Reduction of nitroblue tetrazolium to formazan by folic acid

Article abstract of DOI:10.2478/s11696-013-0495-5

The reduction of nitroblue tetrazolium (NBT) to formazan by folic acid, N-(4-aminobenzoyl) glutamic acid, and other amino acids was studied in this paper. The reduction involves only one of the two tetrazolium rings of NBT. The reaction is considerably mo

Full text of DOI:10.2478/s11696-013-0495-5

Chemical Papers 68 (5) 662–667 (2014)
DOI: 10.2478/s11696-013-0495-5
ORIGINAL PAPER
Reduction of nitroblue tetrazolium to formazan by folic acid
b
a
a
aCesare Achilli, Stefania Grandi, Annarita Ciana, Cesare Balduini,
aGiampaolo Minetti*
aLaboratories of Biochemistry, Department of Biology and Biotechnology “Lazzaro Spallanzani”, University of Pavia,
via Bassi 21, 27100 Pavia, Italy
bDepartment of Chemistry, University of Pavia, viale Taramelli 12, 27100 Pavia, Italy
Received 7 March 2013; Revised 25 July 2013; Accepted 14 August 2013
The reduction of nitroblue tetrazolium (NBT) to formazan by folic acid, N-(4-aminobenzoyl)
glutamic acid, and other amino acids was studied in this paper. The reduction involves only one
of the two tetrazolium rings of NBT. The reaction is considerably more rapid with folic acid and
N-(4-aminobenzoyl) glutamic acid than with the other amino acids under study. The electron donor
moiety appears to be the carboxylic acid in the alpha position. N-ethyl-Nꢀ(3-dimethylaminopropyl)
carbodiimide notably increases the rate of the reaction and promotes the reduction of both tetra-
zolium rings.
c
ꢀ 2013 Institute of Chemistry, Slovak Academy of Sciences
Keywords: nitroblue tetrazolium, formazan, carbodiimide, folic acid, amino acids
Introduction
Many tetrazolium salts, such as 3-(4,5-dimethyl-
thiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide
(MTT), 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxy-
methoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium
(MTS), 4-(3-(4-iodophenyl)-2-(4-nitrophenyl)-2H-5-
tetrazolium)-1,3-benzene disulphonate (WST-1), and
2,3-bis-(2-methoxy-4-nitro-5-suphophenyl)-2H-tetra-
zolium-5-carboxanilide (XTT), are used to investi-
gate cell proliferation and viability in colorimetric
assays based on reductive cleavage of the tetra-
zolium ring by metabolically active cells, resulting
in the production of water-soluble or water-insoluble
dark-blue formazans (Berridge et al., 2005). Further-
more, 2,2ꢀ-bis(4-nitrophenyl)-5,5ꢀ-diphenyl-3,3ꢀ-(3,3ꢀ-
dimethoxy-4,4’-diphenylene) ditetrazolium (nitroblue
tetrazolium, NBT) can be reduced to water-insoluble
monoformazan or diformazan by superoxide anions
(Fig. 1); for this reason, it is widely used in colorimet-
Fig. 1. Structure of NBT and reduction of tetrazolium ring to
ric assays designed to measure the superoxide ions re-
leased during the activation of phagocytic cells (gran-
formazan.
ulocytes, macrophages) (Berridge et al., 2005).
*Corresponding author, e-mail: minetti@unipv.it

C. Achilli et al./Chemical Papers 68 (5) 662–667 (2014)
663
Fig. 2. Structures of molecules tested as NBT-reducing agents: folic acid (FA), N-(4-aminoenzoyl) glutamic acid (4-ABGA), glu-
tamic acid (Glu), glutamine (Gln), glycine (Gly), glycine methyl ester (Gly methyl ester), N,N-dimethyl glycine (N,N-
dimethyl Gly).
