Water Research 267 (2024) 122484

Contents lists available at ScienceDirect

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

Tetracycline and quinolone contamination mediate microbial and

antibiotic resistant gene composition in epiphytic biofilms of
mesocosmic wetlands
Okugbe Ebiotubo Ohore a,1, Jingli Zhang b,1 , Sanji Zhou a , Edmond Sanganyado c ,
Ji-Dong Gu d,e,* , Guojing Yang a,**
a
NHC Key Laboratory of Tropical Disease Control, School of Tropical Medicine, Hainan Medical University, Haikou, Hainan 571199, China
b
Department of Clinical Medicine, Shantou University Medical College, Shantou, Guangdong 515041, China
c
Department of Applied Sciences, Northumbria University, Newcastle Upon Tyne NE2 4PB, UK
d
Environmental Science and Engineering Program, Guangdong Technion - Israel Institute of Technology, 241 Daxue Road, Jinping District, Shantou, Guangdong 515063,
China
e
Guangdong Provincial Key Laboratory of Materials and Technologies for Energy Conversion, Guangdong Technion - Israel Institute of Technology, 241 Daxue Road,
Shantou, Guangdong 515063, China

A R T I C L E I N F O A B S T R A C T

Keywords: The fate and ecological impact of antibiotics on aquatic ecosystems have not been properly elucidated in mes­
Antibiotic resistance genes ocosm wetlands scale. This study explored how tetracyclines (TCs, including tetracycline TC and oxytetracycline)
Bacteria and fluoroquinolones (QNs, including ciprofloxacin CIP and levofloxacin) affect mesocosm wetlands vegetated
Epiphytic biofilm
by V. spiralis, focusing on their impact on epiphytic biofilm microbial communities and antibiotic resistance
Eukaryotes
Wetland
genes (ARGs). Results showed that submerged plants absorbed more antibiotics than sediment. Both TCs and QNs
disrupted microbial communities in different ways and increased eukaryotic community diversity in a
concentration-dependent manner (2-4 mg/L for CIP, 4-8 mg/L for TC). TCs mainly inhibited epiphytic bacteria,
while CIP increased bacterial phyla abundance. TC reduced Cyanobacteriota, Acidobacteriota, and Patescibac­
teria but increased Bacillota, Bacteroidota, and Armatimonadota. In contrast, CIP reduced Bacteroidota, Cya­
nobacteriota, and Gemmatimonadota but increased Bacillota, Planctomycetota, and Acidobacteriota. Significant
differences in ARG profiles were observed between QNs and TCs, with TCs having a more substantial effect on
ARGs due to their stronger impact on bacterial communities. Both antibiotics raised ARG levels with higher
concentrations, particularly for multidrug resistance, tetracyclines, trimethoprim, sulfonamides, aminoglyco­
sides, and fosfomycin, emphasizing their role in antimicrobial resistance. The study suggests that antibiotics can
either stimulate or inhibit ARGs depending on their effects on bacterial communities. This study provides key
evidence on the ecological mechanisms underlying the impact of TCs and QNs on epiphytic microbes of meso­
cosm wetlands.

1. Introduction important ecosystems with enriched microbial communities and plant

species, that create a nature-based ecological remediation platform for
Antibiotics are emerging contaminants of global importance, spark­ emerging contaminants. Recently some countries have adopted CWs for
ing significant attention to their potential adverse public and ecological wastewater and sewage treatment due to their rich microbial commu­
health risks (Sun et al., 2017). Constructed wetlands (CWs), are nity for pollutant degradation (Schierano et al., 2020). However,

* Corresponding author at: Environmental Science and Engineering Program, Guangdong Technion - Israel Institute of Technology, 241 Daxue Road, Jinping
District, Shantou, Guangdong 515063, China.
** Corresponding author at: NHC Key Laboratory of Tropical Disease Control, School of Tropical Medicine, Hainan Medical University, Haikou, Hainan 571199,
China.
E-mail addresses: o.e.ohore@hainmc.edu.cn (O.E. Ohore), jidong.gu@gtiit.edu.cn (J.-D. Gu), yangguojing@hainmc.edu.cn (G. Yang).
1
Okugbe Ebiotubo Ohore and Jingli Zhang contributed equally to this work, (co-first Authors).

