Supplementary data

Effect of refinement and production technology on the molecular composition of edible

cottonseed oils from a large industrial scale production

Yongxin Yea, Jaloliddin Khushvakova, Akmaljon Boboevb, Rano Akramovab, Obidjon Yunusovb,
Dilbar Dalimovac, Shahlo Turdikulovac, Sharafitdin Mirzaakhmedovd, Søren Balling Engelsena,
Bekzod Khakimova,*

a
Department of Food Science, University of Copenhagen, Rolighedsvej 26, 1958, Frederiksberg C, Denmark
b
Tashkent Chemical-Technological Institute, Navoiy street 32, 100011, Tashkent, Uzbekistan
c
Center for Advanced Technologies, Talabalar shaharchasi 3A, 100041, Tashkent, Uzbekistan
d
Institute of Bioorganic Chemistry, Academy of Science of the Republic of Uzbekistan, Mirzo Ulugbek street, 83,
100125, Tashkent, Uzbekistan

*Corresponding author:
Bekzod Khakimov
Office phone: +45 3532-8184
E-mail: [email protected]

Material and methods

Gas Chromatography-Mass Spectrometry (GC-MS)

Cottonseed oils were analyzed using GC-MS as (1) intact oils and (2) derivatization. For the analysis
of intact oil samples, 200 µL of oil was mixed with 800 µL of dichloromethane in 2.0mL glass vials,
vigorously vortexed for 10 sec at ambient temperature and injected into GC-MS. For derivatization
protocol, a single-step hydrolysis followed by methylation was applied using 1.25 M hydrochloric acid
solution in methanol. One hundred µL of oil sample was mixed with 300 µL of 1.25 M hydrochloric
acid in methanol in 2.0 mL Eppendorf tube and incubated in a Thermomixer F2.0 (Eppendorf, Hørsholm,
Denmark) for four hours at 62℃. This thermal treatment with hydrochloric acid and methanol allows
cleavage of ester and ether bonds followed by methylation of labile protons of hydroxyl and carboxylic
acid functional groups of FA and other substances in oil. After the thermal treatment, a mixture was
cooled until ambient temperature, mixed with 400 µL of milli-Q water, and transferred into 15 mL
Falcon tubes. Prior to extracting freed FA and other substances of cottonseed oil from the acid water
mixture, 1.0 mL of ethyl acetate was added into a Falcone tube and vigorously vortexed for one minute
at ambient temperature. Then, 750 µL of the upper phase, the ethyl acetate phase, was transferred into
a fresh 15 mL Falcone tube. Ethyl acetate based extraction was repeated twice and the extracts were
combined. 100 mg of sodium bicarbonate was added to the combined extract in order to neutralize any
residual acid. After the addition of sodium bicarbonate, the mixture was vigorously mixed for 10 sec,
and left in dark for 10 minutes, and then 1.0 mL of ethyl acetate upper phase was transferred into 2.0
mL glass vial and injected into GC-MS. The injection volume was one µL for both sample preparation
protocols, intact oil and derivatized oil. The GC separation was performed on a Zebron ZB 5% Phenyl
95% Dimethylpolysiloxane column (30 m × 250 μm × 0.25 μm) with a 5 m inactive guard column
(Restek Corporation, Bellefonte, PA, US). Hydrogen was used as a carrier gas at the constant column
flow rate of 1.2 ml/min. The initial temperature of the GC oven was 40°C, held for 2 min, heated to
270°C at a rate of 12°C/min, followed by 6°C/min to reach a final temperature of 320°C at which the
oven was kept for 10 min. The mass spectra were recorded in the range of 45 to 500 m/z with a data
acquisition rate of 10 spectra∙s-1. The MS detector and ion source were switched off during the first 6.9
min of solvent delay time. The transfer line and ion source temperature were set to 280°C and 230°C,
respectively.

