LWT - Food Science and Technology 159 (2022) 113197

Contents lists available at ScienceDirect

LWT
journal homepage: www.elsevier.com/locate/lwt

Phospholipase cocktail: A new degumming technique for crude soybean oil

Rafaela Menezes dos Passos a, *, Rúbia Mariana da Silva a, Paula Virginia de Almeida Pontes a,
Marcelo Antônio Morgano b, Antônio J.A. Meirelles a, Christian V. Stevens c,
Marcela Cravo Ferreira d, Klicia Araujo Sampaio a
a
Laboratory of Extraction, Applied Thermodynamics and Equilibrium (ExTrAE), University of Campinas, Faculty of Food Engineering, Department of Food Engineering,
Rua Monteiro Lobato, 80, 13083-862, Campinas, SP, Brazil
b
Instituto de Tecnologia de Alimentos – ITAL, Campinas/SP, CEP, 13070-178, Brazil
c
Department of Green Chemistry and Technology, Faculty of Bioscience Engineering, Ghent University, Campus Coupure, Coupure Links 653, Gent, Belgium
d
Faculty of Technology, University of Campinas (UNICAMP), Rua Pascoal Marmo, 13484-332, Limeira, SP, Brazil

A R T I C L E I N F O A B S T R A C T

Keywords: Enzymatic degumming (EDG) is an emerging alternative process for decreasing the phosphorus content,
Phosphorus content increasing the oil yield, and preserving the oil quality. Purifine® 3G is a cocktail of phospholipases composed of
Oil yield phospholipase A2 (PLA2), phospholipase C (PLC), and phosphatidylinositol phospholipase C (PI-PLC). In this
Phospholipid
study, Purifine® 3G was applied to crude soybean oil, and the optimum degumming conditions (enzyme con­
Enzymatic degumming
Vegetable oil
centration, temperature, and water dosage) were determined using a central composite rotatable design (CCRD).
The contents of diacylglycerols (DAGs) and free fatty acids (FFAs) in the studied system considerably increased at
temperatures below 64 ◦ C and enzyme concentrations above 100 mg/kg, while the phosphorus content
decreased with increasing water amount and enzyme concentration. In particular, EDG with 200 mg/kg of
Purifine® 3G conducted for 120 min at a temperature of 60 ◦ C and water concentration of 3% (w/w) lowered the
residual phosphorus content to 8.9 mg/kg and increased the FFA and DAG concentrations by 0.17% and 0.72%,
respectively. Meanwhile, EDG retained the tocopherol content in crude soybean oil, maintaining its quality.
Hence, Purifine® 3G increases the neutral oil yield (FFA and DAG), decreases the phosphorus content, and
preserves the oil quality, which make it a commercially viable degumming agent.

Degumming is the first and most important step of the refinement

process that separates the majority of phospholipids and gums from oils,
1. Introduction which is imperative for the production of high-quality oils (Jiang,
Chang, Jin, & Wang, 2015a). To ensure high phospholipid removal ef­
Soybeans represent the most commonly produced oilseeds in Brazil. ficiency, an oil with a phosphorus content of less than 10 mg/kg should
Its soybean planted area is expected to expand by 38.5 million hectares be obtained (Jahani, Alizadeh, Pirozifard, & Qudsevali, 2008). Phos­
during the 2020/21 season, which will increase the soybean production pholipids constitute 0.3–0.6% of soybean seeds or 1.5–3.0% of crude
by 131.5 million metric tons (MMT). The forecast for soybean exports in soybean oil (Liu & Ma, 2011). Conventional techniques used for the
2020/21 was 85 MMT, and the high demand for soybean oil is mainly removal of phospholipids include water degumming and acid degum­
driven by its domestic consumption (Ustinova, 2020). Crude soybean oil ming. Water degumming is used to remove hydratable phospholipids,
obtained by oilseed processing must be refined before consumption to while acid degumming removes non-hydratable phospholipids through
remove unwanted components such as free fatty acids (FFAs), metal the addition of acids such as phosphoric and citric ones (Passos et al.,
traces, and phospholipids with minimal losses of useful compounds, 2019).
including acylglycerols, free and esterified sterols, tocopherols, and Enzymatic degumming is an emerging process for the removal of
tocotrienols. According to Sampaio et al. (2015), the refinement of phospholipids from crude oils that utilizes enzymes called phospholi­
edible oils can be performed by either chemical or physical processing. pases. Phospholipase A1 (PLA1) and phospholipase A2 (PLA2) remove
In general, physical refining is preferable due to its environmental fatty acids (FAs) from positions 1 and 2 of phospholipids with respect to
benefits and sustainability.

* Corresponding author.
E-mail addresses: [email protected] (R.M. dos Passos), [email protected] (K.A. Sampaio).

