Water Research 36 (2002) 3398–3403

Research note
Sulphadimethoxine and Azolla filiculoides Lam.: a model for
drug remediation
Cinzia Fornia,*, Antonella Casconeb, Maurizio Fioric, Luciana Migliorea
a
Dipartimento di Biologia, Universita" degli Studi di Roma ‘‘Tor Vergata’’, Via della Ricerca Scientifica I-00133 Roma, Italy
b
ICRAM, Via Casalotti 300, I-00162 Roma, Italy
c
ISS, Lab. Medicina Veterinaria, Viale Regina Elena 266 I-00169 Roma, Italy
Received 8 June 2000; accepted 18 December 2001

Abstract

Plants can be an interesting tool for in situ remediation of drug contaminated waters. In a laboratory model Azolla
filiculoides Lam., an aquatic fern known to absorb pollutants, has been exposed to an environmental persistent
antibiotic commonly used in intensive farming, sulphadimethoxine (S), to test its bioremediation capability. In a 5 week
experiment, plants were cultivated outdoor at four drug concentrations (50, 150, 300 and 450 mg l
1) in N-free mineral
medium. Drug affects growth rate (as biomass yield per week), N2-fixation, heterocyst frequency, but plants are able to
survive. Notwithstanding, at all concentrations tested drug was actively removed from the medium and the
accumulation in the biomass is in order of magnitude up to mg g
1 plant dry weight (1000 ppm). Drug uptake and
degradation rates increase with S concentrations in the culture medium. The efficacy of the model was very high. These
results demonstrated that Azolla can be taken into consideration as a tool for sulphonamides environmental monitoring
and decontamination. r 2002 Elsevier Science Ltd. All rights reserved.

Keywords: Azolla; Bioaccumulation; Bioremediation; Phytoremediation; Sulphonamides; Sulphadimethoxine; Waste management

1. Introduction
Intensive farming implies considerable use of drugs:
among these toxic or potentially toxic compounds,
sulphonamides are regulated in their use as feed
additives, to contemporary treat large number of
animals for the prophylaxis or the therapy of animal
diseases. Sulphonamides could be administered by oral
route (G.U. No. 82/1963, No. 98/1968 and subsequent)
in the case of bacterial infections to cattle, swine and
poultry in concentrations ranging from 25 to 100 mg/kg
body weight in a period of 5–6 days. Sulphadimethoxine
(S) per oral route is not quantitatively absorbed by cattle
[1] and non-absorbed fraction is excreted unchanged in
faeces; phase II metabolites can also be found in dejects,
mainly as acetilated parent compound [1]. S concentra-
*Corresponding author. Tel.: +39-06-72594345; fax: +39-
06-2023500.
E-mail address: forni@uniroma2.it (C. Forni).
tions ranging from 300 to 900 mg/kg were found in fresh
faeces of treated calves [2]. Sulphonamides maintain
significant activity in manure for long time [1].
Normal practices of animal sludge application on soil
can lead to the introduction of these chemicals into the
terrestrial (soil top dressing) and aquatic (run off of top
dressed soils) environment compartments [3]. As a
consequence, not target species can be exposed to
sulphonamides: S is accumulated by Artemia (Crustacea,
Anostraca) and crop plants and weeds [4,2], notwith-
standing its low kow as log P (octanol–water=1.63) [5]
often assumed as index of low accumulation. (In the
cases we studied, the use of Kow as a predictor of a
complex process such as bioaccumulation could be
incorrect, in that Kow is an accurate measure of the
chemical but not of the biological lipophilicity of a
compound, and it does not take into consideration the
life habits of the organisms: plants and Artemia depend
on water for their life: a low Kow will facilitate transfer of
polar compounds.)

