6428 |New J. Chem., 2015, 39, 6428–6436 This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2015… [600249]

6428 |New J. Chem., 2015, 39, 6428–6436 This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2015
Cite this: New J. Chem., 2015,
39,6 4 2 8Application of an optimized electrochemical
sensor for monitoring astaxanthin antioxidantproperties against lipoperoxidation †
Ramona Penu,abSimona Carmen Litescu,*aSandra A. V. Eremia,aIoana Vasilescu,a
Gabriel-Lucian Radu,bMaria Teresa Giardi,cGianni Pezzottidand Giuseppina Reac
An optimized electrochemical sensor was developed to assess the antioxidant capacity of carotenoids,
accumulating during the life cycle of Haematococcus pluvialis cell cultures. The sensor was improved
with a composite renewable surface made of immobilised phosphatidylcholine (PC) on magnetic
nanobeads of iron oxide (Fe 3O4) and PC/Fe 3O4, and it was used to monitor the antioxidant properties of
the ketocarotenoid astaxanthin against in situ generated phosphatidylcholine lipoperoxides. The surface
configuration was able to mimic the natural position and orientation of astaxanthin in the cellularmembrane, conferring to the whole experimental set-up good sensitivity for reactive oxygen species(limit of detection for peroxyl radicals 9.1 /C210
/C010mol L/C01) with a linear response ranging between 10/C08
and 1.6 /C210/C06mol L/C01. The sensor has been proved suitable for both batch and flow measurements.
The accuracy of the flow measurements was unaffected by the magnetic field intensity. Electrochemicalmeasurements confirmed that natural astaxanthin is a more effective antioxidant than synthetic
astaxanthin, vitamin E and lutein and the protective effect of astaxanthin correlates with its
concentration inside the cell. The newly developed sensor is hence useable for in-line monitoring ofwhole-cell based industrial bioprocesses for the production of astaxanthin.
Introduction
Plants and microalgae are a rich source of secondary metabolites
having preventive and/or protective functions against humandiseases. Among these metabolites, carotenoids ( e.g.b-carotene,
lutein, astaxanthin) and vitamin E class compounds (tocopherolsand tocotrienols) are of interest for food and pharmaceuticalindustries. Both the classes of compounds are isoprene derivativesmostly occurring in the photosynthetic membranes, whereinamong other functions, they are involved in antioxidant defence
mechanisms.
1–3Acting mainly as antioxidants, their intake in the
human diet proves to be effective for the prevention of age-related,degenerative and chronic diseases. This therapeutic potentialpromoted a dramatic increase in their consumption as dietarysupplements and as a consequence prompted the developmentof analytical systems capable of dete cting their biolog ical properties
along the entire production chain.
4Astaxanthin is a highly bioactive
red-coloured ketocarotenoid, which naturally occurs in a wide variety
of living organisms. Although synth etic astaxanthin dominates the
world market, currently microalgal cultures are extensively used forthe industrial production of biom ass as an astaxanthin source.
5–7
The unicellular green alga, namely, Haematococcus pluvialis ,i sa
commercial natural source of astaxanthin and it is produced ona large industrial scale.
8–11In specific stressful conditions, such
as the depletion of nutrients or exposure to intense light,
H. pluvialis undergoes a morphological transition from green
active cells to red cyst cells, in which astaxanthin accumulatesin the cyst reaching approximate ly 5% of the cell’s dry weight.
12–14
Thus, the industrial cultivation of H. pluvialis encompasses a first
stage aiming at improving biomass productivity (in physiologicalgrowth conditions), a second phase aiming at increasing astax-anthin accumulation (in stressful conditions), and a final stageof harvesting.
15–17An intense monitoring of these processes
could strongly ameliorate industrial production yields.
In general, antioxidant activity is the ability of a compound
(or mixture) to inhibit oxidative degradation (such as lipidperoxidation). Several experimental data support the idea that theantioxidant capacity of carotenoi ds is related to their structure
18
and in particular to the number of conjugated double bounds.19
aNational Institute of Research and Development for Biological Sciences,
Centre of Bioanalisys, Bucharest, 296 Splaiul Independentei, 060031, Bucharest,
Romania. E-mail: slitescu@gmail.com; Fax: +40212207695; Tel: +40212200900
bFaculty of Applied Chemistry and Materials Science,
University Politehnica Bucharest, 1-7 Gh. Polizu Street, 011061, Bucharest, Romania
cItalian National Research Council, Institute of Crystallography Departments of
Agrofood and Molecular Design, CNR 00015 Monterotondo Scalo, Rome, Italy
dBiosensor SRL, Via Degli Olmetti 44, 00060, Formello, Rome, Italy
†Electronic supplementary information (E SI) available. See DOI: 10.1039/c5nj00457hReceived (in Montpellier, France)
23rd February 2015,Accepted 9th June 2015
DOI: 10.1039/c5nj00457h
www.rsc.org/njcNJC
PAPER

This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2015 New J. Chem., 2015, 39, 6428–6436 | 6429Carotenoids have the capacity to trap both lipid peroxyl radicals
and singlet oxygen species20(that are not quenched by phenolic
antioxidants). Moreover carotenoids have the benefit of avoiding
potential pro-oxidant behaviour due to the presence of the oxo-groups on the side chain. The ketocarotenoid astaxanthin provedto be effective in scavenging hydroxyl-free radicals, which are themost damaging free-radical for humans,
21,22but it is not a
suitable quencher of peroxyl radicals when compared to phenolicantioxidants.
The radical-scavenging properties of astaxanthin rely on its
specific molecular structure, made of thirteen conjugated double
bonds, two oxo-groups in the 4 and 4
0position on the cyclohexene
ring, and two hydroxyl groups at the 3 and 30position, which confer
it with significant efficiency.23–25Astaxanthin proved to be a better
antioxidant than all other carotenoid antioxidants26based on its
capability to disperse on the cellular membrane. The cellularmembrane dispersion occurs by making links from inside tooutside of the cellular membrane.
27This ensures that astaxanthin
has both the ability to scavenge the free radicals from the cell
membrane, viapelectron cloud mobility (from the polyene chain,
as any regular carotenoid), and the added capacity to scavenge thefree radicals from the inner and outer parts of the cellularmembrane due to the presence of hydroxyl and keto groups fromthe two ionone rings.