https://doi.org/10.1016/j.lwt.2022.113197
Received 26 April 2021; Received in revised form 23 January 2022; Accepted 4 February 2022
Available online 8 February 2022
0023-6438/© 2022 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
R.M. dos Passos et al. LWT 159 (2022) 113197

inductively coupled plasma – atomic emission spectroscopy according to

Abbreviations Method Ca 20–99 (American Oil Chemists’Society, 2012a). The pH of
the gums was measured directly with a pH electrode inserted into the
ANOVA analysis of variance gum fraction. FFA content was determined by titration according to the
CCRD central composite rotatable design official Ca 5a-40 method of the American Oil Chemists’ Society
DAG diacylglycerol (American Oil Chemists’Society, 2012b).
EDG enzymatic degumming DAG content was determined according to the AOCS official method
FA fatty acid Cd 11b-91 (American Oil Chemists’Society, 2012c). Approximately
FFA free fatty acid 0.05 g of the analyzed sample was dissolved in 100 μL of tetradecane and
MMT million metric ton 300 μL of N,O-bis((trimethylsilyl)trifluoroacetamide), and the resulting
PA phosphatic acid mixture was homogenized and heated to 70 ◦ C for 20 min. After that, 50
PC phosphatidylcholine μL of the derivatized sample was diluted with hexane (1 mL) and
PE phosphatidylethanolamine injected into a gas chromatograph (Agilent Technologies, model 7890A)
PI-PLC phosphatidylinositol phospholipase C equipped with a DB-5HT capillary column (15 m × 0.32 mm i.d.; film
PLA2 phospholipase A2 thickness: 0.10 μm). Analysis was performed under the following con­
PLC phospholipase C ditions: oven temperature ramp from 50 to 200 ◦ C (15 ◦ C/min), from
200 to 290 ◦ C (3 ◦ C/min; held for 10 min), and from 290 to 360 ◦ C
(10 ◦ C/min; held for 15 min); flame ionization detector (380 ◦ C); and He
carrier gas. DAGs were identified using a diolein standard.
glycerol moieties, respectively. Phospholipase C (PLC) hydrolyzes the Tocopherol content was measured as described by Ansolin, Souza,
bond between the acylglycerol and phosphate groups to liberate diac­ Meirelles, and Batista (2017). The utilized solvents included HPLC grade
ylglycerols (DAGs) (Jiang, Chang, Jin, & Wang, 2015b; Qu et al., 2016; methanol and isopropanol (J. T. Backer, USA), while Milli-Q water and
Sampaio, Zyaykina, Uitterhaegen, Greyt, & Verhé, 2019). To facilitate HPLC grade ammonium hydroxide solution (28–30%) (J. T. Backer,
the enzymatic degumming process, chemical conditioning is typically USA) were used to prepare a mobile phase. Tocopherol standards (α, β, γ,
performed in the previous step. As formulated by Sampaio et al. (2015), and δ) were utilized for tocol quantification (Merck, USA). The mobile
the purpose of chemical conditioning is to increase the hydratability of phase was composed of a methanol: water: ammonia mixture (99:1:0.1,
non-hydratable phospholipids and determine the optimal pH value v/v/v, phase A) and pure isopropanol (phase B). Analysis was performed
corresponding to the maximum enzyme activity. on a Waters Acquity SQD/UPLC System (USA) equipped with a qua­
Purifine® 3G is an enzyme cocktail composed of phospholipase C ternary pump, an automatic injector, a column oven, a photodiode array
(PLC), phosphatidylinositol – specific phospholipase C (PI-PLC), and a detector, and a single quadrupole mass spectrometer with electrospray
minor amount of PLA2, which effectively converts phospholipids into ionization operated in the negative mode. Chromatographic separation
predominantly diglycerides (DAGs), phosphates, FFAs, and lysophos­ was achieved at 25 ◦ C using an Acquity UPLC BEH C18 column (1.7 μm,
pholipids (Nikolaeva et al., 2020). While each phospholipase generates a 2.1 × 100 mm, Waters, USA). The gradient program was 0–6.00 min
specific product, the cocktail not only increases the FFA and DAG con­ 100% A (0.2 mL/min), 6.01–9.00 min 100% B (0.15 mL/min), and
centrations, but also decreases the phosphorus content in a one-step 9.01–15.00 min 100% A (0.2 mL/min). All samples were analyzed in
process. duplicate.
Very few previous studies have focused on increasing the neutral oil
yield, decreasing the phosphorus content, and preserving the tocopherol 2.3. Degumming efficiency
content during degumming vegetable oils with the Purifine® 3G cock­
tail. Nikolaeva et al. (2020) examined the gum mesostructures formed in Degumming efficiency was calculated using the formula proposed by
crude soybean oil after water degumming and enzymatic degumming Lamas, Constenla, and Raab (2016), which was based on the decrease in
with the Purifine® 3G cocktail; however they did not optimize the phosphorus content:
process parameters. Sein, Hitchman, and Dayton (2019) evaluated the
degumming process using the Purifine 3G cocktail from a molecular Efficiency(%) =
Pi − Pr
x 100 (1)
point of view. Guerrand (2017) discussed various enzyme applications Pi
and the use of Purifine 3G for the production of edible oils; nevertheless,
the author did not mention its applicability to a specific vegetable oil. Pi: the total phosphorus content in crude oil (mg/kg);
Therefore, this study aimed to investigate the effects of different process Pr: the residual phosphorus content in the degummed oil (mg/kg)
conditions, such as temperature, water dosage, and enzyme (Purifine®
3G) concentration, on the enzymatic degumming of crude soybean oil.
2.4. Theoretical phospholipid content and increases in DAG and FFA
concentrations
2. Material and methods
The content of phospholipids was determined from the conversion of
2.1. Oil and enzyme phosphorus to phospholipids calculated using the ratio of the phos­
phorus atomic weight (P = 31) to the estimated molecular weight of the
Crude soybean oil was kindly provided by Cargill (Uberlandia-MG/ phospholipids (PL), which was equal to approximately 25 (Galhardo &
Brazil). The cocktail Purifine® 3G (PLC + PI-PLC + PLA2) with an ac­ Dayton, 2021):
tivity of 16900 PLCU/g was supplied by DSM Food Specialties (Delft, the
Netherlands). PLA2 enzyme was produced by a selected strain of PL(%) =
25 x P(mg/kg)
(2)
Aspergillus niger; PLC enzyme was produced by a selected strain of Pichia 10000
pastoris; and PI-PLC enzyme was produced by a selected strain of Pseu­ The increases in the FFA and DAG concentrations were obtained by
domonas fluorescence. the method described in the work of Galhardo and Dayton (2021) with
some modifications. The maximum theoretical increase in the FFA
2.2. Physicochemical analysis content after enzymatic degumming was computed from the phospho­
lipid content in crude oil multiplied by the ratio of the molecular mass of
P, Mg, Fe, and Ca elemental concentrations were obtained by FFA (282 g/mol) to the average molecular mass of phospholipids

