MSc Chemistry (jd)
Analytical sciences
Literature Thesis
The analysis of non-volatile Maillard
reaction products in food matrices.
By
B.Sc. Manish Jhinkoe-Rai
12424625
06-2020
12 ECs
August 2019 – February 2020
Supervisor/Examiner:
Examiner:
Prof. Dr. P. Schoenmakers
Dr. R. Haselberg
2
The Maillard reaction is a non-enzymatic browning reaction that can occur in food products during different stages of its lifecycle. Different parameters are important for the formation of (non-) volatile Maillard reaction products. The Maillard reaction is divided into three stages: initial, intermediate and advanced stage. It became clear that different methods can be used during different stages to detect or identify non-volatile Maillard reaction products. This is depicted in the Hodge scheme. In this report, several analytical methods were discussed for the analysis of non-volatile Maillard reaction products in different stages. These analytical methods included spectroscopic, chromatographic and mass spectrometric features. From the literature, it became clear that separation of an analyte from a sample is more important during the intermediate and advanced stage, due to complex matrices. The main question posed in this report is: ‘What analytical methods have been used for the detection of non-volatile Maillard reactions products in food matrices?’ If the goal of a research is to quantify or to detect a known compound, spectroscopic methods could be sufficient enough. Furthermore, if the goal of a study is structure elucidation, the inclusion of a mass spectrometer in the process is necessary. Liquid chromatography is stated to be more efficient for separation of non-volatile Maillard reaction products than capillary electrophoresis. However, capillary electrophoresis is becoming nowadays more interesting as alternative for liquid chromatography. Combining liquid chromatography with mass spectrometry, it was observed that this is a good option for analysis.
Acknowledgement
I would like to thank Boudewijn Hollebrands and Hans-Gerd Janssen for the opportunity, patience and feedback during the previous months.
3
TABLE OF CONTENT
CHAPTER 1: INTRODUCTION ... 4
CHAPTER 2: THE HODGE SCHEME ... 7
CHAPTER 3: SPECTROSCOPIC DETECTION METHODS ... 9
3.1 INTRODUCTION ... 9
3.2 LAB ... 9
3.3. FTIR SPECTROSCOPY ...10
3.4. FLUORESCENCE SPECTROSCOPY ...14
3.5. NMR SPECTROSCOPY ...17
CHAPTER 4: HYBRID METHODS ...18
4.1 INTRODUCTION ...18
4.2 LIQUID CHROMATOGRAPHY – FLUORESCENCE SPECTROSCOPY ...19
4.3 LIQUID CHROMATOGRAPHY – DIODE ARRAY DETECTION (LC-DAD) ...22
4.4 LIQUID CHROMATOGRAPHY – FOURIER TRANSFORM INFRARED SPECTROSCOPY ...24
4.5 LIQUID CHROMATOGRAPHY – UV-VIS SPECTROSCOPY ...26
CHAPTER 5: DETECTION WITH MASS SPECTROMETRY ...28
5.1 INTRODUCTION ...28
5.2 LIQUID CHROMATOGRAPHY – TANDEM MASS SPECTROMETRY ...28
5.3 LIQUID CHROMATOGRAPHY – FOURIER TRANSFORM -ION CYCLOTRON RESONANCE – MASS SPECTROMETRY ...33
5.4 LIQUID CHROMATOGRAPHY – DIODE ARRAY DETECTION – ELECTROSPRAY IONIZATION – MASS SPECTROMETRY ...34
5.5 TIME-OF-FLIGHT MASS SPECTROMETRY and MALDI-TOF-MS ...35
5.5.1 HPLC-ESI-TOF-MS ...36
5.5.2 MALDI-TOF-MS...37
5.6. CAPILLARY ELECTROPHORESIS – MASS SPECTROMETRY ...39
CHAPTER 6: CONCLUSION ...42
4
CHAPTER 1: INTRODUCTION
The Maillard reaction is a non-enzymatic browning reaction which can take place when food is stored, dried, treated or in vivo in mammals at physiological temperature and pH [01]. This reaction was named after the French chemist Louis Maillard (1878 – 1936). The Maillard reaction corresponds to a set of reactions occurring between amines and carbonyl compounds, especially between reducing sugars and amino acids or proteins [02][03]. The Maillard reaction In food, this reaction is responsible for both changes in colour, flavour and nutritive value as well as for the formation of stabilizing and mutagenic compounds [02]. Therefore, it can be stated that the Maillard reaction is an important chemical reaction in our daily life.
The Maillard reaction is not only of huge importance in the food industry, but also in the medical, tobacco and water industry [03,011]. The Maillard reaction occurs for example: when bread is roasted, when meat is being cooked and when coffee beans are roasted. As consequence of the Maillard reaction, consumers can smell and taste the flavour of a product. The smell of a product is the consequence of the release of volatile compounds, whereas the taste is the consequence of the formation of non-volatile compounds.
The Maillard reaction is a complex network of reactions and can be divided in three stages: the initial stage, the advanced stage and the final stage [03]. The overall reaction mechanism is shown in figure 1.
During the initial stage, an amino group of the amino acids condenses with a carbonyl group of reducing sugars to form a Schiff base. During this reaction step, a N-substituted gycosylamine will be formed, which is then directly followed by either the Amadori or the Heyns arrangement to form a 1-amino-1-deoxy-ketose (1,2-enol form), which is the Amadori reaction product [07]. The formation of 2-amino-2-deoxy-aldose is the Heyns product [07].
During the advanced stage, many reactions can take place, in which the most important is the enolisation. Other reaction possibilities are degradation of the Amadori/Heyns products via deoxyones, fission or Strecker degradation, but also via different reactions types such as: dehydration, elimination, cyclization, fission and fragmentation. During the advanced stage, Amadori or Heyns compounds decompose at high temperatures to form reactive α-dicarbonyls compounds, for example: 4-deoxyhexulose [04, 07]. Fragmentation of these hexuloses and the degradation of the Amadori and/or Heyns products produce reactive α-dicarbonyl compounds, for example: glyoxal. An example of a Maillard reaction intermediate is acrylamide, which originates from the Maillard reaction, in which the Asparagine amino acid reacts with a sugar [04].
During the final stage, amino compounds as well as sugar fragments condense in order to form the brown coloured nitrogenous compounds. These final coloured compounds can be divided into two groups: low and high molecular weight compounds. Low molecular weight compounds are approximately two to four linked rings of heterocyclic compounds [04]. The high molecular weight compounds are called melanoidins. Low molecular weight compounds are typically two to four linked rings and considered as volatile compounds, whereas high molecular weight compounds consist of long polymeric chains considered non-volatile compounds [05,09]. Different factors can influence the development of Maillard reaction products, such as: temperature, time, initial pH, water activity, physical state of the matrix, reactant concentration, type of carbo-hydrate and the ratio of reducing sugar to lysine [08]. The carbo-hydrates involved in the reaction are monosaccharide and reducing disaccharides. These non-reducing disaccharides have to be hydrolysed, before these can take part in the reaction [01].The rate
5
of the Maillard reaction is related to the water activity of the food matrix. During storage under mild temperatures, sugars could crystallize and can therefore cause an increase in the rate of the Maillard reaction due to release of water from the reducing sugar [08]. In order to avoid an increase in the rate of the reaction, food must be stored under low temperatures, low pH and must have low water activity [08]. At neutral or acid pH, Amadori products as well as Heyns products undergo 1,2-enolization producing hydroxymethylfurfural (HMF) and furfural. At high pH, Amadori and Heyns product undergo 2,3-enolization, producing reductones and fission products [010].