19,21,23
Among the overabundance of methods that can be used for the
evaluation of the antioxidant or an tiradical capacity of a compound28
(TEAC – trolox equivalent antioxidant capacity;29,30TRAP – total
radical trapping antioxidant parameter;31LDL – low density
lipoprotein peroxidation;32DMPD – N,N-dimethyl-phenylene-
diamine dihydrochloride;33FRAP – ferric reducing antioxidant
power assay,34ORAC – oxygen radical absorbance capacity
assay;35DPPH – diphenylpicrylhydrazyl free radical;36,37PCL –
photochemiluminescence and b-carotene bleaching38), only a
few (TEAC, DPPH, PCL) are useful for determining the activity
of both hydrophilic and lipophilic species, thus ensuring a
broader range of potential applications.
Despite the fact that all carotenoids contain an important
number of conjugated double bonds, individual carotenoidsdiffer in their antioxidant potential. There are carotenoidsunable to exert measurable antioxidant potential in in vitro
experiments, due to the fact that the experimental conditionsdo not take into consideration the carotenoid’s orientation on
the cellular membrane, which is the main factor affecting the
carotenoid’s efficacy. Currently, rapid screening models able toprovide reliable information based on the proper carotenoidorientation in the membrane are lacking. Some studies provethe ability of carotenoids to protect liposomes (either monolayeror bilayer phospholipid vesicles)
39from singlet oxygen oxidation,
but only a few demonstrate how the distribution in themembrane affects the carotenoid’s efficacy as an antioxidant.
40
Accurate description of the antioxidant capacity of carotenoidsshould be performed only using analytical protocols able to carryout an appropriate discrimination between the mechanisms ofantioxidant action while mimicking the in vivo carotenoid position
and distribution at the cell membrane level due to the modulatoreffect of carotenoids for lipid bilayers.
41,42This is necessarybecause the variability of carotenoid experimental behaviour is
related to the difference between the chain-breaking antioxidant
activity and the radical-scavenger activity. Lipid/phospholipid
membrane or liposome models were used in assessing theantioxidant efficacy of carotenoids
43–45but most of the published
protocols are time-consuming and require heavy equipments(e.g.high performance liquid chromatography with diode array-
mass spectrometry detection
39,44), slowing down the monitoring
and harvesting processes of industrial bioprocesses.
Accordingly, the development of a rapid screening instrument,
taking into account all the abovementioned issues and are able to
provide reliable information on accumulation levels, could be ofparticular industrial interest especially if it could provide suitableinformation on in vivo biological activity. The analytical device
should consider the specific distr ibution (position and orienta-
tion) of carotenoids on the cellular membrane. Electrochemicald e v i c e sa r eu s e f u lt o o l sf o rt h i sp u r p o s ea st h e ya r ea b l et od e t e c teither hydrogen atom transfer (HAT) or electron transfer reaction
(ET) in the antioxidants.
45
The aim of the present study is to propose a model able to
mimic the carotenoid distribution on the cellular membraneapplicable in designing a measuring system based on an electro-chemical sensor. The sensor should provide rapid and reliableinformation on the carotenoids’ efficiency as antioxidants. Themodel developed was obtained by improving a previously reportedelectrochemical sensor
46,47to propose a more useful analytical
tool, providing responses that are unaffected by memory effects –
since a renewable surface is used – and is suitable for industrialin-line applications in terms of stability, sensitivity and versatilityfor both batch and flow measureme nts. The measuring principle
involves an appropriate oxidizabl e substrate (in our current study
phosphatidylcholine (PC)) that is subjected to oxidation by thebiologically relevant free radicals, peroxyl radicals ROO
/C15.48The
process is monitored in the presence and in the absence of
the antioxidant. The case study for the present report was
astaxanthin. The reasons to use phosphatidylcholine as asubstrate include its ability to provide the appropriate conditionsfor the membrane model, the PC oxidation mechanism thatfollows the same oxidative path as low-density lipoproteins (LDL)thus generating lipoperoxides in the presence of peroxyl radicals,and the better lot-to-lot reproducibility of PC with respect to LDL.The optimized electrochemical sensor was applied to assess the
antioxidant properties of astaxanthin, accumulating during the
different stages of growth of H. pluvialis , in order to facilitate
better control of the algal growth and harvesting processes.
Experimental
Reagents
The reagents used were purchased from the following sources.
Phosphatidylcholine (PC) from egg yolk; 2,20-azobis(2-methyl-
propionamidine)dihydrochloride (AAPH) 97%; potassium chloride(KCl); hydrogen peroxide (H
2O2); potassium ferrocyanide;
potassium bromide FT-IR grade; dimethylsulfoxide (DMSO);sodium dodecylsulphate (SDS); potassium phosphate; sodiumPaper NJC

6430 |New J. Chem., 2015, 39, 6428–6436 This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2015borate; hexane; acetonitrile; lutein; a-tocopherol; fluorescein and
synthetic astaxanthin (astaxanthin Z98%, HPLC) were from
Sigma-Aldrich. Magnetic Fe 3O4nanoparticles (nanobeads) were
provided by Polytechnica, University of Bucharest. Astaxanthinfrom H. pluvialis was provided by Algatechnologies Ltd, Israel.
Design of the measuring system based on a surface renewable
electrochemical sensor
The development of the measuring system was based on a
previously reported proof-of-concept electrochemical sensor, and,
considering the anticipated application to in-line measurements,
the following experiment al steps were fulfilled:
(a) Stabilisation of composite-phosphatidylcholine (PC)
bound to Fe
3O4magnetic nanobeads. The procedure is detailed
by Litescu and co-workers.47Briefly, 72 mg of PC was suspended
in 5 mL KCl (0.1 mol L/C01) by slight PC sonication for 3 minutes,
followed by addition of 24 mg of magnetic nanobeads (Fe 3O4)
to the suspension with vortex stirring for 18 hours. After that,
PC-modified nanobeads (PC–Fe 3O4) were separated using a
magnetic separator and repeatedly washed, first time with KCl(0.1 mol L
/C01) and subsequently with water; each supernatant
part was removed while the PC–Fe 3O4nanobeads were dried at
601C for 30 minutes and stored at 4 1C. The efficient PC
deposition on the magnetic nanoparticles and storage stabilityof the PC–Fe
3O4nanobeads were checked by Fourier transform
infrared (FTIR) spectroscopy. The role of the Fe 3O4magnetic
nanobeads is to ensure the surface renewability thus avoiding
memory effects.