2
R.M. dos Passos et al. LWT 159 (2022) 113197

(phosphatidylcholine (PC): 758 g/mol; phosphatidylethanolamine (PE): design (CCRD – 23) was used. Enzyme dosage (X1), temperature (X2),
692 g/mol; phosphatidylinositol (PI): 887 g/mol; phosphatidic acid and water dosage (X3) were treated as independent variables. A total set
(PA): 699 g/mol), which was equal to approximately 759 g/mol. Here, of 17 experiments (= 2k + 2, k + 3 repetitions at the central point),
the authors assumed that all FAs remained in the oil. The estimated FFA where k is the number of independent variables with five different levels
concentration increase is expressed by the following equation: (= 1.68), is provided in Table 1. The relationship between the coded and
actual values used for statistical calculations is described by the
282
FFA (%) = PL x (3) following equation:
759
Ai − A 0
where FFA is the relative amount of FFAs formed during the enzymatic Xi = (5)
ΔA
reaction (%), and PL is the total phospholipid content in crude oil (%).
The maximum theoretical DAG concentration increase was calcu­ where Xi is the coded value of the variable, Ai is the actual value of the
lated as the content of selected phospholipids present in crude oil (PC, variable, A0 is the actual value of Ai at the central point, and ΔA is the
PE, and PI) multiplied by the ratio of the molecular mass of DAG (617 g/ step change of the variable (the obtained design matrix is also listed in
mol) to the average molecular mass of these phospholipids (PC = 758 g/ Table 1). The phosphorus, DAG, and FFA contents were treated as re­
mol; PE = 692 g/mol; PI = 887 g/mol), which was equal to approxi­ sponses. The model for each independent variable was generated using
mately 779 g/mol. In this case, the authors assumed the reaction effi­ the following quadratic equation:
ciency of 85% due to the partial reaction of PE. The estimated DAG
Y = B0 + Σβi Xi + Σβii Xi2 + Σβij Xi Xj (6)
concentration increase is expressed by the following equation:
617 where Y is the predicted response, β0 is the constant, βi is the linear
DAG = (PC + PE + PI) x x 0.85 (4)
779 coefficient, βii is the squared coefficient, βij is the cross-product coeffi­
cient, Xi is the independent variable (uncoded value), and i and j are the
where DAG is the amount of DAGs formed during the enzymatic reaction levels of significance of the terms.
(%), and (PC + PE + PI) is the total amount of phospholipids hydrolyzed The quality of the design was estimated by computing the determi­
by PLC (%). nation coefficient (R2), while the analysis of the variance (ANOVA) was
performed to evaluate the statistical significance of the proposed model
2.5. Water degumming by calculating the regression values and mean squared residual error.
The graphical representation of Eq. (6), known as the response surface,
Water degumming was performed as described previously by Sam­ was applied to determine the interactions between the independent
paio et al. (2015). Crude soybean oil (300 g) was first heated to 80 ◦ C variables and their effects on each response. The obtained data were
followed by water addition (3% w/w), after which the resulting mixture analyzed using the Statistica 5.0 software package (StatSoft Inc., United
was homogenized by mechanical stirring (350 rpm) for 15 min. The States). Statistically significant differences were observed at the p < 0.1
produced gums were separated by centrifugation (2000×g for 15 min). level. The Tukey test was performed to analyze the samples obtained
after aqueous degumming and chemical conditioning as well as crude oil
2.6. Chemical conditioning and tocopherol samples. Here, significant differences were observed at
p ≤ 0.05.
Crude soybean oil (300 g) was heated to 80 ◦ C followed by the
addition of 30% aqueous solution of citric acid at a concentration of 900 3. Results and discussion
mg/kg. The resulting system was high-shear mixed for 1 min at a speed
of 16000 rpm and subsequently agitated for 15 min at 350 rpm. After 3.1. Soybean oil characterization
that, an aqueous solution of sodium hydroxide (14%) was added to the
reaction mixture at a concentration of 500 mg/kg. The oil was subjected The concentrations of FFA, DAG, phospholipids, and minerals in
to high shear mixing (1 min/16000 rpm) to disperse the caustic solution. crude soybean oil, the degummed water, and the chemically conditioned
The produced gums were separated from the degummed oil by centri­ oil are listed in Table 2. The FFA contents in the degummed water and
fugation (2000×g for 15 min) (Sampaio et al., 2015). chemically conditioned oil were lower than that in crude soybean oil as
both processes removed acidic compounds. The DAG concentration was
2.7. Enzymatic degumming reduced by a higher value (56%) during water degumming than that
obtained after chemical conditioning (41%). According to Hitchman
Enzymatic degumming experiments were performed using 300 g of (2009), this phenomenon was caused by the formation of an emulsion
crude soybean oil. In the first step, the solution pH was adjusted to 5.5 by due to the presence of intact phospholipids (such as PC, which was the
chemical conditioning. Subsequently, the oil was heated to 80 ◦ C for 15 strongest emulsifier), which increased the oil content in the gums and,
min under stirring at 350 rpm, after which the temperature of the oil therefore, the oil loss. The estimated phospholipid contents in crude
mixture was decreased to the values established by a central composite soybean oil, the degummed water, and the chemically conditioned oil
rotatable design (CCRD). The amounts of added Purifine® 3G and water were 1.52%, 0.11%, and 0.04%, respectively. The mineral content in
were also determined according to the CCRD. The obtained mixture was soybean oil decreased after both processes. First, water degumming
homogenized under high-shear mixing (16000 rpm) for 1 min and then removed the hydrated forms of phospholipids (PC, PI, and partially PE),
maintained at a required temperature under mixing (350 rpm) for 120
min. The reaction was stopped by heating the oil to 85 ◦ C for 15 min at Table 1
350 rpm. Thereafter, the oil mixture was centrifuged (2000×g for 15 Independent variables and their levels used for the CCRD-23 during the enzy­
min) to separate the degummed oil from the gums. matic degumming of crude soybean oil.
Variables Symbols Levels
2.8. Experimental design and data analysis − 1.68 − 1 0 1 1.68
Enzyme (Purifine® 3G) X1 30 64.4 115 165.6 200
Response surface methodology was adopted to optimize the enzy­ (mg/kg)
matic degumming of crude soybean oil. To obtain the minimum phos­ Temperature (◦ C) X2 50 54 60 66 70
Water dosage (% w/w) X3 1.0 2.2 3.0 4.2 5.0
phorus, maximum FFA, and maximum DAG contents, a central rotatable