Figure 1: A general scheme of the Maillard reaction occurring in food between a sugar and an amino compound [02].
It is important for companies that the Maillard reaction takes place when the food product is being treated or consumed and not when the food product is stored, for example in shops. Due to this reaction, a consumer experiences a particular taste from a product. Companies want their consumers to get the best experience possible. However, when this reaction occurs during storage or before consumption of the product, the food product will eventually lose its quality. This will we be experienced as not good by the consumer, which will a big impact on a company’s imago. Therefore, identifying and detecting non-volatile Maillard reaction products have become an important interest in recent years [06].
Volatile compounds are the first compounds on which a consumer bases its opinion. When consumers do not like the smell of a product, the consumer will likely not continue consuming that product. Therefore at first, the analysis of volatile compounds was studied. The outcome of that literature study was that those compounds could be detected by using a GC-head
6
space. Volatile and non-volatile compounds are both responsible for the consumer’s experience. Since successful methods were developed to analyse the volatile Maillard reaction products, it has also been desired to develop successful methods to analyse the non-volatile Maillard reaction products in food matrices. Therefore this literature study was started to study which analytical methods are used for the analysis of non-volatile Maillard reaction products in different food products. The main question posed in this report is: What analytical methods have been used for the detection of non-volatile Maillard reactions products in food matrices? Reviewing different analytical methods will be done because there will be (many) technical and chemical controllable and uncontrollable factors playing a role. Nevertheless, to answer this question, this report will discuss different types of analytical methods in different chapters. In chapter 2, the Hodge scheme will be presented. Then in chapter 3, the use of spectroscopic methods will be discussed. In chapter 4, chromatographic methods combined with spectroscopic detectors will be discussed. In chapter 5, the inclusion of different mass spectrometers in combination with liquid chromatography will be discussed as well as the use of capillary electrophoresis instead of liquid chromatography. Finally, in chapter 6, the conclusion of this study will be provided.
7
CHAPTER 2: THE HODGE SCHEME
In this chapter, the Hodge scheme will be presented, which is an overview of different analytical methods that can be used to investigate the Maillard reaction during different stages and in different food and drink matrices. This scheme was established by John E. Hodge and is therefore called the Hodge scheme, see figure 2 [29]. Different analytical methods were discussed and proven to be useful for the detection of non-volatile Maillard reaction products [29].
Figure 2: The Hodge scheme [29].
The scheme suggests that during the initial stage: LC-MS, NMR, UV-Vis, MALDI-TOF, LC-UV (λ=280 nm) can be used for detecting Maillard reaction products. This is because these methods cannot produce lots of information when big mixtures of compounds are formed, such as melanoidins.
During the intermediate stage, hybrid methods are used. These are needed because of the more complex matrices. The implementation of liquid chromatography is therefore observed. LC-fluorescence, LC-MS, NMR and LC-UV (λ=320 nm) are commonly used.
During the advanced stage, mainly MALDI-TOF is used. This is done, so compounds can get easily ionized before injected into the mass spectrometer.
So, from the Hodge scheme, three major groups can be made. A fourth group was added for because of the interest of Unilever, so that it was checked whether liquid chromatography could be switched for capillary electrophoresis as a separation method.
8
• Group 1: Detection methods such as LAB, FTIR, fluorescence and NMR. (chapter 3)
• Group 2: Liquid chromatography (LC) with spectroscopic detection methods, such as: LC-fluorescence, LC-DAD, LC-FTIR and LC-UV-Vis
(chapter 4)
• Group 3: Liquid chromatography – mass spectrometry (MS), such as: LC-tandem-MS, LC-FT-ICR-MS, LC-DAD-ESI-MS, LC-TOF-MS, LC-MALDI-TOF-MS
(chapter 5)
• Group 4: Capillary electrophoresis (CE) – mass spectrometry (chapter 5)
Groups 1 and 2 are used for detection of presence of non-volatile Maillard reaction products in matrices, whereas groups 3 and 4 are used for both detection of presence of Maillard reaction products as well as for structure elucidation of non-volatile Maillard reaction products. In group 1, only detection methods are being used, whereas in group 2 also separation methods are being used in combination with the detection methods. In the 3rd group, the
spectroscopic methods have been switched for mass spectroscopic methods.
These different analytical methods will be discussed and it will be seen whether these methods are proven to be useful for the detection of non-volatile Maillard reaction products.
9
CHAPTER 3: SPECTROSCOPIC DETECTION METHODS
3.1 INTRODUCTION
In this section of the report, spectroscopic methods for the detection of Maillard reaction products will be discussed. The discussed methods are: LAB, Fourier-Transform infrared spectroscopy (FTIR), fluorescence and nuclear magnetic resonance (NMR). These methods are considered non-specific, since they lack the use of prior separation and can also not provide specific information about the molecular structures. The detection of compounds is based on spectroscopic features. These methods can be used easily on large scale, so bulk samples can be analysed quickly on macro specific properties, such as colour and specific bonds. The use, advantages and disadvantages of these methods will be discussed for each method.
3.2 LAB
During the Maillard reaction, different products are formed. These products range in colour from pale yellow to golden to cinnamon to brown [1, 6]. LAB is a method that can be used for the detection of the colours of the Maillard reaction products. The information about the colours will be given in three ways with LAB: ‘L’, ‘A’ and ‘B’. ‘L’ stands for the range in black-white colour for components, also referred as luminosity. ‘A’ stands for the range in red to green colours of the components and ‘B’ stands for the range of yellow to blue colours for components, see figure 3 [1, 9].
Figure 3: Representation of LAB, composing out of black-white (L), green-red (A) and blue-yellow (B).
First of all, a distinction must be made whether the colour formation is due to caramelisation or due to the Maillard reaction. It must be realized that caramelisation is only possible at high temperatures (T>120oC). If the temperature during analysis does not reach the 120 oC or more,
brown compounds (melanoidins) formed in the sample will indicate the start of the advanced stage of the Maillard reaction [1].
Quantitative measurements for the brown compounds can be considered as a quick indicator of heat treatment. It is necessary to have a stable reference in that case for the sample, which can be difficult as in the case of milk [1]. These measurements can be performed by measuring the absorbance at a wavelength of 420 nm for melanoidins [1, 2, 3]. These single wavelength measurements of browned solutions at 420 ≤ λ ≤ 460 nm, are frequently used to measure the rate and extent the formation of the coloured compounds during the Maillard reaction. These wavelengths are typically used in the advanced stage. During the intermediate stage, another wavelength of 294 nm is used [2].
10
This method can be used as an indication about the extent of the reaction. However, it is not a reliable way to describe visual colour changes in terms of visual properties of the brown pigments [1]. This method is not providing information about the change of the compounds. So this method is therefore considered as non-specific.
