A predictive model of acrylamide formation influenced by moisture content, lipid oxidation, and methylglyoxal in fried potato slices

Abstract

Acrylamide (AA) is known to be a neurotoxic, genotoxic and carcinogenic compound, which is classified as a probable human carcinogen. The main purpose of this paper was to study the changes in peroxide value (PV), p-anisidine value (PAV), methylglyoxal (MG), and moisture content (MC) in potato slices fried at different temperatures and lengths of time. The determination of AA was carried out using HPLC-MS/MS. Our results showed that AA formation was obviously influenced by MG content and PV; the AA content increases with the increase of PV and MG. The contribution of PV and MG to AA formation was included in a prediction model that could be used as a tool to predict the formation of AA in potato slices, and the model was as follows: AA = −0.200–0.001 × MC + 0.330 × PV + 0.074 × PAV + 0.103 × MG. The predicted AA content value was in good agreement with the actual AA content.

---

1. Introduction

Acrylamide (AA, H₂C=CH–CO–NH₂, CAS no.79-06-1) is known to be a neurotoxic, genotoxic and carcinogenic compound, which is classified by the IARC as a probable human carcinogen.¹ AA was first found in food in April 2002 and it has raised worldwide concerns due to the high levels of AA content in dietary components like processed cereals and potato chips.

The formation mechanism of AA is widely agreed to be through the Maillard reaction (MR) from free asparagine and a carbonyl source.^2–4 In most foods, the main carbonyl source is reducing sugar. However, Frankel⁵ proposed that lipid oxidation may be another carbonyl source that can be substituted for sugar, especially during thermal treatment, and the carbonyl compounds could form AA with asparagine present in food. There is another possible pathway for AA formation based on acrolein as a precursor of AA,^6,7 where the acrolein is oxidized and forms acrylic acid, which would then form AA with asparagine. Acrolein was found in the thermal degradation of lipids.⁸ Analogously, other researchers also found that secondary lipid oxidation products actively participated in the formation of AA in fat-rich food.⁹ It is necessary to investigate the relationship between lipid oxidation and the formation of acrylamide in fried food since the pathways mentioned above are both linked with lipid oxidation. Lipid oxidation products are a family of compounds including aldehydes, ketones, alcohols, epoxides, hydrocarbons and so on.^10,11 Peroxide values are an estimation of peroxides (primary oxidation products), while p-anisidine values are a measure of secondary oxidation products like aldehydes, α- and β-alkenals and of all compounds able to react with p-anisidine.^12,13 Peroxide and p-anisidine values increased in deep frying according to Naz et al.¹⁴

One study suggested that methylglyoxal (MG) was involved in the formation of AA. But other reactions involved in the formation of some small molecules, like formic and acetic acid, could not be excluded.^15–17 A similar study proposed that MG could be an intermediate for the formation of acrylamide.³ Yuan et al.¹⁸ found that MG was one of the most significant α-dicarbonyl compounds formed in the Maillard reaction (MR) and was a representative α-dicarbonyl compound to research the influence on AA formation.

In the process of frying, moisture content varies constantly and this frying process could be regarded as a drying process. Amrein et al.¹⁹ found that moisture content strongly influenced the activation energy of AA formation. With decreasing moisture content, there is a corresponding increase in the activation energy of AA formation. A study by Gökmen et al.²⁰ proposed that during the frying of potato strips, the internal energy of the model systems increased during the heating treatment, and then reached the boiling point of water. After that, the water in food used a large amount of the incoming energy for evaporation, which was an obstacle to internal energy increase. Capuano et al.²¹ illustrated in their study that the water content of food was a key factor in determining the acrylamide level after heating, since two model systems with 4% and 16% water content had different acrylamide contents.

This experiment was conducted on the potato chips fried for different lengths of time and at different temperatures, which was justified since the potato chips obtained similar final conditions. The aim of this study was to determine the contribution of lipid oxidation, MG and moisture content on the formation of AA at different frying times and temperatures, and establish a prediction model using lipid oxidation, MG and moisture content as tools to predict the formation of AA in potato slices. Since it is easy and fast to determine the indexes of lipid oxidation, MG and moisture content and it is very complex to determine AA content in a food matrix, a prediction model could save a lot of time and work and could be helpful to estimate the actual AA content.