(b) PC–Fe 3O4-composite deposition on the surface of the
working electrode of the DRP 220AT DropSens screenprintedelectrochemical cell (working electrode Au-screen printed,AuSPE) was performed depending on the measuring system.For electrochemical batch measurements, 50 mL of PC–Fe
3O4
suspension was deposited on the AuSPE electrode, allowed to
dry for 30 minutes and PC–Fe 3O4-AuSPE cells were stored at
41C, under vacuum. Based on the magnetic properties of the
Fe3O4nanobead support, which works as the PC carrier (in the
PC–Fe 3O4composite), when flow measurements were performed,
the PC–Fe 3O4pumped into the system was stopped on the surface
of AuSPE working electrode by applying a magnetic field to theDelrin
sflow cell.
(c) Assessment of PC peroxidation degree of subsequent
radical attack using PC–Fe 3O4as measuring system and peroxyl
radicals thermally generated from azo-initiators.
The currently reported methods to evaluate ROS effects against
cellular membrane components have the drawback of problematicdirect assay due to extremely short free radical lifetimes; therefore,it is usually preferable to monitor the molecular product of theoxidative stress reaction.
49This methodology is based on the
correlation between the degree of substrate oxidation with ROS
concentration. The design of our measuring system started from
this correlation, and the peroxidation of the used substrate, PC–Fe
3O4, was initiated using peroxyl rad icals obtained at a controlled
rate: an aqueous solution of free radical azo-initiator, AAPH, wasleft at controlled temperature for the time necessary to induce thegeneration at constant rate of peroxyl radicals (ROO
/C15). The ideawas based on the reported data by Niki and co-workers.50The
generated ROS, ROO/C15, are further capable of inducing lipid/
phospholipid peroxidation. The model for electrochemical
measurements was initially develo ped for low-density lipoproteins,46
which was further applied to immobilised PC,47but it needed
optimisation in terms of response range, sensitivity and stability.In our measuring system, thermally-generated ROO
/C15reacts with
an electrochemically inactive substrate (PC deposited on Fe 3O4),
generating the corresponding peroxides (PCOO/C15/Fe3O4)t h a t
prove to be electrochemically active.46,47
(d) Assessment of analytical performance characteristics for
optimised sensor in batch and in flow systems and applicationto antioxidant efficacy assessment.
Electrochemical measurements
Batch measurements. The electrochemical experiments were
performed using an UNISCAN PG 580 potentiostat. The electro-chemical cell consisted of a DRP-220AT DropSens modified as
follows. The gold screen printed working electrode (AuSPE) was
modified with composite material by dropping suspensions ofPC–Fe
3O4nanoparticles. The counter electrode was printed
gold, while the pseudo-reference electrode was screen printedAg/AgCl; supporting electrolyte was 0.1 mol L
/C01KCl.
Flow measurements. Electrochemical experiments were carried
out in a flow system involving a Delrinsflow cell where the
analysed sample solution was driven with a peristaltic pump.
The substrate was introduced into the cell with a constant flow
rate. The value of the applied speed was optimized in order toreach a compromise between the best signal-to-noise ratioswhile avoiding the surface damage of the composite PC–Fe
3O4;
subsequently, the corresponding current–time curves wererecorded. The carrier buffer was KCl 0.1 mol L
/C01.
The techniques used were (i) cyclic voltammetry (CV), in a
potential range 1.2 to /C00.4 V vs.reference electrode, to perform
the system characterisation and to verify the optimum operational
parameters for (ii) chronoamperometric (CA) measurements. TheCA measurements were performed by applying the optimizedreduction potential according to the monitored oxidation product,namely, +0.385 V vs.Ag/AgCl pseudo-reference electrode. In order
to avoid potential electrochemical interferences from gold oxides,prior to AuSPE use, the DRP-220AT cells were slightly sonicatedin ethanol for 30 seconds, washed with bi-distilled water, dried
and stored under vacuum. A DRP-220AT cell was used for
35 consecutive flow measurements due to the benefit of usingthe PC–Fe
3O4renewable surface; however, when batch measure-
ments were performed, the DRP-220AT cell is for single use. Anotheroperational parameter specific for electrochemical measurementsthat was optimized was the reaction time between reactiveoxygen species and the PC–Fe
3O4to ensure a constant produc-
tion of phosphatidylcholine peroxides for both batch and flow
measurements.
FTIR measurements
FTIR spectra were obtained at room temperature using a Bruker
Tensor 27 spectrometer in angular reflectance and transmittancemodes; when necessary, an inert atmosphere was used in theNJC Paper

This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2015 New J. Chem., 2015, 39, 6428–6436 | 6431sample compartment. The spectra were obtained and assigned
using the Opus software. When transmittance measurements
were performed, the spectral range was 4000–400 cm/C01, the
aperture was 4 mm, the spectra resolution was 4 cm/C01and 128 scans
were acquired for each spectrum, at 20 Hz. Samples were pressedinto KBr pellets, and the background spectrum were recordedagainst KBr pellets.
Scanning electron microscopy (SEM) measurements
The SEM measurements were performed on a PC–Fe
3O4nano-
composite using a FEI scanning electron microscope. The
composite nanobeads were deposited on double-conductive Cribbon and the SEM images and EDX spectra were acquired at avoltage of 20 kV.
Astaxanthin extraction from Haematococcus pluvialis cell
cultures and UV-Vis quantification
500 mg of dried H. pluvialis cells were taken in 3 mL DMSO and
mixed with 40 mg of glass beads. The mixture was sonicated for
12 minutes, centrifuged at 4000 rpm for 5 minutes and thesupernatant was collected and eventually, filtered on a PTFEfilter (4.5 mmesh) (if centrifugation is efficient, filtration is not
needed). The amount of dissolved astaxanthin in the collectedsupernatant was easily estimated by UV-Vis spectrophotometry.A stock solution of 100 mgm L
/C01astaxanthin was prepared in
DMSO and a calibration graph for astaxanthin was drawn ( A=f(c))
for the following concentrations of astaxanthin: 0.4 mgm L/C01;
0.8mgm L/C01; 1.6 mgm L/C01;3mgm L/C01;5mgm L/C01; DMSO was
used as the blank; measurements were performed in UV-Viscells with 1 cm path, at 488 nm.
