Advertisement

Analytical and Bioanalytical Chemistry

, Volume 410, Issue 22, pp 5421–5429 | Cite as

Selective labeling for the identification and semi-quantification of lipid aldehydes in food products

  • Boudewijn Hollebrands
  • Eftychia Varvaki
  • Sonja Kaal
  • Hans-Gerd Janssen
Communication
Part of the following topical collections:
  1. Food Safety Analysis

Abstract

Lipid oxidation reactions in foods rich in healthy unsaturated fatty acids result in the formation of a wide range of oxidation products that can have adverse effects on food quality and safety. To improve the understanding of oxidation reactions and methods for their inhibition, detailed information on the type and levels of the oxidation products formed is required. Accurate measurement of lipid oxidation products, especially of the non-volatile aldehyde products, has so far been a challenge due to the low sensitivity and limited specificity of most analytical methods. Here, a novel normal-phase LC method that uses selective labeling of aldehydes and epoxides with 7-(diethylamino)coumarin-3-carbohydrazide (CHH) is described. Labeling of alkanals is quantitative within 10 h. For alkenals, conversion is only around 50% at 24 h reaction time. Detailed MS identification of all aldehydes and epoxides is possible by using high-resolution MS and data-dependent MS2 acquisition. Fluorescence detection was successfully used to quantify groups of oxidation products. Sensitivity was high enough to allow accurate quantification even in fresh mayonnaises, where levels of around only 0.3 g total aldehydes/kg oil were found. Individual species can be quantified by MS if suitable reference standards are available. If no standards can be used, semi-quantification using an average response factor is an option. Clearly, the novel derivatization method is suitable for monitoring secondary lipid oxidation products in the early stages of shelf life. This makes it an important tool for developing improved food products.

Graphical abstract

Selective labeling of low and high molecular weight lipid oxidation products with 7-(diethylamino) coumarin-3-carbohydrazide for identification and semi-quantification

Keywords

Lipid oxidation Semi-quantification Aldehydes Labeling 7-(diethylamino)coumarin-3-carbohydrazide 

Introduction

Lipids, and in particular their main group of constituents, the triacylglycerides (TAGs), are an important and indispensable part of the human diet. In the last decades, food industry has successfully developed products with a healthier lipid profile: less trans fatty acids and more (poly-)unsaturated fatty acids [1]. A potential quality and safety concern of the unsaturated lipids is their higher amenability to oxidation and other reactions. In the initial stages of TAG-oxidation, primary oxidation products, such as hydroperoxides and epoxy/oxo lipids, are formed through radical initiated reactions [2]. These are unstable intermediate reaction products that decay further into stable secondary volatile- and non-volatile oxidation products, responsible for the unpleasant smell and taste of rancid products. From the food safety perspective especially, these secondary lipid oxidation products are considered a potential risk to human health [3, 4]. In particular, low molecular weight aldehydes, such as 2-propenal (acrolein), propanedial (malondialdehyde), and 4-hydroxy-nonenal, are suspected of causing adverse health effects when consumed in excessive amounts [5, 6, 7]. These aldehydes can be formed through the decay of lipid hydroperoxides, but also via further breakdown of the non-volatile secondary lipid oxidation products with an aldehyde functionality [3].

To ensure food safety, it is essential to identify and monitor low and higher molecular weight aldehydes resulting from lipid oxidation reactions. In the last decades, a wide range of methods for the measurement of total aldehyde levels as well as for measuring levels of specific (short chain) aldehydes has been developed. For a review, see [8, 9]. Non-volatile aldehydes have so far received very limited attention, which is probably at least partly due to difficulties in accurately quantifying them. In previous work, we have described a method for measuring relative levels of non-volatile secondary oxidation products, including the 2½-oxo triacylglycerides with aldehyde and epoxide functionalities [10, 11]. That method was based on normal-phase liquid chromatography (NPLC) with atmospheric pressure photo ionization mass spectrometry (APPI-MS). Although the NPLC-APPI-MS method provided valuable information on lipid oxidation and options to delay and control oxidation, for routine use, this method was not sufficiently stable, mainly also because of keto-enol tautomerism [12]. Moreover, MS sensitivity was poor and not quantitative.

