ww.sciencedirect.com

wat e r r e s e a r c h 5 6 ( 2 0 1 4 ) 7 7e8 7

Available online at w

ScienceDirect

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

Toxicity assessment and modelling of Moringaoleifera seeds in water purification by whole cellbioreporter

Ali Adnan Al-Anizi a, Maria Theresa Hellyer b, Dayi Zhang a,c,*aKroto Research Institute, University of Sheffield, Sheffield S3 7HQ, UKbYorkshire Water Education Centre, Esholt, Bradford BD10 0NY, UKc Lancaster Environment Centre, Lancaster University, Lancaster LA1 4YQ, UK

a r t i c l e i n f o

Article history:

Received 26 October 2013

Received in revised form

24 February 2014

Accepted 26 February 2014

Available online 7 March 2014

Keywords:

Moringa oleifera

Cytoxicity

Genotoxicity

Whole cell bioreporter

Acinetobacter baylyi ADP1

* Corresponding author. B27, Lancaster EnvirE-mail addresses: d.zhang@lancaster.ac.u

http://dx.doi.org/10.1016/j.watres.2014.02.0450043-1354/ª 2014 Elsevier Ltd. All rights rese

a b s t r a c t

Moringa oleifera has been used as a coagulation reagent for drinking water purification,

especially in developing countries such as Malawi. This research revealed the cytoxicity

and genotoxicity of M. oleifera by Acinetobacter bioreporter. The results indicated that sig-

nificant cytoxicity effects were observed when the powdered M. oleifera seeds concentra-

tion is from 1 to 50 mg/L. Through direct contact, ethanolic-water extraction and hexane

extraction, the toxic effects of hydrophobic and hydrophilic components inM. oleifera seeds

were distinguished. It suggested that the hydrophobic lipids contributed to the dominant

cytoxicity, consequently resulting in the dominant genotoxicity in the water-soluble

fraction due to limited dissolution when the M. oleifera seeds granule concentration was

from 10 to 1000 mg/L. Based on cytoxicity and genotoxicity model, the LC50 and LC90 of M.

oleifera seeds were 8.5 mg/L and 300 mg/L respectively and their genotoxicity was equiv-

alent to 8.3 mg mitomycin C per 1.0 g dry M. oleifera seed. The toxicity of M. oleifera has also

remarkable synergistic effects, suggesting whole cell bioreporter as an appropriate and

complementary tool to chemical analysis for environmental toxicity assessment.

ª 2014 Elsevier Ltd. All rights reserved.

1. Introduction

As one of the key water purification processes, coagulation

binds the colloidal particles and bacteria, allowing an elec-

trostatic precipitation of contaminates from solution, the

cost-effective consideration of which is distinct in developing

countries, such as Malawi. The seeds ofMoringa oleifera, which

is a tropical tree commonly found in parts of Africa, India,

Malaysia, Sir Lanka and America, has been widely used as a

coagulant in terms of the powder extract (Diaz et al., 1999;

onment Centre, Lancastek, zhangdayi@tsinghua.o

rved.

Ghebremichael et al., 2005; Madsen et al., 1987; Muyibi and

Evison, 1995a; Ndabigengesere et al., 1995). Evidence sug-

gests that the bacterial colloids have been reduced by a per-

centage of 90e99% by the action of M. oleifera seeds

(Sutherland et al., 1990). Further research has shown that M.

oleifera has active diametric cationic proteins and a molecular

weight of 12e14 kDa, a large cation (Ndabigengesere et al.,

1995). The seed extract works by adsorption of colloids and

subsequent charge neutralisation of the resulting compound,

allowing for effective precipitation out of solution. The study

showed that compared to alum, the optimum dosage of

r University, Lancaster LA1 4YQ, UK.rg.cn (D. Zhang).

wat e r r e s e a r c h 5 6 ( 2 0 1 4 ) 7 7e8 778

shelled M. oleifera was similar at 50 mg/L. In the case of seeds

with the shells remaining the effective dosage increases

tenfold (Ndabigengesere et al., 1995). The normal dose used to

treat water of turbidity less than 100 NTU is in the range of

100e200 mg/L (Muyibi and Evison, 1995b; Nkurunziza et al.,

2009; Sutherland et al., 1990). For highly coloured water

(such as commonly found in Malawi) doses of up to 250 mg/L

may be required, which could place the supernatant in the

toxicity range. From Sutherland’s work at Thyolo Water

Treatment Works in Malawi, the dosage of M. oleifera seeds as

a coagulant ranged from 75 to 250 mg/L, reducing the high

turbidity of river water from 270 to 380 NTU down to 4 NTU

(Sutherland et al., 1994).

The bioactive agent was shown to be a steroidal glycoside-

strophantidin and the seed powder reduced total microbial

and coliform counts by 55% and 65% respectively (Eilert et al.,

1981). The 4-(a-L-rhamnosyloxy-benzyl) isothiocyanate was

isolatyed as the active antimicrobial component in Moringa

seeds (Eilert et al., 1981). The chemical constituents of the

seeds contains 4-(a-L-rhamnosyloxy) benzyl isothiocyanate, 4-

(L-rhamnosyloxy) phenylacetonitrile, 4-

hydroxyphenylacetonitrile, 4-hydroxyphenyl-acetamide, 4-

(a-L-rhamnopyranosyloxy)-benzyl glucosinolate Roridin E,

Veridiflorol, 9-Octadecenoic acid, O-ethyl-4-(a-L-rhamnosy-

loxy) benzyl carbamate, niazimicin, niazirin, b-sitosterol,

glycerol-l-(9-octadecenoate), 3-O-(6-O-oleoyl-b-D-glucopyr-

anosyl)-b-sitosterol and b-sitosterol-3-O-b-D-glucopyranoside

(Fahey, 2005). A further study (Santos et al., 2009) isolated the

coagulant ofM. oleifera by separating the lectin from the seeds

to evaluate the hemagglutinating activity in comparison to

alum. This was shown to be unique while the coagulant ac-

tivity was comparable to alum. Other research also demon-

strated that the active component of M. oleifera was neither

protein nor a polysaccharide (Okuda et al., 2001).

Nevertheless, complex components of M. oleifera have also

raised concerns on its potential toxicity (Chivapat et al., 2012;

Kavitha et al., 2012). The toxicity of aqueous M. oleifera

extraction was estimated since 1990’s (Asare et al., 2012;

Awodele et al., 2012) and both acute and chronic impacts

have been reported (Chivapat et al., 2011). Mustard oil glyco-

sinolates are abundant inMoringaceae, which break drown to a

glucose sugar, sulphate, isothiocyanates (mustard oils) or an

organic nitrile. Isothiocyanates are a skin irritant and irritate

the mucous lining of the gastrointestinal tract. Kidney and

liver damage may also result from this fraction along with

inhibiting the uptake of iodine for the thyroid (Fuller and

McClintock, 1986). M. oleifera is also reported to contain three

mustard seed oil glycosides. Glucosinolates are precursors or

organic iosthiocyanates that break down to produce giotro-

genic agents that cause hyperplasis and hypertrophy of the

thyroid gland. In addition, isothiocyanates have shown to

cause gastrointestinal tract lesions in cattle (Majak, 2001).