50 mL from the supernatant were taken and diluted at 3 mL
with DMSO; the absorbency at 488 nm was measured and thevalue obtained was interpolated on the calibration graph todetermine the astaxanthin concentration.
Oxygen radical absorbance capacity measurements
Experiments were performed using an OCEAN Optics QE Pro-FL
fluorescence spectrometer; the diameter of the optical fiber was560mm. The ORAC assay is based on in situ production of peroxyl
free radicals generated viaan azo-compound, 2,2,-azobis(2-methyl-
propionamidine)dihydrochloride (AAPH). The peroxyl radical isable to react with the fluorescent probe on increasing the rate
of fluorescence decay. The fluorescent probe used was fluorescein
(FL): excitation wavelength 490 nm, emission wavelength 515 nm.
Measurements were performed in phosphate buffer 75 /C2
10
/C03mol L/C01, pH = 7.40 at a total volume of 1 mL, using
microemulsions prepared by mixing hexane (0.66% w/w), SDS(4.87% w/w), 2-propanol (6.55% w/w), and 75 /C210
/C03mol L/C01
phosphate buffer (87.93% w/w) to disperse the investigated
lipo-soluble compounds. The micellar environment does not
significantly affect the fluorescence emission and the results
obtained for Trolox are similar to those obtained in aqueousmedia. A blank (FL + AAPH) using a micellar environment wasobtained. Sample curves (fluorescence versus time) were normalized
to the blank curve. The results were obtained by calculatingthe area under the fluorescence curve, and it was expressedas an equivalent of micromoles of Trolox per mg standard
compound.
Results and discussion
Characterisation of designed measuring model based on asurface renewable electrochemical sensor
Stabilisation of PC on magnetic nanobeads. The rationale
for developing a new model applicable to the assessment of the
antioxidant properties of carotenoids is detailed in the introduc-
tion. The basis of our current study is the necessity to ensure anexperimental environment able to mimic, as close as possible,the position and orientation of carotenoids in the phospholipidbi-layer of a cellular membrane in order to provide reliablephysiological information. By exploiting previously reported dataon carotenoids in membrane models and liposomes
41–45and our
group’s experience in developing electrochemical sensors for
ROS and antioxidant monitoring,47we succeeded in designing the
model using a nanobeads composite, PC–Fe 3O4,a sar e n e w a b l e ,
oxidizable surface on a conductive solid support (AuSPE).
The preservation of the structural integrity of the immobilized
oxidizable substrate was the first critical point addressed indeveloping our model. This was necessary to provide an instru-ment able to supply significant data from a physiological pointof view on lipoperoxidation. Consequently, subsequent to PC
immobilization, it is compulsory to preserve the main func-
tional groups of free PC available for radical attack. Structuralcharacterization using FTIR spectrometry was performed forthe PC-magnetic nanobeads; the spectra were obtained forPC-free and PC–Fe
3O4composite nanobeads.
In the case of phosphatidylcholine, the main FTIR absorption
bands, besides the specific bands for methyl and methylenegroups in the region 2800–3000 cm
/C01, are at 1737.4 cm/C01
ascribable to the nCQO stretching vibration for the ester group
of lipids, 1466 cm/C01corresponding to dCH 2/C0deformation vibra-
tion for the methylene group, 1170–1245 cm/C01ascribed to ns
PO2/C0asymmetric stretching vibration for phosphate groups, and
1063–1088 cm/C01ascribed to nsPO2/C0s y m m e t r i cs t r e t c h i n gv i b r a –
tion for phosphate groups (Fig. 1). The OH groups, either free orinvolved in intermolecular bonds, are present and proved bytheir specific absorption bands in the 3280–3300 cm
/C01domain.
By comparing spectra 1 and 2, it could be concluded that the
structural integrity of the PC substrate is preserved followingdeposition, as revealed by the shape and intensity of the specificIR absorption bands. It should be mentioned that the PC quantity
was preserved in both measurements in order to properly assignsignificant changes in absorption bands’ intensities. The stability ofthe obtained composite nanobeads, PC–Fe
3O4, was checked under
various storage conditions (room temperature and 40 1C) for
both dried powder and suspension, the powder proving stable
for 45 days, while the suspension loses stability after 21 days(see ESI, †Fig. S1).
It was concluded that subsequent to the application of the
established procedure for PC immobilization on magneticnanobeads, the recommended storage conditions for compositePaper NJC

6432 |New J. Chem., 2015, 39, 6428–6436 This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2015PC–Fe 3O4nanobeads are dried powder, room temperature, and
no light contact, ensuring in this way a fair stability with respectto anticipated applications of the realised composite, namely,
renewable substrate to study antioxidant preservation against
lipoperoxidation. The PC–Fe
3O4composite so obtained was next
used to test its applicability to monitor the PC peroxidationsubsequent to reactive oxygen species (ROS) attack.
SEM analysis confirmed the PC–Fe
3O4nanobeads formation
and the complete covering of magnetic nanobeads by phospholipids(Fig. 2) and the preservation of nano-sizes even subsequent to PCimmobilisation, despite the fact that SEM was not performed in
dispersed material but as deposited material. EDX confirms the PC
immobilisation on Fe
3O4nanobeads by the presence of P, N and C.
Assessment of immobilized PC peroxidation subsequent to
ROS attack. When cyclic voltammetry experiments are performed
with free PC in solution, no electroactive characteristics wereexhibited, therefore demonstrating that the initial hypothesis wascorrect, meaning that PC shows a similar behaviour to LDL, withno electroactive characteristics in solution or when deposited,
47
while in the presence of AAPH-gener ated peroxidation, a reduction
peak appears (see Fig. 3).
Subsequent to ROO/C15generation by thermic reaction, the
phosphatidylcholine layer from the composite PC–Fe 3O4was sub-
jected to ROS attack, and four redox processes occur, which aredefined by the following peak potentials: lipoperoxides reduction,E
PCOO/C15/PCOOH = +0.385 ( /C60.02) V and EPCOOH/PC =/C00.25 (/C60.02) V;
hydrogen peroxide reduction EH2O2/H2O= +0.630 ( /C60.035) V, the lastone involving superoxide evolution on anodic process at +0.920
(/C60.035) V (see Fig. 3, curve 2).