https://doi.org/10.1016/j.watres.2024.122484
Received 6 July 2024; Received in revised form 10 September 2024; Accepted 18 September 2024
Available online 19 September 2024
0043-1354/© 2024 Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.
O.E. Ohore et al. Water Research 267 (2024) 122484

antibiotic contamination can significantly disrupt biological activities in concentration. Therefore, the objectives of this study were to investigate
CWs, potentially creating hotspots for the spread of antibiotic resistance (1) the fate of antibiotics in water, sediment, and plants in an antibiotics-
genes (ARGs) and posing significant risks to human health and polluted wetland and (2) the mechanism of epiphytic biofilm microbial
ecosystem integrity. Therefore, it is essential to understand the kinetics community changes and ARGs response to tetracyclines and quinolone
and fate of antibiotics in CWs, as well as their impact on nutrient antibiotics exposures. The experiment was conducted in a simulated
removal capacity and microbial communities, to comprehend the wetland, dominated with Vallisneria spiralis and treated with ciproflox­
ecological implications of antibiotic pollution on the wetland micro­ acin (CIP), levofloxacin (LVX), tetracycline (TC), and oxytetracycline
biome and to emphasize the urgency of controlling antibiotic pollution. (OTC). This study provides new insights into the fate of antibiotics in the
Wetland comprises plants, water, sediment, and microorganisms wetland ecosystem.
(biofilm) forming a stable ecosystem. It covers various habitat types,
including marches, floodplains, streams, estuaries, and offshore coastal 2. Materials and methods
areas (Lee et al., 2020). It forms a distinctive ecosystem between the
terrestrial and aquatic environment, and is recognized as an invaluable 2.1. Experimental setup and sample collection
natural resource for human beings due to its high ecological and eco­
nomic value. The component of the wetlands plays a significant role in The experiment was conducted in 2 phases. (1) To investigate the
nutrient removal such as removal of chemical oxygen demand (COD), impact of antibiotics on the nutrient removal capacity of wetlands, a
dissolved organic matter (DOM), suspended solids, total nitrogen, etc. simulated mesocosmic wetland was constructed using 27 plastic barrels
The mechanism of the nutrient removal may be by the oxidative depo­ with lengths, widths, and heights of 67, 48, and 27 cm, respectively, and
sition and biodegradation by microbes. For example, the submerged each barrel contained 10 cm thick sediment, submerged plants, and 85 L
plant releases oxygen, phytochelatins, or phytometallophores produc­ of water (details on materials including plant and chemical purchase are
tion and antimicrobial compounds, from the roots, and can uptake nu­ provided in Text S1). The Sediments were collected from Lake Xuanwu
trients via the roots (Vymazal, 2020), while the microbial component in Nanjing and incubated at room temperature for 7 days. Vallisneria
such as nitrifiers and denitrifiers are responsible for nitrogen removal spiralis plantlets with similar growth were immediately planted and
(Austin et al., 2019). The epiphytic biofilm component is responsible for allowed to acclimatize to the simulated wetlands for 15 days. Simulated
microbial degradation of several organic pollutants such as antibiotics wetlands were divided into two treatment groups and a control group.
and metals, while the water and sediments harbours bacteria and eu­ The first treatment group was classified as the high-concentration group
karyotes for biotrophic interaction (Zhao et al., 2023). In recent years, (H), which received 4 mg/L of ciprofloxacin (CIP), levofloxacin (LVX),
antibiotics have been frequently detected in water and sediment of and 8 mg/L of tetracycline (TC) and oxytetracycline (OTC). The low
wetlands (Li et al., 2019). Unfortunately, the current wastewater treat­ concentration group (L) received 50 μg/L CIP, LVX, TC, and OTC in
ment plants cannot fully remove antibiotics, therefore antibiotics may separate barrels. The control group did not receive any antibiotic
be discharged into wetlands. Our previous studies indicated that tetra­ treatment. The concentration of the high concentration group was based
cycline (Ohore et al., 2021b) and ciprofloxacin (Ohore et al., 2021a) on the intermediate resistant concentration in accordance with the