3
R.M. dos Passos et al. LWT 159 (2022) 113197

decreasing the phosphorus content by 93%. Afterwards, acid condi­ Table 2

tioning removed the non-hydratable forms of phospholipids (PA and Characterization of crude soybean oil, water degummed and chemical condi­
partially PE) that were bound to divalent ions such as Ca, Mg, and Fe, tioned oil.
resulting in a 97% reduction in the phosphorus content. Similar results Analyses Crude soybean Water Chemical
were obtained by Sampaio et al. (2019) who performed the enzymatic oil Degumming Conditioning
degumming of crude corn oil by traditional processes. FFA (%C18:1) 1.24 ± 0.04 1.01 ± 0.06 1.00 ± 0.10
DAG (%) 1.67 ± 0.07 0.73 ± 0.04 0.98 ± 0.02
Phospholipidsa 1.52 0.11 0.04
(%)
3.2. Fitting models
Minerals content (mg/kg)

The experimental runs were performed using the independent vari­ P 608 ± 33.0 45 ± 0.4 19 ± 1.0
Fe 11.7 ± 0.1 0.34 ± 0.01 ND < 0.1
ables and their ranges specified in Table 1. The FFA (Y1), DAG (Y2), and
Ca 52 ± 3.0 23.8 ± 0.4 2.06 ± 0.08
phosphorus (Y3) contents were determined for each condition, and the Mg 58 ± 3.0 10.6 ± 0.2 1.44 ± 0.05
obtained results are presented in Table 3. The data were analyzed using a
Phospholipids estimated by Equation (2).
the Statistica 5.0, software, which generated different models, model
coefficients, R2 values, F-values, and significant probabilities. Using
these parameters, the significance of each experimental variable was the loss of its hydrolytic activity.
obtained. For simplicity purposes, the non-significant terms were The phosphorus content (Eq. (11)), the most important parameter, is
removed from each model. However, the statistically non-significant strongly affected by water addition followed by enzyme concentration.
linear terms were retained when the respective quadratic effects were According to Sampaio et al. (2015), the main portion of the phospho­
statistically significant according to Jahani et al. (2008). Using this lipids present in crude soybean oil is hydratable and thus form micellar
approach, the representativeness of a selected model was improved, the structures with water, which are insoluble in oil due to their polar
importance of the term in its linear and quadratic form, and possible properties and can be easily removed by centrifugation. In this case, the
interactions were determined. According to the obtained regression role of the enzyme is to hydrolyze the phospholipids into FFA or DAG
models, the parameters X1, X12, X2, X22, and X1X3 had significant effects molecules, thereby avoiding their removal in the intact form, which can
on the FFA content; the parameters X1, X2, X3, X32, and X1X2 were sig­ significantly contribute to the oil loss.
nificant for the DAG content; and the parameters X1, X3, and X32 pro­ Table 4 lists the ANOVA results. According to Rodrigues and Iemma
duced significant effects on the phosphorus content. The second-order (2014), parameters with a 90% significance can be considered signifi­
polynomial equations containing the coded factors obtained from the cant due to the variability of bioprocesses involving enzymes and mi­
CCRD model for FFA (%), DAG (%), and phosphorus (mg/kg) content croorganisms. The R2 values obtained for the FFA, DAG, and phosphorus
are listed below: contents using the mathematical model were 70.3%, 92.3%, and 87%,