Derivatization of the aldehyde functionality, preferably with an MS favorable label, could resolve both the tautomerism and sensitivity issues. Milic et al. recently described the use of a derivatization reagent, 7-(diethylamino)coumarin-3-carbohydrazide (CHH), for the analysis of carbonylated compounds in (aqueous) samples of biological origin [13]. Low and high molecular weight carbonylated phospholipids were detected and identified by direct infusion electrospray mass spectrometry [13, 14]. Based on the molecular structure of the CHH reagent, it is expected that the reaction with CHH can also be performed in non-polar solvents making this derivatization method applicable also to lipid matrices. In contrast to other derivatization reagents for aldehydes such as dinitrophenylhydrazone (DNPH), CHH offers the advantage of being detectable both by fluorescence and MS detection. An additional potential advantage of derivatization is that it might make the molar MS responses of different aldehydes more similar, and ideally identical. This would be very attractive because many of the larger aldehydes from lipid oxidation are not commercially available as pure standards.

In this contribution, we describe a novel analytical approach for the identification and quantification of low and high molecular weight aldehydes formed in lipid oxidation reactions in food products. The method uses CHH as the derivatization reagent. Derivatization conditions are optimized and the method is evaluated with regard to sensitivity, information content in the MS and MS/MS spectra, and structure dependence of the compound responses. A separation system employing normal-phase liquid chromatography with parallel fluorescence detection and electrospray ionization mass spectrometry is also described. Finally, the approach is applied for the identification and semi-quantification of aldehydes in aged oil samples in a system using high-resolution mass spectrometry.

Experimental section

Materials

Acetonitrile, isopropyl alcohol (IPA), n-hexane, and methanol (ULC/MS grade) were obtained from Biosolve BV (Valkenswaard, The Netherlands). Chloroform, ethanol, and heptane (HPLC grade) were purchased from Merck (Hohenbrunn, Germany). The derivatization reagent 7-(diethylamino)coumarin-3-carbohydrazide (CHH) and the standard compounds pentane, octane, nonane, decane, butanal, 2-pentenal, pentanal, 2-hexenal, hexanal, hexanal-d12, heptanal, octanal, decanal, deca-2,4-dienal, prop-2-enal, trans-2-butenal, trans-2-heptenal, trans-2-octenal, trans-2-nonenal, 1-pentanol, 1-pentene-3-ol, 1-pentene-3-one, 1-heptanol, and 2-octanol were purchased from Sigma-Aldrich (Taufkirchen, Germany). Nonanal was obtained from Supelco (Bellefonte, PA, USA). Propanal was purchased from Fluka Chemie GmbH (Buchs, Switzerland). All standard compounds had a purity of 95% or higher.

Standards

Stripped medium chain-length triacylglycerides oil (MCT, Fratelli Parodi, Campomorone, Italy) was spiked with the common short-chain lipid oxidation products. The concentrations varied between 8 and 17 mg/kg oil. The summed (molar) concentration levels for alkanals and alkenals were kept identical for reasons outlined below.

Samples

Mayonnaise samples with different degrees of oxidation were obtained from an accelerated shelf life test where small volumes of mayonnaise were stored in head-space vials filled with oxygen and kept at 50 °C for prolonged periods. These samples included two different formulations, one with the antioxidant E385 ethylenediaminetetraacetic acid (EDTA) and one without it. The oil phase of these samples was isolated from the samples using freeze-thaw cycling followed by centrifugation as described by Merkx et al. [15]. Two sets of samples were prepared and measured in two separate batches. In the first week, the time points 0, 4, 11, 18, and 28 days were analyzed. In the second week, the time points 1, 4, 7, 14, and 28 days were analyzed.

