Grzelak et al.: Journal of AOAC International Vol. 94, No.

5, 2011 1567

MICROBIOLOGICAL METHODS

Development of a Novel Direct BioautographyThin-Layer

Chromatography Test: Optimization of Growth Conditions for
Gram-Negative Bacteria, Escherichia coli
Edyta M. Grzelak
University of Maria CurieSkodowska, Department of Chromatographic Methods, Lublin, Poland
Barbara Majer-Dziedzic
University of Life Sciences, Department of Veterinary Microbiology, Lublin, Poland
Irena M. Choma1
University of Maria CurieSkodowska, Department of Chromatographic Methods, Lublin, Poland
With the aim of developing a TLC-direct
bioautography assay using Escherichia
coli as test bacteria, various parameters
influencing the viability of microorganisms
on TLC plates were examined and checked
for flumequine standards. The optimal
times for preincubation and incubation of
bacterial broth were 20 h at 37C and 2 h at
37C, respectively. The optimal viscosity of
the broth was obtained for 0.05% agarose
solution in Mueller-Hinton broth. Various
incubation times of the seeded TLC plates
were also tested (5 h proved to be optimal).
After incubation, the plates were sprayed with
0.2% aqueous [3-(4,5-dimethyldiazol-2-yl)-2,5
diphenyltetrazolium bromide] (MTT) solution
and incubated for 0.5 h at 37C. The precision of
the method was evaluated by the repeatability
(intraday assay) and intermediate precision
(interday assay). The regression coefficients
were 0.9977 and 0.9968, respectively, for
intraday and interday curves. The calibration
curves show good linearity in the range of
0.0050.50 g (0.550.0 g/mL). The established
LOD of flumequine equaled 0.5 g/mL, i.e.,
5ng flumequine in the spot. The developed
direct bioautography test significantly
enhances the sensitivity of the TLC method.

LC-direct bioautography (TLC-DB) combines

planar chromatography with microbiological
detection (1, 2). In this procedure, a developed
TLC plate is dipped in a suspension of microorganisms
growing in a proper broth. The broth medium adheres to
the silica particles on the TLC plate surface and becomes
a source of nutrients for the tested microorganisms. The
Received September 30, 2010. Accepted by AH January 3, 2011.
1
Corresponding authors e-mail: Irena.Choma@umcs.lublin.pl
DOI: 10.5740/jaoac.10-385

TLC plate is then incubated in a humid atmosphere, and

the microorganisms grow directly on it. Visualization
of antibacterial agents is usually carried out using
tetrazolium salts, which are reduced into intensely colored
formazan by dehydrogenase of living microorganisms.
Cream-white spots against a purple background point
to the presence of antibacterial agents on the TLC plate
surface. These are so-called zones of inhibition and
could be interpreted both qualitatively and quantitatively
(3, 4). The main advantage of TLC-DB is that all steps
of this method, such as separation, preconditioning,
incubation, and visualization, are performed directly on
the TLC plate (5). When only one analyte is subjected
to bioautography the procedure can be simplified by
omitting development of the plate. Then, only spotting
the sample followed by microbial detection is needed (6).
The sensitivity of the method enables one to determine
the minimum inhibitory concentration (MIC) or to
detect residues in contaminated food samples even at
maximum residue levels (MRLs; 6, 7). Moreover, the
method is very useful in searching and examining new
antimicrobial agents (8). Therefore, it is crucial to use
optimized conditions for the test microorganisms in
order to enhance the sensitivity of the procedure (9).
The influence of various factors on direct bioautography
was widely discussed by Botz and coworkers (2), who
took into account preconditioning of a TLC plate,
type of adsorbent, the mobile phase and its additives,
living conditions for test microorganisms, and postchromatographic detection. Nagy et al. (10) studied the
viability of Gram-positive bacteria Bacillus subtilis, as
well as Gram-negative bacteria Escherichia coli (11),
using the bioluminescent luciferin/luciferase adenosine
triphosphate assay.
A ready-to-use Chrom Biodip Antibiotics test for direct
bioautography has been commercially available from
Merck since 2000. Various studies on the analysis of
antibiotic residues in milk were done using this test (57).
The obtained results were satisfactory, but unfortunately,
the company resigned from the test production. Since then,

1568 Grzelak et al.: Journal of AOAC International Vol. 94, No. 5, 2011
Table 1. Comparison of bacterial concentration
(CFU/mL) obtained for the various incubation times
(h) in the first, second, and third tested variant
Mean bacterial concentration, CFU/mL (n = 4)
Time of
incubation, h

Variant I

Variant II

Variant III

2.0 104

3.1 105

6.2 106

3.8 104

1.3 106

7.2 107

5.0 10

2.3 108

4.7 108
8.2 108

4.6 10

5.2 10

2.0 10

6.0 107

8.5 107

we have developed new tests for E. coli and B.subtilis.