Considering the significance for lipoperoxide reduction of
the peak potential EPCOO/C15/PCOOH = +0.385 ( /C60.02) V, all further
chronoamperometric experiments were performed at thisspecific value.
Assessment of PC–Fe
3O4renewable sensor performance
characteristics against ROS attack. The optimized sensor was
characterized considering the sensitive response to free radical
attack. In the batch system, the response of the developed
sensor – PC–Fe 3O4dropped on AuSPE – and ROS concentration
were obtained using chronoamperometry. The applied potentialvalue on the working electrode was that which defines thereduction of PCOO
/C15species, +0.385 V vs.reference electrode.
The variation of the current intensity corresponding to lipoper-oxides peak reduction with the concentration of peroxyl radicalswas measured. The dynamic range of response was 10
/C08–1.6/C2
10/C06mol L/C01, the linear dependence on the concentration range
being characterized by the following equation: I(nA) = 9.24 /C2
C(nmol L/C01)/C00.323 ( R2=0 . 9 9 1 3 ) ,w h e r e Cis the concentration
of the generated free radical species calculated considering theNiki constant rate of radical production.
50It is known48that at
371C under air, a concentration of 10 /C210/C03mol L/C01AAPH
produces 1.36 /C210/C03mol L/C01s/C01of free radicals, ROO/C15;1m o l e
of AAPH produces 2 moles of ROO/C15.47In this way, the concen-
tration of ROO/C15existing in the system per unit time can be
calculated. In the measuring conditions, the obtained detectionlimit, calculated as 3 /C2S/N was 9.1 /C210
/C010mol L/C01for ROO/C15;
the limit of detection was calculated considering the level ofnoise registered at +0.385 V in the presence of carrier buffer. Theobtained responses (the average of 3 measurements) proved thatthe designed system is highly sensitive to the extent of peroxylradical attack exerted toward the PC layer of the composite.
At the same time, to be useful in assessing the antioxidant
capabilities, the sensor based on the PC–Fe
3O4composite
should respond specifically to the amount of the PC generatingPCOO
/C15radicals. Therefore, experiments were performed using
bare AuSPE electrodes, both in fr ee PC suspensions and in PC–Fe 3O4
suspensions that were i mmobilised on AuSPE viaam a g n e t i cf i e l d .
Fig. 1 FTIR spectra of nanobeads; PC (1); PC-nanobeads (2).
Fig. 2 (1, 2) SEM images and (3) EDX spectra of PC–Fe 3O4nanobeads.
Fig. 3 Cyclic voltammograms of un-oxidized (1) oxidized (2) PC;
(v= 100 mV s/C01, WE: SPE – Au; KCl 0.1 mol L/C01).NJC Paper

This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2015 New J. Chem., 2015, 39, 6428–6436 | 6433The response of the sensor was highly influenced by the substrate
amount, as noticeable from Fig. 4. The composite modified electrode
acts as an insulator when high amounts of PC are used.
When PC–Fe 3O4nanobeads are stored as powder, as mentioned
before, they are stable for 45 days, and the registered response tolipoperoxidation is stable too. After 60 days, the intensity ofmeasured current decreased to 91% ( /C61.5%), while after 75 days,
the significant loss on PC intensity bands noticed in the FTIRspectra (see ESI, †Fig S1) is accompanied by a loss of the measured
signal, the current intensity decreasing to 35% ( /C61.8%). It can be
concluded that the lifetime of the sensor is at least 50 days.
Design of the measuring flow system. The developed system
was designed to be applied for in-line monitoring of antioxidantproperties; therefore, the response of the sensor in a flow regimewas further tested. Different flow system configurations wereemployed for testing the influence of the magnetic field on theregistered response and the flow rate’s influence on the back-ground (noise associated to measurement). The initial flow
measures, to test the newly designed Delrin cell’s capabilities
of working under a magnetic field, were performed using CA andmeasuring the potassium ferricyanide reduction on AuSPE in thepresence of Fe
3O4magnetic nanoparticles. A ferro-ferricyanide
couple was used as redox probe to test the influence of themagnetic field on the electrochemical response. The switchon/off of the magnetic field faintly influenced the measurement(the associated noise value was less than 0.15 nA for a response
around 1.8 nA). Similarly, the fl ow rate did not affect the measure-
ment (the flow rate ranged between 6 and 250 mLm i n
/C01), see ESI, †
Fig. S2.
The PC–Fe 3O4sensor response with variation of peroxyl free
radical concentration was teste d in flow, by chronoamperometry.
T h ea p p l i e dp o t e n t i a lw a st h es a m ep o t e n t i a la si nt h eb a t c h ,+0.385 V vs.screen printed reference Ag/AgCl, and the results
obtained proved an operational stability of 5 subsequent measure-
ments (see ESI, †Fig. S3); moreover, the linearity of the response
was preserved for AAPH generated peroxyl radicals in the samerange as defined for batch measurements.
Evaluation of magnetic field effect on the registered chrono-
amperometric response. Determinations were performed usingthe significant response of the composite PC–Fe
3O4nanobeads
on peroxidation, in the presence or absence of magnetic field.
The peroxyl concentrations used were in the low concentration
range, 1.6–3.5 nmol L/C01, and the experiments led to the results
presented in Table 1.
As can be noticed, the magnetic field does not influence the
accuracy of the sensor response on p eroxidation, and consequently,
no significant influence is expected when astaxanthin anti-oxidant capacity has to be determined. These data confirm thepreviously performed experiments using the ferro-ferricyanide
redox probe.
The influence of phosphatidylcholine peroxides generation
time. As demonstrated by Niki, the constant rate production of
peroxyl radicals depends on the concentration of the azo-initiatorin the initial solution and the temperature and time of thethermal initiation of free radical formation. Measurementswere performed on the current intensity value correspondingto the reduction of generated PCOOH (PC hydroxy-peroxides) to
establish the optimal incubation time between ROO
/C15and PC
substrate. The CA measurements in the flow were carried outusing a PC–Fe
3O4composite amount able to provide the
equivalent concentration of 40 mgm L/C01PC, while the AAPH
initial concentration was 30 /C210/C03mol L/C01. The results, as
means of triplicate measurements for the entire time range using5 injections per point in two subsequent days of measurementsunder similar conditions, are given in Fig. 5.