Theoretical equations have been made in order to predict the extent of brown pigments as a function of time, but the only disadvantage of this equation is that it cannot be solved because of the amount of variables, such as rate constants of intermediary reactions [1]. Therefore, another equation (1) which is useful is provided:
E = (L²+A²+B²)½ (1)
where E (1) is an index which describes how apart two samples are in the colour space and where L, A and B are values from the measurements [1, 8]. This E index is influenced by the colour lightness, so an increase in E is related to a gain of lightness. E decreases when the system is heated [9]. This makes sense, since it means that the system becomes darker. During the Maillard reaction, a decrease in ‘L’ and in ‘B’ was noticed [1]. This means that the luminosity decreased and yellow compounds were formed. In heated systems, a stronger decrease of ‘L’ was observed, which can partially be explained by the formation of brown products. These products are being formed during the advanced stage of the Maillard reaction. By combining results from both LAB and browning tests, a quick observation of the Maillard products can be realized. This is only possible if the coloured products can be isolated and UV-VIS measurements can take place.
From earlier research, it is known that:
• tryptophan is responsible for the UV absorbance of Maillard reaction products from tryptophan-sugar complexes;
• MRPs prepared from cysteine-sugar complexes show characteristic absorption peaks at 330 nm;
• MRPs prepared from other amino acids show absorbance at around 265 nm; • the type of sugar is important for the UV-VIS absorbance wavelength [2].
It can be concluded that the absorbance wavelength is not specified for Maillard reaction products, since they can differ. Only the browning rate can be scanned quite easily. It provides information about what stage the reaction is, but it does not provide information on the molecules that have been formed. An advantage of LAB is that LAB is not time-consuming at all. On the contrary, the information provided will not be enough for structure elucidation or quantification. In most cases LAB will only confirm that what is being observed. Besides that, other easier and more convenient methods are available. Measuring the browning rate can be effective in order to get a quick idea of the stage of the Maillard reaction.
3.3. FTIR SPECTROSCOPY
Fourier transform infrared spectroscopy (FTIR) can be used for the analysis of melanoidins in food content [3, 8, 61]. It is also important to realize that a blank measurement has to be taken before analysing food samples. Samples containing melanoidins will show peaks at different wave numbers, see table 1. Because of the fact that melanoidins consist of a lot of hydroxyl moieties, the –OH bond will be observed quite easily because the -OH bond is very dominant presented [3]. Also a band at ν=1675 cm-1 will be strongly visible in all melanoidins. The band
11
will appear, some will disappear and some will change slightly in intensity in the spectrum if the Maillard reaction has taken place vs. when the Maillard reaction has not taken place yet. Since melanoidins show a strong shoulder at ν= 1710 cm-1 and show strong vibrations between
ν=1030 cm-1 and ν=1150 cm-1, a quick FTIR-analysis can be performed for the analysis of the
presence of melanoidins by measuring at these wave numbers [3, 61]. It is also known that all melanoidins show vibrations at ν= 1650 cm-1 because of the presence of C=O bonds [3].
Table 1: Wavenumbers with their corresponding bonds measured with FTIR
Wave number (cm-1) Bond Reference
3600-3000 -OH (dominant) 3 3080 -NH+ 3 2943 -CH3 3 2915 -CH2 3 2890-2600 C-H stretch 3 1675 C=O stretch 3 1630 COO-, C=C or C=N 3, 8 1406 RC(=O)R’ 8 1344 -OH 8 1030-1150 C-O 3
There exists a common scheme for structural details formed from some basics sugar complexes and amino acids, as mentioned above in table 1. In degraded products, the content of sp² hybridized carbon is higher, leading to more pi-conjugated systems which enhances absorbance [3]. For these compounds, FTIR measurements should include analysis at ν= 1710 cm-1 and between 1030≤ ν ≤1150 cm-1. Moreover, the fingerprint region must be analysed. The
fingerprint region is a region from approximately 500 cm-1 to 1500 cm-1, which contains very
complex series of absorption. These absorption peaks are observed on a spectrum due to internal vibrations of a molecule. This fingerprint region is different for each molecule and can therefore be used as a discriminating factor for structure elucidation. In particular, the intensity of the band with wavenumber 1710 cm-1 is very important for the differentiation whether
melanoidins are being formed or not yet [3].
According to Mohsin, this method is very rapid and accurate. It can therefore be easily used for the control of non-volatile Maillard reaction products. Besides that, FTIR is also considered very reliable and effective. Therefore, this method is considered as more effective than LAB. However, in complex matrices, sample preparation or analyte separation has to be performed before detection with FTIR can take place. If not, multiple peaks will be presented on a spectrum, causing it to be difficult to understand. This method can be used for amino acids, Amadori compounds and sugars as well. Measurements with this method will be more difficult for proteins and peptides.
This method seems to be very effective and can be easily performed. It is also comfortable to have clear wave numbers (regions) where specific bands for an AGX complex (a complex of arginine, glycine and d-xylose) are shown, see figure 4. By doing this, it can be stated whether Maillard reaction takes place or not. Therefore, this method can be considered useable for detection [4].
In figure 4, it can be seen that intensities of absorbance increase at higher temperatures and with longer reaction time. These are conditions that should be taken into account. In figure 4, colour formation can take place due to the Maillard reaction, but could also take place due to the caramelization reaction. This is because the temperature reached 150 oC. Another reason
12
for the colour formation could be explained by thermal degradation. However, according to figure 5, thermal degradation would not be the case at low temperatures which are usually present during storage of food products [26]. However, thermal degradation still takes a major role in the formation at high temperatures, see figure 5. During the analysis of food content stored in packages under storage circumstances, the temperature will most likely not be above the 80 oC. It can therefore be stated that the changes in absorbance at different wave numbers
will not be due to the caramelization reaction, nor due thermal degradation but due to the Maillard reaction. Nevertheless, it cannot be concluded that the formation of Maillard reaction products will be the only reason for the increase in absorbance.
The region between ν= 1000 cm-1 and ν= 1100 cm-1 was considered as interesting. The
absorbance intensities of a complex formed during the Maillard reaction were measured [4]. The biggest difference was observed between the wavenumbers 900 cm-1 and 1150 cm-1 [3,
4]. Therefore pH and absorbance measurements were performed at 1000 cm-1 and 1043 cm-1
, see table 2. No explanation was provided why the pH of the solution dropped. However this must be at least due to the formation of Maillard reaction products, since these are formed with longer reaction times and with high temperatures, see figure 6 [26].
13
Figure 5: Development of absorbance at 420 nm in Maillard reaction products and thermal degradation products as function of time [26].
Table 2: The IR absorbance values of the AGX complex compound with pH changes of the solution during the Maillard reaction [26].
Figure 6: pH and temperature correlation of Maillard reaction products and thermal degradation products [26]. So, FTIR is mainly useful in the advanced stage. Also now, no structural information is being provided, but only an indication about the stage can be deduced from the data. This is a non-invasive method and it does not require contrast agents [4]. Yet, this technique has two major disadvantages which are: (1) a diffraction limit and (2) difficulties with water absorption signals [4]. A diffraction limit means that there is a maximum spatial lateral resolution that ranges from 3 to 10 μmin the mid-IR spectral range [4]. Even though this method has 2 disadvantages, this
14
Method can still be easily used. It is even possible to perform such an analysis without prior labelling [4].