Labeling of compounds with carbonyl and ketone functionality in oil samples

The reagent solution was prepared by dissolving 5 mg of CHH in 1 ml of n-hexane/chloroform 1:1 (v/v). An internal standard solution was prepared by dissolving hexanal-d12 in n-hexane/chloroform 1:1 (v/v) at a concentration of 5.6 mg/L. Solutions of the standard compounds for quantification were prepared by dissolving accurate amounts of the spiked MCT oil in n-hexane/chloroform 1:1 (v/v) in individual Eppendorf vials to arrive at concentrations between 5 and 100 mg of spiked MCT oil per milliliter solvent. Concentrations of the individual aldehydes in the standards ranged from 50 to 1000 μg/L and the total aldehyde concentration ranged from 1 to 25 mg/L. For analysis of the samples from the shelf life test, small portions (25–35 mg) of the oil phases isolated from a sample were dissolved in 500 μl of n-hexane/chloroform 1:1 (v/v). Next, 10 μl of internal standard solution and 10 μl of the reagent were added. The sample solutions were incubated in the dark at controlled temperature for variable times (see below). Finally, the content of the Eppendorf vials was transferred to amber colored vials prior to analysis.

Normal-phase liquid chromatography-ESI-MS/ddMS2

The HPLC system used in the experiments consisted of an Agilent 1200SL binary pump (Agilent Technologies, Amstelveen, The Netherlands) equipped with an HTC PAL auto sampler (CTC Analytics, Zwingen, Switzerland), an Agilent 1200 series micro vacuum degasser, an Agilent 1200 series thermostatted column compartment, and an Agilent 1200 isocratic pump used for post-column addition of dopant. The mass spectrometer used was a Thermo Fisher Scientific Orbitrap Q-Exactive plus equipped with a HESI source (Thermo Fisher, Breda, The Netherlands). The optimized capillary and aux gas heater temperatures were 320 and 350 °C, respectively. The sheath and auxiliary gas flow were 50 and 10 L/min, respectively. The spray voltage was 3.6 kV. Mass spectra were acquired in data-dependent acquisition at 140 k resolution (FWHM at m/z 200). The consecutive CID fragmentation scans of the top five most abundant peaks were acquired at a resolution of 17.5 k. Both scans were performed in positive ionization mode.

A pursuit XRs 2 Diol column (particle size 3 μm, 250 × 3.0 mm, Agilent, Middelburg, The Netherlands) was used for all analyses. The analytes were eluted using a gradient program at a flow rate of 0.30 ml/min. The initial mobile phase composition was n-hexane/IPA 92:8 (v/v). Composition was then linearly programmed to 80:20 (v/v) in 45 min with a final hold of 15 min. The column temperature was maintained at 35 °C and the injection volume was 5 μl. During the analysis, a solution of 20 mM ammonium formate dissolved in ethanol/methanol 50:50 (v/v) was used to facilitate ionization in the electrospray source. The solution was pumped with a flow rate of 100 μl/min and was mixed with the column-eluent before the ion-source.

Fluorescence detection

A Shimadzu RF-20A XS LC fluorescence detector (Shimadzu, ‘s-Hertogenbosch, The Netherlands) was used for detection. Excitation and emission wavelength were set to 380 and 475 nm, respectively.

Results

The applicability of the CHH reagent for derivatization of lipid oxidation products with a carbonyl functionality was first tested using standard solutions of short-chain aldehydes. Compounds with other chemical functionalities were also tested to obtain an impression of the specificity of the derivatization reaction. Derivatization was done in the presence of stripped MCT oil (free of lipid oxidation products) to make sure the matrix would be similar to that in later analyses of edible oil samples. The spiked MCT oil contained 28 components from various chemical classes including alkanes, alkenes, alcohols, alkanals, and alkenals. Figure 1a shows an overlay of two chromatograms obtained in normal-phase LC with fluorescence detection after CHH derivatization at room temperature with 7 and 24 h incubation time.
Fig. 1

Chromatogram (a) showing the LC-fluorescence analysis of small aldehydes spiked in a medium-chain-length triglyceride oil derivatized with CHH reagent after 7 and 24 h of derivatization and a corresponding MS spectrum (b) acquired at 22.5 min using electrospray ionization in positive acquisition mode