0043-1354/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 4 3 - 1 3 5 4 ( 0 2 ) 0 0 0 1 5 - 5
 C. Forni et al. / Water Research 36 (2002) 3398–3403
The activities of plants may offer a viable means of
drug quantitative reduction in lagoons. Aquatic plants,
such as Azolla, are known to remove from waters
nutrients, organic and inorganic (heavy metals) pollu-
tants [6,7]; they could be utilised also to remove drugs,
common contaminant of intensive farming sludge.
Azolla filiculoides Lam. is a water fern with leaf
cavities containing symbiotic N2-fixing cyanobacterium
Anabaena azollae Strasb. and a bacterial population [8].
This symbiosis has great agronomic interest and showed
resistance to antibiotics [9].
The aim of this study has been to evaluate the survival
of Azolla plants in S contaminated medium and its
efficiency in drug removal.
2. Materials and methods
Plant test. A. filiculoides, from the Botanical Garden
of the Milan University, were kept outdoor, Rome
latitude (September–October: photosynthetic active ra-
diation (PAR), 140–200 mE m
2 s
1), relative humidity
(RH) 70–75% (HANNA Instruments), avoiding rainfall
by plastic cover, for 5 weeks. Nine-gram plants were
inoculated in plastic bowls containing 750 ml N-free
Hoagland medium, pH 6.8 [10] added or not with 50,
150, 300 and 450 mg l
1 (ppm) S (Sigma, Milan). In text
and figures T50, T150, T300 and T450 indicate treated
batches at the different concentrations, namely 50, 150,
300 and 450 mg l
1. A randomised block design was
applied in the experiments; each treatment was per-
formed in triplicate. Test concentrations were nominal
(e.g., not measured analytically). Weekly, plants of each
bowls were collected, wiped off with paper tissue,
weighed and reinoculated; growth medium volume was
measured and refilled to 750 ml with new Hoagland
medium at the same drug concentration.
Medium test. Plastic bowls containing 500 ml N-free
Hoagland medium, pH 6.8, added with 50, 150, 300,
450 mg l
1 S, respectively were kept outdoor and
maintained 5 weeks at the same conditions.
2.1. Biological parameters
Weekly fresh weight (FW) was determined. After five
week nitrogenase activity and heterocyst frequency were
determined. Nitrogen fixation activity was measured by
acetylene reduction activity (ARA) [6]. Data are the
average of three replicates. At the same time, algal
packets were extracted from the cavities of third and
10th leaf. Heterocyst frequency (HF) in freshly isolated
algal packets was determined by counting their number
per 100 total cells [9]. Three or more algal packets per
leaf cavity per plant per treatment were examined.
3399
2.2. Chemical determinations
Drug determinations were performed in the medium
with or without plants and in plants. Every week until
the end of the plant experiment (5 week), 40 ml of
culture media were sampled and kept at
201C until
chemical analysis.
No residues of S were found on plastic bowls.
Preparation of samples. At the end of the experiment,
plants were weighed (FW) and dried at 501C fo 24 h and
weighed to obtain dry weight. Liquid medium was
directly processed. 1 mg sulphametoxipiridazine as inter-
nal standard was added to all samples. Results are the
mean of three replicates.
Extraction procedure. Plants. Five-gram samples were
extracted with 10 ml methylene chloride/acetone (1:1) in
Ultraturrax; added with sodium sulphate and sodium
chloride, vortexed for 1 min and sonicated for 10 min.
After centrifugation (500 g for 5 min) supernatant was
collected, brought to 9.5 ml with methylene chloride/
acetone (1:1) and added with 500 ml acetic acid glacial.
The supernatants were cleaned up on 3 ml SPE-SCX
columns (sulphonic acid solid phase extraction, J.T.
Baker, Holland) previously conditioned with 10 ml
methylene chloride/acetone/acetic acid glacial
(47.5:47.5:5); after washing columns with 3 ml water
and 5 ml methanol, S was eluted by 6 ml NH4OH 2.5%
in methanol. This fraction was collected and dried under
nitrogen stream and the residue suspended in 100 ml
methanol. Culture medium. 10 ml sample were acidified
to pH 5.5 with HCl 1 N and extracted twice with 5 ml
methylene chloride. The organic phase was dried under
nitrogen stream and the residue suspended again in
100 ml methanol.
Chromatographic analysis. The analyses were per-
formed with TLC system. Dilution of the suspended
extracts were: 1/100 for culture medium, T50 and T150
batches; 1/1000 for T300 and T450 batches. 10 ml of each
sample were deposited on a plate (Silica gel 60A, 19
channels; Whatman, USA). Mobile phase was chloro-
form/methanol (95:5). At the end of the run, the plates
were sprayed with fluorescamine 0.1% in acetone,
developed under the dark (10 min) and revealed under
an UV wood lamp.
Drug quantification. The extraction recovery