It can be observed that 20 minutes of incubation between PC
(PC–Fe
3O4nanocomposite suspension, equivalent to 40 mgm L/C01
PC concentration) and AAPH (30 /C210/C03mol L/C01)a t6 0 1C
represents the optimal time to be used for hydroxy-peroxideformation at a constant rate. The maximum measured value ofthe current intensity of the PCOO
/C15/PCOOH reduction peak was
attained in these experimental conditions.
The influence of flow rate on the chronoamperometric response.
When a flow measurement is performed, the fact has to be
taken into account that the flow rate significantly influencesthe method’s sensitivity. Accordingly, the variation of theregistered current intensity with the flow rate was studiedunder the experimental conditions mentioned above. Therecommended flow rate value was 2.5 mL min
/C01(Fig. 6), which
is quite a high value. This flow rate value is, in fact, correlatedto the mechanism of the monitored reaction, which is a fast
one being a radical chain mechanism.
Application of the PC–Fe
3O4developed sensor in the assess-
ment of lipophilic antioxidant efficacy against PC peroxidation(batch and flow measurements).
Fig. 4 The influence of oxidative substrate concentration on the registered
signal.Table 1 Magnetic field influence on chronoamperometric signal (mea-
sured as corresponding to PCOO/C15reduction)a
Peroxyl radical concentration
[ROO/C15] [nmol L/C01]Imean (/C6SD) [nA]
No magnetic field Magnetic field
1.63 3.72 /C60.28 4.33 /C60.09
2.5 8.76 /C60.49 9.60 /C60.59
3.82 14.03 /C61.97 15.27 /C60.85
an= 5 measurements.Paper NJC

6434 |New J. Chem., 2015, 39, 6428–6436 This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2015The optimized system was further applied to assess the
efficacy of several antioxidants against lipoperoxidation. Thedeveloped model was tested for astaxanthin and lutein, to checktheir antioxidant behaviour against PC peroxidation, since theenvisaged applicability of the system is the development of a
measuring model able to mimic the membrane distribution of
different carotenoid classes. The electrochemical response was alsotested for another lipo-soluble antioxidant, namely, a-tocopherol,
considering that the reaction’s anti oxidant-quenched free radical is
as p e c i f i co n e .
Using the previously optimised incubation time, i.e.,2 0m i n u t e s ,
anin situ concentration of 1.63 /C210
/C03mol L/C01and a registered
value of current intensity, iPCOO/C15
FR , which was measured on the
detector level, corresponding to the PC peroxide formation, the
PC–Fe 3O4on the surface of AuSPE in stopped flow was left to react
with peroxyl radicals, ROO/C15. On the second detector, simultaneously,
the same reaction was carried out together with the antioxidant
in order to estimate the preservative effect of tested lipophilic
antioxidants in the same conditions, i.e., in this case, measuring
the current intensity of the reduction peak corresponding to
residual PC peroxides, iPCOO/C15
FRțAox.R e s i d u a lP Cp e r o x i d e sa r et h o s e
peroxides un-quenched by the tested antioxidant. The efficacy of
the antioxidants’ preservation was expressed as a relative per-
centage of the lipoperoxide formation.45In this respect, it should
be mentioned that with increase in the percent of measuredPCOO
/C15/PCOOH (meaning higher the value of the CA measured
current intensity), the antioxidant efficacy decreases.The efficacy of tested antioxidants was expressed as the relative
percentage of lipoperoxide formation using the formula:
%PCOO/C15¼iPCOO/C15
FR
iPCOO/C15
FRțAox/C2100
The results obtained for batch analysis using standard compounds
are given in Fig. 7. The first tested astaxanthin was the synthetic
one. It is noticeable that for the membrane model, the astaxanthinefficacy is higher than the vitamin E efficacy. The percent ofgenerated PC peroxides is higher when the membrane preservationwas ensured by vitamin E with respect to the value obtained forastaxanthin preservation. Our electrochemical results were inagreement with data published by Sowmya et al.
44The astaxanthin
specific orientation and position at the membrane level ensures a
better effect both by the pelectron cloud and the hydroxyl groups
compared to lutein that is effective mainly due to the pelectron
cloud, according to its orientation into the membrane.
The PC–Fe 3O4sensor’s response on assessing the efficacy of
standard synthetic astaxanthin was checked even in flowexperiments (Fig. 8), these experiments proved the PC–Fe
3O4
sensor’s feasibility for potential in-line measurements. The useof two flow lines ensures the possibility to obtain directly the
relative percentage of lipoperoxide formation since the signal
corresponding to peroxidation, i
PCOO/C15
FR , is measured to one
Fig. 5 Variation of the registered current intensity of PCOO/C15/PCOOH
reduction peak with free radical-substrate incubation time.
Fig. 6 Variation of measured PCOOH current intensity with carrier buffer
flow rate.
Fig. 7 Variation of inhibition of PC peroxidat ion with antioxidant concentration.
Fig. 8 Chronoamperograms of PCOO peroxides signal in the absence
and in the presence of astaxanthin registered at +0.385 V vs.screen
printed reference electrode, in flow.NJC Paper

This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2015 New J. Chem., 2015, 39, 6428–6436 | 6435detector, while the signal corresponding to antioxidant preser-
vative effect, iPCOO/C15
FRțAox, is measured simultaneously at the second
detector (see flow system scheme in ESI, †Fig. S4).
The results obtained using the developed sensor were com-
pared to those obtained for the same standard compoundswhen the ORAC assay was used. The trolox equivalent (TE) permicrogram compound being 4.88 mmol L
/C01TE/mg astaxanthin;
3.79 mmol L/C01TE/mg vitamin E and 4.71 mmol L/C01TE/mg lutein.
ORAC assay led to the same trend of antioxidant efficacy as the
results from our developed model.