3.4. FLUORESCENCE SPECTROSCOPY
Fluorescent compounds as well as brown pigments are formed during the advanced stage of the Maillard reaction [5, 6, 7]. This is because the Amadori compounds undergo dehydration and afterwards are considered as fluorophores and as fluorescence compounds [5,6,7]. Fluorophores are precursors of the brown pigments, but are not identical to them [6]. In the last decades, fluorescence has been specified as a more specific method, compared to other spectroscopic approaches. The total amount of fluorescent intermediate compounds (FIC) in a sample consists of: (1) FIC and (2) linked-to-protein fluorescent compounds [5]. FIC is strongly related to the protein content of the sample and fluorescence is mainly linked to the backbone of a protein [7]. For the determination of FIC, it is important to take several variables into account, such as: pH, incubation time, enzyme concentrations, quenching and clarification steps. Addition of different ingredients must also be taken into account, since some additives can induce an increment of the FIC values [7]. At high pH values, especially above a value of 4.6, fluorescent sensitivities decreased [8]. FIC values were determined in different food matrices, such as presented in table 3. These FIC values were measured for different types of cereals, which are consumed during breakfast. As a consequence of the high protein level in the wheat-based group, a significant relationship was found between furosine content and FIC values, which indicated that free fluorescence compounds could be related to blocked lysine [7].
It can be said that fluorescence intensities have been determined by a fluorescence spectrometer, see table 3. The required sample preparation needed, is extensively described [7]. So, fluorescence intensities could be measured quite easily.
The Maillard reaction products formed in the intermediate stage or in the advanced stage of the Maillard reaction show fluorescence. This implies that fluorescent compounds are formed either earlier or later than the brown pigments or other colour formation. The fluorescent compounds are formed earlier than brown pigments under favourable circumstances and are detected even before any colour formation has visibly occurred [6]. This means under neutral pH, with high water activity, with presence of phosphate buffer and with accelerating salts. Consequently, it can be stated that fluorescent compounds are formed later than brown pigment markers under unfavourable circumstances, such as: low pH or/and presence of retardant salts [6].
Fluorescence measurements were performed for glucose-glycine systems, see figure 7 [6]. These systems of glucose-glycine were academic samples and did not origin from any food content. However, this glucose-glycine can easily be formed in food. Because the system was an academic sample, no separation method was needed to prior analysis by fluorescence. The maximum fluorescence intensity is approximately at a wavelength of 435 nm.
Another example contains a fluorescence intensity spectrum of d-glucose-L-arginine (GLA) system measured at room temperature [8]. Also this time, the GLA system was an academic sample. Therefore, no separation methods were necessary. An excitation wavelength of 334 nm was used [8]. These fluorescence intensities and the concentration are in a linear relationship at a specific wavelength.
15
Table 3: Statistical treatment for free and total FIC present in breakfast cereals grouped according the type of cereal, possible consumers, dietary fibre, protein content, presence of honey or cocoa powder and final physical form of the product (factor) [7].
* a) Value represent means +/- SD. | b) Number of samples.
Figure 7: Emission spectra of fluorescent Maillard reaction productsfrom glucose–glycine systems in different buffers with and without salts(excitation wavelength was 365 nm). The measurements were per-formed at 0 days (time = 0) or after 26 days of incubation at 55C forall systems. Ph6: phosphate buffer, 0.1 M, pH 6.84. Ph5: phosphatebuffer, 0.1 M, pH 5. Cit5: citrate buffer, 0.1 M, pH 5. Ac5: acetate buffer, 0.1 M, pH 5. Ac6: acetate buffer 0.1 M, pH 6.84 Mg, Li or Na:samples with MgCl2, LiCl or NaCl,
respectively.426S.B. [6].
Quenching is phenomenon in fluorescence that can occur. It functions as an inner filter effect, which is also dependent on the pH of the solvent, see figure 8. The fluorescence intensities decrease with the introduction of d-isoAA or TA [8]. These are considered as quenchers. No limits of detection and limits of quantification were published, however the recovery was presented including relative standard deviation, see table 4. It can be concluded that a high recovery is perceived as well as a low RSD.
Determination of fluorescent compounds under unfavourable circumstances is not a suitable method, since absorbance in the visible range was detected earlier than the fluorescence [6]. If the requirements for favourable circumstances would be met, the fluorescent compounds could be seen as a good and an early marker for the intermediate stage of the Maillard reaction [6]. But the brown pigments could be observed earlier than the fluorescent markers. Therefore,
16
this method cannot be considered as a sensitive and reliable indicator as well as an universal indicator of the Maillard reaction. Nevertheless, this method can be used for the analysis of the kinetics of the reaction, when no visible changes are observed.
Figure 8: Effect of pH values of buffer solution on the relative fluorescence intensity of GLA-KMnO4 in presence of d-isoAA or TA [8].
Table 4: The determination of d-isoAA and racemic TA mixture, for n=5. N stands for the number of replicates [8].
Typical curves for fluorescence can be observed during analysis. In the beginning, a small amount of fluorescent compounds is detected, which is also the case for browning pigments [6]. This indicates that measurements for both browning as fluorescence are equally sensitive. But as a function of time, the amount of fluorescent compounds increases since no saturation effects take place for fluorescent compounds, as it takes place for the browning pigments [6]. From previous researches, it has become clear that fluorescent compounds from the Maillard reaction have a maximum excitation at wavelengths between the 340 ≤ λ ≤ 370 nm [6]. Also fluorescent wavelengths between the 420 ≤ λ ≤ 450 nm are considered properties of MRPs. Therefore, fluorescence can be measured at λ= 415 nm, with an excitation wavelength of 347 nm. This can be done because there was a good correlation between the fluorescence and the formation of Maillard reaction products [2, 5, 9]. A stoke shift of 50 nm was used in order to eliminate background noise [8]. An advantage is that these wavelengths do not interfere with any other fluorescent amino acid, such as tryptophan [6]. So, fluorescence experiments can be used for the determination of Maillard reaction products in glycated systems, see figure 7. This can be achieved in food matrices, such as was done in cereals, but can also be done for the glycated-amino acids which are academic samples, see figure 7.
Fluorescent Maillard compounds measurements are more specific and provide more information on the extent of the Maillard reaction than other unspecific tools to monitor the reaction. As first approach, this method is also suitable to assess the nutritional quality of food as related to protein damage [5].
This was compared to UV-absorbance and spectrophotometric tri-stimulus colour measurements [5]. This method can also be used for the detection of MRPs, also when food matrices have been put for storage [5, 7]. Furthermore, fluorescent markers are considered to be an early marker of the presence of MRPs [5, 7]. This method is considered easy, rapid and accurate. Also, low quantities are needed for analysis. This method can be seen as an
17
adequate marker of the Maillard reaction [7]. Fluorescence measurements are easy, accurate and cheap. Potentially that the use of a multidetector could improve results, since both absorbance and fluorescence measurements will take place.
