Assessment of the breakdown products of solar/UV induced photolytic degradation of food dye tartrazine

Food and Chemical Toxicology 68 (2014) 307–315

Contents lists available at ScienceDirect

Food and Chemical Toxicology

journal homepage: www.elsevier .com/locate / foodchemtox

Assessment of the breakdown products of solar/UV induced photolyticdegradation of food dye tartrazine

http://dx.doi.org/10.1016/j.fct.2014.03.0250278-6915/� 2014 Elsevier Ltd. All rights reserved.

Abbreviations: MR, mutagenic ratio; MW, molecular weight.⇑ Corresponding author. Tel.: +55 (16) 98260 9946.

E-mail address: [email protected] (T.C. dos Santos).

Tuane Cristina dos Santos a,⇑, Guilherme Julião Zocolo a, Daniel Alexandre Morales b,Gisela de Aragão Umbuzeiro b, Maria Valnice Boldrin Zanoni a

a Chemistry Institute (Unesp), Department of Analytical Chemistry, Rua Francisco Degni 55, Bairro Quitandinha, 14800-900 Araraquara, Brazilb Technology Institute (Unicamp), Rua Paschoal Marmo 1888, 13484-332 Limeira, Brazil

a r t i c l e i n f o

Article history:Received 27 November 2013Accepted 22 March 2014Available online 2 April 2014

Keywords:Dye photolysisTartrazineBreakdown productsMutagenicityLC–MS/MS

a b s t r a c t

The food dye tartrazine (CI 19140) was exposed to UV irradiation from an artificial source, a mercuryvapor lamp, and a natural one, sunlight. It was observed that conditions such as energy dose, irradiationtime, pH and initial dye concentration affected its discoloration. There was 100% of color removal, after30 min of irradiation, when a dye solution 1 � 10�5 mol L�1 was submitted to an energy dose of37.8 J cm�2. Liquid Chromatography coupled to Diode Array Detection and Mass Spectrometry confirmedthe cleavage of the chromophore group and the formation of five by-products at low concentration.Although by-products were formed, the Salmonella/microsome mutagenicity assay performed for both,the dye solution at a dose of 5.34 mg/plate and the solutions obtained after exposure to UV irradiation,did not present mutagenic activity for TA98 and TA100 with and without S9.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

It is one of nature’s inevitabilities that all colored objects dyedwith organic dyes will fade upon exposure to light. In general,the sun’s energy is made up of three distinct spectral components:ultraviolet radiation, visible radiation, and near-infrared radiationand all of them are capable of promoting reactions that lead tofading. Fading is influenced by many environmental factors suchas the dye chemical structure, chemical environment, environmen-tal temperature, humidity and radiation.

Usually, the UV radiation has enough energy to cause unstablebonds to break or transform the dye structure. Putt et al. (2012)showed in their study that some food, drug and cosmetic (FD&C)dyes, including the azo dye Yellow 5 (also known as tartrazine),are suitable for degradation upon irradiation with UV light.Furthermore, they reported dyes were degraded showing simpledecay curves.

The change in the chemical molecule can lead to the formationof new compounds, which can either be more or less harmful tohumans. According to Oancea and Meltzer (2013), the lowerdecrease on the total organic carbonic (TOC) percentage when

compared to tartrazine’s degradation percentage upon the Photo-Fenton process, 43% and 98% respectively, suggested that the dyeis degraded to small molecules, such as aldehydes and organicacids. Even when the irradiation time was prolonged to 120 minthere was only a partial degradation of those molecules once TOCdecrease did not exceed 80%. Therefore, the understanding of thelight effect on dyes, present in the aquatic environment or in foodstuff, is important to predict chemical alterations and toxicologicalproperties (Neppolian et al., 2002).

The interest of this work is to study tartrazine (CI 19140) as amodel of azo dye widely used in food industry. There is a contro-versy about its genotoxic potential and other adverse effects(Hutchinson et al., 1992; Pestana et al., 2010; Lockey, 1959).Watabe et al. (1980) showed that primary metabolization mecha-nism of ingested tartrazine in man is bacterial azo reduction,generating mainly sulphanilic acid. Amin et al. (2010) observedthat rats exposed to high dosage of tartrazine (500 mg/kg bw)exhibited hepatocellular damage. The cytotoxicity study oftartrazine, reported by Mpountoukas et al. (2010), indicates thattartrazine at higher concentrations (4 and 8 mM) has significanttoxic effect on the quality of chromosomes, probably by affectingthe condensation of chromosomes in mitosis. In addition, the acuteoral exposure to tartrazine induced DNA damage in mouse colon asobserved in the in vivo comet assay (Sasaki et al., 2002). Somestudies indicated that tartrazine induced chromosomal aberrations

308 T.C. dos Santos et al. / Food and Chemical Toxicology 68 (2014) 307–315

in fibroblast cells of Muntiacus muntjac (Chung et al., 1981), onbone marrow cells of mice and rats (Giri et al., 1990) and onchromosomes of Allium cepa (Roychoudhury and Giri, 1989). Nev-ertheless, literature also reports studies indicating negative resultsfor genotoxicity/carcinogenicity tests (Kada et al., 1972; Ishidateet al., 1984; Combes, 1986). Therefore, its clear that discussionregarding the toxic potential of tartrazine started long ago and,as it can be seen in the studies cited above, even today there isno clear consensus on this matter. In addition, along with the com-plexity of the by-products generated due azo dye degradation, fewstudies are focusing on the influence of light on dye residues pres-ent in the wastewater and surface water, due to limitation of theavailable analytical methods.