rates ranged from 79% to 85% for the plants and
from 89% to 91% for the medium. They were evaluated
by comparison with SMPD deposited (100, 75 and
50 ng). S was quantified by comparison with
deposited standard (between 225 and 900 ng for
plants, between 90 and 360 ng for medium). Standard
curve (n ¼ 6): y ¼
0:0717x þ 119:7; r2 ¼ 0:9962;
LOD=20 ng. Plates were digitally photographed
(Fig. 1) and fluorescent areas were automatically quan-
tified, as pixel density, by the software molecular analyst
(Biorad).
3400 C. Forni et al. / Water Research 36 (2002) 3398–3403
Fig. 1. Digital photograph of a TLC plate. From left to right
the following samples are shown two SMPD standards, five
SDX standards and three replicates of T50, T150, T300 and
T450 with internal standard.
3. Results
3.1. Azolla biomass yield and morphology
3.1.1. Biomass yield
Fresh weight increase of Azolla plants is shown in
Fig. 2 as weekly net yield. Control plants showed a N-
shaped curve of net yield, demonstrating a logistic
(density-dependent) plant growth: the biomass yield
continuously grew up to the fourth week, the increase
being of about 5 g, after that time density induced a
dramatic decrease in the net yield. A different growth
curve was shown by treated plants, with alternating low
increase/decrease of biomass yield. The response of
plants to 50, 150 and 300 mg l
1 S was quite similar with
overlapping curves: very low biomass increase at the first
week, deep decrease at the second week followed by a
slight steady recover between the third and the fourth
week, and again decrease at the fifth week. At 450 mg l
1
S the yield was almost nihil between the first and the
second week, and a very slight recover was shifted to the
fourth week. At the end of the experiment, the final
plant weight of the treated batches was lower than the
inoculum at all the concentrations tested; no significant
differences among treated batches were found, vice versa
differences were all significant (po0:05) between treated
and control plants.
3.1.2. Morphology and heterocyst frequency (HF)
Treated plants showed progressive morphological
alterations with increasing S concentration: the floating
stem was even more shorter and fragile than controls;
apical leaves were pale green while the oldest ones were
browned; as a rule, roots were detached. In addition, in
treated plants, algal packets within the lobes of the
dorsal leaves were less easily identifiable than in control
and in several samples cyanobacterial filaments were not
found, i.e. oldest leaf cavities.
Anabaena azollae microscopic observation showed
shrunken vegetative cells in the filaments, which were
shorter in treated plants than in the control; this damage
increased with S concentration in the growth medium.
At highest concentrations many detached vegetative
cells were present. HF is reported in Table 1. As a
general rule, S affected heterocyst differentiation, except
for treated batches at 50 and 450 mg l
1.
3.2. Nitrogenase activity
3.2.1. N2-fixation
After 5 weeks, the nitrogenase activity in treated
batches was dramatically reduced to more than 80% of
the control (Table 1). This reduction was directly related
to S concentration in the growth medium, and it fits with
data on the alterations of the cyanobiont population.
3.3. Sulphadimethoxine uptake
S uptake by Azolla from the growth medium at the
end of the experiment is reported in Table 2; the data
showed a clear and step enhancement of uptake with
increasing S concentration in the medium.
Drug amount in the growth media at the end of the
experiment and the total amount of drug supplied are
reported in Table 3. Considering all the data of Tables 2
and 3, we can say that Azolla system in this experimental
test showed a capability of removal ranging from 55.7%
(T50) to 86.3% (T450), in a 5 week period.
The same test was repeated in the absence of plants.
Drug supplied was, respectively, 22.5, 75, 150 and
225 mg; the amount of drug remaining in the growth
media were 21.375, 63.75, 115.5 and 157.5 mg. Abiotic
degradation, in this experimental test, ranged from
about 5% (T50) to 30% (T450), in a 5 week period.
These data were confirmed by a last determination after
12 weeks.
The overall drug remediation activity of Azolla system
is summarised in Fig. 3. Drug remaining in the growth
media in which Azolla was grown ranged from about
44% (T50) to 11% (T450). Drug detected in the plant
structures ranged between 0.6% and 2.1%, respectively,
at T50 and T450. Degradation due to abiotic factors
and/or microorganisms lowered drug concentration
from 5% to 30%, respectively, at T50 and T450,
therefore biotic degradation is able to lower drug
concentration from about 51% to about 56%, respec-
tively, at T50 and T450.
 C. Forni et al. / Water Research 36 (2002) 3398–3403 3401