Besides, further research has revealed that higher level

toxicity of M. oleifera leaves was observed in the ethanolic

extraction, with significant negative effects on rat cells by

inhibiting lipid peroxidation (Ouedraogo et al., 2013).

Significant toxicity effects have been investigated previ-

ously with respective toxicity assay. Oluduro and Aderiye

dosedMoringa seed treatedwater (1e10mg/mL) tomale albino

rats daily for 21 days, and the results suggested that prolonged

consumption of water treated with greater than 2 mg/L of M.

oleifera seed constitute liver infarction (Oluduro and Aderiye,

2009). The oxygen uptake of T. pyriformis was affected by

5mg/LM. oleifera seeds solution, whereas the 96-h LC50 for fish

guppies (Poecilia reticulata) was 196 mg/L (Grabow et al., 1985).

Fed to 33 days old Hooded-Lister (Rowett strain) rats at a

dosage of 50 g and 100 g of seed protein per kg of meal, high

toxicity could be observed at an equivalent intake of 1.9 g ofM.

oleifera per day (Grant et al., 1991). The assessment of coagu-

lantM. oleifera lectin (cMoL) onmoth flour (Anagasta kuehniella)

suggested that cMoL at 1%w/w could increase themortality by

27.6%, indicating that the activity of cMoL is a carbohydrate-

lectin action on the digestive tract (Ramalho de Oliveira

et al., 2011). The effects of M. oleifera on the sexual behaviour

of mice was also found in terms of increasing lumen forma-

tion and epididymal maturity (Cajuday and Pocsidio, 2010). All

the research required toxicity assessment with mammalian

cells or living animals, restricted by the laborious cultivation,

high cost and long test period, raising the demand of fast

assessment methods for the toxicity of M. oleifera.

In this research, a rapid, cheap and easy approach was

explored to estimate the toxicity of M. oleifera, applying

whole cell bioreporter ADPWH_recA (Song et al., 2009). As a

practical tool for drinking water monitoring and quality

assessment in Africa, whole cell bioreporter can effectively

assess the cytoxicity and genotoxicity. The evaluation on

different extractions, including water, ethanol and hexane,

has revealed the impacts of respective components of M.

oleifera on cytoxic and genotoxic behaviour of living bacteria.

The mechanisms of M. oleifera toxicity were also charac-

terised and analysed by a cross-regulated SOS response

model.

2. Material and methods

2.1. Moringa oleifera chemical analysis and extraction

2.1.1. Chemical analysisThe collected seeds of M. oleifera were ground under liquid

nitrogen in a pestle and mortar and then freeze dried for 24 h.

Before chemical analysis and further treatment, the seeds

samples were air dried at room temperature and pulverized

into granules. All the units of the components concentration

were based on the dry weight of the seeds if not specifically

mentioned.

For the protein analysis, 1.0 g ofM. oleifera granules was re-

suspended in 0.8mL phenol solution (equilibratedwith 10mM

Tris HCl, pH 8.0) and 0.8 mL sodium dodecyl sulphite (SDS)

buffer (30% sucrose, 2% SDS, 0.1 M TriseHCl, pH 8.0, 5%

2-mercaptoethanol). After mixing, the phenol phase was

separated by centrifugation at 10,000 rpm for 3 min, and

transferred into fresh tubes with the addition of five volumes

of cold methanol and 0.1 M ammonium acetate. Precipitated

proteins were recovered and then washed with cold meth-

anolic ammonium acetate twice and cold 80% acetone twice.

The final pellet was dried and dissolved in phosphate buffered

saline. The sample was then analysed using Bradford’s assay

against BSA as a standard, to determine total protein

concentration.

wat e r r e s e a r c h 5 6 ( 2 0 1 4 ) 7 7e8 7 79

For chemical components analysis, 1.0 g of M. oleifera

granules was added to 1 mL of concentrated nitric acid and

digested in a microwave for 15 min. Diluted to a total volume

of 10 mL in 10% nitric acid, the suspension subsequently

passed through 0.45 mmfilter (Millipore, USA). ICP-MS analysis

was carried out using a Hewlett Packard 4500 series ICP-MS

against a multi-element standard (1000 ppm in 10% HNO3,

U-ICQ-026, LGC standards, Teddington, UK).

2.1.2. Water extractionFor water extraction, determined amounts of M. oleifera

granules were suspended in 250 mL deionized water, and the

final concentrations of granular M. oleifera were 400, 200, 100,

40, 30, 20, 4.5, 1.0, 0.5, 0.25, 0.10 and 0.01 g/L respectively. For

the soluble fraction of M. oleifera seeds, the granular suspen-

sion passed through 0.2 mm filter (Millipore, USA).

2.1.3. Solvent extractionThe hydrophobic fraction of M. oleifera was obtained by n-

hexane and ethanolic-water extraction methods. The n-hex-

ane extraction of M. oleifera followed previous Soxhlet

method, in which about 5 g granules of M. oleifera seeds were

mixed with 100 mL n-hexane and the extraction process was

carried out at room temperature for 8 h (Nguyen et al., 2011).

The ethanolic-water extraction was applied by dissolving 10 g

of M. oleifera seeds in ethanol/water (80/20, v/v) and was

shaken for 48 h at room temperature (Ouedraogo et al., 2013).

The extractions therefore passed through a 0.2 mm filter (Mil-

lipore, USA) and were subsequently evaporated at 100 �C (n-

hexane extraction) and 50 �C (ethanolic-water extraction)

until a constant weight was achieved. A further extraction

was carried out using the respective methods on 5 g granules

of seeds by ethanolic-water extraction followed by the n-

hexane extraction on the same sample.

2.2. Real drinking water sample simulation and test

Real drinking water samples were collected from an open

shallow well that serves a population of 120 people at Kum-

ponda 2 well, Blantyre, Malawi (co-ordinates 714374E,

8270070W), in which the drinking water was purified with M.

oleifera. The key water purification process using M. oleifera in

Malawi included grounding the seeds with a pestle and

mortar, creating a paste and suspension with untreated raw

water and rapidly stirred by hand.

M. oleifera shelled seeds were ground and sieved to particle

size of less than 2 mm. 20 g ground seeds were wetted to a

paste and added to 1.0 L of raw water sample and mixed well.

The suspension was then stirred at 50 rpm for a duration of

25 min and then left for the solids to settle for 30 min. Sub-

sequently, the supernatant was transferred into fresh glass

bottle as M. oleifera seeds stock solution.