Thus, the aim of this work is to investigate the tartrazine degra-dation under UV and solar light irradiation by using High Perfor-mance Liquid Chromatography coupled to Diode Array Detection(HPLC–DAD) and Liquid Chromatography coupled to MassSpectrometry (LC–MS–MS), mimicking experimental situationswhere tartrazine is submitted to artificial and sunlight irradiation.Our particular interest was to focus on evidence of harmful by-products generation for a food dye exposed to different irradiationconditions. Understanding the dye biotransformation route underlight influence can help us to predict the toxicity of by-productsgenerated. The original dye solution and the breakdown productsgenerated from photolysis were evaluated by using Salmonellatyphimurium strains TA98 (hisD3052, rfa, Dbio, DuvrB, andpKM101) and TA100 (hisG46, rfa, Dbio, DuvrB, pKM101) with andwithout exogenous metabolic activation.

2. Material and methods

2.1. Apparatus and procedure

The photolytic experiments were performed using indoor and outdoor reactors.In the indoor equipment was used a 300 mL single compartment reactor,maintained at 20 �C using a thermostatically controlled water bath (Nova Técnica,Brazil). The dye solution was illuminated using a 125 W Philips medium pressuremercury lamp (I = 21 mW/cm2), without the glass tube and inserted into a quartzbulb, as the UV light source (315–400 nm). The solution was constantly stirred.

The outdoor photolytic experiments were carried out using a solar photochem-ical reactor, previously described in literature (Trovo et al., 2008). Aliquots weretaken after 0, 1, 2, 3, 4, 5, 6 h from the beginning of photochemical treatment.The reaction vessels were held in a rack at a 22� angle from horizontal at the Chem-istry Institute, UNESP (Araraquara, São Paulo, 21�500S, 48�000W). The experimentswere performed in sunny days of October, between 10 am and 4 pm. The ambienttemperature ranged between 28 and 36 �C.

UV irradiation, from both artificial and solar source, was measured with an UV-A radiometer Solar Light Co (PMA2100). For comparison of solar and artificial lighttest results, the UV radiation was evaluated in terms of irradiance, which is definedas the rate at which radiant energy is incident on a surface per unit area of surface(W cm�2) and the energy dosage, received on the illuminated reactor surface(J cm�2) (Eq. (1)) (Trovo et al., 2005):

Energy dose ðJ cm�2Þ ¼Z t2

t1Irradiance ðW cm�2Þ � dtðsÞ ð1Þ

The solar irradiance measured was 0.95 mW cm�2 and the artificial light irradi-ance was 21 mW cm�2.

All pH measurements were taken using a Corning 555 pH meter. In experimentswhere the pH was to be kept constant, 0.1 mol L�1 NaOH solution or H2SO4 solutionwas added to the cell in order to control the pH.

2.2. Analytical methods

2.2.1. UV–VisThe UV–Vis absorption was recorded with a Hewlett Packard 8453 spectropho-

tometer operating from 200 to 800 nm and a quartz cell. The discoloration percent-age of tartrazine (k = 427 nm) was calculated using Eq. (2):

% Discoloration ¼ ðAt0 � At1ÞAt0

� 100 ð2Þ

where At0 is the absorbance of the initial tartrazine solution and At1 the absorbancemeasured at different photolysis time intervals.

2.2.2. HPLC–DAD determinationThe High Performance Liquid Chromatography with photodiode array detector

(HPLC–DAD) was carried out using a Shimadzu SCL-10AVP equipped with an auto-sampler. Analysis was performed in a reversed-phase column Shimadzu CLC-ODS(C18) (25 cm � 4.6 mm i.d. � 5 lm, 100 Å). All solutions were previously filteredthrough a 0.45 lm PTFE filter. The best experimental conditions optimized weremobile-phase consisting of 50 mmol L�1 aqueous solution of ionic liquid tetrafluo-roborate 1-butyl-3-metylimidazolium, BMIm-BF4, and acetonitrile (80:20) v/v, flowrate of 0.4 mL min�1 and a column temperature of 40 �C. The chromatograms wereinvestigated between 200 and 800 nm, and the maximum wavelengths selected toanalyze the tartrazine was 427 nm.

Tartrazine concentration was determined using a calibration curve constructedin the linear range of concentrations between 1 and 9 lmol L�1. Analysis were per-formed using the HPLC–DAD optimized condition, and each peak area was relatedto the respective concentration, at 427 nm, through the following equation:Area = 1.03 � 105 + 2.06 � 1010 concentration, n = 7, r = 0.9962, with detection andquantification limits equal 0.12 and 0.39 lmol L�1, respectively.