1.1. Proton nuclear magnetic resonance (1H NMR) spectroscopy

All 1H NMR measurements were performed at the Department of Food Science (University of
Copenhagen). A Bruker Avance III 600 MHz NMR spectrometer was used. The spectrometer was
equipped with a 5 mm broadband inverse RT (BBI) probe, automated tuning and matching accessory
(ATMA) and cooling unit BCU-05, and an automated sample changer (SampleJet, Bruker BioSpin)
with sample cooling (5 °C) and preheating stations (25 °C). Data acquisition and processing were
carried out using a software TOPSPIN 3.6.2 PL5 (Bruker BioSpin, Rheinstetten, Germany). Sample
was pre-heated at 25 °C for 60 s in SmapleJet and kept inside the NMR probe head for 3 min in order
to reach temperature equilibrium at 27 ± 0.1 °C. Then, automated tuning and matching, automated
locking, and automated shimming (TOPSHIM routine) were performed. The One-dimensional 1H NMR
spectra were acquired using the standard pulse sequence noesy1d from the Bruker pulse program library.
Sixty-four scans were acquired, after two dummy scans, and the generated free induction decays (FIDs)
were collected into 75k data points using a spectral width of 21 ppm. The acquisition time, relaxation
delay, and mixing time were set to 2.99, 5.0 and 0.01 s, respectively. The receiver gain was set to 2.25
for all samples. The automation program used to control sample measurements included, together with
the acquisition routines for locking, automated tuning and matching, and shimming, the 90° hard pulse
calibration, and optimized presaturation power for each sample, as well as automated data processing
including Fourier transformation of FID, with a line-broadening of 0.3 Hz, automated phasing, and
baseline correction.
Figures S1-S4
 Intensity (a.u.) Intensity (a.u.)

B
A

0
0.2

0.1
0
0.1
0.2
0.3
1-decyne
undecane

10
10
dodecane
2,5-dimethyl-benzaldehyde
n-undecanal
2-decenal
2,4-decadienal
copaene

are summarized in Table S1.

β-caryophyllen
1-dodecanol (Z)-pentadec-6-en-1-ol
2,6-di-tert-butylphenol α-ylangene
1-iodo-2-methylundecane
palustrol

15
15

2-methylhexadecane β-bisabolol
methyl myristate (Z)-nonadec-5-ene
methyl pentadecanoate 2-methyl-e-7-hexadecene
methyl palmitoleate
butyl octadeca-9,12-dienoate
methyl (Z)-heptadec-10-enoate ethyl palmitate
methyl margarate
methyl linoleate
linoleic acid (standard)
methyl (Z)-nonadec-10-enoate

20
20

methyl ricinoleate isopropyl linoleate

methyl arachidate
oleamide
isopropyl linoleate
methyl ricinoleate 3-hydroxypropyl palmitate
methyl behenate α-monopalmitin

RT (min)
RT (min)

methyl lignocerate
glyceryl monooleate

squalene squalene
25
25

γ-tocopherol γ-tocopherol
α-tocopherol stigmastan-3,5-diene
α-tocopherol (standard)
campesterol
β-sitosterol
β-sitosterol
30

30
Intact oil

Derivatized oil
35

35

Figure S1. The representative total ion count (TIC) chromatograms of the Gas Chromatography-Mass Spectrometry (GC-MS) data acquired
on a refined press a cottonseed oil sample (A) as is after dilution with dichloromethane and (B) after hydrolysis. The assignments of all signals
 A Process Stages effect (GC-MS Press oil) B Process Stages effect (GC-MS Extraction oil) C Process Stages effect (NMR Press oil) D Process Stages effect (NMR Extraction oil)
12 8 8
8
Press Oil (18 × 140) SC1 SC1
SC1 6 SC2
exp.var% = 35% 6
p-val < 0.001
8 4 4
4
2
2
ASCA score

4 0 0
0
-2
0 -4 -2
-4
-4
-4 -8 Extract Oil (12 × 138) -6 Press Oil (18 × 140) Extract Oil (12 × 138)
exp.var% = 27% exp.var% = 47% -6 exp.var% = 41%
SC1 -8
p-val = 0.001 p-val < 0.001 p-val = 0.001
SC2
-8 -12 -10 -8
Unrefined- Refined- Refined Unrefined- Refined- Unrefined- Refined-
Unrefined- Refined- Refined
Press-Oil Press-Oil Deodorized Oil Extract-Oil Extract-Oil Extract-Oil Extract-Oil
Press-Oil Press-Oil Deodorized Oil

E Process Stages effect (GC-MS Press oil) F Process Stages effect (GC-MS Extraction oil) G Process Stages effect (NMR Press oil)
H Process Stages effect (NMR Extraction oil)
0.2 0.2 β-sitosterol SC1
SC1 δ-tetradecalactone SC1 SC1 sn-1,2/2,3dg bin (gossypol)
bin2
SC2 stigmasterol trimethyl citrate 0.3 sn-1,2-DG
linoleic acid β-sitosterol2 β-sitosterol SC2 β-sitosterol bin (gossypol) bin12
0.2 α-monopalmitin methyl arachidate bin4 unk2 bin (oleanolic acid, maslinic acid)
unk2 isopropyl linoleatemethyl margarate unk4
β-himachalene palmitic acid unk5
2-methyloctadec-7-yne 2-methylhexadecane α-tocopherol 0.2 β-sitosterol bin2 bin (gossypol) 0.1
0.1 2-cyclohexyl-eicosane methyl margarate squalene β-EA bin (sn-1,2/2,3 DG
methyl caprylate n-undecanal α-springene
ASCA loading