3.5. NMR SPECTROSCOPY
NMR spectroscopy was used in order to determine the structure of Amadori compounds [47, 48]. Solid state NMR spectroscopy can also be used in the study of melanoidins [61]. Prior to NMR spectroscopy, separation by HPLC has to be performed for non-volatile compounds in a complex matrix [47, 48, 49, 50, 61]. Good separation by HPLC causes better resolution in spectra. NMR was used in order to determine the structure of Amadori compounds [47]. 13
C-NMR of non-volatile compounds can be performed in ²H2O [47, 48]. Besides the 13C-NMR, it
is also possible to perform 1H-NMR, 19F-NMR or 17O-NMR [49]. An important detail in 13C-NMR
is that assignments of chemical shifts to the sugar moiety could not be made with any precision. So, the same chemical shifts were assigned to different carbon atoms in Amadori compounds. This did not lead to a unique discrimination of data. NMR spectroscopy on its own is not sufficient enough for complete structure elucidation of melanoidins [61]. There is a need of another supporting detection method. Most of the time, (high resolution) mass spectrometry was used next to NMR [49, 50]. However, NMR can be used in order to detect the presence of Maillard reaction products, see figure 9 [4].
Figure 9: The comparison of 1H NMR spectra of AGX mixture solution at 150 °C for 0 and 60 minutes [4].
From these data, it can be seen that different peaks are observed at different chemical shifts. As can be seen, the area between δ=3 ppm and δ=6 ppm did not differ that much. Therefore, interested regions 1 and 2 are used as reference points. The composition of the sample changed after 60 minutes, see figure 9. This is most likely due to the Maillard reaction, which causes new formations of bonds. However, the temperature was set at 150 °C in this case. In the case of storage, the temperature will most likely not reach the 150 °C. In that case, it might be interesting to check whether NMR peaks will also be different as seen in figure 9. If a match could be found between a sample and a reference compound, identification could take place. So based on the information above, NMR spectroscopy could be used in the study. It was not stated that the interested regions where always the same for different Maillard reaction products, so these can differ for each Maillard reaction product. It also takes time to separate the mixture. It is preferred to use high resolution mass spectrometry for structure elucidation. However, it must be stated that NMR spectroscopy can be used afterwards as validation method for the provided structure. Not so many studies had included NMR spectroscopy into their research.
18
CHAPTER 4: HYBRID METHODS
4.1 INTRODUCTION
In the previous chapter, several spectroscopic methods were used for the detection of Maillard reaction products. Yet, these methods provide some insight in complexes that are formed during storage on molecular level in products. It can be useful to use these methods when an indication of the formation of Maillard reaction products is desired. Non-specific methods do not gain enough of information about the molecular structure of Maillard reaction products. Even more, large quantities of food matrices can be difficult to analyse when no separation has been performed. This can be due to a high limit of detection of a certain method for a particular analyte. However, sometimes the inclusion of separation methods can be useful or necessary in order to get more specific details about the composition of the Maillard reaction products. This is a reason to perform separation before starting detection.
Hybrid methods are separation methods which are connected to detection methods. High performance liquid chromatography (HPLC) is a separation method that is commonly used for the separation of Maillard reaction products. HPLC can separate non-volatile water-soluble compounds without prior derivatization, which is considered as an advantage [11]. However, a disadvantage of liquid chromatography is that in case a gradient is used, the column must be conditioned at the beginning of the each analysis. So that good chromatograms as well as no irregular baselines (noise) can be observed [15]. Furthermore, some variables that are important for HPLC are: temperature, pressure, time and pH. Nevertheless, HPLC has a limited ability to resolve and provide structure sensitive information. On the contrary, HPLC is very robust and easy. The best is that HPLC can be coupled to several different detectors, such as mass spectrometers or spectrometers.
HPLC combined with spectroscopy versus spectroscopy alone are both used for the detection of presence of Maillard reaction products. HPLC-spectroscopy vs. spectroscopy, shows that spectroscopy alone is not that sensitive and selective enough. A HPLC method was used for the detection of HMF as well as a spectrophotometric white method [52, 53]. It became clear that spectrophotometric white method works well for small concentrations according to the data, see tables 5 and 6 [52].
Table 5: Recovery of HMF added to honey and determined by HPLC and spectroscopic white methods [52].
At low concentrations, more deviations were found for the recovery percentages by HPLC- spectroscopy. At higher concentrations of Maillard reaction products, HPLC-spectroscopy becomes more sensitive and more selective [52, 53]. It becomes clear that the linear range (mg/L) is more stable for HPLC-spectroscopy than for the spectrophotometric white method, 1-20 mg/L vs 1-5 mg/L, see table 6.
19
Table 6: Performance data of the HPLC-spectroscopy vs. spectroscopy [52].
Also the limit of detection for aqueous solution by HPLC-spectroscopy is very low, compared to spectroscopic method, 0.09 mg/L vs. 0.23 mg/L for HMF in honey solutions (table 6). The same applies to the limit of quantification, which is lower for aqueous solutions measured by HPLC-spectroscopy, 0.27 mg/L vs. 0.70 mg/L for HMF in honey (table 6). Besides that, the correlation coefficient is very high for aqueous solutions measured by HPLC-spectroscopy. From these quantitative data, the conclusion can be drawn that: HPLC-spectroscopy has the preference for the determination of Maillard reaction products in aqueous solutions. No information about intra- and interday precision was presented.
Because it is now known that HPLC-spectroscopy can be very effective in the study to detect Maillard reaction products, the next sections will discuss different spectroscopic detection methods combined with prior separation of samples with HPLC. These are: LC-fluorescence, LC-DAD, LC-FTIR and LC-UV-VIS. These methods are commonly used for the detection of non-volatile compounds which are formed during the Maillard reaction and are therefore chosen for this chapter [10–27].
4.2 LIQUID CHROMATOGRAPHY – FLUORESCENCE SPECTROSCOPY
The formation and accumulation of fluorescent compounds are observed during the early, but also during the advanced stage of the Maillard reaction (Stecker degradation) at low temperatures [10, 13, 24]. The formation of fluorescent compounds can be described as a zero-order reaction [13]. However, the calculation of the amount of fluorescent compounds cannot be realized, since several factors such as rate constants of the intermediary reactions are still unknown [13]. Examples of fluorescent compounds are glucose/caseinate and lactose/sodium caseinate [53]. Caseinate is a protein. In most cases, this method is used for the detection of proteins-bounded fluorescence compounds [53].
Liquid chromatography- fluorescence spectroscopy seems to be a good method for the separation and determination of the non-volatile Maillard reaction products [10, 11, 12, 13]. This method is gaining more importance, since it has been known that this method was used for the determination of fluorescent Maillard reaction products in milk [10]. This method is also used as an indicator of the level of advanced glycosylation end products [13].
A critical point which has to be made, is that this method also measures fluorescence of the fluorescent amino acid tryptophan. In order to correct for this, a blank measurement can be made. This can be done by dividing the fluorescence of the Maillard products by that of tryptophan [10]. Tryptophan is frequently excited at 290 nm, and shows emit light at 340 nm [10]. However, it is not known whether the Maillard reaction products are attached to the tryptophan, or that these MRPs are detached from the tryptophan.
20
Another critical point is that the choice of solvent is very important, since particular solvents can influence the fluorescence quantum yield of the Maillard product, for example by quenching [10].