The chromatogram in Fig. 1a clearly shows that the CHH reagent successfully labels the carbonyl compounds spiked to the MCT oil. No derivatives of alkanes, alcohols, or ketones were detected. Most peaks could be assigned by comparison with single standards, but the complexity of the chromatogram required the use of mass spectrometry for unambiguous peak annotation. MS detection, coupled in series to the NPLC-fluorescence setup, was performed in positive ionization mode. Post-column addition of a steady flow of ammonium formate dissolved in a mixture of ethanol/methanol was applied to facilitate ionization. The derivatized compounds primarily appeared as [M + H]+ ions in the mass spectra. The successful identification of the aldehyde derivates is demonstrated in Fig. 1b. Molecular formulas were tentatively assigned. For example, m/z 358.2126 and 398.2438 were here assigned to the ions C20H28O3N3+ and C23H32O3N3+. These are the protonated molecular ions of CHH-derivatized hexanal and 2-nonenal, respectively.

From the comparison of the two chromatographic profiles in Fig. 1a, obtained after 7 and 24 h of derivatization, two important shortcomings of the approach become apparent. Firstly, the derivatization reaction is not complete within 7 h. This can be seen by the increase in fluorescence signal between the two samples. Further studies indicate that for the alkenals, that were spiked to the MCT oil in our standard, the reaction was not even complete after 24 h (data not shown). Secondly, the peak profiles at 7- and 24-h derivatization time are different. The relative peak areas of the alkanal and alkenal peaks change depending on the derivatization time. This indicates that there is a difference in reaction rate between these two groups of aldehydes. The slow derivatization reaction and the differences in the reaction rates for the different aldehydes make it challenging to make the method quantitative. An attempt was made to further optimize the derivatization procedure.

Several parameters might affect the rate and efficiency of the derivatization reaction. These could include factors such as reaction time, temperature, solvent choice, presence of a catalyst, etc. To obtain an impression of the relative importance of these factors, a design-of-experiments (DOE) screening experiment was performed to determine the key parameters that influence the reaction. Special care was paid to the selection of the solvent. The CHH derivatization reaction was previously used for the derivatization of phospholipids in aqueous environments [13]. Due to the poor solubility of our samples in water or other polar solvents, the use of non-polar solvents was required. Since the derivatization reaction is a nucleophilic addition-elimination reaction where water is one of the products, it is important to adapt the derivatization conditions for use with organic solvents. The parameters included in the design are listed in Table 1. Aniline was used as catalyst following its successful use in other studies where nucleophilic addition-elimination reactions played a role [16].
Table 1

Parameters considered in the screening DOE

Parameter

Min.

Max.

Incubation time (h)

7

32

Temperature (°C)

20

60

Solvent system

Hexane/propanol

Hexane/chloroform

Catalyst concentration (mM)

0

70

Molar ratio sample: reagent

1:1000

1:5000

From the screening DOE, it could be concluded that the rate of the derivatization reaction is not, or hardly, influenced by the concentration of the reagent nor by the addition of a catalyst. The main factors determining the derivatization rates were the temperatures and reaction time. In a second DOE step, these parameters were further optimized using a response surface DOE with a total of 24 individual experiments. In Fig. 2, response surface plots are shown for different incubation temperatures, incubation times, and solvent systems. Maximum conversion was obtained at elevated temperatures using hexane/propanol as the solvent mixture. However, experimentally, these conditions were not ideal. At elevated temperatures, additional peaks appeared in the chromatograms which were identified as degradation products that were not present when using low reaction temperatures. Using the most optimal conditions, i.e., hexane/chloroform as solvent and incubation at room temperature, resulted in complete derivatization of all alkanals within 10 h without the formation of degradation products. The alkenals unfortunately were found to react much slower. For these compounds, only approximately 50% of the starting material is converted after 24 h. The reaction rates of the different alkenals appeared similar and independent of the alkyl chain length. Because of the incomplete derivatization reaction and because determination of the exact percentage of conversion of the alkenals is difficult, accurate quantification of these species is not possible. Still, if reaction times are constant and approximate conversion percentages are determined, semi-quantification is feasible. In all later experiments, a reaction time of 24 h was used for practical reasons and to obtain a better sensitivity for the alkenals.
Fig. 2