0.1 sn-1,2-DG keto-epoxy-monoenes1

9-cis,11-trans CLA
0.1 β-sitosterol bin (HPO-c(Z,e)dEs)

0 0 0 0

2-decenal -0.1
-0.1 γ-Toco
α-ylangene ethyl palmitate methyl ricinoleate α-ylangene linoleic acid bin (linoleic acid)
2-cyclohexyl-eicosane -0.1 (Z)-nonadec-5-ene sn-1,2/2,3 DGunk4 unk10 -0.1 bin (punicic acid, HO-c(Z,E)dEs)
cis-β-farnesene trimethyl citrate α-selinene 3-hydroxypropyl palmitate -0.2 bin (HPO-c(Z,e)dEs) linoleic acid triglycerides
β-bisabolene ethyl palmitate2 methyl ricinoleate unk15 bin (β-sitosterol ) triglycerides γ-Toco unk12
-0.2 copaene β-sitosterol
β-eleostearic acid
α-selinene -0.3 bin4
-0.2 -0.2
0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140 0 10 20 30 40 50 60 70 80 90 0 10 20 30 40 50 60 70 80 90
Variables Variables Variables Variables

Figure S2. Process stage effect tested using ANOVA-simultaneous component analysis (ASCA) with permutation test (1000 permutations) in the
cottonseed oils obtained from pressing and extract technologies, separately. Boxplots shows the simultaneous components from ASCA performed
on GC-MS data for the press oil (A) and for the extract oil (B). Panels (C) and (D) correspond to ASCA performed on NMR data for the press and
the extract oil, respectively. Corresponding loadings plots are showed in panels E-H. exp.var% = explained variation; p-val = p-value (FDR
corrected); SC = simultaneous component.
 8
A 10
GC-MS
B NMR Unrefined Press Oil
8 6 Unrefined Extract Oil
6 4
4
2
2

PC2 (20%)
PC2 (20%)

0 0
-2 -2
-4
-4
-6
-8 -6
-10 -8
-10 -5 0 5 10 -6 -4 -2 0 2 4 6 8 10 12
PC1 (31%) PC1 (28%)
0.2
C D bin (gossypol)
γ-Toco
trimethyl citrate GC-MS 0.2 NMR fatty acid
methyl caprylate oleamide methyl ricinoleate bin (gossypol)
dodecane α-Toco 0.15
steroids
α-selinene methyl gadoleate
0.1 tocopherols
β-bisabolene
pentadecane copaene squalene 0.1 sn-1,2-DG
β-Toco δ-cadinene β-sitosterol β-sitosterol terpenes
β-caryophyllen
γ-Toco α-humulene 22-ketocholesterol 0.05 TG α-EA
PC2 (20%)

PC2 (20%)
β-EA hydrocarbons
FA (β-CH2)
0 methyl behenate 0 gossypol
TG linoleic acid
β-sitosterol
2,4-decadienal alloaromadendrene TG
linoleic acid
-0.05 linoleic oxidation compounds
acid β-sitosterol
2-decenal SFA and ω-9 MUFA KO-c(E,Z)dEs others
oleic acid β-himachalene methyl pentadecanoate
-0.1 FA
-0.1 stearic acid methyl pelargonate FA (-CH3) HPO-c(E,E)dEs squaleneEE&GG unknown
squalene -0.15 β-sitosteryl acetate linolenic acid sn-1,3-DG
myristic acid cis-β-farnesene
palmitoleic acid
octylacetylene ethyl palmitate -0.2 β-sitosterol
dec-1-yne
-0.2
-0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2 0.25 0.3 -0.3 -0.2 -0.1 0 0.1 0.2
PC1 (31.1%) PC1 (28%)