Separation of the big complexes formed during the Maillard reaction is quite challenging, due to the hydrophobicity of the compounds. Therefore, reversed-phase liquid chromatography is used as separation method [10]. In order to get improve retention times, a gradient can be applied [11]. Afterwards, fluorescence is being measured at λ= 420 nm, whereas the excitation wavelength that commonly is used is 350 nm [10, 13]. The maximum excitation and the maximum emission wavelengths are different for each protein and compound and therefore also for each Maillard reaction product.
Frequently, advanced glycosylation end products are excited at wavelengths of 340-370 nm to measure fluorescence, whereas emission wavelengths of 420-440 nm are observed [13, 53]. Data showed that the amount of fluorescence increased as a function of incubation time [10].
An advantage of measuring at this wavelengths is that the emission intensities do not come from the fluorescence of tryptophan, which has a maximum excitation wavelength of 290 nm and a maximum emission wavelength of 336 nm [53]. Also, the maximum emission of the fluorescent compounds shifted at more severe treatment conditions [53]. The fluorescence intensities were measured for both the free fluorescent compounds, which means not being bound to proteins, and the fluorescent compounds bound to proteins [53]. This is high likely due to the polymerization reactions between fluorescent intermediates which were not bound to a protein. These fluorescent intermediates formed a more chemical complex. However this hypothesis has not been confirmed and more study is required for confirming this statement [53].
Another advantage of this method is that it can also be used at temperatures starting of 60 oC.
A distinction between fluorescent compounds produced in the early stage and in the advanced stage of the Maillard reaction can be detected, if a particular temperature range is chosen [11]. Moreover, this method is also suitable for long heat treatments [10]. The advanced stage Maillard reaction products have stronger temperature dependence than reaction products formed in the early stage [13].
Also glycosylated products, such as: proline, morpholine and Amadori-tryptophan, are being detected by fluorescence with prior separation of liquid chromatography, see figure 10. This mixture of compounds were academic samples.
21
Figure 10: Simultaneous on-line monitoring of chromatographic data from a single injection of the standard mixture containing D-glucose D-fructose, Amadori-proline, Amadori-morpholine and Amadori-tryptophan. Mobile phase: 70% acetonitrile, 30% phosphate buffer (0.04M, pH 6.5). Arp-Mor, Amadori-morpholine; Arp-Pro, Amadori-proline; Arp-Trp, Amadori-tryptophan; Glu, D-glucose; Fru, D-fructose [11].
So, it is proven that liquid chromatography-fluorescence spectroscopy can be used for the analysis of Maillard reaction product in different stages. However as the Maillard reaction continues, the amount of fluorescent compounds decrease as the formation of brown compounds increases [11, 13]. This relationship between decrease of fluorescent compounds and increase of brown compounds is not completely understood [13]. However, the decrease of fluorescent compounds in time does not imply that this method is only useful during the early stages. Contrariwise, this method is also used in advanced stages of the Maillard reaction, until the analyte shows a decrease in fluorescence intensity [10, 12]. During the advanced stages of the Maillard reaction, fluorescence can be measured, since the formation of smaller fluorescent active molecules is observed earlier than the formation of brown compounds [13]. Even though this is impossible, performing fluorescence measurements without separation will be difficult.
A suggestion is to use a multidetector, so that both fluorescence and absorption spectra can be obtained simultaneously. This combination provides more information [11]. Fluorescence mode of detection is higher in selectivity compared to UV-Vis and can even be performed without the prior need of derivatization [11]. This is important because at low concentrations of analytes, high signal-to-noise ratios are preferred. An implementation of derivatization would mean more steps and therefore a lower detection signal.
So, liquid chromatography fluorescence can be easily performed for the detection of Maillard reaction products during the early, intermediate and even during the advanced stage of the Maillard reaction. Temperature and reaction time are key variables for fluorescence spectroscopy. Fluorescence is used in studies studying proteins, but also small fluorescent compounds such as furosine [10]. Also, the formation of brown compounds presupposes the fluorescence intensities. Even though, many compounds have their own specific emission
22
wavelength, a specific range of wavelengths is observed for Maillard reaction products. Liquid chromatography fluorescence is considered as reliable method for separation and detection of Maillard reaction products [25]. This method is not generally used for quantification.
4.3 LIQUID CHROMATOGRAPHY – DIODE ARRAY DETECTION (LC-DAD)
As stated before, reversed-phase liquid chromatography can be used for the separation of many Maillard reaction products at a pH of 5 [14, 15, 16, 17, 18, 19, 20]. Beside reversed-phase LC, ion-exchange chromatography can also be used for the separation of Maillard reaction products [14, 16, 17]. Furthermore, it is even possible to use two-dimensional liquid chromatography (2D-LC), with different stationary phases [14]. The preference is to use reversed phase in the second dimension of the separation. For example hydroxymethylfurfural (HMF), a furanic compound which forms during the Maillard reaction, tends to retain strongly on ion-exchange stationary phases [14, 18].
With high temperatures and with long heating time, the intensity of the peak at 420 nm will increase rapidly [17]. The diode array detection range was set from 190 to 400 nm. Peaks were observed at λ= 254 nm and λ= 280 nm [14, 15, 16, 18, 21, 22]. These wavelengths corresponded to colourless compounds [15]. Besides the wavelengths of 254 and 280 nm, detection for coloured compounds was taken at λ= 360 and λ= 460 nm [15, 16, 17, 19, 21, 22]. These peaks at λ= 360 nm and λ= 460 nm are mostly observed when the pH is not kept constant at 5. Most details can be obtained from spectra at λ= 254 nm and λ= 280 nm, since a lot of well resolved peaks were noticed in this region [19, 20, 21]. It is important to know that for this specific wavelengths, the pH does not have to be controlled [19]. If the pH is not controlled, small differences in maximum absorbance wavelengths will be observed. These small differences were no problem for the diode array detector and therefore spectral matching was neither a problem [19].
By using reversed-phase liquid chromatography diode array detection, 4 types of peaks can be observed [19, 21, 22].
1. Peaks with low retention times with unretained materials
2. Resolved peak starting from to till tM; tm = migration time of an analyte
3. Convex broad bands 4. Tailing broad bands
The peaks with unretained materials absorbed light in the visible range (λabs=360 nm) and
these materials had a yellow-brownish colour [15, 19, 21, 22], which could imply the formation of melanoidins. Due to hydrodynamic chromatography (HDC) and/or size exclusion chromatography (SEC), these big bulky unretained materials could elute before t0 and could
therefore be separated by SEC. Resolved peaks could easily be detected with diode array detection. Convex broad bands and tailing broad bands are considered unresolved bands. Convex broad bands can appear at λ= 360 and λ= 460 nm [15, 21]. The tailing peak did not show broad bands of unresolved materials at λ= 460 nm [15, 22]. No information was provided in order to decrease the convex and tailing broad bands.