The initial experiments with food samples show that the
sensitivity of the method is very high, and the test enables
determination of the residues of antibiotics on the MRL
level or even lower.
This paper describes the process of optimizing growth
conditions for the Gram-negative test bacteria, E. coli,
used in the TLC-DB method. The test was checked for
flumequine standards applied in various amounts on
TLC plates without developing the plates, according
to the method described elsewhere (6, 12). Influence
of various factors on the viability of the bacteria was
studied: the time of preincubation and incubation of
the microorganisms, preconditioning of the TLC plates,
viscosity of the culture broth, and incubation time of the
seeded TLC plates. As a result, a TLC-DB test based on
Gram-negative bacteria, E. coli, was established.
Experimental
Materials and Equipment
Precoated TLC Si60 and Si60 F254 glass-backed plates,
both 10 20 cm, as well as a TLC sprayer, were purchased
from Merck (Darmstadt, Germany). Planimeter was
obtained from PZO (Warsaw, Poland). Reprostar 3 Video
Camera was from CAMAG (Muttenz, Switzerland).
Chemicals
[3-(4,5-dimethyldiazol-2-yl)-2,5 diphenyltetrazolium
bromide] (MTT) used to visualize antimicrobial activity,
as well as Hepes and Triton X-100, laboratory grade, were
obtained from Sigma (St. Louis, MO). Flumequine was
purchased from Polfa (Tarchomin, Poland). Methanol of
analytical grade was from POCH (Gliwice, Poland).
Microorganism and Medium
The test strain E. coli (ATCC 25922) was obtained from
the American Type Culture Collection. Mueller-Hinton

(M-H) broth, M-H agar, and agarose were purchased from

Biocorp (Warsaw, Poland).
Preincubation and Incubation of Culture Broth
Preincubation.A bacterial colony was taken from an
E. coli culture, put into 10 mL M-H broth (pH 7.2 0.2,
adjusted with Hepes), and incubated at 37C for 20 h, the
first hour in a shaker incubator and then on a magnetic
stirrer. After that time, the bacterial mean concentration
9
equaled 4 10 CFU/mL (the value was a mean of 12
values obtained from 12 preincubations).
Incubation.To acquire the bacteria concentration
within 107108 CFU/mL, which is related to the log phase,
three variants of incubation following preincubation were
tested (Table 1).
(1) The bacterial suspension obtained from the
preincubation was diluted with M-H broth (pH 7.2 0.2)
in 1:100 proportions. Then, 0.1 mL diluted suspension
was added to 20 mL M-H broth (pH 7.2 0.2) and placed
on the magnetic stirrer at 37C for 15 h.
(2) As in the first variant, the bacterial suspension
from the preincubation was diluted with M-H broth
(pH 7.2 0.2) in 1:100 proportions. Then, 1.0 mL of
the diluted suspension was added to 20 mL M-H broth
(pH 7.2 0.2) and placed on the magnetic stirrer at 37C
for 15 h.
(3) In the third variant, 20 mLM-H broth (pH 7.2 0.2) was
inoculated with 1 mL bacterial suspension obtained directly
from the preincubation (containing 4 109 CFU/mL)
and placed on the magnetic stirrer at 37C. Again, the
incubation times from 15 h were tested.
Preparation of Antibiotic Solutions
Flumequine stock solution (1 mg/mL) was prepared
by dissolving an appropriate amount of the antibiotic in
methanol. Working standard solutions at concentrations
of 50.0, 10.0, 5.0, 1.0, and 0.5 (g/mL) were prepared
from the stock solution by dilutions in methanol.
Preconditioning of the TLC Plates
The preconditioning was carried out by heating the
TLC plates at 120C for 2 h and cooling them for 1 h in
the desiccator.
Application of the Standards
A 10 L volume of the flumeqine working standard
solutions was applied to the TLC plates using a Linomat5
CAMAG applicator. The samples were spotted in two
tracks at the following concentrations for each track: 50.0,
10.0, 5.0, 1.0, and 0.5 g/mL. The amounts of flumequine
per spot were 0.5, 0.1, 0.05, 0.01, and 0.005 g. The plates
were then subjected to direct bioautography (without