To replicate conditions of dosing for water purification in

Malawi, the M. oleifera stock solution was diluted by raw

Malawi water at differing concentrations of 10, 20, 40, 60, 100,

150, 200, 300, 500 and 1000 mg/L. The water suspension fol-

lowed fast stirring at 200 rpm for 30 s, slow stirring at 50 rpm

for 15 min and settlement for 30 min. Both influent and

effluent were taken and stored under 4 �C, and the toxicity

was analysed at the University of Sheffield. Kinetic toxicity

tests were carried out on the samples from both the stock

solution and the raw water samples dosed with the M. oleifera

seeds stock solution. Samples were taken at 30 min in the first

3 h, and then daily for 4 days at room temperature.

2.3. Toxicity assessment

2.3.1. Toxicity assessment strains and culture conditionWhole cell bioreporter ADPWH_recA (Song et al., 2009), con-

structed with the luxCDABE gene fused on the chromosome of

Acinetobacter baylyi ADP1 host, was utilised to assess the

cytoxicity and genotoxicity of M. oleifera. It has been reported

that A. baylyi ADP1 based bioreporter is an effective biological

device to estimate the specific chemicals and relative cytox-

icity/genotoxicity in environmental samples (Zhang et al.,

2013). The bioreporter strain was grown at 30 �C overnight in

LuriaeBertani (LB) medium with 10 mg/mL Kanamycin as

antibiotic selection pressure, harvested by 10 min centrifu-

gation at 3000 rpm. The cell pellets were subsequently

resuspended in deionized and sterile water of the same vol-

ume as the stock solution, which can then be stored for at

least 45 days with sufficient activity for toxicity assessment.

The ADPWH_recA bioreporter applicable solution was pre-

pared by diluting the stock solution 10 times in the fresh LB

medium and was ready for bioluminescent measurement. All

chemicals were analytical grade reagents from

SigmaeAldrich.

2.3.2. Bioluminescence measurementFor all sample measurement, 180 mL of ADPWH_recA bio-

reporter applicable solution were transferred into the well of

black clear-bottom 96-well microplate (Corning Costa, USA),

with additional 20 mL of specific sample solution in relative

wells. Three measurement replicates and two biological rep-

licates were undertaken for each sample. The bioluminescent

signal of the microplate wells was determined by Synergy 2

Multi-mode Microplate Reader (BioTek Instruments, Inc., UK).

Incubated at 30 �C during the whole measurements process,

the bioluminescence and OD600 were measured every 10 min,

and 30 s vertical shaking was applied before each measure-

ment for cell and chemicals dispersion.

2.3.3. Toxicity estimationThe relative bioluminescence was applied to estimate the

cytoxicity and genotoxicity of water samples by dividing

induced bioluminescence by OD600 with Gene5 software. The

bioluminescent induction was obtained by averaging the data

from 250 min to 300 min, which was the best responsive time

for ADPWH_recA. Bioluminescence induction ratio was eval-

uated by dividing relative bioluminescence of samples by

relative bioluminescence of controls (non-induced samples).

2.4. Toxicity model

2.4.1. GenotoxicitySOS response represents the global response of bacterial cells

to carcinogens and the subsequent process of DNA repair

(Sancar, 1996). Previous investigations have focused on the

simulation of key steps within the SOS response process,

including mutagenesis, single stranded DNA (ssDNA)

wat e r r e s e a r c h 5 6 ( 2 0 1 4 ) 7 7e8 780

stimulation and DNA repair activation (Foster, 2007; Krishna

et al., 2007). Some gene regulation models are investigated

based on the specific chemical-DNA-protein interaction in

each response steps in Escherichia coli (Foster, 2007). However,

there is no lexA gene encoding the LexA repressor enzyme

associated in SOS response process in A. baylyi ADP1. Most of

the models are on ultraviolet irradiation (Gilchrest and Eller,

1999), which is not the same as the methylation and alkyl-

ation DNA damage mechanisms of M. oleifera extraction. A

new mathematical model is therefore conducted for the spe-

cific SOS response process in Acinetobacter. For the two

methylation and alkylation processes, the equilibrium in

Eq. (1) determined the existence of methylated or alkylated

(KMO) double stranded DNA (dsDNA), resulting in certain

amount of ssDNA synthesis rate (kMO) when exposed to M.

oleifera extraction. The recognition of ssDNA will conse-

quently result in the cleavage of LexA-like SOS repressor (LSR,

cell�1), the dynamics of which is expressed in Eqs. (2) and (3).

½M:oleifera� þ ½dsDNA�5KMO½damaged�DNA��/kMO½ssDNA� (1)

½dLSR� ��!ssDNA$kssDNA2$½sLSR� (2)

½dLSR� þ 2$½sLSR� ¼ ½LSR�total (3)

Here, dLSR (cell�1) and sLSR (cell�1) refer to LSR dimer and

monomer respectively, showing unique repression and acti-

vation on the SOS box. [LSR]total (cell�1) represents the total

amount of LexA-like SOS repressor in terms of monomer.

Similar to a previous investigation (Zhang et al., 2012), the

multi DNA damage mechanism of respective genotoxins and

the subsequent recognition of SOS repair on ssDNA and SOS

box promotion can be expressed by the transcriptional cross-

regulation model. With kssDNA representing the cleavage re-

action constant of LSR dimer by RecA protein, the equilibrium

of LSR dimer (KdLSR) andmonomer (KsLSR) determines dynamic

gene expression (SOS response) level, which is kdSLR and ksLSRrespectively, as illustrated in Eqs. (4) and (5). The SOS response

level when exposed to different types of genotoxins is shown

in Eq. (6).

½dLSR� þ ½SOS� box�5KdLSR½dSLR� SOS�/kdSLRSOS response repression

(4)

½sLSR� þ ½SOS� box�5KsLSR½SOS�/ksLSRSOS response expression (5)

SOSs ¼�kssDNA$kdSLR

1þ kssDNA$½LSR�total

�$

½dLSR�K�1dLSR þ ½dLSR�

þ�

kdSLR

2$ð1þ kssDNAÞ$½LSR�total�$

½sLSR�K�1sLSR þ ½sLSR�

(6)

SOSr;s ¼ 1þ�

kdSLR

2$ð1þ kssDNAÞ$½LSR�total�$

½M:oleifera�ðKsLSR$KMO$kssDNA$kMOÞ�1 þ ½M:oleifera�

(7)

Here, [S] (cell�1) refers to the concentration of genotoxins

inside the cells, and SOSs represents the intensity of SOS

response (cell�1). Compared with the bioluminescent baseline

absence of genotoxins, the relative SOS response ratio, SOSr,s,

is expressed in Eq. (7). The SOS response coefficient is defined

as KsLSR$KMO$kssDNA$kMO, referring to the combined efficiency

of genotoxin DNA damage, ssDNA recognition and SOS box

promotion. KGenotoxicity is the SOS responsive intensity of

specific carcinogens and refers to the coefficient of genotox-

icity impacts, representing kdSLR/2$(1 þ kssDNA)$[LSR]total.