FFA content (%C18 : 1) = 1.08 + 0.086X1 − 0.044X12 + 0.037X2 − 0.058X22 − 0.044X1 X3 (9)

respectively. Although the R2 of the FFA content (70.3%) was not very
high, Fcalc.(regression/residuals) > Ftab, indicating a linear regression.
Furthermore, Fcalc.(lack of fit/pure error) < Ftab, suggesting that the
DAG content (%) = 1.38 + 0.51X1 − 0.18X2 + 0.022X3 − 0.16X32 + 0.21X1 X2
(10)
Table 3
Phosphorus content(mg / kg) = 12.56 − 11.00X1 − 34.52X3 + 38.25X32 (11) Independent, dependent variables, and results for each condition of CCRD 23.
Independent variables Dependent variables
where X1, X2, and X3 are the coded variables for the enzyme concen­
tration, temperature, and water dosage, respectively. Trials X1a X2a X3a Y1 Y2 Y3
As indicated by the obtained statistical models, the enzyme con­ 1 − 1.00 − 1.00 − 1.00 1.00 1.26 37.8
centration produced the strongest effects on the FFA (Eq. (9)) and DAG 2 − 1.00 − 1.00 1.00 1.00 1.32 15.3
(Eq. (10)) contents followed by temperature. Enzyme activity can be 3 − 1.00 1.00 − 1.00 0.94 0.92 52.7
4 − 1.00 1.00 1.00 1.07 0.95 35.6
affected by several factors, including enzyme concentration, tempera­ 5 1.00 − 1.00 − 1.00 1.01 1.66 30.0
ture, reaction time, and solution pH. According to De Greyt (2013), 6 1.00 − 1.00 1.00 1.01 1.60 9.30
enzyme concentration depends on the enzyme type and phospholipid 7 1.00 1.00 − 1.00 1.08 1.63 35.1
content in the oil. Phospholipases are enzymes that cleave the bonds 8 1.00 1.00 1.00 1.07 1.65 15.5
9 − 1.68 0.00 0.00 0.90 0.92 27.2
between the glycerol backbone and FAs, DAGs, and phosphate ester. The
10 1.68 0.00 0.00 1.17 1.70 8.90
enzymes used in the present study existed in the form of a cocktail 11 0.00 − 1.68 0.00 0.98 1.44 21.0
known as Purifine® 3G. According to Gupta (2017, pp. 60–76), this 12 0.00 1.68 0.00 1.05 1.15 13.0
cocktail is specific to all four common phospholipids (PC, PE, PA, and 13 0.00 0.00 − 1.68 1.10 1.02 129.0
PI). It contains PLC that interacts with PC and PE and PI-PLC that reacts 14 0.00 0.00 1.68 1.07 1.08 11.4
15 0.00 0.00 0.00 1.07 1.34 21.0
with PI to produce DAG and phosphate ester. The other enzyme present 16 0.00 0.00 0.00 1.10 1.35 28.7
in Purifine® 3G is PLA2, which is selective for PC, PE, and PA. It cleaves 17 0.00 0.00 0.00 1.10 1.47 16.0
the FA from position 2 of the glycerol backbone to produce FFA and
*X1 - Enzyme concentration (mg/kg); X2 – Temperature (◦ C); X3 – Water dosage
lysophospholipids. Note that enzyme activity is high within a narrow
(%w/w).
range of temperatures, where the increase in temperature can increase **Y1 – FFA content (as % C18:1); Y2 – DAG content (%) and Y3 – Phosphorus
the reaction rate; however, the temperatures higher than the optimal content (mg/kg)/.
value can cause a partial denaturation of the enzyme accompanied by a
For detailed information of the process conditions see Table 1.

4
R.M. dos Passos et al. LWT 159 (2022) 113197

Table 4
ANOVA results for the responses: FFA, DAG and Phosphorus content.
Independent variable Variation Source Sum of Squares Degrees of Freedom Mean Square F calc. F tab. p-value

FFA content (%C18:1) Regression 0.05 5 0.01 5.25 2.45 0.00

Residuals 0.02 11 0.00
Lack of Fit 0.02 9 0.00 7.60 9.38 0.12
Pure Error 0.00 2 0.00
Total 0.07 16
R2 70.3%
DAG content (%) Regression 1.20 5 0.24 26.25 2.45 0.00
Residuals 0.10 11 0.01
Lack of Fit 0.09 9 0.01 1.92 9.38 0.40
Pure Error 0.01 2 0.01
Total 1.30 16
R2 92.3%
Phosphorus content (mg/kg) Regression 9398.70 3 3132.90 24.94 2.56 0.00
Residuals 1632.91 13 125.61
Lack of Fit 1551.05 11 141.00 3.45 9.39 0.24
Pure Error 81.86 2 40.93
Total 11031.60 16
R2 87.0%

*Significance at p < 0.1.

Fig. 1. Experimental and predicted values for A) FFA content, B) DAG content and C) Phosphorus content.

generated equation can adequately describe the analyzed data. Ac­ phosphorus content of 129 mg/kg. These results indicate that the higher
cording to Rodrigues and Iemma (2014), a suitable model must satisfy is the water content, the larger is the amount of hydratable phospho­
both the Fcalc.(regression/residuals) > Ftab and Fcalc.(lack of fit/pure error) < Ftab lipids removed from the oil after centrifugation (Costa, Almeida,
requirements for all responses. Alvim-Ferraz, & Dias, 2018).
As shown in Fig. 1A, the FFA content exhibits low absolute variations
(ranging from 0.9% to 1.27%), which may increase the difference be­
tween the experimental and predicted values. In Fig. 1C, the water 3.3. Analysis of generated response surfaces
contents between 3 and 5% correspond to the phosphorus contents of
8.9–35.6 mg/kg, and with the water content of 1% produced a By using the obtained models, it was possible to draw surfaces that
described the influences of enzyme concentration, temperature, and

5
R.M. dos Passos et al. LWT 159 (2022) 113197

Fig. 2. Response surface of interaction among A) Temperature (◦ C) and Enzyme conc. (mg/kg) for FFA content; B) Temperature (◦ C) and Enzyme conc. (mg/kg) for
DAG content; C) Water dosage (%) and Enzyme conc. (mg/kg) for FFA content; D) Water dosage (%) and Enzyme conc. (mg/kg) for FFA content; and E) Water dosage
(%) and Enzyme (mg/kg) for Phosphorus content.