Grzelak et al.: Journal of AOAC International Vol. 94, No. 5, 2011 1569

Figure 1. TLC-DB of flumequine standard solutions on Si60F254 plateculture broth with addition of
(A) 0.1% agarose; the amounts of flumequine/spot are as follows: left track: 1.0, 10.0, 100.0 (g); right track:
100.0, 10.0, and 1.0 (g; going down); the applied volume: 10 L and (B) 0.05% agarose; the applied volumes
of flumequine standard at 0.01 g/L concentration are left track, 0.5, 1.0, 5.0, 10.0, 50.0, 100.0 (L); right track,
100.0, 50.0, 10.0, 5.0, 1.0, and 0.5 (L; going down).

developing with a mobile phase; 6, 12). Each TLC plate

was also monitored under UV light at 254 nm.
Bioautography
The TLC plates spotted with flumequine standard
solutions were manually immersed (for about 10 s) in the
bacterial suspension obtained as described above. Then,
they were placed in a moistened plastic box lined with
wetted paper and incubated for various times, i.e., 3, 4, 5,
and 15 h at 37C.
Visualization of the TLC Plate
After incubation, the TLC plates were sprayed with
0.2% MTT aqueous solution and incubated again at 37C
for 0.5 h. Dehydrogenase of the bacteria reduces MTT
into purple formazan. One drop of Triton X-100/10 mL
aqueous MTT solution was found to enhance the intensity

of the color. In places where flumequine standards were

applied, cream-white inhibition zones were observed.
Results and Discussion
E. coli (ATCC 25922) was chosen as a representative of
Gram-negative bacteria. The TLC-DB test was established for
flumequine standards, which were applied to TLC plates in
various amounts. The method of analyte quantitation without
developing the TLC plate was chosen, based on research by
Dhenasar (12), who used such a method for densitometric
detection. The usefulness of the Dhenasar method in a case of
bioautographic detection was proven in our laboratory (6). In
order to optimize the growth conditions of the microorganisms
on the TLC plate, many factors have been studied.
Incubation of the Culture Broth
Incubation.The bacteria growing on the silica surface

1570 Grzelak et al.: Journal of AOAC International Vol. 94, No. 5, 2011
Table 2. Mean areas of growth inhibition zones
for flumequine standard solutions from 1 day of
measurements (intraday)
Concentration
Dose of
of flumequine, flumequine/spot,
g/mL
g
0.5

0.005

Mean area SD

RSD,
%

0.158 0.015 (n = 8)

9.74

1.0

0.01

0.404 0.011 (n = 8)

2.91

5.0

0.05

1.216 0.019 (n = 4)

1.58

10.0

0.1

1.616 0.057 (n = 4)

3.57

50.0

0.5

2.4410.185 (n = 4)

7.59

of the TLC plate should be in a logarithmic growth

phase, when the culture is the most homogeneous and
gives numerous multiplications (13). As a consequence,
well-colored plates with clear inhibition zones are
obtained as a result of bioautography. To acquire the
bacteria concentration within 107108 CFU/mL, which
is related to the log phase, three variants of incubation
following preincubation were tested as described in the
Experimental section. When the third variant was applied,
the mean concentration of E. coli on the TLC plate was
sufficient after only 2 h of incubation. Therefore, in order
to shorten the time of the experiment, preincubation and
the third variant of incubation were chosen for further
experiments.
Viscosity of the Culture Broth
Viscosity of the culture broth seems to be one of
the most important factors influencing bioautographic
detection. When the broth is too fluid it can run of from
the TLC plate surface, resulting in unclear inhibition
zones. On the other hand, too high viscosity makes it
difficult to cover the TLC plates with a plain layer of
bacterial suspension. For this reason, various gelatinizing
agents are usually added to obtain optimal viscosity of
the culture broth, e.g., bacterial polysaccharide, casein, or
starch (14). In preliminary experiments the broth without
addition of gelatinizing agent was used. However, it was
too fluid and did not adhere well to the silica gel surface,
which resulted in poor viability of the bacteria on the
TLC plates. Hence, 0.1% agar solution in M-H broth was
tested, but it also gave insufficient effects. The best results
were obtained with addition of agarose as a gelatinizing
agent. Two concentrations, 0.1 and 0.05%, were tested.
When 0.1% agarose solution in M-H broth was used,
the viscosity of the culture bullion was too high, giving
smudges on the plate surface. The optimal results were
obtained with 0.05% agarose solution; the TLC plates
were well colored by MTT and smooth (Figure 1).