2.2.3. LC–ESI–MS/MSAll experiments using Liquid Chromatography coupled to Mass Spectrometry

were recorded in an electrospray ionization tandem Mass Spectrometry (ESI–MS/MS) QTrap system (Applied Biosystems 3200 QTrap). The LC system was Agilent1200 equipped with an autosampler. The Mass Spectrometry measurements wereperformed in negative ionization mode, ESI(�). Nitrogen was the collision gas. Ionelectrospray voltage, ion source temperature and desolvation potential were�4.5 kV, 650 �C and �50 V, respectively. The scan rate was set at 4000 UMA s�1

for 400 ms and collision energy of 10, 30 and 50 V. Entrance potential was 8 Vand only fragments with m/z between 50 and 600 Da were taken. The previous sep-arations were performed using an Agilent Zorbax C18 column (150 � 4.6 mm i.d.,5 lm). The mobile phase system consisted of methanol and an aqueous solution1 mmol L�1 ammonium acetate 70:30 (v/v). The elution was on isocratic mode,the flow rate 0.5 mL min�1, the injection volume 10 lL and the oven temperature40 �C. After elution, samples were introduced directly into the mass spectrometerwithout flow splitting. All samples were prepared by dilution in the mobile phaseand filtered through a 0.22 lm membrane.

2.3. Chemicals and reagents

The tartrazine dye 85% (CI 19140) was obtained from Sigma–Aldrich and usedwithout purification. HPLC grade methanol and acetonitrile were purchased fromTedia and ultrapure water was prepared by the Milli-Q system (Millipore). Forthe irradiation experiments tartrazine was dissolved only in ultrapure water.

2.4. The Salmonella/microsome mutagenicity assay

The mutagenicity of the dye and its by-products, obtained after exposure to UVirradiation, was evaluated using the Salmonella/microsome assay according to theISO 16240:2005 (2005) with the S. typhimurium strains TA98 (hisD3052, rfa, Dbio,DuvrB, and pKM101) and TA100 (hisG46, rfa, Dbio, DuvrB, pKM101) with andwithout exogenous metabolic activation. The metabolic activation was providedby Arochlor 1254 induced Sprague Dawley rat liver S9 mix (MolTox, Boone, NC),which was prepared at a concentration of 4% (v/v). For TA98 and TA100, the positivecontrols were 4-nitroquinoline-1-oxide (4NQO) (Sigma–Aldrich) at 0.5 lg plate�1

and 2-aminoanthracene (2AA) (Sigma–Aldrich) at 2.5 lg plate�1, both dissolvedin dimethylsulfoxide. Sterile water was used as negative control. The test was con-ducted using a single dose for each treatment. For each plate, 0.5 mL of sodiumphosphate buffer or S9 mix and 1 mL of sample solution were added. Sample solu-tions were sterilized by filtration (0.45 lm). The test was performed in single dosesof 5.34 lg per plate using two replica plates. After 66 h incubation at 37 �C, the col-onies were manually counted under a stereoscope. Before each test, samples werechecked for contamination using nutrient agar plates. Single maximum doses weretested in duplicates because the volumes of the treated solutions were not enoughfor dose response experiments.

Toxicity was also evaluated by visual observation under a stereoscope. Resultsare expressed as the number of revertants lg�1 equivalent of the sample per plate.The ratio mutagenic was calculated for each sample, as the average number ofrevertants per plate divided by the average number of revertants per plate in thenegative control. A sample was considered positive when the ratio mutagenicwas P2.

3. Results and discussion

3.1. Dye discoloration

Fig. 1a shows the absorption spectra obtained for 1 � 10�5

mol L�1 dye solution before and after exposure to UV irradiationfor 4 h (302.4 J cm�2). The characteristic absorption bands of the

0.0

0.1

0.2

0.3

d b

a

Abso

rban

ce (u

A)

λ (nm)

c

A

-1.2

-1.0

-0.8

-0.6

-0.4

-0.2

0.0 B

ln (A

bs/A

bs0)

Irradiation time (min)200 400 600 800 0 2 4 6 8 10

2 4 6 8 10

0

1

2

3

4

abcdefghijk Abs

at 4

27 n

m

x104 u

A

Retention time (min)

C

Fig. 1. (A) UV/Vis spectra of 1 � 10�5 mol L�1 aqueous tartrazine solution, pH 6.0, after exposure to UV irradiation (I = 21 mW cm�2) for: (a) 0, (b) 15, (c) 30 and (d) 240 min.(B) Initial degradation rate curve, at 427 nm. (C) HPLC chromatograms obtained for the elution of 1 � 10�5 mol L�1 tartrazine solution, pH 6.0, after (a) 0 (b) 2 (c) 4 (d) 6 (e) 10(f) 15 (g) 30 (h) 60 (i) 120 (j) and 180 (k) 240 min of exposure to light. Chromatographic conditions: mobile phase consisted of 50 mM aqueous solution of ionic liquid BMIm-BF4 and acetonitrile (80:20) v/v, flow rate of 0.4 mL min�1 and a column temperature of 40 �C.

0

20

40

60

80

100

120

disc

olor

atio

n ef

ficie

ncy

(%)

a

b

cd

T.C. dos Santos et al. / Food and Chemical Toxicology 68 (2014) 307–315 309

chromophore and aromatic groups (Silva et al., 2008) at 427 nmand 257 nm, respectively, are successively reduced duringphotolytic exposition. The inset of Fig. 1a indicates that dye con-centration is not detected by spectrophotometric technique after30 min of irradiation. The decrease in tartrazine concentration asa function of time was plotted as ln (At/A0) vs. time (t), Fig. 1b,and a linear relationship was obtained where the slope indicatesvalues of k = 0.150 min�1. This is a typical behavior of a first-orderreaction in dye consumption (Silva et al., 2008). The influence oflight irradiation on tartrazine degradation was also confirmed byHPLC–DAD (Fig. 1c). Tartrazine initially presents a well-definedpeak at retention time of (tr) 8 min that decreases successivelywith time. Concomitantly, after 2 min of light exposition an extrapeak appears, tr = 7 min, probably due to formation of some inter-mediary compound.