disrupt the epiphytic microbial community, but the precise mechanisms guidelines of Clinical and Laboratory Standards Institute (CLSI) (2015),
of action between these antibiotic classes remain unclear. Tetracyclines whereas the low concentration group was based on the level of occur­
(TCs) and fluoroquinolones (QNs) are among the widely used antibiotics rence of antibiotics in the natural environment according to Danner
in global livestock, with a high potential for broad dissemination of et al. (2019). Pollution was induced in all treatment groups by adjusting
ARGs (Liu et al., 2021). Resistance to tetracyclines can occur through to 10 mg/L CODCr, 5 mg/L TP, 50 mg/L TN, and 40 mg/L NH4 using
efflux, ribosomal protection, or enzymatic inactivation (Park et al., glucose, KH2PO4, KNO3, and NH4CL stock solutions, respectively. The
2017), while QNs resistance can occur through alterations in the drug wetland design and pollution inducement were according to design
target enzymes and alterations in access to the drug target enzymes parameters from a previous study (Yan et al., 2019). The concentrations
(DNA gyrase and mutations in topoisomerase IV) (Liu et al., 2021). The of nutrients in the overlying water were adjusted only on the initial day
release of TCs and QNs residues is toxic to plants and aquatic organisms of the experiment and monitored every week during the 14 days of
and may exert a stronger inducing effect on biofilms, thus promoting antibiotic exposure. Sampling was performed on Days 2, 4, 7, and 14.
bacterial resistance (Liu et al., 2021). WWTPs are considered hotspots Samples for determination of the fate of the antibiotics were collected
for the release of antibiotics into wetlands (Michael et al., 2013). A from surface water, sediment, and plants. Water samples were collected
recent study has shown that antibiotics, especially tetracyclines, are in 1 L amber glass bottles at days 4 and 14 sampling points. The
persistent, with long half-lives (30–180 days) and high adsorption experiment was conducted in triplicate.
ability in sediment (Tang et al., 2020). The adsorption and transport (2) To further investigate the impact of antibiotics on the microbial
processes of antibiotics may directly or indirectly influence the degra­ community, we created 21 separate microcosmic wetlands following
dation and bioavailability of nutrients in the wetlands. However, a previously described methods. TC and CIP were selected as represen­
comprehensive investigation of the kinetics of antibiotics in the wetland tatives of the tetracycline and fluoroquinolone antibiotic classes. High
is lacking, especially investigating the potential of submerged macro­ dosages of TC, 4 mg/L (sample name: TA), 8 mg/L (TB), and 16 mg/L
phytes in antibiotics treatment for ecological remediation. Additionally, (TC), and CIP, 2 mg/L (CA), 4 mg/L (CB), and 8 mg/L (CC), were
understanding the characteristics of antibiotics in wetlands could pro­ administered to the wetland mesocosms and left undisturbed for 30
vide important insights into the mechanisms of dissolved organic matter days. The concentrations were based on the susceptible (TA and CA),
(DOM) composition and transformation in wetland’s surface water. intermediate (TB and CB), and resistant (TC and CC) standards (Clinical
It is well established that antibiotics can exert a bacteriostatic effect and Laboratory Standards Institute (CLSI), 2015) for antimicrobial sus­
on the epiphytic biofilm microbial community in wetlands, but the un­ ceptibility testing. The control group (CO) did not receive any antibiotic
derlying mechanisms of microbial responses remain poorly understood. treatment. All experiments were conducted in triplicate. Finally,
Besides, the selection pressure imposed on the microbial community epiphytic biofilms were extracted on the 30th day after post-antibiotic
may stimulate the proliferation and transmission of ARGs among bac­ exposure, to determine the impact of high levels of antibiotic contami­
teria, potentially promoting a stimulatory effect on certain bacterial nation on the epiphytic biofilm microbial community and ARGs.
populations. This study proposes that antibiotics can have both stimu­
latory and inhibitory effects on specific microorganisms, significantly
impacting the ecological function of the wetland and the profile of ARGs,
with the effect dependent on the antibiotic’s mode of action (MOA) and