Figure 2. TLC-DB calibration curves, along with the

reported SDs for flumequine standards: (A) intraday
assay and (B) interday assay.

Incubation Time of TLC Plates

Too-short, as well as too-long, incubation time leads
to an imperfect formation of inhibition zones. When the
incubation time is too short and bacteria concentration too
low, zones of inhibition are not sharp enough. Conversely,
when the time is too long, zones can be overgrown by
bacteria. Therefore, it is crucial to find an optimal time of
incubation. In our experiments the incubation time of 5 h
was found to be optimal. The TLC plates were smoothly
covered with bacteria and well colored after visualization.
Other Factors Influencing Bacterial Growth
The TLC silica gel plates, both containing and not
containing fluorescent indicators, were used in the
experiments. Bacterial growth on TLC plates depends on
the chemical character of the sorbent surface; therefore, the
influence of additional substance, i.e., an indicator in silica
gel, should be tested (2). It seemed that, independently
from the presence of the indicator, the results of TLC-DB
tests were the same. Hence, further experiments were
performed using more universal TLC silica gel plates
with F254 factor. Additionally, we checked to determine
if preconditioning of the TLC plate influences bacterial
growth. According to Botz et al. (2), better effects were
obtained when TLC plates were preconditioned, because
then the layers became more resistant to be detached
when soaked. Our experiments, performed on both
preconditioned and not-preconditioned TLC plates, did
not support this observation. Therefore, the plates were

Grzelak et al.: Journal of AOAC International Vol. 94, No. 5, 2011 1571

the same experimental conditions (Table 2). The spotted

TLC plates were subjected to direct bioautography with
E. coli. The TLC plates after DB were scanned with a
scanner, and the areas of inhibition zones were measured
with a planimeter. As seen from Table 2, the highest RSD
values were obtained for the outermost concentration
values, both the lowest and the highest. The problem
with very low concentrations is that they give very small
inhibition spots, which makes measuring them difficult.
On the other hand, too high concentrations of antibiotics
per spot sometimes leads to the deformation of inhibition
zones. These outermost concentrations estimate the range
of the method from 5 to 500 ng/spot.
TLC-DB measurements were repeated three times
during a 2-month period; the interval between each
day of measurements was about 1 month. The interday
and intraday precisions were evaluated using one-way
analysis of variance. Except for the lowest dose, the F
statistic does not allow rejection (at significance level of
5%) of the null hypothesis, that the areas of the analyzed
inhibition zones belonging to 3 days (groups) are drawn
from the same population.
Linearity

Figure 3. TLC-DB of flumequine standard solutions

on Si60F254 plate. The applied volume: 10 L. The
flumequine concentrations applied are: left track
(going down) 0.5, 1.0, 5.0, 10.0, and 50.0 (g/mL);
right track, 50.0, 10.0, 5.0, 1.0, and 0.5 (g/mL). The
amounts of flumequine/spot are: left track (going
down) 0.005, 0.01, 0.05, 0.1, 0.5, and 1.0 (g); right
track, 1.0, 0.5, 0.1, 0.05, 0.01, and 0.005 (g).

not preconditioned, which simplified and shortened the

procedure.
Validation of the Method
Precision
The repeatability of the method (intraday precision)
was evaluated by analyzing the area of inhibition zones
corresponding to flumequine standards, which were
applied in 10 L volumes at five different concentrations
on two different plates during the same day and under

The linearity of the calibration curves was determined

for intraday and interday precision experiments (Figure2).
The curves were constructed by plotting mean areas of
inhibition zones (cm2) versus the logarithm of doses (g)
of the antibiotic. The calibration curve for intraday assay
was obtained on the basis of the data from Table 2. The
calibration curve for interday assay was obtained for
the means from all inhibition zone areas at a given dose
obtained as a result of TLC-DB performed during all
measurement days. The shape of the curves is typical for
bioautography, i.e., there is a linear dependence between
the logarithm of applied doses and the mean areas of
inhibition zones (3, 4). The calibration curves show good
linearity in the range of 0.0050.50 g (0.550.0 g/mL).
The regression coefficients were 0.9977 and 0.9968,
respectively, for intraday and interday plots.
LOD
The LOD was determined as the lowest concentration of
the antibacterial agent that caused recognizable bacterial
growth inhibition on each of the TLC plates subjected to
bioautography (Figure 3). In direct bioautography, the
LOD is equivalent to the LOQ. If it is possible to see the
inhibition zone, it is possible to measure its area, as well.
For this method, LOD established for flumequine equaled
0.5g/mL, i.e., 5 ng flumequine in the spot. This value
corresponds to the MIC of flumequine toward E. coli
bacteria (15). While flumequine MRLs levels range from
50 g/kg in milk to 1500 g/kg in bovine and porcine
kidney (16), the obtained results pointed to possible