2.4.2. CytoxicityDifferent from genotoxicity, cytoxicity is caused by direct in-

hibition effects of cytoxic compounds, consequently resulting

in the suppression of cell activities. Various types of cytoxic

effect can be identified, such as membrane integrity loss as

the result of cell lysis and protein activity inhibition. Particu-

larly, with the dynamic cytoxic coefficient (kcytoxicity), the

protein inhibition can be expressed in the following Eq. (8).

½M:oleifera� þ ½protein� 5kcytoxicity½inhibited protein� (8)

KCytotoxicity ¼ k�1cytoxicity

k�1cytoxicity þ ½M:oleifera� (9)

Cell activities ¼ 1� KCytotoxicity ¼ ½M:oleifera�k�1cytoxicity þ ½M:oleifera� (10)

Considering the equilibrium reaction state, the cytoxic in-

hibition ratio is therefore associated with the exposed cytoxic

chemicals, as expressed in Eq. (9). Consequently, the cytox-

icity also affects cell metabolism, causing damage to enzymes

and cell activities, or even apoptosis (Joiner et al., 1999). The

remaining cell activities, as defined in Eq. (10), are charac-

terised by the constant expressed luxCDABE gene on the

chromosome of A. baylyi ADP1, which is linked to the function

of cellular proteins.

2.4.3. Hydrophobic components release dynamicsThe release process of hydrophobic components in M. oleifera

seeds is comparable with classical polymeric drug delivery

processes and regarded as the slow release dynamics (Zhang

et al., 2014). Considering the time-dependence process in

power law,kineticsofcarbonreleaseprocesscouldbedescribed

in simple form by Eq. (11) (Ritger and Peppas, 1987).

Mt

MN

¼ ktn (11)

Here, Mt is the amount of hydrophobic molecules released

at time t, MN is the theoretical total amount of released mol-

ecules, and k is the release coefficient. Compared with Peppas

Model (Brazel and Peppas, 2000), the relative anomalous

diffusion can be described by respective n value as Fickian

(n ¼ 0.5), Case II (n ¼ 1.0) and Super Case II (n > 1.0) transport.

3. Results

3.1. Components analysis of Moringa oleifera seeds

M. oleifera seeds were determined to contain 1.75% moisture

and 187 mg/g protein. As illustrated in Table 1, the element

concentrations in M. oleifera seeds showed that no significant

Table 1 e Element components in M. oleifera seeds.

Element Concentration(mg/g dry weight)

Standard deviation(mg/g dry weight)

Li 6.628 � 10�6 1.445 � 10�7

Be <LoD e

Na 1.342 5.296 � 10�1

Mg 1.506 2.967 � 10�2

Al < LoD e

K 2.552 1.166 � 10�1

Ca 0.143 2.431 � 10�3

V <LoD e

Cr 2.655 � 10�4 3.385 � 10�5

Mn 3.002 � 10�3 1.183 � 10�4

Fe 1.113 � 10�2 1.525 � 10�3

Co <LoD e

Ni 0.251 � 10�4 1.832 � 10�6

Cu 1.239 � 10�3 2.435 � 10�5

Zn 1.107 � 10�2 2.823 � 10�4

Ga <LoD e

As <LoD e

Se 4.977 � 10�4 5.405 � 10�5

Rb 5.438 � 10�4 2.132 � 10�5

Sr 1.538 � 10�3 2.322 � 10�5

Ag <LoD e

Cd <LoD e

In <LoD e

Cs <LoD e

Ba 3.597 � 10�4 1.030 � 10�4

Tl <LoD e

Pb 0.068 � 10�5 1.095 � 10�6

Bi <LoD e

wat e r r e s e a r c h 5 6 ( 2 0 1 4 ) 7 7e8 7 81

heavy metals or other elements exist, indicating the potential

cytoxicity and genotoxicity of the seeds granules or suspen-

sion were derived from the organic matter. Compared with

previous research (Olaymei and Alabi, 1994) in which M. olei-

fera seeds contain 34.1% protein, 15.5% lipids and 15% carbo-

hydrates, the seeds in this study have less protein and

carbohydrates. The key active components to coagulate dur-

ing water purification process have been suggested as a non-

proteinaceous compound with a molecular mass of <3 kDa

Fig. 1 e Bioluminescent response of ADPWH_recA to soluble (A)

control.

(Okuda et al., 2001). Other research suggested the active

components to be soluble proteins with a net positive charge

and molecule weight of 13 kDa with the coagulant activity

comparable to alum (Ndabigengesere et al., 1995; Sutherland

et al., 1990). The 9.1 kDa flocculent M. oleifera protein MO2.1

has also been shown for its effective coagulation properties

(Pavankumar et al., 2014). Due to the synergetic effects of

complex components in M. oleifera seeds, further GCeMS

analysis failed to distinguish the dominant organic compo-

nents contributing to the toxicity effects (data not shown

here), indicating whole cell bioreporter as an appropriate

approach to estimate the synergetic toxicity of different

water-soluble portion and granules.

3.2. Toxicity of water-soluble Moringa oleifera

Previous research has addressed the cytoxicity and genotox-

icity of various components of M. oleifera seeds (Araujo et al.,

2013), whereas limited work attempted to reveal the respec-

tive toxic effects in hydrophilic and hydrophobic fractions at

the same time. As shown in Fig. 1, significant toxicity effects

were observed for thewhole cell bioreporter after 200e400min

exposure to the solubleM. oleifera seedswater extraction at the

lowest1mg/Lconcentration.CytoxicitywasdominantwhenM.

oleifera concentration was 50 mg/L and genotoxicity was

obvious at concentration over 100 mg/L. Compared with the

equivalent toxicity ofmitomycin C,water soluble proportion of

1.0 gM. oleifera is equal to 1.2mgmitomycinC (Song et al., 2009),

showing significant toxic impacts on water quality. Limited

research has suggested that hydrophobic lipids are the main

cytoxic components in M. oleifera seeds and the dominant

genotoxic components are anti-nutritional factors such as

toxic alkaloids or amino acids (Eilert et al., 1981). Due to limited

dissolution of hydrophobic components, the results demon-

strated that the toxicity of the water extraction and soluble

proportion mainly represent as genotoxicity but slight cytox-

icity. Similar to the conclusion of previous investigation (Grant

et al., 1991), the results suggested that the water-soluble

and granules (B) seeds with mitomycin C (MMC) as positive

wat e r r e s e a r c h 5 6 ( 2 0 1 4 ) 7 7e8 782

fraction ofM. oleiferahas low toxicity and thedominant toxicity

come from hydrophobic components.