Surface response plot of derivatized aldehyde products as a function of the incubation temperature, incubation time, and solvent system

In addition to knowledge on the degree of conversion in the derivatization reaction, accurate quantification of the oxidation products also requires that the response factors of the analytes are known. For compounds that are available as pure standards, these can be determined, but if no standard materials are available, they need to be estimated. For fluorescence detection, the response of the various derivatized aldehydes was found to be equimolar. This was tested by applying the derivatization procedure to several single standard solutions with known alkanal concentrations (Fig. S1 in the Electronic Supplementary Material, ESM). MS molar responses unfortunately are not constant. The experimentally determined response factors for MS detection are summarized in Table 2. Similar analytes have similar responses, but chain length and degree of unsaturation clearly affected the molar responses. In general, response factors versus the hexanal-d12 internal standard used in our experiments ranged from 0.5 to 1.5 for the alkanals and from 0.3 to 0.6 for alkenals. The latter value being lower because of the incomplete derivatization of the alkenals. Since there is no universal MS response for the CHH-labeled aldehydes, MS-based quantification requires the use of multiple standards. Because of the large number of aldehydes that could be formed, this might not be realistic. Fortunately, for alkanals, response factors are not extremely different and the use of an average (molar) response factor is possible. In our analyses of alkanals, we used an average response factor of 1.00 vs hexanal-d12 for all aldehydes. For alkenals, the situation is more complex because the reaction for these compounds is not complete. For absolute quantification, a correction is hence needed both for differences in MS response factor, as well as for incomplete conversion. Alternatively, labeled alkenals can be added to the samples prior to derivatization to nullify the effect of incomplete reactions. This route was not pursued due to the limited commercial availability and stability of these compounds.
Table 2

MS response factors versus internal standard hexanal-d12. Data were obtained after complete conversion of single standard solutions

Alkanals

Response factor

Alkenals

Response factor

Propanal

1.52

Decadienal

0.33

Butanal

1.15

2-heptenal

0.28

Pentanal

1.43

non-2-enal

0.54

Hexanal

1.00

2-octenal

0.56

Heptanal

0.83

2-butenal

0.44

Octanal

0.82

  

Nonanal

0.85

  

Decanal

0.58

  

Despite the previously mentioned limitations, the new method has several advantages over the existing LC-MS method [11]. Firstly, detection sensitivity is much higher thanks to the very high sensitivity of both MS and fluorescence for the labeled species. Secondly, the chromatographic difficulties caused by keto-enol tautorism are eliminated by the derivatization. Thirdly, sensitive and molar-based detection of derivatized products is possible by fluorescence detection if conversion is quantitative. Finally, high molecular weight 2½-oxo triacylglycerides with aldehyde and epoxide functionalities and small volatile aldehydes such as hexanal can be analyzed with the same analytical method as will be shown below.