There is no need for derivatization of analytes when diode array detection is used [14,15]. This simplifies the sample preparation method. Moreover, diode array detectors are very useful, since many of the compounds in the Maillard reaction absorb in the visible region [20]. An advantage of diode array detection is that it detects the presence of colourless, yellow and brown compounds, whereas fluorescence only detects the colourless and yellow compounds [15]. Detection can take easily place when a sample is separated by liquid chromatography. However, it is possible that some peaks will only be present after the reaction has taken place. In that case, appointing compounds to a peak will be difficult. Peaks that contain Maillard
23
reaction products can be identified and analysed with diode array detection [20, 22]. By discriminating unknown analytes on both their retention time and their excitation wavelength, identification of compounds can take place, see tables 7 and 8 [15, 16, 21]. The highest wavelength observed is 315 nm. This is in line with the observation earlier stated: The peaks with unretained materials absorbed light in the visible range (λabs=360 nm) and these materials
had a yellow-brownish colour [19, 21, 22], which could imply the formation of melanoidins. Table 7: HPLC retention times and λmax values of standard reference compounds [15,21].
Table 8: Retention times and absorption maxima of furanic compounds in fruit juice concentrates and drinks are compared with those from table 7. When a match was found between retention time and absorption maximum, compounds were identified [15].
The lowest limit of detection was 0.003 mg/L for 5-HMF. On the contrary, a limit of detection of 3 mg/L was found for furfuryl alcohol, which is a 1000 times higher. Besides that, not all compounds were identified, see table 8. This method was then tested to identify and to quantify different Maillard reaction products (for example: 5-HMF, furfural, DMHF, furoic acid and 2-acetylfuran) in different fruit juices, see table 9. Several compounds were detected, whereas others were not detected in these different matrices, see table 9. The compounds that were detected were at higher concentrations than the limit of detection, see tables 8 and 9.
So, it has been stated that liquid chromatography diode array detection can be used for the separation and detection of non-volatile Maillard reaction products. Different small molecules were detected and quantified as can be seen in tables 7, 8 and 9. These small molecules originated from different matrices. However, this method can also be used for academic samples. The implementation of detection wavelengths of λ= 254nm and λ= 280 nm seem to satisfy in several different analyses. This method can be useful for analysis of small molecules formed during the Maillard reaction, but also for the proteins (free and sugar bound) [15].
24
Besides that, this method can be useful for analysis of compounds from a range of carbohydrate-based foods [22]. This method is quite rapid and simple, since no derivatization procedure is required [14, 22].
Table 9: Analysis results of furanic compounds in fruit juice concentrates and drinks (mg/L) [15].
4.4 LIQUID CHROMATOGRAPHY – FOURIER TRANSFORM INFRARED
SPECTROSCOPY
Liquid chromatography- Fourier transform infrared spectroscopy (LC-FTIR) seems to be a useful method for detection of Maillard reaction products in the nutrition and pharmaceutical industry, since this method does not require extensive sample preparation [23]. No quantitative data was provided for this statement. LC-FTIR can be used for qualitatively analysis of Maillard reaction products in the early stage of the reaction [24]. This method is not used for quantification of Maillard reaction products. During this phase of the Maillard reaction, structure sensitive information is provided with LC-FTIR [24]. This is because this method has a limited ability to analyse Maillard reaction products. It can be difficult to appoint vibrations to certain compounds formed during late stages of the Maillard reaction. An ion exchange column can be used for separation so that the spectrum. This method is also quite interesting in cases where the substances are chemically reactive or unstable [23].
During the advanced stage of the Maillard reaction, a certain amount of complexes will be formed, which all have approximately the same frequency and causing therefore overlap of different peaks of different vibrations. This problem leads to difficulties for identification of vibrations [23, 24]. This could be the reason why not so many studies have incorporated the use of LC-FTIR during studies performed in the advanced stage of the Maillard reaction. Academic samples were made by adding some amounts of asparagine with glucose. These academic samples were then identified by LC-FTIR [23, 24]. Spectra of LC-FTIR were taken in a range starting with a wave number of 900-1800 cm-1 with a resolution of 8 cm-1 [23]. It is
also possible to start at 800 – 1800 cm-1 with a resolution of 4 cm-1 [24]. There was also a
background spectrum was taken before the sample was injected. The peaks seem to increase in intensity when the temperature was set for longer time intervals at 140 oC, see figure 11
[24]. It seems that the Schiff base is being detected later than the amino acid Asn. So, this method can be used for the study of amino acids and small molecules, but also for amino acid bound to a sugar.
A FTIR spectrum was made for each peak, see figure 12. It can be seen that the peak at 1502 cm-1 disappeared after a reaction time of 2 hours at a temperature of 140 oC [24]. This is high
likely due to the Maillard reaction. This is high likely due to the successive decarboxylation reaction at the backbone amino group, which occurs during the Stecker degradation, see figure 1, which is the intermediate stage [24].
25
Figure 11: HPLC chromatograms of the time and temperature (T= 25-140 oC) - dependent evolution of Maillard
reaction products at pH = 8.0 [24].
Figure 12: HPLC-FTIR spectra of the reaction mixture of asparagine and fructose (pH 8.0) at 140°C. (a) Room temperature (b) Heated (c) Difference spectra [24].
During the intermediate stage of the Maillard reaction, several changes can be expected. These are for instance: the formation of the Schiff base Imine group, the enaminol group and the decarboxylated Amadori products with distinct C=O vibrations in the 1700 cm-1 region. So,
a change can be expected in the region of 1700 cm-1. If that is known and observed in the
spectra, the stage of the Maillard reaction can be appointed as the intermediate stage. Imines can be converted to an Amadori product under high temperatures and under strong acidic conditions [23]. Also, the loss of a peak at 1502 cm-1 occurs during the Maillard reaction. This
peak originates from the backbone amino and/or carboxylate group [23].
Different vibrations can be detected at several specific wavelengths and can therefore be linked to specific vibrations, see table 10. By identifying these specific vibrations in the spectrum, partly more information about the molecular structure can be gained. Also, spectra could be taken from the food sample before the Maillard reaction could take place. A change in wave number and thus a change in vibrations will be observed at certain moment, in which it could be possible to check what bond changes has occurred. Therefore, it can be deducted
26
which bonds have been broken and which have been formed, see table 10. Even though these numbers are listed in table 2, wave numbers of different vibrations can be downshifted, or even diminished, due to cleavage of a moiety of a compound.
Table 10: A list with their different wavenumbers with their corresponding vibrations [23, 24].
Wave number (cm-1) Vibration Remarks
1675, 1681 C=O Stretching
1700-1620 COO-, NH
2, C=O Extensive coupling
1390-1310 OH, CO2 Extensive coupling
1575, 1593, 1618 NH2 Backbone amino group
1358-1421 C-N=C Imine group
1660 C=N Schiff base
1388, 1589 C(fru)-NH=CH(Asn) Stretching
So, it can be stated that this method will be useful during the early and intermediate stages of the Maillard reaction to check which vibrations will be present and which vibrations will be diminished due to the Maillard reaction. However during later stages of this reaction, the spectrum will become harder to read and understand, due to the fact that downshifting as well as overlap is playing a major role. This means that this method has a limited applicability. This method could be useful if the structure of a compound is already known. So, it could be easy to identify whether the compound is still present. If the structure of the compound is unknown, this method will be slightly useful, in a way that it can provide information about which vibrations will be present after some reactions steps and which will not. No information was provided about whether this method was also used for the determination of Maillard reaction products in food products. It is possible to use this method to detect the Maillard reaction products in academic samples. This means adding glucose and amino acids in a vial. And therefore, it is proven that this method is useful for the detection of amino acids and small molecules.