3.3. Toxicity of Moringa oleifera granules andhydrophobic components

Exposing M. oleifera seeds granules to the toxicity bioreporter,

strong cytoxicity impacts were found when the concentration

isabove1mg/L, as illustrated inFig. 2. Thesignificant inhibition

of the granules reduced the bacterial activities, representing a

lower bioluminescent signal. It was interesting that a higher

responsive ratio was observed when the seed granules’ con-

centration was from 50mg/L to 450 mg/L, indicating genotoxic

induction was dominant in higher dosage ofM. oleifera seeds.

To distinguish the unique impacts of cytoxicity and geno-

toxicity of M. oleifera, the toxicity model was simulated

considering respective toxicity mechanisms, as demonstrated

in Fig. 3. The dynamic cytoxic coefficient (kcytoxicity) of water

soluble fraction of M. oleifera seeds is 0.04, remarkably higher

than 0.01 as the coefficient of seeds granules. As for geno-

toxicity, KGenotoxicity, the coefficient of genotoxicity impacts,

was 0.98 and 1.50 for water soluble fraction and seeds gran-

ules respectively, and the SOS response coefficient of gran-

ules, KsLSR$KMO$kssDNA$kMO, is about 10 times higher than that

of the water soluble proportion. The results of model simu-

lation matched the data from the toxicity test (Figs. 4 and 5),

indicating the toxicity effects of M. oleifera seeds were the

combination of cytoxicity and genotoxicity. The different

toxicity coefficients came from the respective dissolution of

hydrophilic components as in previous findings that thewater

soluble fraction of M. oleifera seeds behaved cytoxicity only

when the concentration was below 20 mg/L but genotoxicity

from 20 mg/L to 10,000 mg/L. This was due to high cytoxic

coefficient (less cytoxicity inhibition) and low genotoxic co-

efficient (less genotoxicity effects). Though the cytoxicity ef-

fect of granule seeds is remarkable due to the low cytoxic

coefficient, the high genotoxic coefficient and SOS response

coefficient can promote strong genotoxic response at a certain

level of granule dosage (10 mg/L to 500 mg/L), in accordance

with experimental results.

Fig. 2 e Bioluminescent response of ADPWH_recA to soluble (

It was obvious that different toxicity behaviour of M. olei-

fera seeds was determined by the dissolution of seeds com-

ponents into the aqueous phase. Extracted by ethanol and

hexane respectively, different toxic effects were found for the

hydrophobic components of M. oleifera, as illustrated in Fig. 6.

Ethanol-water extraction performedmainly with genotoxicity

and the highest bioluminescent response was 1.41 when the

seeds concentration was 2000 mg/L (Fig. 6A). The dramatic

bioluminescence signal drop in higher concentration was due

to the impacts of cytoxicity. No significant induction existed in

hexane extraction from 0.1 to 1000 mg/L (Fig. 6B), suggesting

the similar promotion and inhibition effects caused by geno-

toxicity and cytoxicity, followed by strong inhibition and

suppression of bioluminescence at higher concentration. For

the residual fractions, the seeds granules after ethanol treat-

ment behaved similar to the ethanolic-water phase and the

genotoxic response was from 1 to 1000mg/L (Fig. 6C), whereas

the hexane residuals had a peak response (1.32) at 2000 mg/L

(Fig. 6D). M. oleifera is reported to contain three mustard seed

oil glycosides, which were all hydrophobic components. Glu-

cosinolates are precursors or organic iosthiocyanates that

break down to produce giotrogenic agents that cause hyper-

plasis and hypertrophy of the thyroid gland. Isothiocyanates

have shown to cause gastrointestinal tract lesions in cattle

(Majak, 2001). Earlier research isolated the glycosidic mustard

oil from M. oleifera seeds, such as the derivative 4-(a-L-rham-

nosyloxy) benzyl isothiocyanate, from the supernatant and no

significant toxicity were observed (Grabow et al., 1985).

Further research has also demonstrated that less than 2 mg/L

ofM. oleifera seed in water treatment process would not cause

obvious mustard oil leakage to constitute liver infarction

(Oluduro and Aderiye, 2009).

4. Discussion

4.1. Toxicity mechanism of Moringa oleifera

In accordance with model analysis, the toxicity of M. oleifera

can be categorised into the following three types, soluble

A) and granules (B) seeds under different concentrations.

Fig. 3 e Model simulation of respective water-soluble cytoxicity (A), total cytoxicity (B) and genotoxicity (C).

wat e r r e s e a r c h 5 6 ( 2 0 1 4 ) 7 7e8 7 83

cytoxicity, insoluble cytoxicity and genotoxicity. Each toxicity

has respective model parameters with unique biological im-

pacts. From the model prediction illustrated in Figs. 4, 5 and 6,

it is obvious that the bioluminescent response of ADPWH_-

recA to M. oleifera extraction was affected by both cytoxicity

and genotoxicity.

Based on cytoxicity and genotoxicity model, the LC50 and

LC90 of M. oleifera seeds were 8.5 mg/L and 300 mg/L respec-

tively and their genotoxicity is equivalent to 8.3mgmitomycin

C per 1.0 g dry M. oleifera seeds. The results matched previous

investigation on the powdered M. oleifera seeds under

different development stages of a malarial vector Anopheles

stephensi mosquitoes, in which the LC50 and LC90 were

57.79e78.93 mg/L and 125.93e143.20 mg/L, respectively

(Prabhu et al., 2011).

Related research also had the similar opinion towards the

toxicity of M. oleifera with different toxicity assays (Ayotunde

et al., 2011; Kavitha et al., 2012). Ayotunde et al. studied im-

pacts of aqueousM. oleifera extract on the adultNile tilapia, and

they found the toxicity from 300 mg/L with total kidney and

skin cell degeneration, and gill arch and filaments by 350mg/L

resulting in death of all the fish (Ayotunde et al., 2011). Kavitha

exposed M. oleifera to Cyprinus carpio fish cell line and showed

toxicity LC50 at 124.0 mg/L (Kavitha et al., 2012). Further

studies focused on the dichloromethane extract of M. oleifera

seeds indicating their cytoxicity on human cancer cells lines,

non-small cell lung adenocarcinoma (A549), colon carcinoma

(HCT 116) and a non-cancer cell line Chinese hamster ovary

cells (AA8) using the MTT cytoxicity assay. The LC50 values

were evaluated between 2.6 and 3.0 mg/L, and the key cytoxic

components were identified as 4-(40-O-acetyl-a-L-rhamnosy-

loxy) benzyl isothiocyanate and 4-(a-L-rhamnosyloxy) benzyl

isothiocyanate against the non-small cell lung adenocarci-

noma (A549) and colon carcinoma (HCT 116) (Ragasa et al.,

2012).