0 50 100 150 200 250 300 350

Energy Dose (J cm )-2

Fig. 2. Effect of the initial dye concentration on discoloration during exposure to UVirradiation for 240 min: (a) 5 � 10�6 (b) 1 � 10�5 (c) 8 � 10�5 (d) 1 � 10�4 mol L�1.

3.2. Effect of the initial dye concentration

Fig. 2 illustrates the effect of initial dye concentration on degra-dation rate, performed by varying the following concentrationvalues, 1 � 10�4, 8 � 10�5, 1 � 10�5 and 5 � 10�6 mol L�1, andstudying the changes on the solutions coloration. The photodecom-position of the dye was followed by measuring the absorbancedecay at k = 427 nm over 240 min. The absorbance decay,expressed as dye concentration, showed 100% of color removal inall experiments. The degradation rates, obtained from the linearrelationship of ln (At/A0) vs. t, decreased from 0.388 min�1 (forthe most diluted solution, 5 � 10�6 mol L�1) to 0.015 min–1 (forthe most concentrated, 1.0 � 10�4 mol L�1), showing that the effectof UV irradiation is stronger on diluted solution conditions, wherethe light absorption is higher. Therefore, the UV irradiation iscapable of discoloring tartrazine solutions, indicating the probable

cleavage of the azo bond due photolysis and the formation ofby-products.

3.3. Effect of pH

The effect of pH on the discoloration of a 1 � 10�5 mol L�1 tar-trazine solution, promoted by UV irradiation, was monitored frompH 2.0 to 10.0 (the initial pH value was adjusted with 0.1 mol L�1

H2SO4 solution or 0.1 mol L�1 NaOH solutions). Fig. 3a shows thatdiscoloration reached 100% after 20 min of irradiation. The rate

0

20

40

60

80

100 A

disc

olor

atio

n ef

feci

ency

(%)

Irradiation time (min)

pH2.0 pH6.0 pH8.0 pH10.0

0 50 100 150 200 250 0 2 4 6 8 10 12 14 16-3.0

-2.5

-2.0

-1.5

-1.0

-0.5

0.0

ln (A

bs/A

bs0)

Irradiation time (min)

a

bcd

B

Fig. 3. (A) Discoloration percentages and (B) discoloration rates, under UV irradiation, of a 1 � 10�5 mol L�1 tartrazine solution at pH values (a) 2.0, (b) 10.0, (c) 6.0 and (d) 8.0,according to first-order kinetics.

310 T.C. dos Santos et al. / Food and Chemical Toxicology 68 (2014) 307–315

constants were calculated from a first-order kinetic, Fig. 3b, andpresent values of 0.306, 0.150, 0.122 and 0.270 min�1, respectivelyfor pH 2.0, 6.0, 8.0 and 10.0. Maximum values were obtained atacidic pH probably due to the molecule hydrolytic desulfonation(Braun et al., 1991), where the aromatic rings of sulfonic acid mol-ecules are not protonated in its ground state. Protonation mustthen occur upon electronic excitation, being the lowest singletexcited state (S1) the reactive excited state.

Considering the environmental focus of this study, it is impor-tant to estimate the dye half-life time (t1/2) under extreme UV irra-diation condition. It is known that the larger the value ofdegradation rate constant (k) the faster is the disappearance ofdye present in rivers, for example. Thus, it can be established arelation between k and half-life time for a given substance: consid-ering the characteristic equation of the first-order kinetics, ln (At/A0) = �kt, and the fact that the t1/2 is defined as the time requiredfor the concentration of the substance decrease to half of its initialvalue, some mathematical operations can be made (Souza et al.,2008; Atkins and Jones, 2007) resulting in the half-life time equa-tion: t1/2 = ln2/k. Taking into consideration the rate constantk = 0.150 min�1 (for the solution 1 � 10�5 mol L�1, pH 6.0) previ-ously obtained, its found the value t1/2 = 5 min, meaning it takes5 min to decrease tartrazine concentration to half of its initialvalue. Even though the results obtained here show energy dosagehigher than the one present in environmental conditions, they stilljustify the concern of this study.

3.4. The effect of solar light irradiation on aqueous dye solution

The UV portion of solar irradiation is attenuated when it passesthrough Earth’s atmosphere, being scattered and absorbed by oxy-gen and ozone. All UV-C radiation (100–290 nm) is absorbed by theozone layer, while 90% of the radiation with wavelength between290 and 315 nm reaches the Earth’s surface. Thus, the environment

Table 1Energy doses calculated from measurements of the solar UVA irradiation and solutiontemperature observed during the discoloration of a 1 � 10�5 mol L�1 (pH 6.0)tartrazine solution, between 10 am and 4 pm in Araraquara.

