitting vial cap or a defective vial. Ensure the caps fit snugly into the sample vials to prevent this error.
Traditionally, chromatographic methods such as GC-MS, GC-FID, or HPLC have been used to detect SCCAs in a wide range of matrices, including air, water, soil, and plant samples14. While effective, a major drawback in these methods is the need to derivatize SCCAs before detection. Derivatization uses hazardous reagents the limit of detection is often affected by the efficiency of the derivatization reaction13,14. The detection of non-derivatized SCCAs through CZE avoids analyte loss during processing, exposure to hazardous derivatizing agents, and reduces sample preparation time. These benefits can be seen in the high percent recovery rates (ranging between 91.94-106.08%) calculated for all SCCAs measured in this study. In addition, CZE quantification of SCCAs achieves limits of detection comparable to those reported in literature for GC-MS, GC-FID, and HPLC, in the range of parts per million to parts per billion13,14. When combined with high-speed run times, high-resolution separation power, the straight-forward sample processing protocols employed by this method of analysis, the sensitivity and convenience of CZE make it an attractive alternative to traditional GC or HPLC methods.
However, it is important to note that, while CZE provides several advantages over GC-MS HPLC, or LC-MS/MS based methods of SCCA analysis, it is not appropriate for all SCCA analyses. For example, since the CZE method described here relies solely on spectrophotometric detection, it does not allow for the identification of unknown SCCAs present in the sample, or the confirmation of SCCA identity via mass spectral profiling. Because of this, GC-MS or LC-MS/MS methods would be better suited for SCCA analyses in samples containing large numbers of unknown SCCAs, or samples in which it is critical to confirm the identity of all SCCAs present. Additionally, as previously reported LC-MS/MS based methods allow for lower limits of detection (LODs)-on average, 0.05 µg/ml for LC-MS/MS17 vs. the 1 µg/ml for CZE obtained using the method presented here-LC-MS/MS based quantification of SCCAs may be more appropriate for samples in which SCCAs are present in very low or trace amounts.
The protocol presented here focuses on the extraction and quantification of short chain carboxylic acids (SCCAs) from plant tissues. Obtaining proficiency in this technique will open the door to a wide range of applications of this technology for the individual researcher. This technique, with only slight variations in methodology, has been successfully employed in the analysis of SCCAs in a diverse population of samples and substrates, ranging from food products and tissue samples to environmental water and soil samples8. Additionally, gaining familiarity with the chemical principals underlying this technology though this protocol will provide researchers with the knowledge base needed to attempt the analysis of other compounds, such as carbohydrates or metals8,13. The analytical flexibility of capillary electrophoresis makes it an extraordinarily powerful tool suitable for a myriad of experimental applications.
The authors have nothing to disclose.
The authors would like to acknowledge the financial support of this project by The J.M. Smucker company.
Ceramic Moarter and Pestle | Coorstek | 60310 | |
Beckman Coulter P/ACE MDQ CE system | Beckman Coulter | Various | |
Glass sample vials | Fisher Inc. | 033917D | |
1.5 ml microcentrifuge tubes | Fisher Inc. | 02-681-5 | |
LC/MS grade water | Fisher Inc. | W6-1 | Milli-Q water (18.2 MΩ.cm) is also acceptable |
15 ml glass tube/ Teflon lined cap | Fisher Inc. | 14-93331A | |
Parafilm M | Fisher Inc. | 13-374-12 | |
CElixirOA detection Kit pH 5.4 | MicroSolv | 06100-5.4 | |
BD Safety-Lok syringes | Fisher Inc. | 14-829-32 | |
17 mm Target Syringe filter, PTFE | Fisher Inc. | 3377154 | |
32 Karat, V. 8.0 control software | Beckman Coulter | 285512 | |
capillary electrophoresis (CE) sample vials | Beckman Coulter | 144980 | |
caps for CE vials | Beckman Coulter | 144648 | |
Liquid Nitrogen | N/A | N/A | Liquid nitrogen is available from most facilities services |