6
R.M. dos Passos et al. LWT 159 (2022) 113197

water dosage on the FFA, DAG, and phosphorus contents. It is well two parameters reduces both the FFA and DAG concentrations because
known that the FFA group can be removed by PLA2 from position 2 of a the enzyme requires a certain amount of water for its activation and the
phospholipid molecule with respect to glycerol during the enzymatic hydrolysis of phospholipids. Jahani et al. (2008) studied the enzymatic
degumming process. Furthermore, PLC hydrolyzes the bond between degumming of rice bran oil using PLA1 and noticed that the FFA content
acylglycerol and the phosphate group to release DAGs. In both cases, the was reduced after decreasing the amount of water in the reaction
enzymes generate compounds that are present in the oil, thus increasing mixture. Regarding the DAG content, Jiang, Chang, Jin, and Wang
the oil yield. (2014) observed that higher water dosages increased the DAG content
Fig. 2A and B shows that increasing the enzyme concentration (expressed in terms of the oil recovery efficiency) during soybean gum
(>100 mg/kg) and reaction temperature (<64 ◦ C) increases the FFA and deoiling by PLC.
DAG contents, confirming the existence of synergistic effects between Fig. 2E shows that the phosphorus content decreases with increasing
the variables. According to Eq. (3), an FFA increase of 0.56% was ex­ enzyme concentration and water dosage because the enzyme promotes
pected only for the PLA2 component. However, because the cocktail had the hydrolysis of phospholipids and water directly contributes to the
an unspecified small amount of PLA2, a higher FFA content was detected removal of hydratable phospholipids. The chemical conditioning pro­
in trial 10, which corresponded to only 31% of the total amount. Lamas cess was characterized by the degumming efficiency of 97%, and trials 6
et al. (2016) performed the enzymatic degumming of sunflower oil using and 10 exhibited a degumming efficiency of 98.5% (Eq. (1)) with P < 10
MAXAPAL (PLA2) as a result, the system acidity increased by 1.05 (mg mg/kg, which was required for physical refining. According to De Greyt
KOH/g) due to the phospholipase action. and Kellens (2005), efficient removal of phospholipids during vegetable
To calculate the theoretical DAG increase, the phospholipid oil refining is essential for the production of high-quality oils because
composition of vegetable oil should be considered. For crude soybean residual phospholipids can cause oil darkening and produce off-flavors.
oil, the ranges of the PC, PE, and PI contents relative to the total phos­ In trial 10 (Table 3) in comparison to the water degumming and
pholipids content are equal to 25–33%, 19–31%, and 10–18%, respec­ chemical conditioning processes (Table 2), increased the FFA content by
tively (Dayton & Galhardo, 2014; Jiang et al., 2015b; Sampaio et al., 0.16% and 0.17% and the DAG content by 0.97% and 0.72%, respec­
2015). Thus, by applying these values to the total phospholipid content tively. The phosphorus content decreases observed during enzymatic
in crude soybean oil (1.52%), a range of 0.82–1.24% was obtained for degumming (trial 10), water degumming, and chemical conditioning
the mixture of PC, PE, and PI. This was utilized to calculate the were equal to 98.5%, 92.6%, and 96.8%, respectively. The Purifine 3G
maximum theoretical DAG increase via Eq. (4), which varied between exceeded those of traditional degumming methods.
0.55 and 0.83%. After comparing the results of trial 10 with the chem­
ical conditioning data (1.00% FFA), a 0.17% increase in the FFA content
3.4. Tocopherol profiles obtained for different soybean oil samples
and 0.72% increase in the DAG content were observed, indicating that
almost all phospholipids (98.8%) have been converted. Nikolaeva et al.
Edible oils are major natural sources of tocopherols and tocotrienols
(2020) used Purifine® 3G and PLA2 from DSM to perform the enzymatic
known as tocols. These compounds are among the most important lipid-
degumming of crude soybean oil. For Purifine® 3G, the authors ob­
soluble antioxidants present in foods as well as in human and animal
tained 2% (w/w) of FFA and for PLA2 –7% (w/w) of FFA, indicating that
tissues. According to Shahidi and De Camargo (2016), the most abun­
the fraction of PLA2 in the cocktail strongly influenced the FFA con­
dant types of tocopherols in soybean oil include α-tocopherol and
centration. Sampaio et al. (2019) performed the enzymatic degumming
γ-tocopherol. The concentration of these tocopherols in crude vegetable
of crude corn oil using Purifine® PLC; as a result, the DAG content
oils vary considerably due to climatic and agronomic factors, fruit
increased by 0.54% at a temperature of 60 ◦ C and enzyme concentration
quality, seed origin, and the utilized oil extraction system (Cert, Moreda,
of 200 mg/kg.
& Pérez-Camino, 2000). Hence, the influences of enzyme concentration,
Fig. 2C and D depict the effects of enzyme concentration and water
reaction temperature, and water dosage on the enzymatic degumming of
dosage on the FFA and DAG contents. They show that decreasing these
soybean oil were studied in this work. The effects of the degumming

Fig. 3. Tocopherol profile for different soybean oil samples: CSBO (crude soybean oil), trials 9, 10, 11, 12, 13 and 14. Different letters in the same tocopherols type
indicate significant differences (p < 0.05).