Various mayonnaise samples obtained from an accelerated shelf life test were analyzed with the method to demonstrate its applicability. Mayonnaise formulations with- and without the anti-oxidant E385 that had been stored at 50 °C for different periods of time (up to 28 days) were analyzed using NPLC with fluorescence and MS detection after the newly developed derivatization approach. Figure 3 shows the chromatographic profiles of the formulations recorded with fluorescence detection. Very complex patterns are obtained, where the intensities of the CHH-labeled lipid oxidation products show a strong increase at increasing storage times. The rather wide elution window in the NPLC chromatogram indicates that different classes of compounds covering a considerable range of polarities are present. Clearly, the intensities of the CHH-labeled lipid oxidation products in the formulations without E385 are significantly higher, showing that this formulation is more prone to oxidation. In the product without E385, the main oxidation products are the smaller aldehydes (retention time window 20 to 50 min). Many of these compounds are available as pure standards, so in principle they can be quantified accurately. Next to the small aldehydes, other components are present that elute earlier from the column. The components that elute before 16 min decrease at increasing storage times, but do not differ between the formulations tested here. These compounds were identified as epoxides using the high-resolution mass spectra shown in Fig. 4. From the product quality and safety perspective, they are not extremely relevant as their levels only decrease over time. Spectra of the components eluting between 16 and 22 min are also shown in Fig. 4. These are identified as 2½-TAG aldehydes. The MS spectra in Fig. 4a, b for the epoxides and 2½-TAG aldehydes show the presence of high molecular mass compounds with masses between 900 and 1200 Da. The most abundant peak in the mass spectrum of the epoxides (Fig. 4a) can be assigned to the [M + H]+ ion of C71H115N3O9 (mass error of 0.82 ppm). Characteristic fragment ions, such as m/z 894.7529, and fragment ions between m/z 550 and 650 (Fig. 4c) provide useful information on the fatty acid composition and the structure of the CHH-labeled TAG epoxide. The fatty acid chain on which the CHH label is attached has two additional double bonds, and on the glycerol backbone probably two oleic acids are present as evidenced by m/z 603.5330. The spectra clearly show the mass fragments of the CHH label at m/z 244.0968 and 262.1071. These fragments were also used in previous studies to identify CHH-labeled components [13]. The MS and MS/MS spectra of the 2½-TAG aldehydes, Fig. 4b, d, were used in a similar way to identify the labeled molecules. In this example, the CHH label is attached to an C9 alkyl chain. At the glycerol backbone, probably also two oleic acids are present. Clearly, the high-resolution accurate mass data in combination with data-dependent MS/MS acquisition is a strong tool for the identification of unknown aldehydes and epoxides in the complex lipid oxidation samples. Note that the exact location of the double bonds could not be determined. Most likely, they are still at their original positions.
Fig. 3

NPLC-Fluorescence chromatograms showing the separation of CHH-labeled oxidation products in two mayonnaise formulations with (a) and without E385 (b)

Fig. 4

MS spectra of a CHH-labeled epoxides (rt 15 min) and b 2½-TAG aldehydes (rt 19.5 min) and the corresponding MS/MS spectra, c epoxides d 2½-TAG aldehydes, of the most abundant peak in each MS spectra

Studies into the safety implications of lipid oxidation products have so far focused both on total levels of compounds, e.g., total aldehydes, as well as on the occurrence of specific analytes [3, 4]. Using our method, information on both these parameters can be obtained from one single analysis. For quantitative analysis of the total levels of aldehydes, either the MS or the fluorescence signal can be used. Fluorescence has the advantage of equimolar responses, but because of differences in degree of derivatization, class-specific standards are still needed. From literature, it is known that for rape seed oil, the main oil in our samples, the ratios between alkenals and alkanals is approximately 1:1 [17]. For that reason, our calibration standard contained these two classes in equal molar amounts. Figure 5 gives the concentration levels of epoxides, 2½-TAG aldehydes, and small aldehydes quantified based on the fluorescence responses. Using the newly developed method, even the low total aldehyde concentrations present in freshly prepared mayonnaises, around 0.6 mMol/kg isolated oil (or 0.3 g/kg isolated oil), could be readily detected. The results in Fig. 5 clearly show that lipid oxidation in the formulation with E385 (Fig. 5a) is much slower than in the product without the additive E385 (Fig. 5b).
Fig. 5

Bar-plots showing the concentration levels of TAG epoxides, 2½ TAG aldehydes, and small aldehydes in oil extracted from mayonnaise formulated with (a) and without E385 (b)