4.5 LIQUID CHROMATOGRAPHY – UV-VIS SPECTROSCOPY
Liquid chromatography UV-Vis (LC-UV-Vis) detection can be used for separation and detection of Maillard reaction products in all stages. This method was proven to be very successful in the advanced stage, in which UV-Vis spectra could be taken of melanoidins. Also separation of Amadori compounds containing an aromatic ring is possible with liquid chromatography [11]. Detection at λ= 190 nm and λ= 210 nm provides poor sensitivity as well as poor selectivity [11]. On the contrary, a wavelength of 220 nm was used for the detection of peptides and proved to be better [26].
Analysis was performed with academic samples, such as furosine and lysine [25]. Besides that, measurements were also performed on dry matter during backing of bread [25]. Both bread crumb and bread crust were analysed for presence of Maillard reaction products. During the first 16 minutes, only browning indices were measured in the bread crust, rather than in the bread crumb [25]. During these measurements, the pH of the sample was between 5.2-6.0 [25]. Discrete and different absorption maxima were found at λ= 234, 277, 294 and 307 nm [24]. Non-volatile intermediate products tend to absorb at a wavelength of 294 nm [24]. Absorption at this wavelength reflects the amount of uncoloured intermediate Maillard reaction products. Pre-melanoidins, compounds which are formed during the advanced stage of the Maillard reaction absorb particularly in a range of 320-350 nm, whereas melanoidins absorb in the range of 420-450 nm, see figure 5 [24, 26, 64]. Different absorption wavelengths have different electronic transition in different bonds, see table 11.
27
Table 11: Absorption wavelengths with their corresponding electronic transitions [24]
Wavelength (nm) Bond or property Transition
265 Carbonyl π→π*
270-300 Carbonyl n→π*
280 (forbidden) ketone n→π*
300 C=O, C=N n→π*
Maillard reaction products formed in the intermediate and advances stage seem to absorb at λ= 360 and 420 nm [25, 26]. Absorbance at λ= 420 nm was used as an indicator of the colour formation in the final stage of the Maillard reaction [26]. The intensity of the absorbance measured at λ= 420 increased with increasing temperatures [26]. This increase in absorbance could be totally related to the Maillard reaction and not to the caramelization, especially if the system does not contain high levels of fructose-glycine solutions [26]. Melanoidins, that were formed in different samples, had experienced enzymatic digestions. However these melanoidins showed no big difference in absorbance [25]. This indicates that measuring absorbance is not a specific tool for quantification. Liquid chromatography UV-Vis absorbance is not as reliable as liquid chromatography fluorescence measurements, since in fluorescence excitation specific wavelengths are needed [25]. Besides that, Amadori compounds absorb poorly in the UV-Vis range [27]. This is due to the instability at high temperatures as well as the presence of nucleophilic and oxidation agents. Also modifying these intermediates is not always possible.
Different furanic compounds, which are considered small molecules, were identified in fruit juices concentrates [14]. Different compounds showed different absorption maxima and detection limits, see table 8.
So, LC-UV-Vis is a method that can be used for the detection of Maillard reaction products in food samples as well as in academic samples. This method is useful for the detection of small compounds, amino acids and polymers of amino acids and sugars [25]. Analysis with Vis at a wavelength of 280 nm can be used in the early stage of the Maillard reaction. LC-UV-Vis with a wavelength of 320 nm seems useful in the intermediate stage of the Maillard reaction. Unfortunately, the LC-UV-Vis method is not as powerful in the advanced stage of the Maillard reaction as in the initial and intermediate stage, but can be used at a wavelength of 420 nm [27]. As mentioned before, LC-UV-Vis is not specific enough and cannot be used for quantification. This method provides an insight in which stage of the Maillard reaction a sample is. However, LC-fluorescence is more specific and more reliable compared to LC-UV-Vis.
28
CHAPTER 5: DETECTION WITH MASS SPECTROMETRY
5.1 INTRODUCTION
Several methods were discussed in chapters 3 and 4, in which it became clear that those methods do not provide enough specific molecular information for structure elucidation. Nevertheless, those methods provide an indication of the stage of the Maillard reaction product, which can be deducted from the colour formation. New analytical methods will be discussed in this chapter for the possibility of structure elucidation of the Maillard reaction products. Therefore, the implementation of the mass spectrometer will be used as detection method. This detection method seems to be important since advanced glycated end products do not show fluorescence nor absorbance in the UV-Vis region [39]. Mass spectrometry has been appreciated as the most powerful technique for structure identification of many Maillard reaction products [29, 33, 44]. This is due to its high sensitivity and high mass resolution [33]. No official methods were recommended before the utility of the mass spectrometer was discovered for this work field [29].
Liquid chromatography (LC) is a useful method for the separation of non-volatile methods. Depending on the characteristics of an analyte, different modes of stationary phases of the LC-column can be used. Most of the times, reversed phase column was used due to the characteristics of the hydrophobic moieties of the Maillard reaction products. But, also cation exchange chromatography is used for the separation of Amadori compounds [28, 34]. This can also be used for the removal of salts/ions from a sample so that the mass spectrometer will not be damaged by salts [34]. Simultaneously studying the structure of non-volatile compounds is difficult due to high diversity in chemical properties [45]. Another disadvantage of LC is, that is stated to lack high resolution compared to capillary electrophoresis (CE) [16, 20, 22]. Therefore, CE gains more interest over time. Because of the applicability of CE, the application of liquid chromatography and capillary electrophoresis will be both discussed as separation method
In this chapter, the next mass spectrometers will be discussed: tandem-mass spectrometry, Fourier-transform ion cyclotron resonance mass spectrometry (FT-ICR-MS) and time-of-flight mass spectrometry (TOF-MS).
5.2 LIQUID CHROMATOGRAPHY – TANDEM MASS SPECTROMETRY
Liquid chromatography tandem mass spectrometry (LC-MS/MS) is considered as a golden standard in the analysis of Maillard reaction products [29]. This is because more structural information can be gained by tandem mass spectrometry [35]. It was also possible to determine the position of modified amino acids bound which formed a protein [31]. Next to that, it is also possible to determine the sequence of peptides or proteins or even amino acids bound to sugars [28]. Next to that, glycation sites of lysozyme were identified with MS/MS [66]. LC-MS/MS was considered better than MALDI-TOF-MS for the identification of the glycation sites of lysosome [66]. Masses can be detected with 2 digits after the comma, providing relatively detailed information. On the contrary, high resolution mass spectrometry provides data with 4 digits after the comma.
However, for tandem mass spectrometry applies that the gain in information is only possible when the target compound is known. This method is very specific and sensitive and can be used for analysis in food products as well as in academic sample mixtures [28]. Therefore, this method is being used frequently to measure and to quantify Amadori compounds in the food industry as well as in the tobacco industry, see table 12 [33]. Amino acids bound to a sugar derivative have been detected and identified [33, 34]. Low quantities of different Amadori compounds were identified by tandem mass spectrometry.