1572 Grzelak et al.: Journal of AOAC International Vol. 94, No. 5, 2011

application of the developed TLC-DB method for the

analysis of flumequine residues in food, assuming higher
applied volumes of contaminated samples, e.g., 5070L
and/or sample pretreatment, e.g., SPE or matrix solidphase dispersion shown in our previous papers (4, 7, 17).
Conclusions
The optimal growth conditions for E. coli as test bacteria
for TLC-DB assay were found to be as follows: 20h of
preincubation at 37C to obtain bacterial concentration
at 4 109 CFU/mL; 2 h of incubation at 37C to obtain
bacterial concentration at 7.2 107 CFU/mL; and 5 h
of incubation of the seeded TLC plates at 37C, then
visualization by spraying the TLC plates with 0.2%
aqueous MTT solution, followed by incubation of the TLC
plates for 0.5 h at 37C.
The TLC-DB method was tested for flumequine
standards. The obtained results show good linearity,
precision, and excellent sensitivity. Moreover, the
experiments indicate that it should be possible to determine
the residues of flumequine on its MRL level. In the future,
we will present studies of optimal growth conditions for
B. subtilis as well as to apply the method in analyzing
various classes of antibiotics in contaminated food.
References
(1)Betina, V. (1973) J. Chromatogr. 78, 4151.
doi:10.1016/S0021-9673(01)99035-1
(2)Botz, L., Nagy, S., & Kocsis, B. (2001) in Planar
Chromatography, Sz. Nyiredy (Ed.), Springer,
Budapest, Hungary, pp 489516
(3)Kdr-Pauncz, J., & Harsnyi, I. (1980) J. Chromatogr.
195, 251256. doi:10.1016/S0021-9673(00)96816-X
(4)Choma, I.M., Choma, A., Komaniecka, I., Pilorz, K.,

& Staszczuk, K. (2004) J. Liq. Chromatogr. Relat.

Technol. 27, 20712085. doi:10.1081/JLC-120039419
(5)Choma, I.M. (2005) LC-GC Eur. 18, 482488
(6)Choma, I.M., Kowalski, C., Lodkowski, R.,
Burmanczuk, A., & Komaniecka, I. (2008) J.
Liq. Chromatogr. Relat. Technol. 31, 19031912.
doi:10.1080/10826070802194815
(7)Choma, I. (2006) J. Liq. Chromatogr. Relat. Technol.
29, 20832093. doi:10.1080/10826070600759942
(8)Hostettmann, K., Terreaux, Ch., Marston, A., & Potterat,
O. (1997) J. Planar Chromatogr.-Mod. TLC 10,
251257
(9)Rios, J.L., Recio, M.C., & Villar, A. (1988) J.
Ethnopharmacol. 23, 127149. doi:10.1016/03788741(88)90001-3
(10)Nagy, S., Kocsis, B., Koszegi, T., & Botz, L. (2002) J.
Planar Chromatogr.-Mod. TLC 15, 132137
(11)Nagy, S., Koszegi, T., Botz, L., & Kocsis, B. (2003) J.
Planar Chromatogr.-Mod. TLC 16, 121126
(12)Dhanesar, S.C. (1999) J. Planar Chromatogr.-Mod. TLC
12, 114119
(13)Cole, M.D. (1994) Biochem. Systematics Ecol. 22,
837856. doi:10.1016/0305-1978(94)90089-2
(14)Eymann, R., & Hauck, H.E. (2000) in Proc. Intern.
Symp. Planar Separations-Planar Chromatogr., Sz.
Nyiredy (Ed.), Budapest, Hungary, pp 6775
(15)WHO Food Additives Series 39 (1992) Forty-Eighth
Meeting of the Joint FAO/WHO Expert Committee on
Food Additives, Geneva, Switzerland
(16)EMEA/MRL/823/02-FINAL (2002) Committee
for Veterinary Medical ProductsFlumequine,
Summary Report, http://www.ema.europa.eu/pdfs/vet/
mrls/082302en.pdf
(17)Choma, I.M. (2006) J. Planar Chromatogr.-Mod. TLC
19, 104108

Copyright of Journal of AOAC International is the property of AOAC International and its content may not be
copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written
permission. However, users may print, download, or email articles for individual use.