4.2. Toxicity releasing of Moringa oleifera seeds

Calculated and shown in Table 2, the dynamic cytoxic co-

efficients (kcytoxicity) of the entire extraction and residual

fraction were equal, the value of which is 0.01, suggesting

their similar cytoxic impacts. For genotoxicity, n-hexane

extraction had a higher coefficient of genotoxicity (KGenotoxicity,

0.85) and lower SOS response coefficient

Fig. 4 e Bioluminescent response ratio of ADPWH_recA

against respective concentration of seeds (filtered). Dot for

experimental data and line for model simulation.

Cytoxicity is the dominant toxic impacts when

concentration is below 20 mg/L, whereas genotoxicity can

be observed above the threshold.

Fig. 5 e Bioluminescent response ratio of ADPWH_recA

against respective concentration of seeds (granules). Dot

for experimental data and line for model simulation.

Cytoxicity can be observed when the concentration is at

1 mg/L, and significant cytoxic impacts cover the whole

scale. Genotoxicity can be found when the concentration is

between 10 and 1000 mg/L.

wat e r r e s e a r c h 5 6 ( 2 0 1 4 ) 7 7e8 784

(KsLSR$KMO$kssDNA$kMO, 0.0018) than those of residual (0.30 and

0.10 respectively), indicating the main genotoxic components

in the extraction fraction. In contrast, higher coefficient of

genotoxicity and lower SOS response coefficientwere found in

the residual fraction (0.70 and 0.002 respectively), compared

with those in the extraction fraction (0.05 and 0.10). It could be

explained by different extraction components via the two

methods. N-hexane is better for oil extraction from plant

seeds than ethanol (Ferreira-Dias et al., 2003), and the residual

fraction contains more hydrophilic lipids and lectins, which

exist in the ethanolic-water extraction phase when treated by

ethanol. Since hexane extraction contains more hydrophobic

components (Nguyen et al., 2011), the results suggested that

the hydrophobic components, such as mustard oil glyco-

sinolates, were the main genotoxic compounds during the

seeds treatment process.

The toxicity dynamics of M. oleifera seeds release process

showed the key mechanism of Fickian diffusion, when real

drinking water from Malawi was tested for the cytoxicity and

genotoxicity change after long term purification treatment by

the seeds (Fig. 7). During the first 2 days, the bioluminescent

response of ADPWH_recA to M. oleifera seeds extraction con-

verted rapidly from inhibition (Day 1) to induction (Day 2). Of

all the hydrophobic fractions, mustard oil glycosinolates are

abundant and the potential toxic components in M. oleifera,

which break drown to glucose sugar, sulphate, iso-

thiocyanates or an organic nitrile (Fuller and McClintock,

1986). Since isothiocyanates are a skin irritant and irritate

the mucous lining of the gastrointestinal tract, genotoxicity

became dominant at low seed concentration levels (10 mg/L),

performing a positive response at 2.71. The further toxic

trends of M. oleifera seeds showed that more genotoxicity was

observed in the drinking water with longer treatment (Day 3

and Day 4), representing the releasing equilibrium of geno-

toxic components. The releasing coefficient was estimated as

0.48, highly related to the treatment process, specifically the

stirring time, indicating the existence of Fickian diffusion

during release process. The macromolecule framework of M.

oleifera seeds resisted the hydrophobic molecule trans-

portation (Westedt et al., 2007), which was significantly

affected by interface reaction and water sorption (Hopfenberg

and Frisch, 1969). Associated with pretreatment and granule

scales, the Fickian diffusion coefficient (k, mg (g d1/2)�1) was

1.08, representing the high toxicity releasing capacity of M.

oleifera seeds during water purification process. It is remark-

able that higher risk of carcinogenic effect might exist in the

drinking water samples after long-term treatment due to the

metabolites of mustard oil glycosinolates. For example, the

isothiocyanates might cause the inhibition of iodine uptake

for the thyroid, consequently resulting in kidney and liver

damage due to the treated drinking water (Fuller and

McClintock, 1986).

4.3. Synergetic toxicity effects of M. oleifera seeds

Comparing the different cytoxicity and genotoxicity of hexane

or ethanol extraction and residuals, it was clear that the total

toxicity of M. oleifera seeds was not equal to the sum of the

respective solvent extraction and residuals, indicating that

the synergetic effects existed. As illustrated in Table 2, various

Fig. 6 e Toxicity dynamics of M. oleifera seeds extraction (A for ethanol and B for hexane) and residuals (C for ethanol and D

for hexane) with respective hexane and ethanol solvent.

wat e r r e s e a r c h 5 6 ( 2 0 1 4 ) 7 7e8 7 85

types of M. oleifera seeds granules have their unique cytoxic

and genotoxic effects. The dynamic cytoxic coefficients of

either extraction or residual fractions in n-hexane and

ethanol treatment were lower (higher cytoxicity) than that of

water-soluble fraction of M. oleifera seeds. The repression ef-

fects therefore existed in seeds granules to reduce the integral

cytoxicity of the seeds, resulting in higher cytoxicity of water-

soluble compounds in the aqueous phase without such sup-

pression (Araujo et al., 2013). In contrast, genotoxicity had the

acceleration effects that the coefficients of genotoxicity in all

Table 2 e Parameters of genotoxicity and cytoxicity in differen

Treatment kcytoxicity

Water soluble fraction 0.04

Seeds granules 0.01

Hexane extraction 0.01

Hexane residual 0.01

Ethanol extraction 0.01

Ethanol residual 0.01

the fractions were significantly lower than water-soluble

fraction or seeds granules, hinting that the interaction be-

tween hydrophobic and hydrophilic components could pro-

mote the genotoxicity of the seeds. Chemical analysis faced

challenges to distinguish such repression or acceleration ef-

fects of toxicity in complex samples, such as the M. oleifera

seeds, since it could only provide the existence and level of

each component, not linking individual molecule and their

compound with any toxic effects. To estimate the synergetic

impacts of cytoxicity and genotoxicity, the bioreporter is an

t treatments.

KGenotoxicity KsLSR$KMO$kssDNA$kMO

0.98 0.004

1.50 0.05

0.85 0.0018

0.30 0.1

0.05 0.1

0.70 0.002

Fig. 7 e The toxicity of M. oleifera treated water sample

from Malawi. Assessed on the Day 1, 2, 3 and 4 after

treatment, the water samples behaved cytoxicity and

genotoxicity change, caused by the occurrence of

biodegradation and release of metabolites.

wat e r r e s e a r c h 5 6 ( 2 0 1 4 ) 7 7e8 786

effective tool and has the capability to represent the compli-

cated toxicity effects via SOS regulated bioluminescence

response.