2. Materials and methods

2.1. Regents

All solvents used were of HPLC grade and other chemicals were of analytical grade. AA (>99.5%) and [¹³C₃]AA (99% isotopic purity) were purchased from Sigma-Aldrich (St. Louis, Mo., USA) and Cambridge Isotope Laboratories (Andover, MA, USA), respectively. MG (40% in water) and o-phenylenediamine (OPD) were obtained from Fluka Biochemika Company (Beijing, China). Phenolphthalein, potassium iodide, chloroform, glacial acetic acid, sodium thiosulfate, isooctane, p-anisidine and other chemicals were obtained from Beijing Chemicals Co. (Beijing, China). HPLC-grade methanol was purchased from Fisher Scientific International Inc. (Beijing, China). SPE cartridges Oasis MCX (3 ml, 60 mg) were supplied by Waters (Milford, MA, USA).

Stock solutions of 1 mg kg⁻¹ of AA and [¹³C₃]AA were prepared in methanol. The stock solution was diluted with water to give a series of standard solutions (1.0, 2.0, 5.0, 8.0, 10.0, 15.0 and 20.0 μg kg⁻¹). 90 μg kg⁻¹ of [¹³C₃]AA was used as the internal standard for quantification.

2.2. Preparation of potato slices by frying

A batch of potatoes was purchased from the local market. Potato tubers were washed, peeled and cut from the medulla into identical slices (2.0 cm × 2.0 cm × 2.0 mm). Slices were rinsed immediately for 1 min in water to eliminate some starch adhering to the surface, and blanched at 100 °C for 1 min. Thirty slices were put into a heating basket, which was large enough to enable free movement of the slices in the frying oil. The slices were fried at 140 ± 1 °C for up to 5 min, 160 ± 1 °C for up to 4 min, and 180 ± 1 °C for up to 3 min in fresh vegetable oil (Palm oil, Fujian Fuyuan Panpan Food Co. Ltd.). Five liters of palm oil were used in order to keep a constant frying temperature in an electrical fryer (HH–S, Ronghua Instrument Co., Ltd., Jintan, Zhejiang). The potato chips were processed by frying until they had the same final condition. Then the slices were drained over a wire screen for 5 min. The slices were cooled to room temperature and homogenized in a blender (HR 2094, Philips Instruments Co., Ltd., Zhuhai, China) for further analysis.

2.3. Analysis of AA by HPLC-MS/MS

The sample preparation was performed according to the method desribed by Liu et al.²² A 1.00 g portion of the sample was transferred to a 50 ml centrifuge tube. Then, 1 ml [¹³C₃]–AA solution of 90 μg kg⁻¹ and 9 ml water were added into the tube. The mixture was incubated on a horizontal shaker (150 rpm) at 25 °C for 20 min, then 10 ml of acetonitrile, 4 g of anhydrous magnesium sulfate and 0.5 g of sodium chloride were added consecutively. The tube was immediately sealed and shaken vigorously for 1 min, and then centrifuged at 5000 rpm for 5 min at 4 °C. The salt combination induced the separation of water and acetonitrile layers, and forced the majority of AA into the acetonitrile layer. Three layers were obtained as follows: an acetonitrile layer containing AA at the top, a matrix layer in the middle and a water layer with the excessive salts at the bottom. The acetonitrile solution (9 ml) was transferred into a glass test tube and evaporated to dryness under a stream of nitrogen, in a water bath at 40 °C. The dried acetonitrile extract on the wall of the glass tube was re-dissolved in 0.5 ml water under vortexing, so that some of the highly lipophilic co-extractives were excluded again. The aqueous extract was filtered through a 0.45 μm syringe filter for further purification by an Oasis MCX SPE cartridge. The aqueous extract (0.5 ml) passed through the SPE cartridge was conditioned consecutively with 2 ml methanol and 2 ml water, and the effluent containing AA was collected. Subsequently, the AA retained on the SPE column was eluted with 0.5 ml water. The initial effluent and this wash were combined and filtered through a 0.22 μm syringe filter for LC-MS/MS analysis.