Fig. 2. Growth of Azolla association, as net biomass yield, at five S concentrations. The yield was measured as g fresh weight added in
a 1 week period. T50, T150, T300 and T450 mean treated batches at the different concentrations, namely 50, 150, 300 and 450 mg l
1.

Table 1 Table 2
Heterocyst frequency (HF), as %, and nitrogenase activity of Uptake of sulphadimethoxine by Azolla in the model system;
Anabaena azollae, as % of control, after 5 weeks of both the amount per gram of plant fresh weight and the total
sulphadimethoxine treatment. The activity was measured as amount found in the samples are reported
nmol C2H4 h
1 g
1 fresh weight
Treatment Plant uptake S. E. Total amount
Treatment HF Nitrogenase activity (mg g
1) (mg)

Third leaf Tenth leaf T50 58.3579.64 5.52 1.33

T150 160.86712.60 7.27 3.18
Control 3.1 13.4 — T300 992.457211.07 121.86 19.58
T50 11.4 16.3 20 T450 2012.097169.90 98.09 45.00
T150 0.94 5.4 18.4
T300 5.8 7.7 15.1
T450 4.3 16.9 10.8 Table 3
Amounts of sulphadimethoxine supplied to the media and
detected in the media after 5 weeks
4. Discussion Treatment Total amount Amount detected at
supplied (mg) the end (mg)
S, at concentrations between 50 and 450 mg l
1,
T50 205.75 89.83
significantly affects plant growth and morphology, the
T150 627.75 132.67
alteration increases with drug concentration; never-
T300 1288.5 202.42
theless plants are able to survive to a 5 week period of T450 2092.5 240.89
treatment.
Nitrogenase activity of the cyanobiont dramatically
decreases in all treated plants and this decrease is (450 mg l
1) an increased heterocyst differentiation is
inversely related to S concentration. On the contrary found. This double behaviour of the organisms can be
heterocyst frequency shows a double trend: at lower driven by two opposite stimuli:
concentration (50 mg l
1) heterocyst differentiation is
stimulated; at higher concentrations (150–300 mg l
1) * a positive one, related to the need of higher N
the toxic effect prevail. At the highest concentration amount in stressed conditions, and
3402 C. Forni et al. / Water Research 36 (2002) 3398–3403

Fig. 3. S remediation activity of the Azolla system. Percentage of drug reduction in the system due to fern activity (biotic degradation,
in figure) and due to effects of abiotic factors and/or microorganisms (abiotic degradation, in figure). Percentage of drug residues in
plants (in plants, in figure) and in the media (in growth media, in figure).