Any extracellular molecules that act as electron-
donors can potentially reduce the tetrazolium ring and
significantly increase the signal of the assay, yield-
ing a false positive result. Many components of cell
culture media, such as antioxidants, vitamins, and
serum proteins (albumin), can react with MTT and
XTT (Natarajan et al., 2000; Chakrabarti et al., 2001;
Bernhard et al., 2003; Funk et al., 2007). Numerous
compounds, such as ketones, α-hydroxyketones, thi-
ols, quinines, and polyhydric phenols, can also reduce
NBT to formazan (Manni & Sinsheimer, 1961; Sin-
sheimer & Salim, 1965; Graham et al., 1975; Oteiza et
al., 1977).
jority of commonly used organic solvents. Mixtures of
NBT and FA, each at a final concentration of 0.5 mM,
were prepared using ethanol/water (ϕr = 1 : 1), and
incubated at 25◦C in the dark. After 72 h, 0.5 vol-
ume (with respect to the sample volume) of abso-
lute ethanol was added, the samples were sonicated
to completely solubilise the formazan and then anal-
ysed by TLC on silica gel 60 plates (Merck, Germany)
by elution with 1-butanol, acetic acid (100 %), and
water, (ϕr = 78 : 5 : 17) as described in the litera-
ture (Altman, 1974). Fig. 3a shows a typical TLC elu-
tion of a sample, where the purple spot corresponds
to monoformazan, as indicated by comparison with a
standard mixture of monoformazan and diformazan
loaded in the adjacent lane. To further confirm its
identity, the spot was scratched from the plate, ex-
tracted with absolute ethanol and analysed by elec-
trospray ionisation-mass spectrometry (ESI-MS, LCQ
FLEET ion trap mass spectrometer, Thermo Fisher
Scientific, USA). The positive ion mass spectra indi-
cated a molecular mass of 748 Da, consistent with the
monoformazan structure (Fig. 3b).
This study found that folic acid (FA; Fig. 2) can
reduce the NBT to formazan.
Experimental
The main reagents used were purchased from
Sigma–Aldrich (Italy). 1 mM stock solution of NBT
(dichloride salt) in absolute ethanol and 20 mM aque-
ous stock solution of FA (disodium salt) were pre-
pared. FA could only be brought into solution by sal-
ification with a stoichiometric quantity of NaOH, be-
cause the free acid is not soluble in water or in the ma-
To understand which of the many functional moi-
eties of FA was involved in this reaction, and to charac-
terise the underlying mechanism of the reaction, sev-

664
C. Achilli et al./Chemical Papers 68 (5) 662–667 (2014)
Fig. 3. TLC for detection of formazans (a). A mixture of monoformazan (MF) and diformazan (DF) as reference standard and
the sample prepared by incubation of NBT with FA were loaded on the plate. The positive ion mass spectra (ESI-MS) of
the monoformazan purified by TLC (b).
eral other molecules with structures and functional
groups similar to those of FA were tested for their
ability to reduce NBT: 4-ABGA, Glu, Gln, Gly, Gly
methyl ester, N,N-dimethyl Gly (Fig. 2).
The reaction mixtures were prepared as follows.
For the solubilisation of FA, 20 mM stock solutions of
the other derivatives, irrespective of their hydrosolu-
bility, were prepared in water as sodium salts, with a
stoichiometric quantity of NaOH. The final concen-
tration of NBT and of each compound tested was
0.5 mM. After incubation in the dark at 25◦C for
144 h, 0.5 volume of absolute ethanol was added to
the samples, which were then sonicated. 100 L of
the sample was then used for quantitative determi-
nation of the formazan content by reading the ab-
sorbance at 550 nm, using a 96-well microtitre plate
reader (Tecan, Switzerland). Fig. 4 shows the results
obtained. The sample containing Gly methyl ester did
not show any appreciable absorbance reading, indicat-
Fig. 4. Reduction of NBT by different compounds; in the y-
ing the absence of formazan. Conversely, all the other
axis, absorbance at 550 nm is reported: cont – NBT
molecules exhibited a significant production of for-
mazan, albeit with different intensities: 4-ABGA ex-
hibited a reactivity similar to that of FA, which was
approximately 60 % higher than those exhibited by
Glu, Gln, Gly, and N,N-dimethyl Gly. For these other
compounds, TLC analysis also confirmed formation of
the monoformazan only (not shown).
only. Values are the mean
SD of three experiments.