5. Conclusion

With the bioreporter estimation and model analysis, sig-

nificant cytoxicity and genotoxicity of M. oleifera seeds are

investigated in this research. The whole cell bioreporter is a

cost-effective, easy-operation and reliable tool, as an alter-

native assessment technique for analysis of toxicity levels

within drinking water, particularly for developing countries.

The results indicated that the main toxicity is from the

insoluble fatty acidic components of M. oleifera, which would

remain in the supernatant. The high risk of M. oleifera seeds

for water purification suggests further investigation of

appropriate water purification techniques for rural areas in

developing countries. Having target characteristics of low

cost and ecological friendly methods for water treatment,

other potential approaches are the pretreatment of M. olei-

fera seeds to reduce the toxicity, such as seeds extraction

heating at 80 �C for toxic enzymes deactivation (Ferreira

et al., 2009) or advanced precipitation to remove the resid-

ual proteins.

Acknowledgements

The authors would like to thank National Natural Science

Foundation of China (41301331) for financial support. Dr Jamie

Young (Sheffield Hallam University) helped in chemical

analysis of M. oleifera seeds components.

r e f e r e n c e s

Araujo, L.C.C., Aguiar, J.S., Napoleao, T.H., Mota, F.V.B.,Barros, A.L.S., Moura, M.C., Coriolano, M.C., Coelho, L.,Silva, T.G., Paiva, P.M.G., 2013. Evaluation of cytotoxic andanti-inflammatory activities of extracts and lectins fromMoringa oleifera seeds. PLoS One 8 (12).

Asare, G.A., Gyan, B., Bugyei, K., Adjei, S., Mahama, R., Addo, P.,Otu-Nyarko, L., Wiredu, E.K., Nyarko, A., 2012. Toxicitypotentials of the nutraceutical Moringa oleifera at supra-supplementation levels. J. Ethnopharmacol. 139 (1), 265e272.

Awodele, O., Oreagba, I.A., Odoma, S., da Silva, J.A.T.,Osunkalu, V.O., 2012. Toxicological evaluation of the aqueousleaf extract of Moringa oleifera Lam. (Moringaceae). J.Ethnopharmacol. 139 (2), 330e336.

Ayotunde, E.O., Fagbenro, O.A., Adebayo, O.T., 2011. Histologicalchanges in Oreochromis niloticus (Linnaeus 1779) exposed toaqueous extract of Moringa oleifera seeds powder. Turk. J. Fish.Aquat. Sci. 11 (1), 37e43.

Brazel, C.S., Peppas, N.A., 2000. Modeling of drug release fromswellable polymers. Eur. J. Pharm. Biopharm. 49 (1), 47e58.

Cajuday, L.A., Pocsidio, G.L., 2010. Effects of Moringa oleifera Lam.(Moringaceae) on the reproduction of male mice (Musmusculus). J. Med. Plants Res. 4 (12), 1115e1121.

Chivapat, S., Sincharoenpokai, P., Saktiyasuthorn, N.,Shuaprom, A., Thongsrirak, P., Sakpetch, A., Rungsipipat, A.,2011. Acute and chronic toxicity of Moringa oleifera Linn leavesextracts. Thai J. Veterinary Med. 41 (4), 417e423.

Chivapat, S., Sincharoenpokai, P., Suppajariyawat, P.,Rungsipipat, A., Phattarapornchaiwat, S.,Chantarateptawan, V., 2012. Safety evaluations of ethanolicextract of Moringa oleifera Lam. Seed in experimental animals.Thai J. Veterinary Med. 42 (3), 343e352.

Diaz, A., Rincon, N., Escorihuela, A., Fernandez, N., Chacin, E.,Forster, C.F., 1999. A preliminary evaluation of turbidityremoval by natural coagulants indigenous to Venezuela.Process Biochem. Amst. Neth. 35 (3e4), 391e395.

Eilert, U., Wolters, B., Nahrstedt, A., 1981. The antibiotic principleof Moringa oleifera and Moringa stenopetala. Planta Medica 42 (1),55e61.

Fahey, J.W., 2005. Moringa oleifera: a review of the medicalevidence for its nutritional, therapeutic, and prophylacticproperties. Trees Life J. 1 (5).

Ferreira-Dias, S., Valente, D.G., Abreu, J.M.F., 2003. Comparisonbetween ethanol and hexane for oil extraction from Quercussuber L. fruits. Grasas Y Aceites 54 (4), 378e383.

Ferreira, P.M.P., Carvalho, A.F.U., Farias, D.F., Cariolano, N.G.,Melo, V.M.M., Queiroz, M.G.R., Martins, A.M.C., Machado-Neto, J.G., 2009. Larvicidal activity of the water extract ofMoringa oleifera seeds againstAedes aegypti and its toxicity uponlaboratory animals. An. Acad. Bras. Ciencias 81 (2), 207e216.

Foster, P.L., 2007. Stress-induced mutagenesis in bacteria. Crit.Rev. Biochem. Mol. Biol. 42 (5), 373e397.

Fuller, T.C., McClintock, E., 1986. Poisonous Plants of California.University of California Press, Berkeley, California.

Ghebremichael, K.A., Gunaratna, K.R., Henriksson, H., Brumer, H.,Dalhammar, G., 2005. A simple purification and activity assayof the coagulant protein from Moringa oleifera seed. Water Res.39 (11), 2338e2344.

Gilchrest, B.A., Eller, M.S., 1999. DNA photodamage stimulatesmelanogenesis and other photoprotective responses. J.Investig. Dermatol. Symp. Proc. 4 (1), 35e40.

Grabow, W.O.K., Slabbert, J.L., Morgan, W.S.G., Jahn, S.A.A., 1985.Toxicity and mutagenicity evaluation of water coagulatedwith Moringa oleifera seed preparations using fish, protozoan,bacterial, coliphage, enzyme and Ames Salmonella assays.Water SA 11 (1), 9e14.

wat e r r e s e a r c h 5 6 ( 2 0 1 4 ) 7 7e8 7 87

Grant, G., More, L.J., McKenzie, N.H., Dorward, P.M., Stewart, J.C.,Telek, Pusztai, A., 1991. A survey of the nutritional andhaemagglutination properties of several tropical seeds. Livest.Res. Rural Dev. 3 (3).

Hopfenberg, H.B., Frisch, H.L., 1969. Transport of organicmicromolecules in amorphous polymers. J. Polym. Sci. Part B-Polymer Lett. 7 (6), 405e409.

Joiner, M.C., Lambin, P., Marples, B., 1999. Adaptive response andinduced resistance. Comp. Rend. Acad. Des. Sci. Ser. III-Sci.Vie-Life Sci. 322 (2e3), 167e175.

Kavitha, C., Ramesh, M., Kumaran, S.S., Lakshmi, S.A., 2012.Toxicity ofMoringa oleifera seed extract on some hematologicaland biochemical profiles in a freshwater fish, Cyprinus carpio.Exp. Toxicol. Pathol. 64 (7e8), 681e687.