Time (h) Irradiation time (h) Energy dose (J cm�2) Temperature (�C)

10:00 0 – 32.0011:00 1 1.56 36.0012:00 2 2.28 38.0013:00 3 3.88 39.0014:00 4 5.34 40.0015:00 5 6.66 42.0016:00 6 7.69 42.00

is mainly exposed to UV-A (315–400 nm), UV-B (290–315 nm), vis-ible and infrared radiation.

Table 1 shows the energy doses calculated after each exposuretime and the respective solution temperatures observed during thediscoloration of a 1 � 10�5 mol L�1 (pH 6.0) tartrazine solutionexposured to sunlight. The UV-A irradiation measured on the dayof the experiment was 0.95 mW cm�2. It is important to mentionthat the energy dose shown in the x-axis of the plots correspondsto the effective energy dose, which differs from the measuredenergy dose, since the volumes of the reactor and the solutioneffectively irradiated by the sunlight are not same in theexperiments. The effective energy dose was calculated accordingto Eq. (3):

Energy doseeffective ¼ Energy dosemeasured �Volumeirradiated

Volumetotalð3Þ

where Volumeirradiated is the volume of the glass tubes (433 mL) andVolumetotal corresponds to the reactor reservoir volume, 1000 mL.

The results obtained by UV–Vis spectrophotometric and HPLC–DAD analysis indicate that under solar irradiation there was no dyemodification within 6 h of exposure. There was no decreasing inabsorbance in the UV–Vis spectra and no extra peak, or decreasingin tartrazine peak, in the HPLC chromatograms. Similar experi-ments were carried out for tartrazine solutions at pH values equalto 2.0 and 10.0. The results also pointed no degradation.

According to Neppolian et al. (2002) only 5% of the solar radia-tion has a great energy vale. Furthermore, solar energy is moreintense in the wavelength range between 450 and 550 nm, visibleregion, and less intense in the UV region, especially at 254 nm, andthat is probably why no degradation occurred during the period oftime studied.

So, considering the performance obtained for a dye solution1 � 10�5 mol L�1 irradiated by a mercury vapor lamp (125 W), itwould take an energy dose of 37.8 J cm�2 (30 min irradiation) todegrade the chromophore moiety (total discoloration). Taking intoaccount that the energy dose achieved by sunlight exposure, dur-ing 6 h, was 7.69 J cm�2 it should be necessary about 30 h of expo-sure to sunlight to promote the total solution discoloration,considering the same period of irradiation and similar climaticconditions.

Moreover, as previously mentioned in the text, the UV portionof solar irradiation is attenuated when it passes through Earth’satmosphere, being scattered and absorbed by oxygen and ozone.Therefore, its light irradiance will be smaller when compared tothe one from an artificial source, as observed with reported data.The study was performed for a determined range of time (6 hmaximum for solar irradiation). When we think of a wider period

Fig. 4. LC–ESI (�)–MS chromatograms obtained for tartrazine solutions exposed to UV light for (A) 0 (B) 2 (C) 4 (D) 6 (E) 10 (F) 15 (G) and 30 (H) 240 min, in an acquisitionrange 50–600 Da.

T.C. dos Santos et al. / Food and Chemical Toxicology 68 (2014) 307–315 311

of time the effects caused by solar irradiation would probably bethe same as the ones caused by artificial irradiation. In anotherwords, there would be the formation of breakdown products,what justifies the relevance of this work. Another factor to beconsidered is that the solar irradiance changes depending on cli-matic conditions and location, so it could be higher in anothertime of the year or if the studies were performed at a differentlocation.

3.5. Analysis by LC–ESI–MS/MS

LC–ESI–MS/MS analyze were also conducted aiming to diagnosethe main products generated for tartrazine solution exposure to UVirradiation. Analysis are based on retention times (tr), molecularmass (MM) and ionic products obtained. According to Holcapeket al. (1999), higher signal stability and sensitivity can be achievedusing electrospray ionization (ESI) in negative mode, so that was

312 T.C. dos Santos et al. / Food and Chemical Toxicology 68 (2014) 307–315

chosen for this work. Fig. 4 shows the chromatograms obtained byESI(�) MS analysis of 1 � 10�5 mol L�1 tartrazine solution exposedto UV irradiation during 240 min. It can be seen that concentrationof tartrazine, peak at tr = 2.25 min, decreases slowly until it reachesnegligible values after 30 min of irradiation (Fig. 4g). At the sametime it is noticed the presence of new peaks associated with theby-products formed (Fig. 4g and h), agreeing with HPLC–DAD data(shown in Fig. 1c).

The peak, in the chromatograms of Fig. 4, is associated with thefragment (m/z = 198.2) which arises due to the reduction of the dyemolecule occurred in the equipment ionization source. This is char-acteristic of the pyrazolone ring cleavage and the peak is associatedwith isocyanate [O3SAC6H4AN@C@O]� (Rafols and Barcelo, 1997).The same fragmentation behavior was observed in Feng’s et al.(2011) recent work. After acquiring the chromatograms, fragmen-tation studies were performed using the QTRAP tool. The frag-mented ions are present in low relative abundances, beingtherefore necessary the use of MS/MS experiments in which theprecursor ion is isolated by ion mass analyzer and collides withN2 gas generating ions fragmented by CID (collision induced disso-ciation) (Vanerkova et al., 2006). Table 2 summarizes the fragmen-tation studies performed. The results indicate the formation ofProduct I (m/z = 467.5), as resulting from the loss of three sodiumions plus the gain of two hydrogens, generating the deprotonatedion identified by MS/MS (Rafols and Barcelo, 1997). Furthermore,it is also confirmed the consecutive neutral loss of two water mol-ecules (m/z = 449.4) and (m/z = 431.5), characteristic of hydroxyl-ated compounds (Holcapek et al., 2007).