The analysis of AA in the potato slices was performed by an Alliance 2695 Separation Module (Waters, Milford, MA, USA) coupled to a Micromass Quattro Micro triple-quadrupole mass spectrometer (Micromass, Manchester, UK) with MassLynx software. The final tested solution (20.0 μl) was injected onto a reversed ODS-C₁₈ column (250 × 4.6 mm, 5 μm, Hypersil, Thermo, USA) maintained at 30 °C. The elution mode was isocratic using a mixture of 10% acetonitrile and 90% water containing 0.1% formic acid as a mobile phase at a flow rate of 0.4 ml min⁻¹. AA was detected by MS/MS using electrospray ionization in the positive ion mode. The multiple reaction monitoring (MRM) of degradation patterns m/z 72 → 55 for AA and m/z 75 → 58 for [¹³C₃]–AA respectively, was used for the quantification of AA. The optimized MS instrument parameters obtained by tuning were as follows: capillary voltage, 1 kV; cone voltage, 20 V; source temperature, 110 °C; desolvation temperature, 400 °C; desolvation gas flow, 600 l h⁻¹; cone gas flow, 50 l h⁻¹; argon collision gas pressure for MS/MS, 2 × 10⁻³ mbar; collision energy for each transition in the MRM mode, 13 eV. In the MRM transitions, the dwell and interscan times were 0.4 and 0.1 s, respectively. Each determination was performed in triplicate.

2.4. Analysis of MG by HPLC

The analysis of MG was based on the method described by Bednarski et al.²³ and De Revel et al.²⁴ A 2.00 g portion of the sample was transferred to a 50 ml centrifuge tube and 18 ml water was added into the tube. The mixture was shaken vigorously for 1 min, and 5 ml of the sample solution was taken, and 100 μl of 0.1 mmol l⁻¹ OPD (1.08 g OPD was dissolved in 10 ml methanol) was added into the solution and mixed thoroughly for the derivatization reaction. The pH value of the sample solution was adjusted to 8.0 with 0.1 mol l⁻¹ sodium hydroxide and the solution was heated at 60 °C for 3 h for the derivatization of MG to a quinoxaline derivative. After the derivatization, the pH of the sample was adjusted to 3.0 with 1.0 mol l⁻¹ hydrochloric acid. 2 ml chloroform was added and mixed vigorously to extract the quinoxaline derivative of MG. The sample was centrifuged at 4500 rpm for 10 min, and the bottom chloroform layer was taken. The extraction was repeated three times, all of the chloroform layers were combined and evaporated under N₂ at 40 °C, and the resulting residue was dissolved in 1.0 ml HPLC-grade methanol. The solution was filtered through a 0.45 μm filter (Millipore Corporation, Bedford, MA, USA) and then used for HPLC analysis.

The derivatives were analyzed using a LC-2010A HPLC system (Shimadzu Co., Japan). 20 μl of the sample was injected onto a C₁₈ column (250 × 4.6 mm, 5 μm, Shimadzu, Japan), and eluted isocratically at 35 °C with 68% methanol in water, at a flow rate of 1.01 ml min⁻¹. The detection wavelength was 313 nm. MG was quantified by external calibration, in which the concentrations of the MG standard solution ranged from 10 to 500 μg ml⁻¹.

2.5. Determination of peroxide value (PV)

PV was determined by using conventional iodometric titration with thiosulfate.²⁵ 2 g of the sliced potato sample was weighed and placed in an iodine flask. The liposoluble constituents in the sample were dissolved completely with a 30 ml mixture of chloroform and glacial acetic acid (2 : 3), and 1.0 ml saturated KI was added to the solution. The iodine flask was lightly shaken for 0.5 min, and left to stand in the dark for 3 min. The solution was titrated with a standard hyposulfite solution at a concentration of 0.002 mol l⁻¹. Each measurement was carried out in triplicate.

2.6. Determination of p-anisidine value (PAV)

PAV was determined by using a spectrophotometer (Beijing Puxi universal instrument Co., Beijing, China).²⁶ 2 g of the sliced potato sample was weighed and placed in a 25 ml volumetric flask. The liposoluble constituents in the sample were dissolved with isooctane and 1 ml p-anisidine reagent was added accurately, then left to stand for 10 min. The isooctane solvent was used as the blank to determine the absorbance without the addition of p-anisidine. One milliliter of the p-anisidine reagent with 25 ml isooctane was used as the blank to determine the absorbance after the addition of p-anisidine. The absorbance was determined at a wavelength of 350 nm. The PAV was determined according to the following equation:
A_s being the absorbance at 350 nm before adding p-anisidine, A_b, the absorbance at 350 nm after adding p-anisidine, and m, the weight of the sample. Each measurement was carried out in triplicate.