7
R.M. dos Passos et al. LWT 159 (2022) 113197

process on the tocopherol profiles obtained for different soybean oil acknowledge Cargill for kindly supplying the samples of crude soybean
samples are shown in Fig. 3. The concentrations of all tocopherol ho­ oil and DSM Food Specialties for providing the phospholipase cocktail.
mologues were significantly different (p < 0.05) from those in crude
soybean oil due to the process conditions. In trials 9 and 10, increasing References
the enzyme concentration from 30 to 200 mg/kg retained 6.3%, 10.9%,
and 11.4% of β/γ-tocopherol, α-tocopherol, and δ-tocopherol species, American Oil Chemists’ Society. (2012a). AOCS official method Ca 20-99. Analysis of
phosphorus in oil by inductively coupled plasma optical emission spectroscopy
respectively, owing to the stronger enzyme action and lower degree of
(ICP–AES). In Official methods and recommended practices of the AOCS (6th ed.). AOCS
oil entrainment. A comparison of trials 11 and 12 revealed that Press.
increasing the reaction temperature from 50 to 70 ◦ C reduced the American Oil Chemists’ Society. (2012b). AOCS official method Ca 5a-40. Free fatty
acids. In Official methods and recommended practices of the AOCS (6th ed.). AOCS
β/γ-tocopherol, α-tocopherol, and δ-tocopherol concentrations by 4.0%,
Press.
6.0%, and 6.7%, respectively, which was likely caused by the hydrolysis American Oil Chemists’ Society. (2012c). AOCS official method Cd 11b-91. Mono- and
and/or oxidation reactions. In trials 13 and 14, when the water con­ diglycerides by capillary gas chromatography. In Official methods and recommended
centration increased from 1% to 5%, the concentration of all tocopherol practices of the AOCS (6th ed.). AOCS Press.
Ansolin, M., Souza, P. T., Meirelles, A. J. A., & Batista, E. A. C. (2017). Tocopherols and
species varied between 1.2% and 2.1%. Therefore, the amount of water tocotrienols: An adapted methodology by UHPLC/MS without sample pretreatment
slightly affected the tocopherol concentration. steps. Food Analytical Methods, 10, 2165–2174. https://link.springer.com/article/10
.1007/s12161-016-0768-z.
Cert, A., Moreda, W., & Pérez-Camino, M. C. (2000). Chromatographic analysis of minor
4. Conclusion constituents in vegetable oils. Journal of Chromatography A, 881, 131–148. https://
doi.org/10.1016/S0021-9673(00)00389-7
The enzyme concentration, temperature, and water content strongly Costa, E., Almeida, M. F., Alvim-Ferraz, M. C., & Dias, J. M. (2018). Effect of Crambe
abyssinica oil degumming in phosphorus concentration of refined oil and derived
influenced the FFA and DAG contents, and the phosphorus content biodiesel. Renewable Energy, 124, 27–33. https://doi.org/10.1016/j.
significantly decreased after degumming. By optimizing the process renene.2017.08.089
conditions, it was possible to maximize the FFA and DAG concentrations Dayton, C. L. G., & Galhardo, F. (2014). Enzymatic degumming. Green vegetable oil
processing: Revised first edition. AOCS Press. https://doi.org/10.1016/B978-0-
and minimize the phosphorus content in the degummed oil (P < 10 mg/ 9888565-3-0.50009-1
kg). In addition, the enzymatic degumming process conducted at higher De Greyt, W., & Kellens, M. (2005). Deodorization. In Bailey’s industrial oils and fat
enzyme (200 mg/kg) and water (5%) concentrations and low tempera­ products (6th ed., pp. 341–383). New Jersey: Hoboken.
Galhardo, F., & Dayton, C. (2021). Enzymatic degumming. Retrieved from https://lipidlibr
ture of 50 ◦ C resulted in tocopherol concentrations that significantly
ary.aocs.org/edible-oil-processing/enzymatic-degumming/. (Accessed 11 March
differed from those in crude soybean oil (p < 0.05), indicating a high 2021).
degree of tocopherol retention. Therefore, the application of the Puri­ Greyt, W. De (2013). Edible oil refining: Current and future technologies. In W. Hamm,
fine® 3G cocktail simultaneously increased the FFA and DAG contents R. J. Hamilton, & G. Calliauw (Eds.), Edible oil processing (pp. 127–151).
Guerrand, D. (2017). Lipases industrial applications: Focus on food and agroindustries.
and decreased the phosphorus content in soybean oil. In addition, the OCL, 24, D403. https://doi.org/10.1051/ocl/2017031
degumming alternative preserved the original tocopherol content in the Gupta, M. K. (2017). Practical guide to vegetable oil processing. Elsevier.
oil, indicating its potential applicability for the oil industry. The com­ Hitchman, T. (2009). Purifine® PLC: Industrial application in oil degumming and
refining. Oil Mill Gazetteer, 115.
bination enzymatic degumming with enzymatic transesterification Jahani, M., Alizadeh, M., Pirozifard, M., & Qudsevali, A. (2008). Optimization of
using soft processes and non-refined inexpensive vegetable oils would be enzymatic degumming process for rice bran oil using response surface methodology.
a promising way for biodiesel production, for this approach may LWT – Food Science and Technology, 41, 1892–1898. https://doi.org/10.1016/j.