4-ABGA exhibited a reducing capacity only slightly
lower than that of FA, despite the absence of the
pterin moiety in its structure. Glu, an amino acid in
which the 4-aminobenzoyl group is also absent, had
a reactivity considerably diminished but still present.
Glu presents three possible reactive sites: α-amino, α-
carboxyl, and γ-carboxyl groups. Gln, a derivative of
Glu in which the γ-carboxylic acid is blocked by an
amide bond, displayed a reactivity level superimpos-
Results and discussion
Taking the structural differences of the various
compounds into consideration, it is worth noting that

C. Achilli et al./Chemical Papers 68 (5) 662–667 (2014)
665
Fig. 5. Reduction of NBT by different compounds in the presence of EDC (a). In the y-axis, the absorbance at 550 nm is reported;
cont – mix of NBT and EDC. The values are the mean
of reaction with carboxylic acid are represented (b).
SD of three experiments. The structure of EDC and the scheme
able to that of Glu. On this basis, it was hypothesised
that the γ-carboxyl group is not involved in the reac-
tion.
incubation in the dark at 25◦C for 144 h were sup-
plemented with 0.5 volume of absolute ethanol and
homogenised by sonication prior to reading the ab-
sorbance at 550 nm, by a 96-well microtitre plate
reader (100 L aliquots). Fig. 5 shows the results ob-
tained. EDC considerably increased the rate of the re-
action for most compounds, except for N,N-dimethyl
Gly, where the reactivity was unchanged, and for
Gly methyl ester, which remained unreactive. The in-
crease in formazan production was fifteen-fold for the
FA sample, nine-fold for 4-ABGA, and approximately
two-fold for Glu, Gln, and Gly. TLC analysis of FA
and 4-ABGA samples revealed the production of both
monoformazan and diformazan in approximately com-
parable amounts (not shown). Furthermore, a moder-
ate increment in the rate of NBT reduction by FA
was observed in the presence of the coupling agent
N-hydroxysuccinimide (not shown).
FA and the other amino acids tested in this study
have different characteristics concerning their ionis-
able groups (amino and carboxylic) that could influ-
ence the pH value of the mixture and, potentially,
modify the rate of the reaction. For this reason, the
initial pH of the different samples prepared with EDC
was measured (pH-meter: InLab Expert Pro, Mettler-
Toledo, USA). The values ranged from pH 6 to pH
9 (FA, pH 7.3; 4-ABGA, pH 8.9; Glu, pH 8.9; Gln,
pH 8.5; Gly, pH 8.8; Gly methyl ester, pH 6.0; N,N-
In addition, Gly, the simplest amino acid, en-
dowed with α-amino and α-carboxyl groups only, had
a reactivity superimposable to that of Glu and Gln.
By blocking the α-amino moiety of Gly by methy-
lation (N,N-dimethyl Gly), the formazan production
remained stable, but after blocking the α-carboxylic
acid by methylation (Gly methyl ester), the reactivity
was suppressed, becoming similar to the control (cont;
Fig. 4). These results strongly suggest that the moi-
ety implicated in the reduction of NBT by FA is the
α-carboxylic acid.
Also tested was the reduction of NBT by FA and
by the other compounds in the presence of N-ethyl-
Nꢀ-(3-dimethylaminopropyl) carbodiimide (EDC), a
coupling agent that reacts selectively with carboxylic
acids forming an O—C bond. The literature reveals
that both carboxyl groups of FA can react with car-
bodiimides and other coupling agents, although the
reaction favours γ-carboxyl group (Wang et al., 1996;
Gabizon et al., 1999; Wei et al., 2005; Viola-Villegas
et al., 2008). For the experiments, 5 mM of stock solu-
tion of EDC hydrochloride was prepared in water and
used immediately at a final concentration of 0.5 mM.