2.7. Determination of moisture content (MC)

The MC of the potato slices was measured by the air dry oven method.²⁷ 2 g of the sliced potato sample was weighed accurately into an oblate weighing bottle and dried until it reached a constant weight in an air dry oven (Shanghai Boxun medical instrument Co., Shanghai, China) at 100 °C. Each measurement was carried out in triplicate.

2.8. Statistical analysis

Statistical analysis was performed with SPSS 12.0 software, and a t-test was used for comparison of means from multiple samples in different treatments. Graphs were drawn with OriginPro 7.5 software.

3. Results and discussion

3.1. AA content of potato slices after frying

The potato slices were fried at 140 °C for up to 5 min, 160 °C for up to 4 min and 180 °C for up to 3 min. These were maximum time periods and measurements were made at intervals up to them. The changes in AA formation are shown in Fig. 1(a). With an increase in the frying time and temperature, the AA content increased accordingly. The highest AA content (0.901 ± 0.011 μg g⁻¹) was found when potato slices were heated by frying at 180 °C for 3 min. In order to evaluate the effect of frying temperature on AA formation, the activation energy (E_a) was used and the data are shown in Table 1. The lines in the figures follow a given first order kinetics model at a given time or temperature shown in Table 1. With higher values of E_a, more energy was provided for the formation of AA. The relationship between AA and frying temperature was analyzed with the first order kinetics equation and the data are also shown in Table 1. The data of the first regression were obtained at a given time and temperature. The first order kinetics equations illustrated that the AA content had a good correlation with frying temperature and the AA formation was significantly affected by the E_a. The formula used to compute E_a is as follows:

ln k = ln A − E_a/RT (1)

Fig. 1 The changes in (a) acrylamide content, (b) peroxide value, (c) p-anisidine value, (d) methylglyoxal and (e) moisture content of potato chips fried at 140 °C for up to 5 min, 160 °C for up to 4 min and 180 °C for up to 3 min.

Table 1 The first order kinetics equations of the indexes and frying times

		First order kinetics equation	R^2	Ea	SE^a
a SE is the standard error to the regression curve. P < 0.05.
AA	140 °C	y = 1.081 − 1.020 × 10^−x/6.340	0.994	6.344	0.276
	160 °C	y = 0.837 − 0.832 × 10^−x/8.730	0.998	7.805	0.151
	180 °C	y = 0.920 − 0.920 × 10^−x/10.800	0.999	8.965	0.033
PV	140 °C	y = 0.303 − 0.273 × 10^−x/3.6500	0.997	4.447	0.105
	160 °C	y = 0.262 − 0.233 × 10^−x/4.864	0.999	5.697	0.037
	180 °C	y = 0.261 − 0.233 × 10^−x/5.580	0.997	6.477	0.124
PAV	140 °C	y = 1.799 − 1.731 × 10^−x/1.983	0.996	2.352	0.116
	160 °C	y = 1.744 − 1.678 × 10^−x/3.037	0.991	4.011	0.079
	180 °C	y = 1.706 − 1.641 × 10^−x/4.606	0.998	5.754	0.049
MG	140 °C	y = 0.948 − 0.913 × 10^−x/2.505	0.987	3.154	0.151
	160 °C	y = 0.647 − 0.614 × 10^−x/2.892	0.983	3.824	0.028
	180 °C	y = 0.716 − 0.683 × 10^−x/3.700	0.999	4.929	0.051
MC	140 °C	y = −36.100 + 97.67 × 10^−x/5.485	0.913	5.846	0.457
	160 °C	y = 4.380 + 54.300 × 10^−x/7.230	0.976	7.124	0.092
	180 °C	y = 4.470 + 54.412 × 10^−x/8.910	0.997	8.240	0.048

---

This is similar to the results from another study,²⁸ which found that the maximum AA concentration was significantly lower in potato crisps processed at the lowest frying temperature (170 °C) compared to the highest temperature (185 °C). The study by Pedreschi et al.²⁹ showed that high concentrations of AA in fried products could be avoided at low temperatures. The results of Kotsiou et al.³⁰ also showed that the content of AA increased obviously with prolonged heating times and at higher temperatures. A similar result was found in the paper by Romani et al.³¹