2
O.E. Ohore et al. Water Research 267 (2024) 122484

2.2. Scanning electron microscopy (SEM) and determination of at 80 ◦ C pending DNA extraction. DNA was extracted using the E.Z.N.A.
chlorophyll concentration ® Soil DNA extraction kits (OMEGA, U.S.A) following the manufac­
turer’s protocol. The quality of the extracted DNA was determined by
For SEM, fresh plant leaves of similar growth size were randomly electrophoresis on 1 % agarose gel. DNA quantification and concentra­
selected for SEM analysis and the biofilm was extracted according to the tion were performed using a NanoDrop 2000 UV–Vis spectrophotometer
methods described by Yan et al. (2019). The chlorophyll was determined (Thermo Scientific, USA) and TBS-380.
according to methods described by Nwankwegu et al. (2023). See Text The V4-V5 region 16S rRNA and 18S rRNA were used to determine
S2 for details. the abundance of bacteria and eukaryotic communities according to the
methods described in our previous study (Ohore et al., 2022b). See Text
2.3. Determination of antibiotics concentration S8 for details. The sequencing information was archived in the Sequence
Read Archive of NCBI under accession numbers PRJNA625661 and
The methods for determining the concentration of antibiotics in PRJNA675349.
water, sediments, and plants were based on the methods described in, The procedures for library construction and metagenomic
Zhou et al. (2012) with modifications for plants. Briefly, 20 g of plant sequencing were conducted following the procedures outlined in our
samples were obtained and dried under sunlight to remove chlorophyll, prior research (Ohore et al., 2021a). Functional annotation of ARGs was
followed by freeze-drying at 100 ◦ C. The plants were then ground and performed by conducting BLASTP searches (version 2.2.28) against the
filtered through a sieve with a 1 mm mesh size before extraction. In­ Comprehensive Antibiotic Resistance Database (CARD) (Wang et al.,
ternal standard methods were used for quality assurance, and recovery 2020). The metagenomic dataset was archived in the NCBI Sequence
rates ranged from 95 % to 105 %. Detailed information is provided in Read Archive under the accession numbers PRJNA633302 and
Text S3–S6. PRJNA625849.
Antibiotic degradation in surface water was assessed using expo­
nential data with a pseudo-first-order equation (Kiki et al., 2020). Eq. 2.6. Statistical analysis
(1).
( ) The Inferential statistical analysis was conducted based on one-way
C0 ANOVA, and Fisher’s exact test was used to check for statistical signif­
ln = kt (1)
Ct icance using STAMP bioinformatics software. Statistical significance
Where Co is the concentration of the antibiotics on the initial day and was set at P < 0.05. FastTree (version 2.1.3) and R language were used
Ct is the concentration of the antibiotics at time “t” intervals. The for phylogenetic evolutionary sequencing. Origin software (version
degradation rate/constant (k) and half-life (t1/2) were determined based 9.65) was used to generate graphs. ANOSIM, hierarchical cluster anal­
on Eqs. (2) and (3), respectively. ysis, principal component analysis, and Partial Least Squares Discrimi­
( )/ nant Analysis (PLS-DA) were conducted to investigate the similarities or
C0 distinctions in the bacterial and eukaryotic community structures at the
k = − ln t (2)
Ct OTU level using the Bray-Curtis distance algorithm (See Text S9 for
details).
ln (2)
t1/2 = (3)
k 3. Results and discussion
The accumulation kinetics of the antibiotic was according to the
model described previously by Hu et al. (2012). Eqs. (4), (5) 3.1. The behaviour of antibiotics in wetland ecosystems

Cs = Cmax ⌊1 − exp (− kt)⌋ (4)

The environmental behaviour of antibiotics in wetlands has not been
( ) thoroughly studied. In this study, we used a first-order pseudo-kinetic
− ln 1 − Cs
Cmax model to track the biodegradation and bioaccumulation of antibiotics in
k= (5) a simulated wetland system (Figs. 1 and S1). The results showed that the
t
concentration of antibiotics in the water column decreased significantly
where Cmax is the concentration of a pollutant in the sediment/plant over time. Indicating effective removal from the surface water. How­
when the transfer of the pollutant reaches the equilibrium and Cs is the ever, antibiotic levels increased significantly (P < 0.05) in plants and
concentration of a pollutant in the sediment/plant at time t. Based on the sediment as time progressed, suggesting that these environments facil­
constant k, the half-life (T0.5) of antibiotic contamination was expressed itate the adsorption of antibiotics and play crucial roles in their fate
as Eq. (6). within wetlands (Mu et al., 2024).
Among the QNs treatment, CIP concentrations were higher than LVX
ln (2) in the water, sediment, and plants. For TCs, TC concentrations were
T0.5 = (6)
k higher than oxytetracycline (OTC) in these same environments. A study
by Mu et al. (2024) demonstrated that H. verticillate could adsorb nor­
2.4. Determination of dissolved organic matter and PARAFAC modelling floxacin (NOR) and ofloxacin (OFL) from the surface water of wetlands.
Building on this study, our research demonstrated that submerged
The DOM in the water samples was determined according to the macrophyte growth can create new surfaces for the adsorption of anti­
methods described in Yang et al. (2020). The staRdom and ggplot2 biotics, including tetracyclines and quinolones, thereby acting as a
packages in R were used to create the PARAFAC model of EEMS. See nature-based concentrator. Nevertheless, However, the accumulation of
Text S7 for details. these antibiotics poses potential risks to plant health.
To investigate the dynamics of antibiotic pollutants, we used the
2.5. Biofilm collection, DNA extraction and amplicon and metagenomic pseudo-first-order kinetic parameters, including the rate constant k/
sequencing d and half-life t1/2:d (Fig. S1 and Table S1). The physical and chemical
properties are listed in Table S2. Our result showed that the biodegra­
The epiphytic biofilm was extracted from 40 g of fresh leaves on the dation constant k or removal rate of antibiotics in water decreased over
30th day of antibiotic exposure following the method described by Yan time. Generally, higher concentrations led to a higher biodegradation
et al. (2019). The biofilm was weighed and stored in a collection tube rate, likely due to increased competition among microbial species and