All the other reagents were also used at a final con-
centration of 0.5 mM. The reaction mixtures, after

666
C. Achilli et al./Chemical Papers 68 (5) 662–667 (2014)
This would not favour the reaction of the α-carboxyl
group with NBT as much as hypothesised. However,
the reaction may still be promoted, thanks to elec-
tronic effects induced by the 4-aminobenzoyl moiety
present in FA and 4-ABGA. In effect, the resonance
between the amino group in the para position of the
aromatic ring and the carbonyl group of the amide
does not involve the nitrogen atom of the amide. This
could confer on it an electron-donor character suffi-
ciently pronounced to promote the reaction of the α-
carboxyl group, contrary to what would occur with
other aromatic or aliphatic amides. The α-nitrogen
could equally act as an electron-donor sufficiently to
promote the formation of a Schiff base during the re-
arrangement of FA. These electronic effects may also
partially explain the consistent differences in kinetics
observed between the simple amino acids and the com-
pound in which the α-amino group is engaged in an
amide bond.
Fig. 6. Mechanism proposed for reduction of NBT by FA and
related compounds.
The increased rate of the reaction in the presence
of EDC could be due to the coupling of carbodiimide
with the α-carboxyl group, which may promote the
nucleophilic attack on the tetrazolium ring, further
supporting the evidence concerning the involvement
of a carboxyl group in the mechanism of the reaction.
It cannot be excluded that EDC also reacts with the
γ-carboxyl group, even at a higher rate than with the
α-carboxyl group, as already known from the litera-
ture; however, this event could not be detected in the
experimental system in this study, because it would
not be productive in terms of NBT reduction.
The presumed mechanism is also partially sup-
ported by ESI-MS analysis of the reaction mixture
prepared with FA in the presence of EDC, in which a
molecule with a molecular mass compatible with the
carboxylated form of NBT is clearly detectable, as a
possible intermediate product of the reaction (Fig. S2
in the Supplementary materials). Although the cor-
responding Schiff base derived from the rearrange-
ment of FA was not detected, it should be noted that
FA was also not detected by ESI-MS analysis, even
though it was still present in significant amounts (at
dimethyl Gly, pH 8.7). To assay the potential effect
of the pH on the reaction, a set of samples containing
NBT, FA, and EDC (each at the final concentration of
0.5 mM) at different pH values (pH 6, pH 7, pH 8, and
pH 9) was prepared, using sodium phosphate as buffer
(5 mM, final concentration). The samples were incu-
bated for 72 h and processed for the measurement of
formazan production as detailed above. The amounts
of formazan produced did not differ substantially in
the reactions carried out at pH 7, pH 8, and pH 9.
Conversely, the sample at pH 6 gave a yield of for-
mazan 25 % higher than the other buffered samples
(Fig. S1 in Supplementary materials). For this reason,
the different reactivity observed in unbuffered samples
containing FA and the other amino acids may be ex-
cluded as being due to the different pH. Furthermore,
the lack of reactivity in the sample containing Gly
methyl ester which, by contrast, has a pH value that
could favour the reaction, corroborates involvement of
the carboxylic group in the reaction mechanism.
It is hypothesised that the reaction occurs as nu-
cleophilic attack by the carboxylic oxygen on the ni-
trogen in position four of the tetrazolium ring, the
electron acceptor proposed for the reduction of tetra-
zolium by hydroxy ketones (Oteiza et al., 1977). The
following step could be the break of the O—N bond
with the donation of the electrons to the tetrazolium
nitrogen, followed by the break of the bond between
the nitrogen atoms in positions two and three of the
ring, with the formation of formazan. The probable
rearrangement of FA (and of the other compounds)
could be the formation of a Schiff base between the α-
carbon and nitrogen, followed by the release of carbon
dioxide, as represented in Fig. 6.
least 125
M) at the end of the incubation, as de-
termined by RP-HPLC (data not shown). This phe-
nomenon could be due to the high propensity of FA
and of its derivative(s), to fragment during ionisation,
generating a complicated and often non-reproducible
ionisation pattern, and to form neutral salts or zwit-
terion species which suppress the MS signal.