The levels of specific aldehydes are more reliably determined using MS detection. The time trends for some key markers as determined by the derivatization-NPLC-MS method are given in Fig. 6. The hexanal levels shown here (Fig. 6a) are obtained using hexanal-d12 as the internal standard. The response factors given in Table 2 were used to determine the concentration of other small common aldehydes. The limit of quantification of the labeled aldehydes present in our standard was 5 mg/kg. For a 70% fat mayonnaise, this translates to approximately 7 mg/kg. All analytes show comparable trendlines, but absolute levels differ. For analytes for which no standard was available, only semi-quantitative information could be obtained. Here, we used a response factor 1.00 relative to hexanal-d12. Based on a comparison of the total levels of aldehydes determined using fluorescence detection on the one hand, and summed levels of the main individual species obtained from MS signals on the other hand, we estimate the error in this semi-quantification to be less than a factor 2. Although far from perfect, this is a significant improvement over existing methods in which for unknown analytes only relative information could be obtained. Moreover, low levels of oxidation products that so far escaped analysis are now clearly detectable.
Fig. 6

Comparison of various labeled lipid oxidation products in two different mayonnaise formulations. The small aldehyde hexanal (a) was absolutely quantified by using an internal standard, and the levels of hydroxy nonenal (b) were determined semi-quantitative

Conclusions

A normal-phase LC method with prior CHH derivatization and fluorescence or MS detection was developed. In comparison with the currently available methods, the new methods is much more sensitive. Moreover, artifacts due to keto-enol tautomerism are eliminated. Using MS and MS/MS spectra, unambiguous compound identification is possible. A wide variety of volatile short-chain aldehydes, large 2½-TAG aldehydes, and epoxides was identified in fresh and aged mayonnaise samples. Quantification using fluorescence detection has the advantage that all analytes have equimolar responses. Correction using representative samples of known composition is needed for samples containing alkenals because of the incomplete derivatization of these analytes. The new method was successfully used in a comparative study of different lipid stabilization technologies. Total levels of 2½-TAG aldehydes increased from 0.33 mMol/kg in fresh samples to 3.4 mMol/kg after 28 days of accelerated aging. Epoxides decreased from 0.85 to 0.1 mMol/kg over the same time period. MS quantification requires pure standards, although the use of an estimated average response factor will most likely not result in dramatic errors. With MS quantification, individual species could be detected down to a level of 1 μMol/kg oil.

Notes

Acknowledgements

The authors thank Herrald Steenbergen for preparing the standards used in this research.

Compliance with ethical standard

Conflict of interest

At the time of writing, all authors were employed by Unilever, a major manufacturer of the food products studied in the manuscript.

Supplementary material

216_2018_1101_MOESM1_ESM.pdf (1.1 mb)
ESM 1 (PDF 1.14 MB)