3.2. Effect of fat oxidation on AA formation

The formation of peroxides is correlated with the level of fat oxidation. Peroxides are transient chemical compounds, and they are not always directly correlated with sample oxidation, of mostly heated samples. The PAV is well correlated with the level of secondary oxidation products, namely aldehydes, which are far more stable than peroxides. Therefore, for an accurate estimation of the oxidation status, both parameters should be interpreted simultaneously.³² The PV is the estimation of peroxides (primary oxidation products), and it indicates the initial reaction products of oil oxidation.¹² The PAV is typically used to determine the content of further degradation products of fat oxidation. In this study, we discussed the effect of PV and PAV on the formation of AA, and the changes in the PV and PAV of potato slices are shown in Fig. 1(b) and (c), respectively. As expected, the PV sharply increased with the frying time and temperature. When the potato slices were fried at 140 °C for 5 min, the PV increased from 0.094 ± 0.006 g per 100 g to 0.238 ± 0.005 g per 100 g, which is an increase of 60.5%. The same trend was also found when the potato slices were fried at 160 °C and 180 °C; the PVs increased by 46.6% and 34.8%, respectively. Debnath et al.³³ found that the PV of fresh rice bran oil (RBO) was 0.021 g per 100 g. The PV was found to increase from 0.028 g per 100 g to 0.046 g per 100 g when heated. Since peroxides are transient chemical compounds, different PVs were obtained in different experiments, but the formation trends were similar. The relationship between PV, frying time and AA content is shown in Table 1 and 2 and Fig. 2(a). The first order kinetics equations illustrated that peroxides affected the formation of AA. The E_a of the reaction also showed that with higher frying temperatures, the frying system obtained more energy, which is good for oil oxidation. The peak PV and AA content occurred at 180 °C for 3 min. The correlation of PV and AA content was significant (R² = 0.999, p < 0.05). Peroxides were the result of oil oxidation, and oil oxidation could be an important variable influencing AA formation,^34,35 because it may influence the surface tension between food and oil as well.

Table 2 The first order kinetics equations of the indexes and AA content

		First order kinetics equation	R^2	Ea	SE^a
a SE is the standard error to the regression curve. P < 0.05.
PV	140 °C	y = −0.762 + 0.710 × 10^−x/3.950	0.982	4.720	0.373
	160 °C	y = −0.911 + 0.900 × 10^−x/4.480	0.998	5.402	0.259
	180 °C	y = 1.108 + 1.402 × 10^−x/5.250	0.999	6.424	0.031
PAV	140 °C	y = −0.133 + 0.130 × 10^−x/9.770	0.999	7.828	0.052
	160 °C	y = −0.230 + 0.223 × 10^−x/11.540	0.996	8.809	0.106
	180 °C	y = −1.166 + 1.145 × 10^−x/28.700	0.998	12.647	0.165
MG	140 °C	y = −0.456 + 0.446 × 10^−x/9.680	0.995	7.797	0.082
	160 °C	y = −0.134 + 0.126 × 10^−x/13.320	0.978	9.324	0.014
	180 °C	y = −2.976 + 2.944 × 10^−x/26.030	0.999	12.282	0.155
MC	140 °C	y = 1.809 − 1.179 × 10^−x/22.000	0.975	10.617	0.126
	160 °C	y = −0.073 + 1.069 × 10^−x/29.460	0.930	12.183	0.089
	180 °C	y = 2.266 − 1.280 × 10^−x/41.130	0.999	14.004	0.069

---

Fig. 2 The relationships between acrylamide and (a) peroxide value, (b) p-anisidine value, (c) methylglyoxal and (d) moisture content.