The results given in this paper indicate preliminary
data concerning a new method for the reduction of
NBT to formazans. It cannot be excluded that other
tetrazolium salts may be reduced in the same manner.
The growth media in common use for the in vitro
culture of eukaryotic cells have a total concentration
of free amino acids in the millimolar range. On the ba-
sis of the results given in this paper, these relatively
high levels of amino acids in a cellular sample could
FA and 4-ABGA have the α-amino group engaged
in an amide bond, a condition that generally makes
the nitrogen atom electron-poor due to its involvement
in the resonance with the carbonyl group of the amide.

C. Achilli et al./Chemical Papers 68 (5) 662–667 (2014)
667
interfere with metabolic assays based on tetrazolium
References
reduction, affording false positive responses. FA is also
present in cell culture media although at nanomolar
levels, too low to significantly interfere with the assay.
However, in recent years, many FA-derivatives have
been designed and in vitro tested for the selective de-
livery of therapeutic and diagnostic molecules into ma-
lignant cells that express FA-receptors at higher lev-
els than normal cells (Low & Kularante, 2009). Fur-
thermore, FA-derived molecules such as amethopterin
(methotrexate) and aminopterin are proposed for the
treatment of cancer, auto-immune diseases, and infec-
tions due to their inhibitory effects on FA metabolism
(Takimoto, 1996). These FA-based agents are prelim-
inarily tested with cells grown in vitro to assess their
efficacy, safety, and biological properties, at millimolar
concentrations (Walker et al., 2000), thus potentially
interfering with tetrazolium-based colorimetric assays.
FA and its metabolites are important biological
components of the cell, essential for several processes
such as DNA and amino acid synthesis, and their de-
pletion induces many cellular disorders and diseases.
Hence, the reactivity with FA could be the cause of
the toxicity of tetrazolium compounds found in several
studies (Nineham, 1955). This phenomenon requires
further investigation, to clarify the mechanism of the
interaction of tetrazolium compounds with biological
systems.
Altman, F. P. (1974). Studies on the reduction of tetrazolium
salts. Histochemie, 38, 155–171. DOI: 10.1007/bf00499663.
Bernhard, D., Schwaiger, W., Crazzolara, R., Tinhofer, I.,
Kofler, R., & Csordas, A. (2003). Enhanced MTT-reducing
activity under growth inhibition by resveratrol in CEM-C7H2
lymphocytic leukemia cells. Cancer Letters, 195, 193–199.
DOI: 10.1016/s0304-3835(03)00157-5.
Berridge, M. V., Herst, P. M., & Tan, A. S. (2005). Tetrazolium
dyes as tools in cell biology: New insights into their cellular
reduction. Biotechnology Annual Review, 11, 127–152. DOI:
10.1016/s1387-2656(05)11004-7.
Chakrabarti, R., Kundu, S., Kumar, S., & Chakrabarti, R.
(2001). Vitamin A as an enzyme that catalyzes the re-
duction of MTT to formazan by vitamin C. Journal of
Cellular Biochemistry, 80, 133–138. DOI: 10.1002/1097-
4644(20010101)80:1<133::aid-jcb120>3.0.co;2-t.
Funk, D., Schrenk, H. H., & Frei, E. (2007). Serum albumin
leads to false-positive results in the XTT and the MTT assay.
BioTechniques, 43, 178–186. DOI: 10.2144/000112528.
Gabizon, A., Horowitz, A. T., Goren, D., Tzemach, D.,
Mandelbaum-Shavit, F., Qazen, M. M., & Zalipsky, S. (1999).
Targeting folate receptor with folate linked to extremities of
poly(ethylene glycol)-grafted liposomes: In vitro studies. Bio-
conjugate Chemistry, 10, 289–298. DOI: 10.1021/bc9801124.
Graham, R. E., Biehl, E. R., Kenner, C. T., Luttrell, G. H.,
& Middleton, D. L. (1975). Reduction of blue tetrazolium
by corticosteroids. Journal of Pharmaceutical Sciences, 64,
226–230. DOI: 10.1002/jps.2600640207.