The spectrum reported for Product II (m/z = 298.2) highlightsthe azo bond cleavage AN@N leading to the neutral loss of N2

(28 Da), present as the deprotonated ion [II-SO3HAN2]�

Table 2Summary of the fragmentation study of tartrazine photochemical breakdown products.

MM (Da) Deprotonated ion MS (m/z – structure) Ionic product M

Tartrazine 198.2 – [M-H]� 134 – [M-H-SOMW = 534.37

Product I 467.50 – [M-H]� 449 – [M-H-H2

MW = 467.41 431 – [M-H-2H

Product II 298.30 – [M-H]� 217 – [M-H-SOMW = 298.25 189 – [M-H-SO

184 – [H2CNC6H

Product III 283.30 – [M-H]� 269 – [M-H-N]MW = 283.24 229 – [M-H-C3H

201 – [M-H-C3H189 – [M-H-C4H

Product IV 239.20 – [M-H]� 223 – [M-H-O]MW = 239.23 195 – [M-H-NH

157 – [M-H-C3H

Product V 185.20 – [M-H]� 171 – [M-H-N]MW = 185.18 155 – [M-H-H2

131 – Unknown122 – Unknown

(m/z = 189.0); the presence of the ionic product [H2CNC6H4SO3]�

(m/z = 184.1), resulting from pyrazolone ring rupture and the lossof ASO3H groups noticed on [II-SO3H]� (m/z = 217.1) and (m/z = 189.0) ions. Such losses are common to polisulfonatedcompounds and can be explained by the destabilization causedby sulfates groups to the deprotonated ion (Rafols and Barcelo,1997; Vanerkova et al., 2006; Holcapek et al., 2007).

The same approach was used to identify Products III, IV and V.MS/MS spectra allowed identification of ionic fragments resultingfrom the five member ring rupture: [III-N]� (m/z = 269 0),[III-C3H4N]� (m/z = 229.0), as well as the complete loss of thisring: [III-C3H2N2O]� (m/z = 201.0), [III-C4H2N2O]� (m/z = 189.0),[IV-C3H3N2O]� (m/z = 157.1). The presence of fragments character-istic of oxygen [IV-O]� (m/z = 223.2) and nitrogen losses [V-N]�

(m/z = 171.0), [V-H2N2]� (m/z = 155.0) was also noticed. As forthe fragment Dm/z = 44.0, related to by-product IV, the presenceof the ion [IV-NH2CO�]� (m/z = 195.2) can be justified by the occur-rence of an isobaric neutral loss, corresponding to radical NH2CO�,common in some molecules containing heterocyclic nitrogen andcarbonyl groups near the nitrogen (Holcapek et al., 2010).

Based on the results shown so far and on fragmentation data ofanalogous compounds (Rafols and Barcelo, 1997), it is possible topropose the tartrazine degradation route promoted by UV irradia-tion (Fig. 5).

3.6. Ames assay

Results for the negative and positive controls and testedsamples are presented in Tables 3 and 4. TA98 and TA100 strainsadequately responded to the positive and negative controls. Both,dye solution at a dose of 5.34 mg/plate and corresponding

S/MS (m/z – structure) Proposed structure

2]�

N N NC

N

C

C

COONa

SO3NaOH

NaO3S

O]�

N N NC

N

C

C

COOH

SO3-OH

HO3S2O]�

3H]�

NH2 NC

N

C

C

COOH

OHSO3

-3H-N2]�

4SO3]�

�

NC

N

CH

C

COOH

OHSO3

-4N]�

2N2O]�

2N2O]�

�

NCH

N

CH2

CO

S

O

O

O-2CO�]�

3N2O]�

�

N SO3-

NH

N2]�

Fig. 5. Scheme of photochemical degradation of tartrazine under UV irradiation based on structures identified by LC–ESI–MS/MS.

Table 3Mutagenicity results of the dye (T0) and samples after different UV exposure times, for TA98 with and without S9.

Sample Without S9 With S9

Plate 1 Plate 2 Average MR Plate 1 Plate 2 Average MR

Negative control 21 18 20 – 31 36 34 –T0 37 31 34 1.7 29 17 23 0.7T2 min 8 27 18 0.9 12 19 16 0.5T4 min 20 23 22 1.1 18 15 17 0.5T6 min 28 24 26 1.3 35 28 32 0.9T10 min 39 40 40 2.0 29 30 30 0.9T15 min 33 33 33 1.7 23 33 28 0.8T30 min 27 27 27 1.4 38 35 37 1.1T1 h 48 13 31 1.6 32 23 28 0.8T2 h 34 29 32 1.6 24 29 27 0.8T3 h 24 24 24 1.2 30 31 31 0.9T4 h 26 17 22 1.1 31 43 37 1.1Positive controls 270 303 287 15 996 869 933 27.8

T.C. dos Santos et al. / Food and Chemical Toxicology 68 (2014) 307–315 313

solutions after exposure to UV irradiation, did not present muta-genic activity for TA98 and TA100 with and without S9 (Tables 3and 4) because the obtained mutagenic ratios (MR) were below 2for all tested conditions. Although some of the treated solutionspresented MR around 0.5, when the background was inspected

no signs of toxicity were observed. Therefore the results indicatethat no mutagenic activity is expected at lower doses. Martinset al. (2006) also tested the degradation products of CI Basic Red9 using TA98 and TA100 at 500 lL/plate after ozonation and nomutagenic activity was observed both with and without S9.