The PAV is typically used to determine the content of further degradation products of oil oxidation. As can be observed in Fig. 1(c), the first order kinetics equations showed that the PAV increased with an increase in frying time and temperature. With the increase in frying time, aldehydes and other secondary oxidation products are produced. The highest PAV was detected at 180 °C for 3 min (R² = 0.998, p < 0.05). These results were similar to those found by other authors. The PAVs were low before frying, and increased gradually after frying with different kinds of oils as the frying time increased.³² All the experimental points of AA content and PAV are analyzed in Table 2 and Fig. 2(b). The correlation between the PAV and AA content was obviously significant. The AA concentration increased with the increase in PAV. The PAV indicates the aldehyde, ketone and quinine content in products, which is one of the indexes of the degree of oil oxidation. These results are in agreement with earlier published data.^34,35

3.3. Effect of MG on AA formation

MG is a typical α-dicarbonyl compound formed by the Maillard reaction in potato chips, and plays an important role in the formation of AA.^18,36,37 Changes in the MG content of potato slices at different times and temperatures are shown in Fig. 1(d). With an increase in heating time and temperature, the MG content increased accordingly. The highest MG content could be found when the potato slices were fried at 140 °C for 5 min, reaching 0.840 ± 0.008 μg kg⁻¹. As for frying at 160 °C and 180 °C, a similar increase trend was also found, and the MG contents were 0.683 ± 0.001 μg kg⁻¹ and 0.718 ± 0.001 μg kg⁻¹, respectively. The results of the study by Yuan et al.³⁶ suggested that the MG formation rate increased linearly with heating time. In our experiments, a higher increase rate at 140 °C may be due to the longer frying time. The longer frying time could be helpful to the formation of MG. MG was involved in the formation of AA as an example of an α-dicarbonyl intermediate. In the study by Arribas-Lorenzo and Morales,³⁸ the MG content was detected in cookies baked at 190 °C for different lengths of time. The baking process markedly accelerated the formation of α-dicarbonyl compounds, such as glyoxal and methylglyoxal, which was confirmed by the results of our study. The previous studies all kept a watchful eye on the linear equation of MG and AA. This experiment focused on the relationship between MG, frying time and AA content with the first order kinetics equation. The results showed that the correlation was more obvious and the E_a of the system increased gradually (Tables 1 and 2, Fig. 2(c)). Yuan et al.³⁶ proved that MG was the main intermediate to form AA and implied that MG could not only take part in the formation of AA, but could also form some intermediates or other small molecules during the Maillard reaction.

3.4. Effect of MC on AA formation

The changes in the MC of potato slices at different frying times and temperatures are shown in Fig. 1(e). The moisture content decreased gradually with prolonged frying times at the same temperature. The MC also decreased with higher temperatures at the same frying time. Since the potato chips could absorb oil during the frying process, oil-absorbing sheets were used to reduce the oil in the fried potato chips, and the oil could then be ignored when calculating the moisture content. The results showed that the frying process decreased the MC of the potato slices, but the experimental MC evolution at 140 °C doesn't fit very well with the model shown in Table 1. This is probably because at that temperature, the time needed to begin the evaporation process is higher. In a study by Bull et al.,³⁹ an increase in frying temperature resulted in a decrease in the MC, which could be an advantage in the storage of food as well as lengthening shelf-life. Similarly, MacMillan et al.⁴⁰ found that the process of cooking by immersion frying decreases the moisture content through evaporative losses. Another study showed that an increase in both pre-frying and final frying time decreased the moisture content of French fries.⁴¹ The decrease in water content has been certainly more related to AA formation. Amrein et al.¹⁹ found that MC strongly influenced the E_a of AA formation. With the decrease in MC, there is a corresponding increase in the E_a of AA formation and the study showed that the formation of AA occurred in the last phase of the frying process. Gökmen et al.²⁰ studied the effect of water vapor on the formation and elimination of AA in potato systems, and the results showed that although a fraction of the AA formed during frying is removed with vaporized water, it appears to be insignificant. In our study, we also investigated the effect of moisture on the formation of AA. With the decrease in moisture content, the formation of AA increased accordingly. From the current results, we found that the formation of AA had some relationship with the decrease in the moisture content. The vaporization of the water will enhance the concentration of compounds in the potato chips. The changes in the amounts of AA over time followed a kinetic trend as shown in Tables 1 and 2 and Fig. 2(d).