Application of PC–Fe 3O4sensors in assessing astaxanthin
antioxidant efficacy during accumulation in H. pluvialis
The developed sensors were applied to flow measurements
to test the efficacy of astaxanthin during its accumulation inH. pluvialis at different growth stages. Astaxanthin was
extracted after algal cell disruption using the protocol givenin Section 2.5 and the content of extracted astaxanthin was
determined using UV-Vis spectrophotometry. The amount of
astaxanthin so determined was used to calculate the volume of
the extract injected in the measuring system in order to ensure theequivalent astaxanthin amount a s used in the standard solution,
namely, 4.10 /C210
/C09mol L/C01. Measurements were performed
after 20 minutes of incubation of free radicals with PC–Fe 3O4
oxidizable substrate in the pres ence of astaxanthin. The results
obtained are given in Table 2 and considering the SD values, the
electrochemical measurements are accurate. However, it should
be emphasised that at least in the first phase of H. pluvialis
growth (green-olive stage), the obtained values could be due to thepresence of other carotenoids acting sinergistically.
Accordingly, it should be emphasized that the results
obtained give an overall estimation of the total antioxidantefficiency against lipoperoxidation, the method not being able todiscriminate between the components. The observed antioxidant
activity is fairly well correlated to overall carotenoid accumulation
during growth. In light stress (as was the case in our samplesprovided by Algatech, Kibutz Ket urra), the astaxanthin amount
increases, whereas the amounts of chlorophylls and astaxanthinprecursors decreases. The carotenoid fraction in the green stage isquite exclusively composed of lutein and b-carotene, while in the
red stage (cysts), the major caroten oid is astaxanthin. Moreover, it
s h o u l db em e n t i o n e dt h a tt h ed e v e l o p e dm o d e lb a s e do no x i d a –
tion of PC–Fe
3O4composite provides resul ts regarding astaxanthinantioxidant efficacy comparable wi th those previously reported by
Capelli et al.51
To properly discuss on the antioxidant efficiency of meta-
bolites of plant origin and deliver a correct estimation of theireffect, the equivalent antioxidant capacity (EAC) was calcu-lated. The value was expressed per mg dry weight of algal massused to obtain the same inhibition of PCOOH formationinduced by 1 mg synthesized astaxanthin. Despite the factthat synthetic astaxanthin was less efficient than astaxanthinextracted from H. pluvialis , the synthetic compound was used
to calculate the equivalence since it is less affected by lot-to-lot
variability.
Conclusions
An electrochemical sensor based on phosphatidylcholine (PC)immobilization on magnetic nanoparticles mimicking membrane
lipid layers was developed and optimised in order to ensure the
suitable experimental conditions in terms of carotenoid distri-bution in membranes (position and orientation). The obtainedsensor was applied for the determination of the antioxidantefficacy of astaxanthin and proved to be useful even for otherlipophilic antioxidants (vitamin E class). The PC–Fe
3O4-based
sensor proved to have high sensitivity in assessing equally thereactive oxygen species, ROO
/C15(LoD = 9.1 /C210/C010mol L/C01)a n d
in situ generated PCOO/C15. The sensor provided reliable informa-
tion both for batch and flow measurements. The designed flowsystem demonstrated its applicability for the in-line control ofastaxanthin accumulation subsequent to algal encystment, oneof the main advantages of the constructed device is its capabilityto supply a renewable surface in flow measurements thusavoiding false responses due to oxidizable substrate damage. Itwas proved that the storage stability of obtained PC–Fe
3O4
composite is longer than 45 days.
The developed electrochemical sensor was further applied
to assess the astaxanthin properties during accumulation inH. pluvialis ; the registered responses being correlated with
the expected content of carotenoids in various cyst stages.Moreover, considering the reported distribution of lutein andb-carotene at the membrane level and based on registered
sensor response variability between the green stage and the
red stage, it is obvious that the developed membrane model
using immobilized PC is feasible and supplies information witha better match between in vitro and in vivo measurements. It
was proved that natural astaxanthin is a more effective anti-oxidant than synthetic astaxanthin, and that astaxanthin ismore efficient than vitamin E. Our results are comparable withthose reported in the literature.
49
Acknowledgements
The authors acknowledge FP7 Sensbiosyn project 232522/2009for financial support and COST TD 1102 action for providingthe opportunity of improving knowledge by fruitful debates.Table 2 Electrochemical evaluation of astaxanthin efficacy against PCOOH
formation, flow rate 2.5 mL min/C01, astaxanthin equivalent concentration 4.10 /C2
10/C09mol L/C01a
Sample% of generated
PCOO radicals ( /C6SD)EAC ( /C6SD) equivalent synthetic
astaxanthin per mg dry weight
Astaxanthin std. 70.8 ( /C60.08) 1.00 ( /C60.05)
Green stage 81.4 ( /C60.37) 0.35 ( /C60.07)
Olive stage 80.4 ( /C60.10) 0.88 ( /C60.02)
Brown stage 72.7 ( /C60.39) 0.94 ( /C60.13)
Red stage 43.1 ( /C60.05) 1.65 ( /C60.05)
aSD – standard deviation calculated for 5 repeated determinations.Paper NJC

6436 |New J. Chem., 2015, 39, 6428–6436 This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2015References
1 G. Marrazzo, I. Barbagallo, F. Galvano, M. Malaguarnera,
D. Gazzolo, A. Frigiola, N. D’Orazio and G. Li Volti, Crit. Rev.
Food Sci. Nutr. , 2014, 54, 1599.
2 R. G. Fassett and J. S. Coombes, Mar. Drugs , 2011, 9, 447.
3 J. Falk and S. Munne ´-Bosch, J. Exp. Bot. , 2010, 61, 1549.
4 T. Lavecchia, G. Rea, A. Antonacci and M. T. Giardi, Crit.
Rev. Food Sci. Nutr. , 2013, 53, 198.
5 G. Goswami, S. Chaudhuri and D. Dutta, J. Microbiol. Bio-
technol. , 2010, 26, 1925.
6 I. Po-Fung and C. Feng, Process Biochem. , 2005, 40, 733.
7 N. Sun, Y. Wanga, Y.-T. Li, J.-C. Huang and F. Chen, Process
Biochem. , 2008, 43, 1288.
8 M. Wan, J. Zhang, D. Hou, J. Fan, Y. Li, J. Huang and
J. Wang, Bioresour. Technol. , 2014, 167, 276.
9 W. Zhang, J. Wang, J. Wang and T. Liu, Bioresour. Technol. ,
2014, 158, 329.