Table 4Mutagenicity results of the dye (T0) and samples after different UV exposure times, for TA100 with and without S9.

Sample Without S9 With S9

Plate 1 Plate 2 Average MR Plate 1 Plate 2 Average MR

Negative control 156 117 137 – 67 97 82 –T0 120 141 131 1.0 101 112 107 1.3T2 min 82 139 111 0.8 126 132 129 1.6T4 min 105 121 113 0.8 144 146 145 1.8T6 min 81 132 107 0.8 132 99 116 1.4T10 min 137 102 120 0.9 134 105 120 1.5T15 min 105 88 97 0.7 131 115 123 1.5T30 min 70 71 71 0.5 123 107 115 1.4T1 h 98 113 106 0.8 127 128 128 1.6T2 h 117 98 108 0.8 123 112 118 1.4T3 h 99 111 105 0.8 121 107 114 1.4T4 h 94 79 87 0.6 94 105 100 1.2Positive controls 1011 1800 1406 10.3 1304 1500 1402 17.1

314 T.C. dos Santos et al. / Food and Chemical Toxicology 68 (2014) 307–315

Matsushita et al. (2006) used the Ames test to evaluate the photol-ysis of an organophosphorus insecticide and they observed that themutagenicity was decreased with the TA100 related strains(YG1026 and YG1029 but increased with the TA98 related strains(YG1021 and YG1024). They only performed the experimentswithout metabolic activation, but they were able to show thatphotolysis in this case generated mutagenic compounds, whichwas not the case in our study.

4. Conclusions

Our findings indicated that the complete discoloration of theazo dye tartrazine was observed after 30 min of exposure to anenergy dose of 37.8 J cm�2 at pH 6.0. The process followed afirst order kinetic, with a degradation rate constant of0.150 min�1 and at least 5 subproducts diagnosed by HPLC–DAD and HPLC–DAD–MS/MS analysis. The azo bond cleavageand subsequent rupture of the five-membered ring evidencedthat drastic change in the molecule was observed due photo-lytic degradation.

Under solar irradiation, there was no dye modification within6 h of exposure. Considering the necessary time and energy doseirradiated from a 125 W mercury lamp, 30 min and 37.8 J cm�2

respectively, to achieve dye’s total discoloration and the energydose achieved by sunlight after 6 h of exposure, 7.69 J cm�2, itwas estimated that under sunlight irradiation breakdown productswould probably be generated after 30 h of exposure.

Degradation products did not show mutagenic activity in theSalmonella/microsome mutagenicity assay (Ames test) with TA98and TA100 strains in the presence and absence of metabolic activa-tion. Therefore, this research has shown significant results con-cerning the degradation of tartrazine. Even though the dye itselfand its breakdown products did not present mutagenic activitywhen tested by Ames assay at low concentrations, further studiesdealing with in vitro genotoxicity assay should be necessary andattention should be paid to prevent future environmental contam-ination and human health damage.

Conflict of Interest

The authors declare that there are no conflicts of interest.

Transparency Document

The Transparency document associated with this article can befound in the online version.

References

Amin, K.A., Abdel 2nd, A.H.H., Abd, E.A.H., 2010. Effects of azo dyes tartrazine andcarmoisine on biochemical parameters related to renal, hepatic function andoxidative stress biomarkers in young male rats. Food Chem. Toxicol. 48, 2994–2999.

Atkins, P., Jones, L., 2007. Cinética Química. In: Atkins, P., Jones, L. (Eds.), Princípiosde Química, third ed. Bookman, São Paulo, pp. 577–624.

Braun, A.M., Maurette, M.T., Oliveros, E., 1991. Photochemical Technology, first ed.Wiley, Chichester.

Chung, K.T., Fulk, G.E., Andrews, A.W., 1981. Mutagenicity testing of somecommonly used dyes. Appl. Environ. Microbiol. 42, 641–648.

Combes, R.D., 1986. On the mutagenicity of tartrazine (FD & C Yellow No. 5). Arch.Toxicol. 59, 67–68.

Feng, F., Zhao, Y., Yong, W., et al., 2011. Highly sensitive and accurate screening of40 dyes in soft drinks by liquid chromatography–electrospray tandem massspectrometry. J. Chromatogr. B 879, 1813–1818.

Giri, A.K., Das, S.K., Talukder, G., et al., 1990. Sister chromatid exchange andchromosome aberrations induced by curcumine and tartrazine on mammaliancells in vivo. Cytobios 62, 111–117.

Holcapek, M., Jandera, P., Prikryl, J., 1999. Analysis of suphonated dyes ansintermediates by electrospray mass spectrometry. Dyes Pigm. 43, 127–137.