* a negative one, due to the presence of N atoms in the
structure of the S molecule.
The first is induced Azolla, pushing the cyanobiont
towards higher N2-fixation, to overcome stress condition
(as it happens at 50 mg l
1, at this concentration HF
strongly increases). The second is induced by the
cyanobiont itself: two N sources are available, i.e. those
produced by the nitrogenase activity and those carried
by the S molecules (4 atoms/molecule; The Merck Index,
12th Edition, 1996). Finally, these enter into the leaf
cavities and are partly degraded and partly accumulated
into plant biomass. These sources could be able to
inhibit nitrogenase activity. This hypothesis is confirmed
by the presence of cyanophycin granules in the
vegetative cells, indicating that Anabaena is not N-
depleted. This unbalanced situation can lead to a high
HF in contrast to the low N2-fixing activity.
A. filiculoides system is able to absorb S from the
growth media. The drug entering into this biological
system partly remains into the biomass: uptake is high,
in the order of magnitude between mg g
1 and mg g
1,
and comparable to other uptake rates found in other
plants [2]. Notwithstanding, a great amount of drug
cannot be found in the growth media at the end of the
experiment, due to both the activity of the fern and to
the effects of abiotic factors and/or microorganisms.
The comparison between growth media at the same
environmental conditions with and without plants
showed that a higher proportion of drug was degraded
if plants were present (from 50.7% at T50 to 56.3% at
T450), being low the drug reduction in the absence of the
fern (from 5% at T50 to 30% at T450). Furthermore, in
this small Azolla bioremediation model (three bowls of
750 ml per concentration), the total amount of drug that
had been removed in a 5 week period is very high,
corresponding to a percentage ranging from 56.3% at
T50 to 88.5% at T450. These results are surprising for
the direct relationship between amount of drug in the
medium and removal capability.
Furthermore, the use of a thin layer chromatography
coupled with a software for image analysis transforms
the qualitative and easy to do test in a quite accurate
quantitative method: this tool can be routinely per-
formed also in the sewage waste treatment plants,
needing no sophisticated technical apparatus or very
high technical specialisation for the staff (as in the case
of HPLC).
5. Conclusions
In conclusion, Azolla association survives in very high
S concentration, although toxic effects on the plants
 C. Forni et al. / Water Research 36 (2002) 3398–3403
were detected. Nevertheless great amounts of S are
degraded by the Azolla bioremediation system, and this
technique, based on the high trapping/degrading capa- [2]
city of Azolla is a sound starting point to set up a
suitable and easy tool for drug decontamination in
manure lagoons. This laboratory model, being very high
the concentrations utilised and extremely wide the [3]
concentration range of the tests, clearly evidentiate the
capability of Azolla as bioremediation technique, to be [4]
applied to animal waste lagoons. Research in this
hitherto little explored field should be made.
Acknowledgements [5]
The authors are grateful to Dr. Lorenzo Tancioni of [6]
our Department, for preparing a sort of house for our
Azolla stock and for its kind and valid help in many
occasions. We wish to thank also Dr. Gianfranco
Brambilla, Laboratory of Veterinary Medicine, ISS, [7]
for the help during all the phases of the work and Dr.
Stefano Morabito, Laboratory of Veterinary Medicine,
ISS, for its help with image analysis hardware and [8]
software. Work supported by MURST 60%, ISS, and
CNR Project Biotechnology (No. 99.00326. PF49)
grants.
[9]
References
[10]
[1] Bevill RF. Sulfonamides. In: Boot NH, Mcdonald LE,
editors. Veterinary pharmacology and therapeutics, 6th ed.
3403
Iowa State University Press, Ames, IA. Science Press,
Ephrata, PA, 1988. p. 785–95.
Migliore L, Civitareale C, Brambilla G, Cozzolino S,
Casoria P, Gaudio L. Effect of sulphadimethoxine on
cosmopolitan weeds (Amaranthus retroflexus L., Plantago
major L., Rumex acetosella L.). Agric Ecosystem Environ
1997;65(2):163–8.
J
rgensen SE, Halling-S
rensen B. Drugs in the environ-
ment. Editorial. Chemosphere 2000;40:691–9.
Migliore L, Brambilla G, Grassitelli A, Dojmi di Delupis
G. Toxicity and bioaccumulation of sulphonamides
utilized for intensive farming in Artemia salina (L.)
(Crustacea, Anostraca). Int J Salt Lake Res 1993;2(2):
141–52.
Hansch C, Leo A, Hoekman D. Exploring QSAR.
Hydrophobic electronic and steric constance. Washington,
DC: ACS, 1995.
Forni C, Chen J, Tancioni L, Grilli Caiola M. Evaluation
of the fern Azolla for growth, nitrogen and phos-
phorus removal from wastewater. Water Res 2001;35(6):
1592–8.
Sela M, Garty J, Tel-Or E. The accumulation and effect of
heavy metals on the water fern Azolla filiculoides. New
Phytol 1989;112:7–12.
Grilli Caiola M, Forni C. The hard life of Prokaryotes
in leaf cavities of Azolla. In: Seckbach J, editor.
Enigmatic microorganisms and life in extreme environ-
ments. Netherlands: Kluwer Academic Publishers, 1999. p.
629–39.
Forni C, Tel-Or E, Bar E, Grilli Caiola M. Effects of
antibiotic treatment on Azolla–Anabaena and Arthrobac-
ter. Plant Soil 1991;137:151–5.
Van Hove C, Diara H, Godard P. Azolla en Afrique de
l’Ouest—in West Africa. Ed. Impr. E. Oleffe, Belgium,
1983.