Krishna, S., Maslov, S., Sneppen, K., 2007. UV-inducedmutagenesis in Escherichia coli SOS response: a quantitativemodel. PLoS Comput. Biol. 3 (3), 451e462.

Madsen, M., Schlundt, J., Omer, E.E., 1987. Effect of watercoagulation by seeds of Moringa oleifera on bacterialconcentrations. J. Tropical Med. Hyg. 90 (3), 101e109.

Majak, W., 2001. Review of toxic glycosides in rangeland andpasture forages. J. Range Manag. 54 (4), 494e498.

Muyibi, S.A., Evison, L.M., 1995a. Moringa-oleifera seeds forsoftening hardwater. Water Res. 29 (4), 1099e1104.

Muyibi, S.A., Evison, L.M., 1995b. Optimizing physical parametersaffecting coagulation of turbid water with Moringa-oleiferaseeds. Water Res. 29 (12), 2689e2695.

Ndabigengesere, A., Narasiah, K.S., Talbot, B.G., 1995. Activeagents and mechanism of coagulation of turbid waters usingMoringa-oleifera. Water Res. 29 (2), 703e710.

Nguyen, H.N., Gaspillo, P.A.D., Maridable, J.B., Malaluan, R.M.,Hinode, H., Salim, C., Huynh, H.K.P., 2011. Extraction of oilfrom Moringa oleifera kernels using supercritical carbondioxide with ethanol for pretreatment: optimization of theextraction process. Chem. Eng. Process. 50 (11e12), 1207e1213.

Nkurunziza, T., Nduwayezu, J.B., Banadda, E.N., Nhapi, I., 2009.The effect of turbidity levels andMoringa oleifera concentrationon the effectiveness of coagulation in water treatment. WaterSci. Technol. 59 (8), 1551e1558.

Okuda, T., Baes, A.U., Nishijima, W., Okada, M., 2001. Isolationand characterization of coagulant extracted from Moringaoleifera seed by salt solution. Water Res. 35 (2), 405e410.

Olaymei, A.B., Alabi, R.O., 1994. Studies on traditional waterpurification using Moringa oleifera seeds. Afr. Study Monogr. 15(3), 135e142.

Oluduro, A.O., Aderiye, B.I., 2009. Effect of Moringa oleifera seedextract on vital organs and tissue enzymes activities of malealbino rats. Afr. J. Microbiol. Res. 3 (9), 537e540.

Ouedraogo, M., Lamien-Sanou, A., Ramde, N., Ouedraogo, A.S.,Zongo, S.P., Goumbri, O., Duez, P., Guissou, P.I., 2013. Protectiveeffect of Moringa oleifera leaves against gentamicin-inducednephrotoxicity in rabbits. Exp. Toxicol. Pathol. 65 (3), 335e339.

Pavankumar, A.R., Kayathri, R., Murugan, N.A., Zhang, Q.,Srivastava, V., Okoli, C., Bulone, V., Rajarao, G.K., Agren, H.,

2014. Dimerization of a flocculent protein from Moringaoleifera: experimental evidence and in silico interpretation. J.Biomol. Struct. Dyn. 32 (3), 406e415.

Prabhu, K., Murugan, K., Nareshkumar, A., Ramasubramanian, N.,Bragadeeswaran, S., 2011. Larvicidal and repellent potential ofMoringa oleifera against malarial vector, Anopheles stephensiListon (Insecta: Diptera: Culicidae). Asian Pac. J. Trop. Biomed.1 (2), 124e129.

Ragasa, C.Y., Levida, R.M., Don, M.J., Shen, C.C., 2012. Cytotoxicisothiocyanates from Moringa oleifera Lam seeds. Phillip. Sci.Lett. 5 (1), 46e52.

Ramalho de Oliveira, C.F., Luz, L.A., Guedes Paiva, P.M.,Breitenbach Barroso Coelho, L.C., Marangoni, S., RodriguesMacedo, M.L., 2011. Evaluation of seed coagulant Moringaoleifera lectin (cMoL) as a bioinsecticidal tool with potential forthe control of insects. Process Biochem. Amst. Neth. 46 (2),498e504.

Ritger, P.L., Peppas, N.A., 1987. A simple equation for descriptionof solute release. I. Fickian and non-fickian release from non-swellable devices in the form of slabs, sphere, cylinders ordiscs. J. Control. Release 5 (1), 23e26.

Sancar, A., 1996. DNA excision repair. Annu. Rev. Biochem. 65,43e81.

Santos, A.F.S., Luz, L.A., Argolo, A.C.C., Teixeira, J.A.,Paiva, P.M.G., Coelho, L.C.B.B., 2009. Isolation of a seedcoagulant Moringa oleifera lectin. Process Biochem. Amst. Neth.44 (4), 504e508.

Song, Y., Li, G., Thornton, S.F., Thompson, I.P., Banwart, S.A.,Lerner, D.N., Huang, W.E., 2009. Optimization of bacterialwhole cell bioreporters for toxicity assay of environmentalsamples. Environ. Sci. Technol. 43 (20), 7931e7938.

Sutherland, J.P., Folkard, G.K., Grant, W.D., 1990. Naturalcoagulants for appropriate water treatment: a novel approach.Waterlines 8 (4), 30e32.

Sutherland, J.P., Folkard, G.K., Mtawali, M.A., Grant, W.A., 1994.Affordable Water Supply and Sanitation, Colombo, Sri Lanka.

Westedt, U., Kalinowski, M., Wittmar, M., Merdan, T., Unger, F.,Fuchs, J., Schaller, S., Bakowsky, U., Kissel, T., 2007. Poly(vinylalcohol)-graft-poly(lactide-co-glycolide) nanoparticles forlocal delivery of paclitaxel for restenosis treatment. J. Control.Release 119 (1), 41e51.

Zhang, D., Ding, A., Cui, S., Hu, C., Thornton, S.F., Dou, J., Sun, Y.,Huang, W.E., 2013. Whole cell bioreporter application forrapid detection and evaluation of crude oil spill in seawatercaused by Dalian oil tank explosion. Water Res. 47 (3),1191e1200.

Zhang, D., Zhang, X., Wang, Y., Zhou, G., Li, G., 2014. Innovativeslow-release organic carbon-source material for groundwaterinsitu denitrification. J. Environ. Sci.-China Publ..

Zhang, D., Zhao, Y., He, Y., Wang, Y., Zhao, Y., Zheng, Y., Wei, X.,Zhang, L., Li, Y., Jin, T., Wu, L., Wang, H., Davison, P.A., Xu, J.,Huang, W.E., 2012. Characterization and modeling oftranscriptional cross-regulation in Acinetobacter baylyi ADP1.ACS Synth. Biol. 1 (7), 274e283.