Figure S3. Global comparison of 14 cottonseed oil samples using PCA. Scores plots from (A) a Gas Chromatography-Mass Spectrometry (GC-
MS) variable table (14 samples × 140 variables) and (B) a global SigMa metabolite table (14 samples × 92 variables), colored by process
stages and numbered by different batches. The corresponding loading plots from (C) GCMS data and (D) NMR data were colored by 9
metabolite classes, which are fatty acid (esters), steroids, tocopherols, terpenes, hydrocarbons, gossypol, oxidation compounds, others and
unknown. PC1 versus PC2 scores and loadings represent PCA model developed on the autoscaled metabolite table. 1-iodo-2-M = 1-iodo-2-
methylundecane; OA = oleic acid; LA = linoleic acid; PA = palmitic acid; Mel-PD=methyl pentadecanoate; ArDen = alloaromadendrene; Mel-
PG = methyl pelargonate; Mel-R = methyl ricinoleate; NoNa = nonacosane; PD = pentadecane; Mel-C = methyl caprylate; Mel-A = methyl
arachidate; Toco = tocopherol; sn-1,3-DG = sn-1,3-diacylglycerols; SFA and ω-9 MUFA = saturated and monounsaturated ω-9 fatty chain;
EE&GG = esters of phytol & geranylgeraniol; α-EA = α-eleostearic acid; β-EA = β-eleostearic acid; tyr. der. = tyrosol derivatives; TG =
triglycerides; 9-cis,11-trans CLA = (9-cis, 11-trans) conjugated linoleic acids.
 A 10 GC-MS
B 10 NMR Refined Press Oil
8 8 Refined Extract Oil
6 6
4 4
PC2 (17.7%)

PC2 (23%)
2 2
0
0
-2
-2
-4
-6 -4

-8 -6
-10 -8
-10 -5 0 5 10 -8 -6 -4 -2 0 2 4 6 8
PC1 (32.7%) PC1 (34%)
0.2
C α-Toco D
methyl gadoleate GC-MS 0.2 NMR fatty acid
0.15 methyl ricinoleate
steroids
squalene methyl arachidate 0.15
0.1 methyl caprylate sn-1,2/2,3 DG
HPO-c(E,E)dEs tocopherols
α-selinene
0.1 α-EA FA FA (β-CH2)
terpenes
β-caryophyllen SFA and ω-9 MUFA
PC2 (17.7%)

0.05 α-humulene

PC2 (23%)
β-Toco γ-Toco 0.05
KO-c(E,Z)dEs bin (gossypol) hydrocarbons
δ-cadinene acetyl-beta-sitosterol FA (-CH3)
0 (Z)-pentadec-6-en-1-ol 0 sn-1,2-DG β-sitosterol gossypol
16-kaurene pentadecane β-sitosteryl acetate sn-1,3-DG TG
TG TG
-0.05 EE&GG β-sitosterol β -sitosterol oxidation compounds
-0.05 copaene oleamide dodecane squalene
cis-β-farnesene 2-decenal γ-Toco β-sitosterol others
α-ylangene oleic acid -0.1
bin (gossypol) linolenic acid
-0.1 nonacosane linoleic acid unknown
heptacosaneoctylacetylene -0.15 β-EA linoleic acid
squalene palmitoleic acid β-sitosterol
-0.15 β-bisabolene -0.2 9-cis,11-trans CLA

-0.2
-0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2 0.25 0.3 -0.3 -0.2 -0.1 0 0.1 0.2
PC1 (32.7%) PC1 (34%)

Figure S4. Global comparison of 12 cottonseed oil samples using principal component analysis (PCA). Scores plots from (A) a global Gas
Chromatography-Mass Spectrometry (GC-MS) variable table (12 samples × 140 variables) and (B) a global SigMa metabolite table (12
samples × 92 variables), colored by process stages and numbered by different batches. The corresponding loading plots from (C) GCMS data
and (D) NMR data were colored by 9 metabolite classes, which are fatty acid (esters), steroids, tocopherols, terpenes, hydrocarbons, gossypol,
oxidation compounds, others and unknown. PC1 versus PC2 scores and loadings represent PCA model developed on the autoscaled metabolite
table. 1-iodo-2-M = 1-iodo-2-methylundecane; OA = oleic acid; LA = linoleic acid; PA = palmitic acid; Mel-PD=methyl pentadecanoate;
ArDen = alloaromadendrene; Mel-PG = methyl pelargonate; Mel-R = methyl ricinoleate; NoNa = nonacosane; PD = pentadecane; Mel-C =
methyl caprylate; Mel-A = methyl arachidate; Toco = tocopherol; sn-1,3-DG = sn-1,3-diacylglycerols; SFA and ω-9 MUFA = saturated and
monounsaturated ω-9 fatty chain; EE&GG = esters of phytol & geranylgeraniol; α-EA = α-eleostearic acid; β-EA = β-eleostearic acid; tyr. der.
= tyrosol derivatives; TG = triglycerides; 9-cis,11-trans CLA = (9-cis, 11-trans) conjugated linoleic acids.