Low, P. S., & Kularatne, S. A. (2009). Folate-targeted therapeu-
tic and imaging agents for cancer. Current Opinion in Chem-
ical Biology, 13, 256–262. DOI: 10.1016/j.cbpa.2009.03.022.
Manni, P. E., & Sinsheimer, J. E. (1961). Reactivity of some
hydroxyketones with blue tetrazolium. Analytical Chemistry,
33, 1900–1903. DOI: 10.1021/ac50154a034.
Natarajan, M., Mohan, S., Martinez, B. R., Meltz, M. L., &
Herman, T. S. (2000). Antioxidant compounds interfere with
the 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bro-
mide. Cancer Detection and Prevention, 24, 405–414.
Nineham, A. W. (1955). The chemistry of formazans and
tetrazolium salts. Chemical Reviews, 55, 355–483. DOI:
10.1021/cr50002a004.
Oteiza, R. M., Wooten, R. S., Kenner, C. T., Graham, R. E., &
Biehl, E. R. (1977). Kinetics and mechanism of blue tetra-
zolium reaction with corticosteroids. Journal of Pharmaceu-
tical Sciences, 66, 1385–1388. DOI: 10.1002/jps.2600661008.
Sinsheimer, J., & Salim, E. F. (1965). Reactivity of blue tetra-
zolium with nonketol compounds. Analytical Chemistry, 37,
566–569. DOI: 10.1021/ac60223a031.
Conclusions
The reduction of NBT to formazan by FA and
other amino acids is described in this article. The
reduction, which involves only one of the two tetra-
zolium rings of NBT, is considerably more rapid with
folic acid and N-(4-aminobenzoyl) glutamic acid than
with the other amino acids examined. The electron-
donor moiety appears to be the carboxylic acid in
the alpha position. The relatively high concentrations
(millimolar) of amino acids present in cell samples and
in vitro biological systems such as cell culture media
could produce false positive results when metabolic
assays based on tetrazolium reduction are used. Al-
though FA is present in biological fluids at sub-
micromolar concentrations, many FA-derivatives are
used in vitro as therapeutic and diagnostic molecules.
Accordingly, attention should be paid to the poten-
tial reduction of tetrazolium salts when high doses of
amino acids, FA and its derivatives are present in cel-
lular samples.
Takimoto, C. H. (1996). New antifolates: Pharmacology and
clinical applications. Oncologist, 1, 68–81.
Viola-Villegas, N., Vortherms, A., & Doyle, R. P. (2008). Tar-
geting gallium to cancer cells through the folate receptor.
Drug Target Insights, 3, 13–25.
Walker, T. M., Rhodes, P. C., & Westmoreland, C. (2000). The
differential cytotoxicity of methotrexate in rat hepatocyte
monolayer and spheroid cultures. Toxicology in Vitro, 14,
475–485.
DOI: 10.1016/s0887-2333(00)00036-9.
Wang, S., Lee, R. J., Mathias, C. J., Green, M. A., & Low,
P. S. (1996). Synthesis, purification, and tumor cell uptake
of 67Ga-deferoxamine-folate, a potential radiopharmaceuti-
cal for tumor imaging. Bioconjugate Chemistry, 7, 56–62.
DOI: 10.1021/bc9500709.
Wei, W. H., Fountain, M., Magda, D., Wang, Z., Lecane, P.,
Mesfin, M., Miles, D., & Sessler, J. L. (2005). Gadolin-
ium texaphyrin-methotrexate conjugates. Towards improved
cancer chemotherapeutic agents. Organic & Biomolecular
Chemistry, 3, 3290–3296. DOI: 10.1039/b503664j.
Acknowledgements. This work was funded by the CARIPLO
Foundation, Project no. 2009-2440 ”Development and safety
assessment of nanostructured compounds applicable to boron
neutron capture therapy”. We wish to thank Dr Federica
Corana (Centro Grandi Strumenti, University of Pavia) for
the mass spectrometry analysis.