References

  1. 1.
    Andrés SC, Zaritzky NE, Califano a N. Innovations in the development of healthier chicken sausages formulated with different lipid sources. Poult Sci. 2009;88:1755–64.  https://doi.org/10.3382/ps.2008-00495.CrossRefPubMedGoogle Scholar
  2. 2.
    Yin H, Xu L, N a P. Free radical lipid peroxidation: mechanisms and analysis. Chem Rev. 2011;111(10):5944–72.  https://doi.org/10.1021/cr200084z.CrossRefPubMedGoogle Scholar
  3. 3.
    Vieira SA, Zhang G, Decker EA. Biological implications of lipid oxidation products. J Am Oil Chem Soc. 2017;94:339–51.  https://doi.org/10.1007/s11746-017-2958-2.CrossRefGoogle Scholar
  4. 4.
    Kanner J. Dietary advanced lipid oxidation endproducts are risk factors to human health. Mol Nutr Food Res. 2007;51:1094–101.  https://doi.org/10.1002/mnfr.200600303.CrossRefPubMedGoogle Scholar
  5. 5.
    Kubow S. Routes of formation and TOXIC consequences. Free Radic Biol Med. 1992;12:63–81.CrossRefPubMedGoogle Scholar
  6. 6.
    Csala M, Kardon T, Legeza B, Lizák B, Mandl J, Margittai É, et al. On the role of 4-hydroxynonenal in health and disease. Biochim Biophys Acta - Mol Basis Dis. 2015;1852:826–38.  https://doi.org/10.1016/j.bbadis.2015.01.015.CrossRefGoogle Scholar
  7. 7.
    Guillén MD, Goicoechea E. Toxic oxygenated α,β-unsaturated aldehydes and their study in foods: a review. Crit Rev Food Sci Nutr. 2008;48:119–36.  https://doi.org/10.1080/10408390601177613.CrossRefPubMedGoogle Scholar
  8. 8.
    Barriuso B, Astiasarán I, Ansorena D. A review of analytical methods measuring lipid oxidation status in foods: a challenging task. Eur Food Res Technol. 2013;236:1–15.  https://doi.org/10.1007/s00217-012-1866-9.CrossRefGoogle Scholar
  9. 9.
    Spickett CM. The lipid peroxidation product 4-hydroxy-2-nonenal: advances in chemistry and analysis. Redox Biol. 2013;1:145–52.  https://doi.org/10.1016/j.redox.2013.01.007.CrossRefPubMedPubMedCentralGoogle Scholar
  10. 10.
    Steenhorst-Slikkerveer L, Louter A, Janssen H-G, Bauer-Plank C. Analysis of nonvolatile lipid oxidation products in vegetable oils by normal-phase high-performance liquid chromatography with mass spectrometric detection. J Am Oil Chem Soc. 2000;77:837–45.  https://doi.org/10.1007/s11746-000-0134-1.CrossRefGoogle Scholar
  11. 11.
    Hollebrands B, Janssen H-G. Liquid chromatography—atmospheric pressure photo ionization-mass spectrometry analysis of the nonvolatile precursors of rancid smell in mayonnaise. LC-GC Eur. 2017;30(9):470–83.Google Scholar
  12. 12.
    Keeffe JR, Kresge a J, Schepp NP. Keto-enol equilibrium constants of simple Monofunctional aldehydes and ketones in aqueous solution. J Am Chem Soc. 1990;112:4862–8.  https://doi.org/10.1021/ja00168a035.CrossRefGoogle Scholar
  13. 13.
    Milic I, Hoffmann R, Fedorova M. Simultaneous detection of low and high molecular weight carbonylated compounds derived from lipid peroxidation by electrospray ionization-tandem mass spectrometry. Anal Chem. 2013;85:156–62.  https://doi.org/10.1021/ac302356z.CrossRefPubMedGoogle Scholar
  14. 14.
    Ni Z, Milic I, Fedorova M. Identification of carbonylated lipids from different phospholipid classes by shotgun and LC-MS lipidomics. Anal Bioanal Chem. 2015;407:5161–73.  https://doi.org/10.1007/s00216-015-8536-2.CrossRefPubMedGoogle Scholar
  15. 15.
    Merkx DWH, Hong GTS, Ermacora A, van Duynhoven JPM (2018) Rapid quantitative profiling of lipid oxidation products in a food emulsion by 1H NMR. Anal Chem.  https://doi.org/10.1021/acs.analchem.8b00380
  16. 16.
    Dirksen A, Dawson PE. Rapid oxime and hydrazone ligations with aromatic aldehydes for biomolecular labeling. Bioconjug Chem. 2008;19:2543–8.  https://doi.org/10.1021/bc800310p.CrossRefPubMedPubMedCentralGoogle Scholar
  17. 17.
    Guillén MD, Ruiz A. Monitoring the oxidation of unsaturated oils and formation of oxygenated aldehydes by proton NMR. Eur J Lipid Sci Technol. 2005;107:36–47.  https://doi.org/10.1002/ejlt.200401056.CrossRefGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany, part of Springer Nature 2018

Authors and Affiliations

  • Boudewijn Hollebrands
    • 1
  • Eftychia Varvaki
    • 1
  • Sonja Kaal
    • 1
  • Hans-Gerd Janssen
    • 1
    • 2
  1. 1.Unilever R&D VlaardingenVlaardingenThe Netherlands
  2. 2.Van ‘t Hoff Institute for Molecular SciencesUniversity of AmsterdamAmsterdamThe Netherlands

Personalised recommendations