3.5. Analysis of the contribution rate of PV, PAV, MG content, and MC on AA formation

In our present study, we found that the PV, PAV, MG content and MC could influence the formation of AA in potato slices when the potato slices were fried at different temperatures and lengths of time. The contribution rates of each response variable were analyzed according to the degree of the correlation of the factors and the AA content.β is the contribution rate, Sj is the sum of the squares of the indexes of regression analysis which were not listed in the paper and fj is the degree of freedom of Sj. Se is the sum of the squares of standard deviation to the regression curve and fe is the degree of freedom of Se. S is the total sum of the square of deviations. Sj, Se, fj, and fe were not listed in the paper. The contribution rate of each factor is listed in Table 3. The results showed that the PV, PAV, MG content and MC could influence the formation of AA, while the PV and MG content mainly affected the formation of AA. The two factors contributed to nearly 62.50% of the AA formation. The MG content contributed the highest rate to AA formation, which was confirmed by our previous study.^18,36,37 In order to confirm the relationships between the responses and the independent variables, the data pertaining to the independent and response variables were analyzed to get a regression equation as follows:

AA = −0.200 – 0.001 × MC + 0.330 × PV + 0.074 × PAV + 0.103 × MG (p < 0.05)

Table 3 Analysis of the contribution rates of response variables on the formation of AA

Factor	Sum of squares	Degree of freedom	Contribution rate (%)
MC	16.7972	1	11.59
PV	0.0002	1	28.99
PAV	0.0058	1	22.90
MG content	0.0015	1	33.51
Overall error	0.0317	3	3.01

---

More experiments of the products from the factory (Fujian Fuyuan Panpan Food Co. Ltd., Jingjiang City, Fujian) were conducted to confirm the regression equation above. 50 slices of potato chips fried at 170 °C for 2 min were analyzed. The results are shown in Fig. 3. The linear equation was y = 0.005 + 0.905x (R² = 0.882, p < 0.05). The x value was the actual value of AA content and the y value was the predicted value of AA content. Since it is a complex procedure to extract and determine the AA content, this prediction model could provide a convenient and fast method to estimate AA content as long as the four parameters above are provided. The predicted AA content value was in good agreement with the actual AA content.

Fig. 3 Correlation between the actual data and the predicted data of acrylamide formation.

4. Conclusion

In our present study, the formation of AA with different frying times and temperatures was studied. The concentration of AA increased with prolonged frying time and increased temperature. The PV, PAV, MG content and MC were involved in the process of AA formation. The PV and MG content mainly affected the formation of AA. The PAV was the secondary factor and the MC only slightly affected AA formation. A prediction model could be used as a tool to predict the formation of AA in potato slices, and the model was as follows:

AA = −0.200 − 0.001 × MC + 0.330 × PV + 0.074 × PAV + 0.103 × MG (p < 0.05)

The predicted AA content value was in good agreement with the actual AA content.

Acknowledgements

This work was supported by the Fund of National Basic Research Program of China (“973” Program, 2012CB720805), National High Technology Research and Development Program of China (“863” Project, 2011AA100806), and the Fund of Projects in the National Science & Technology Program during the Eleventh Five-Year Plan Period (No. 2009BADB9B07). Accordingly, the authors gratefully acknowledge the funds supports.