3
O.E. Ohore et al. Water Research 267 (2024) 122484

Fig. 1. The concentration of the antibiotics (a-i) detected in water, sediment, and plants on the 4th day and 14th day post-antibiotics exposure. The initial con­
centration for the high-concentration group (H) were 4 mg/L for ciprofloxacin (CIP) and levofloxacin (LVX); and 8 mg/L for tetracycline (TC) and oxytetracycline
(OTC). While the low-concentration group (L) received an initial concentration of 50 μg/L of CIP, LVX, TC, and OTC.

the evolution of more efficient degradation pathways (Han et al., 2020). particularly in high-pollution environments, shows its potential as a
Within the first four days, the biodegradation constant for natural and cost-effective solution for antibiotic treatment and
high-concentration levofloxacin (H-LVX) was 1.89, higher than cipro­ remediation.
floxacin (CIP) at 1.53. For the low-concentration groups, the constants The persistence of antibiotics in various environmental media was
were 1.44 for low-concentration levofloxacin (L-LVX) and 1.27 for investigated. There was a significant difference (p < 0.05) in the half-life
low-concentration ciprofloxacin (L-CIP), indicating that LVX degraded of antibiotics in the water, sediment, and plants (Table S1). In water, the
faster than CIP. However, after 14 days, the degradation constant was half-life (days) was highest in the H-CIP (43.06 days) group, followed by
higher in the low-concentration group compared to the H-LVX (28.69)> L-CIP (10.19)> H-OTC (9.34) >L-LVX (9.13)>L-OTC
high-concentration group. (7.53)> L-TC (3.95)>H-OTC (3.29). QNs generally had longer half-lives
Our study also showed that antibiotics were adsorbed by both sedi­ than TCs, indicating greater persistence in aquatic environments. This
ment and plants. Antibiotic accumulation in sediment increased as extended persistence can lead to prolonged exposure for aquatic or­
concentration decreased, while the opposite trend was observed in ganisms, potentially disrupting ecosystems and promoting the devel­
plants. Comparison of the bioaccumulation constant demonstrated that opment of antibiotic resistance genes (ARGs).
the adsorption constant in H-CIP (0.11 vs 0.02), H-LVX (0.26 vs 0.05), H- In the sediment, the half-life was also the highest in the H-CIP (45.25
TC (0.07 vs 0.03), H-OTC (0.09 vs 0.04), and L-CIP (0.09 vs 0.07) d) group, followed by the H-TC (20.48) > H-OTC (17.93) > H-LVX
treatment groups were higher in plant compared with sediment. The (14.13) > L-OTC (12.25) >L-OTC (10.29) > L-CIP (9.58) > L-LVX
reverse trend was observed for L-TC (0.05 vs 0.07) and L-OTC (0.03 vs (7.58). The half-lives of the antibiotics followed the order L-OTC (19.85)
0.06). This suggests that Vallisneria spiralis exhibits a higher affinity for > L-TC (13.67)> H-TC (10.00) > L-CIP (7.70) > H-OTC (7.45) > L-LVX
antibiotic uptake compared to sediment. Its high adsorption rate, (7.37) > H-CIP (6.53) > H-LVX (2.71). Overall, CIP had a higher half-life