Holcapek, M., Volna, K., Vanerkova, D., 2007. Effects of functional groups on thefragmentation of dyes in electrospray and atmospheric pressure chemicalionization mass spectra. Dyes Pigm. 75, 156–165.

Holcapek, M., Jirasko, R., Lisa, M., 2010. Basic rules for the interpretation ofatmospheric pressure ionization mass spectra of small molecules. J.Chromatogr. A 1217, 3908–3921.

Hutchinson, A.P., Carrick, B., Miller, K., et al., 1992. Adverse reactions to syntheticfood colours: interactions between tartrazine and muscarinic acetylcholinereceptors in isolated guinea-pig ileum. Toxicol. Lett. 60, 165–173.

Ishidate, M., Sofuni, T., Yoshikawa, W., et al., 1984. Primary mutagenicity screeningof food additives currently used in Japan. Food Chem. Toxicol. 22, 623–636.

ISO 16240, International Organization for Standardization: Water Quality –Determination of Genotoxicity of Water and Waste Water-Salmonella/Microsome Test (Ames Test), 2005.

Kada, T., Tutikawa, K., Sadaie, Y., 1972. In vitro and host-mediated ‘‘rec-assay’’procedures for screening chemical mutagens; and phloxine, a mutagenic reddye detected. Mutat. Res. 16, 165–174.

Lockey, S.D., 1959. Allergic reactions due to F D and C Yellow No. 5, tartrazine, ananiline dye used as coloring and identifying agent in various steroids. Ann.Allergy 17, 719–721.

Martins, A.O., Canalli, V.M., Azevedo, C.M.N., et al., 2006. Degradation ofpararosaniline (C.I. Basic Red 9 monohydrochloride) dye by ozonation andsonolysis. Dyes Pigm. 68, 227–234.

Mpountoukas, P., Pantazaki, A., Kostareli, E., et al., 2010. Cytogenetic evaluation andDNA interactions studies of the food colorants amaranth, erythrosine andtartrazine. Food Chem. Toxicol. 48, 2934–2944.

Neppolian, B., Choi, H.C., Sakthivel, S., et al., 2002. Solar/UV-induced photocatalyticdegradation of three commercial textile dyes. J. Hazard. Mater. B89, 303–317.

Oancea, P., Meltzer, V., 2013. Photo-Fenton process for the degradation of tartrazine(E102) in aqueous medium. J. Taiwan Inst. Chem. Engine 44, 990–994.

Pestana, S., Moreira, M., Olej, B., 2010. Safety of ingestion of yellow tartrazine bydouble-blind placebo controlled challenge in 26 atopic adults. Allergol.Immunopathol. 38 (3), 142–146.

Putt, K.S., Kernick, E.R., Lohse, B.K., et al., 2012. The use of chromophore andfluorophore degradation to quantitate UV dose: FD&C dyes as chemicalidenticators for UV sterilization. J. Microb. Meth. 91, 215–221.

Rafols, C., Barcelo, D., 1997. Determination of mono- and disulphonated azo dyes byliquid chromatography–atmospheric pressure ionization mass spectrometry. J.Chromatogr. A 777, 177–192.

T.C. dos Santos et al. / Food and Chemical Toxicology 68 (2014) 307–315 315

Roychoudhury, A., Giri, A.K., 1989. Effects of certain food dyes on chromosome ofAllium cepa. Mutat. Res. 223, 313–319.

Sasaki, Y.F., Kawagushi, S., Kamaya, A., et al., 2002. The comet assay with 8 mouseorgans: results with 39 currently used food additives. Mutat. Res. 519, 103–119.

Silva, C.P., Marmitt, S., Haetinger, C., et al., 2008. Avaliação da degradação do corantevermelho bordeaux através de processo fotoquímico. Eng. Sanit. Ambiental. 13(1), 73–77.

Souza, K.V., Zamora, P.P., Zawadski, S.F., 2008. Imobilização de Fe(II) em matriz dealginato e sua utilização na degradação de corantes têxteis por processo Fenton.Quim Nova 31 (5), 1145–1149.

Matsushita, T., Matsui, Y., Matsui, Y., 2006. Estimating mutagenic compoundsgenerated during photolysis of fenitrothion—by HPLC fractionation followed bymutagenicity testing and high-resolution GC–MS analysis. Chemosphere 64,144–151.

Trovo, A.G., Villa, R.D., Nogueira, R.F.P., 2005. Utilização de reações foto-fenton naprevenção de contaminações agrícolas. Quím Nova 28 (5), 847–851.

Trovo, A.G., Melo, S.A.S., Nogueira, R.F.P., 2008. Photodegradation of thepharmaceuticals amoxicillin, bezafibrate and paracetamol by the photo-Fenton process-application to sewage treatment plant effluent. Photochem.Photobiol. 198 (2–3), 215–220.

Vanerkova, D., Sakalis, A., Holcapek, M., et al., 2006. Analysis of electrochemicaldegradation products of sulphonated azo dyes using high-performance liquidchromatography/tandem mass spectrometry. Rapid Commun. Mass Spectrom.20, 2807–2815.

Watabe, T., Ozawa, N., Kobayashi, F., 1980. Reduction of sulphonated water-solubleazo dyes by micro-organisms from human faeces. Food Cosmet. Toxicol. 18,349–352.