References

1. International Agency for Research on Cancer (IARC), Monographs on the Evaluation of Carcinogenic Risks to Humans, 60, 1994, Lyon, France, http://monographs.iarc.fr/ENG/Monographs/vol60/mono60-16.pdf.
2. D. S. Mottram, B. L. Wedzicha and A. Dodson, Nature, 2002, 419, 448.
3. R. H. Stadler, I. Blank, N. Varga, F. Robert, J. Hau and P. A. Guy, Nature, 2002, 419, 449.
4. V. A. Yaylayan, A. Wnorowski and C. Locas Perez, J. Agric. Food Chem., 2003, 51, 1753.
5. E. N. Frankel, Lipid oxidation, The Oily Press, Dundee, Scotland, 1998.
6. A. Becalski, B. P. Y. Lau, D. Lewis and S. W. Seaman, J. Agric. Food Chem., 2003, 51, 802.
7. C. Gertz and S. Klostermann, Eur. J. Lipid Sci. Technol., 2002, 104, 762.
8. K. Umano and T. Shibamoto, J. Agric. Food Chem., 1987, 35, 909.
9. G. Arribas-Lorenzo, V. Fogliano and F. J. Morales, Eur. Food Res. Technol., 2009, 229, 63.
10. A. Kamal-Eldin and L. A. Appelqvist, Grasas Aceites, 1996, 47, 342.
11. D. B. Min and J. M. Boff, Chemistry, nutrition, and biotechnology, Marcel Dekker, New York, 2nd edn, 2002, pp. 365–411.
12. P. R. Jannalagadda, R. V. Bhat, R. V. Sudershan and A. N. Naidu, Food Qual. Preference, 2001, 12, 223.
13. M. D. Guillén and N. Cabo, Food Chem., 2002, 77, 503.
14. S. Naz, R. Siddiqi, H. Sheikh and S. A. Sayeed, Food Res. Int., 2005, 38, 127.
15. T. Hofmann, W. Bors and K. Stettmaier, J. Agric. Food Chem., 1999, 47, 379.
16. A. Hollnagel and L. W. Kroh, J. Agric. Food Chem., 2000, 48, 6219.
17. S. I. F. S. Martins, W. M. F. Jongen and M. A. J. S. Van Boekel, Trends Food Sci. Technol., 2001, 11, 364.
18. Y. Yuan, G. H. Zhao, X. S. Hu, J. H. Wu, J. Liu and F. Chen, Eur. Food Res. Technol., 2008, 226, 1301.
19. T. M. Amrein, A. Limacher, B. Conde-Petit, R. Amadò and F. Escher, J. Agric. Food Chem., 2006, 54, 5910.
20. V. Gökmen, T. K. Palazoğlu and H. Z. Şenyuva, J. Food Eng., 2006, 77, 972.
21. E. Capuano, T. Oliviero, O. Ç. Açar, V. Gökmen, V. Fogliano, Food Res Int, http://www.elsevier.com/locate/foodres.
22. J. Liu, G. H. Zhao, Y. Yuan, F. Chen and X. S. Hu, Food Chem., 2008, 108, 760.
23. W. Bednarski, L. Jedrychowski, E. G. Hammond and Z. L. Nikolov, J. Dairy Sci., 1989, 72, 2474.
24. G. De Revel, L. Pripis-nicolau, J. C. Barbe and A. Bertrand, J. Sci. Food Agric., 2000, 80, 102.
25. GB/T 5009.37—2003. Method for analysis of hygienic standard of edible oils.
26. GB/T 24304—2009/ISO6885:2006. Animal and vegetable fats and oils—Determination of anisidine value.
27. GB 5009.3—2010. National food safety standard determination of moisture in foods.
28. T. Wicklund, H. Østlie, O. Lothe, S. H. Knutsen, E. Bråthen and A. Kita, LWT–Food Sci. Technol., 2006, 39, 571.
29. F. Pedreschi, K. Kaack and K. Granby, LWT–Food Sci. Technol., 2004, 37, 679.
30. K. Kotsiou, M. Tasioula-Margari, E. Capuano and V. Fogliano, Food Chem., 2011, 124, 242.
31. S. Romani, M. Bacchiocca, P. Rocculi and M. D. Rosa, J. Food Compos. Anal., 2009, 22, 582.
32. S. Casal, R. Malheiro, A. Sendas, B. P. P. Oliveira and J. A. Pereira, Food Chem. Toxicol., 2010, 48, 2972.
33. S. Debnath, N. K. Rastogi, A. G. G. Krishna and B. R. Lokesh, Food Bioprod. Process., 2012, 90, 249.
34. C. Gertz, Eur. J. Lipid Sci. Technol., 2004, 106, 736.
35. C. Gertz, S. Klostermann and S. P. Kochhar, Ol., Corps Gras, Lipides, 2003, 10, 297.
36. Y. Yuan, G. H. Zhao, F. Chen, J. Liu, J. H. Wu and X. S. Hu, Food Chem., 2008, 108, 885.
37. H. Q. Ye, Y. T. Miao, C. C. Zhao and Y. Yuan, Int. J. Food Sci. Technol., 2011, 46, 1921.
38. G. Arribas-Lorenzo and F. J. Morales, J. Agric. Food Chem., 2010, 58, 2966.
39. R. J. Bull, M. Robinson, R. D. Laurie, G. D. Stoner, E. Greisiger, J. R. Meier and J. Stober, Cancer Res., 1984, 44, 107.
40. B. MacMillan, H. Hickey, B. Newling, M. Ramesh and B. Balcom, Food Res. Int., 2008, 41, 676.
41. T. Sanz, C. Primo-Martín and T. van Vliet, Food Res. Int., 2007, 40, 63.

---