lwt.2007.12.007
noticeably reduce the reaction time (less steps), energy consumption Jiang, X., Chang, M., Jin, Q., & Wang, X. (2014). Effect of ultrasound treatment on oil
(low temperature) and increase productivity (DAG and FFA increase), recovery from soybean gum by using phospholipase C. Journal of Cleaner Production,
which represents an important topic for future studies. 69, 237–242. https://doi.org/10.1016/j.jclepro.2014.01.060
Jiang, X., Chang, M., Jin, Q., & Wang, X. (2015a). Application of phospholipase A1 and
phospholipase C in the degumming process of different kinds of crude oils. Process
Declaration of interests Biochemistry, 50, 432–437.
Jiang, X., Chang, M., Jin, Q., & Wang, X. (2015b). Application of phospholipase A1and
phospholipase C in the degumming process of different kinds of crude oils. Process
The authors declare that they have no known competing financial
Biochemistry, 50, 432–437. https://doi.org/10.1016/j.procbio.2014.12.011
interests or personal relationships that could have appeared to influence Lamas, D. L., Constenla, D. T., & Raab, D. (2016). Effect of degumming process on
the work reported in this paper. physicochemical properties of sunflower oil. Biocatalysis and Agricultural
Biotechnology, 6, 138–143. https://doi.org/10.1016/j.bcab.2016.03.007
Liu, D., & Ma, F. (2011). Soybean phospholipids. Recent trends for enhancing the diversity
CRediT authorship contribution statement and quality of soybean products. Retrieved from http://www.intechopen.com/books
/recent-trends-for-enhancing-the-diversity-and-quality-of-soybean- products/so
Rafaela Menezes dos Passos: Writing – original draft, Conceptu­ ybean-phospholipids.
Nikolaeva, T., Rietkerk, T., Sein, A., Dalgliesh, R., Bouwman, W. G., Velichko, E., et al.
alization, Methodology, and, Formal analysis. Rúbia Mariana da Silva: (2020). Impact of water degumming and enzymatic degumming on gum
Methodology, and, Formal analysis. Paula Virginia de Almeida mesostructure formation in crude soybean oil. Food Chemistry, 311, Article 126017.
Pontes: Methodology, and, Formal analysis. Marcelo Antônio Mor­ https://doi.org/10.1016/j.foodchem.2019.126017
Passos, R., Ferreira, R., Batrista, E., Meirelles, A. J. A., Maximo, G. J., Ferreira, M. C.,
gano: Methodology, and, Formal analysis. Antônio J.A. Meirelles: et al. (2019). Degumming alternatives for edible oils and biodiesel production. Food
Supervision. Christian V. Stevens: Writing – review & editing. Marcela and Public Health, 9, 139–147. https://doi.org/10.5923/j.fph.20190905.01
Cravo Ferreira: Writing – review & editing. Klicia Araujo Sampaio: Qu, Y., Sun, L., Li, X., Zhou, S., Zhang, Q., Sun, L., et al. (2016). Enzymatic degumming of
soybean oil with magnetic immobilized phospholipase A2. LWT – Food Science and
Supervision, Conceptualization, Writing – review & editing, and, Technology, 73, 290–295. https://doi.org/10.1016/j.lwt.2016.06.026
Funding acquisition. Rodrigues, M. I., & Iemma, A. F. (2014). Estratégia experimental para fatoriais
fracionados e delineamento composto central rotacional (DCCR). In Planejamento de
experimentos & otimização de processos (3a, p. 149). Campinas: Cárita.
Acknowledgments Sampaio, K. A., Zyaykina, N., Uitterhaegen, E., Greyt, W. De, & Verhé, R. (2019).
Enzymatic degumming of corn oil using phospholipase C from a selected strain of
The authors are thankful to the National Council for Scientific and Pichia pastoris. LWT – Food Science and Technology, 107, 145–150. https://doi.org/
10.1016/j.lwt.2019.03.003
Technological Development (CNPQ) [grant number 429873/2018-2],
Sampaio, K. A., Zyaykina, N., Wozniak, B., Tsukamoto, J., Greyt, W. De, & Stevens, C. V.
São Paulo Research Foundation (FAPESP) [grant numbers 2014/21252- (2015). Enzymatic degumming: Degumming efficiency versus yield increase.
0, 2016/10636-8], and to Espaço da Escrita – Pró-Reitoria de Pesquisa – European Journal of Lipid Science and Technology, 117, 81–86. https://doi.org/
UNICAMP for language editing services. This study was financed in part 10.1002/ejlt.201400218
Sein, A., Hitchman, T., & Dayton, C. L. G. (2019). Enzymes in vegetable oil degumming
by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior – processes. Retrieved from . (Accessed 28 October 2021) https://books.google.com.
Brasil (CAPES) – Finance Code 001. The authors also want to br/books?id=eJGfDwAAQBAJ&lpg=PA342&ots=B6uqIEr8fe&dq=%22Enzymes%

8
R.M. dos Passos et al. LWT 159 (2022) 113197

20in%20Vegetable%20Oil%20Degumming%20Processes%22&hl=pt-BR&p Ustinova, E. (2020). Oilseeds and products update – Brazil. https://www.fas.usda.gov/da

g=PA342#v=onepage&q=%22Enzymes%20in%20Vegetable%20Oil%20Degumm ta/brazil-oilseeds-and-products-update-25/. (Accessed 4 June 2021).
ing%20Processes%22&f=false.
Shahidi, F., & De Camargo, A. C. (2016). Tocopherols and tocotrienols in common and
emerging dietary sources: Occurrence, applications, and health benefits. International
Journal of Molecular Sciences, 17, 1–29. https://doi.org/10.3390/ijms17101745