10 P. Pe ´rez-Lo ´pez, S. Gonza ´lez-Garcı ´a, C. Jeffryes, S. N. Agathos,
E. McHugh, D. Walsh, P. Murray, S. Moane, G. Feijoo andM. T. Moreira, J. Cleaner Prod. , 2014, 64, 332.
11 L. B. Brentner, M. J. Eckelman and J. B. Zimmerman, Environ.
Sci. Technol. ,2 0 1 1 , 45, 7060.
12 R. T. Lorenz and G. R. Cysewski, Trends Biotechnol. , 2000, 18, 160.
13 P. Z. Margalith, Appl. Microbiol. Biotechnol. , 1999, 51, 431.
14 C. Hagen, S. Siegmun and W. Braune, Eur. J. Phycol. , 2002,
37, 217.
15 R. Sarada, S. Bhattacharya and G. A. Ravishankar, World
J. Microbiol. Biotechnol. , 2002, 18, 517.
16 J. Steinbrenner and H. Linden, Plant Mol. Biol. ,2 0 0 3 , 52,3 4 3 .
17 S. Boussiba, W. Bing, A. Zarka, J. P. Yuan and F. Chen,
Biotechnol. Lett. , 1999, 21, 601.
18 Y. M. A. Naguib, J. Agric. Food Chem. , 2000, 48, 1150.
19 A. Vershinin, Biofactors , 1999, 10, 99.
20 G. Britton, S. Liaaen-Jensen and H. Pfander, Carotenoids
Hand Book , Birkha ¨user, Basel, Switzerland, 2004.
21 A. Mortensen, L. H. Skibsted, J. Sampson, C. Rice-Evans and
S. A. Everett, FEBS Lett. , 1997, 418, 91.
22 I. Higuera-Ciapara, L. Felix-Valenzuela and F. M. Goycoolea,
Crit. Rev. Food Sci. Nutr. , 2006, 46, 185.
23 R. R. Ambati, S. M. Phang, S. Ravi and R. G. Aswathanarayana,
Mar. Drugs , 2014, 12,1 2 8 .
24 M. Guerin, M. E. Huntley and M. Olaizola, Trends Biotechnol. ,
2003, 21,2 1 0 .
25 K. Takahashi, T. Takimoto, K. Sato and Y. Akiba, Anim. Sci.
J., 2011, 82, 753.
26 J. P. Yuan, J. Peng, K. Yin and J. H. Wang, Mol. Nutr. Food
Res., 2011, 55, 150.27 E. Yamashita, Funct. Foods Health Dis. , 2013, 3, 254.
28 S. C. Litescu, S. A. V. Eremia and G. L. Radu, Adv. Exp. Med.
Biol., 2010, 698, 241.
29 N. J. Miller and C. A. Rice-Evans, Clin. Sci. , 1995, 41, 1789.
30 R. Re, N. Pellegrini, A. Proteggente, A. Pannala, M. Yang and
C. A. Rice-Evans, Free Radical Biol. Med. , 1999, 26, 1231.
31 D. D. M. Wayner, G. W. Burton, K. U. Ingold and S. Locke,
FEBS Lett. , 1985, 187, 33.
32 J. M. DeLong, R. K. Prange, D. M. Hodges, C. F. Forney,
M. C. Bishop and M. Quilliam, J. Agric. Food Chem. , 2002,
50, 248.
33 B. Ou, D. Huang, M. Hampsch-Woodill, J. A. Flanagan and
E. K. Deemer, J. Agric. Food Chem. , 2002, 50, 3122.
34 G. H. Cao, H. M. Alessio and R. G. Cutler, Free Radical Biol.
Med. , 1993, 14, 303.
35 C. Sanchez-Moreno, Food Sci. Technol. Int. , 2002, 8, 121.
36 W. Brand-Williams, M. E. Cuvelier and C. Berset, LWT –
Food Sci. Technol. , 1995, 28, 25.
37 W. Bors, C. Michel, M. Saran and E. Lengfelder, Biochim.
Biophys. Acta , 1978, 540, 162.
38 A. A. Woodall, G. Britton and M. J. Jackson, Biochim.
Biophys. Acta , 1997, 1336 , 575.
39 T. Iyama, A. Takasuga and M. Azuma, Int. J. Vitam. Nutr.
Res., 1996, 66, 301.
40 A. Cantrell, D. J. McGarvey, T. G. Truscott, F. Rancon and
F. Beohm, Arch. Biochem. Biophys. , 2003, 412, 47.
41 A. Kosteka-Gugala, D. Latowski and K. Strzalka, Biochim.
Biophys. Acta , 2003, 1609 , 193.
42 Y. Matsushita, R. Suzuki, E. Nara, A. Yokoyama and
K. Miyashita, Fish. Sci. , 2000, 66, 980.
43 K. Fukuzawa, Y. Inokami, A. Tokumura, J. Terrao and
A. Suzuki, Biofactors , 1998, 7, 31.
44 R. Sowmya and N. M. Sachindra, Food Chem. , 2012,
134, 308.
45 M. F. Barroso, N. de-los-Santos-A ´lvarez, C. Delerue-Matos
and M. B. P. P. Oliveira, Biosens. Bioelectron. , 2011, 30, 1–12.
46 S. C. Litescu, N. Cioffi, L. Sabbatini and G. L. Radu, Electro-
analysis , 2002, 14, 858.
47 S. C. Litescu, S. A. V. Eremia, M. Diaconu, A. Tache and
G. L. Radu, in Environmental Biosensors , ed. V. Somerset,
InTech, 2011, pp. 95–116.
48 M. Valko, D. Leibfritz, J. Moncol, M. T. D. Cronin, M. Mazur
and J. Telser, Int. J. Biochem. Cell Biol. , 2007, 39, 44.
49 I. Dalle-Donne, R. Rossi, R. Colombo, D. Giustarini and
A. Milzani, Clin. Chem. , 2006, 52, 601.
50 E. Niki, Methods Enzymol. , 1990, 186, 100.
51 B. Capelli, D. Bagchi and G. R. Cysewski, Nutrafoods , 2013,
12, 145.NJC Paper