4
O.E. Ohore et al. Water Research 267 (2024) 122484

in sediment compared to water and plants. LVX showed longer half-lives thrive (Ohore et al., 2021a; Sridhar et al., 2021). In contrast, the resis­
in water than in sediment or plants, while tetracyclines and oxytetra­ tant group (TC and CC) exhibited opposite patterns suggesting that the
cyclines had longer half-lives in sediment compared to water or plants. bacterial populations in the TC group have developed or acquired a
The extended half-lives of antibiotics in sediment indicate that these wider range of resistance mechanisms that allow them to persist (Fu
compounds can accumulate in sediments, potentially impacting et al., 2017). The reason for this differential impact may be related to the
epiphytic microbes. This highlights the importance of considering the MOA of the antibiotics. Therefore, based on the study findings and the
persistence of environmental pollutants when assessing their ecological potential mechanisms at play, this study revealed that TC disrupted the
effects (Albero et al., 2018). bacterial community structure more significantly.
For the eukaryotic community, the alpha diversity showed a signif­
3.2. Impact of antibiotics on epiphytic biofilm microbial community and icant (P < 0.05) increase in their alpha diversity in the antibiotic
nutrient removal capacity of wetlands treatment group compared with the control group (Table S3). This
increased diversity could indicate a more varied eukaryotic community
3.2.1. Tetracycline and ciprofloxacin disrupted the epiphytic microbes using taking advantage of the newly available resources (Yan et al., 2020).
different modes of action However, there was a slight decrease in eukaryotic diversity in the TC
We adopted SEM to examine the aggregation of epiphytic biofilms 30 and CC groups, likely due to antibiotic toxicity to eukaryotic, demon­
days post-antibiotics exposure. The results showed that increasing con­ strating a concentration-dependent stimulation of the eukaryotic com­
centrations of TC and CIP caused a drastic disruption of the epiphytic munity. This concentration-dependent response suggests that at certain
biofilm microbial communities (Fig. 2). This suggests that antibiotics levels, antibiotics may negatively impact eukaryotic organisms, poten­
affected the aggregation of microbes and the composition of epiphytic tially due to direct toxicity or indirect effects through altered in­
biofilms, although the specific mechanisms and modes of action (MOA) teractions with bacterial communities.
on these microbes remain unclear. TC and CIP, representing the tetra­ The ANOSIM, hierarchical cluster, principal component analysis,
cyclines (T) and quinolones (C) antibiotics classes based on the sus­ and Partial Least Squares Discriminant Analysis (PLS-DA) of the bacteria
ceptible (TA and CA), intermediate (TB and CB), and resistant (TC and and eukaryotic communities (Fig. 3) also showed significant differences
CC) concentration by Clinical and Laboratory Standards Institute (CLSI) in the sample groupings and clustering. CIP and TC appeared to selec­
(2015) standards for microbial susceptibility testing were selected for tively target specific bacterial species, affecting the overall bacterial
further microbial analysis. community composition and indirectly influencing eukaryotic diversity.
We employed alpha and beta diversity analysis including; richness This demonstrates that the TC and CIP impacted the epiphytic bacteria
(sobs, ace, chao 1, and jack index), diversity (Shannon, Simpson, community through different modes of action. For instance, Tang et al.
invsimpson, qstat, and pd), evenness (heip, simpsoneven, and (2024a) differences in antibiotic MOA can lead to significant shifts in
smithwilson), and coverage to unravel the impact of the antibiotics on microbial community composition, as specific microbes targeted by
the epiphytic microbes (Table S3). The results showed that there was a antibiotics may create competitive advantages for other species. How­
significant difference among the antibiotic treatment groups (P < 0.05), ever, the specific effects of TC and CIP on the epiphytic microbial
with both groups also differing significantly from the control group (P < communities of wetlands remain unclear. Shifts in microbial community
0.05) (Fig. 3). CIP treatment groups (CA and CB) showed higher bac­ structure can affect key ecological functions, such as nutrient cycling
terial abundance compared to TC groups (TA and TB). This higher and organic matter decomposition, potentially altering wetland health
abundance in CIP-treated groups may be due to the differential sus­ and resilience. Understanding these shifts can help predict changes in
ceptibility of bacterial species to CIP, which allows certain species to ecosystem dynamics due to antibiotic presence and guide strategies for

Fig. 2. Scanning electron microscopy (SEM) images of epiphytic biofilm demonstrating the impact of tetracycline at 0 mg/L (a), 4 mg/L (b), 8 mg/L (c), and 16 mg/L
(d), and ciprofloxacin at 0 mg/L (e), 2 mg/L (f), 4 mg/L (g), and 8 mg/L (h) on the aggregation of microbial communities in biofilms. The antibiotics significantly
inhibited the formation of epiphytic biofilms and microbial communities on the leave surfaces.

5
O.E. Ohore et al. Water Research 267 (2024) 122484

Fig. 3. Beta diversity analysis of epiphytic biofilm (a-e) bacterial and (g-k) eukaryotic community treated with antibiotics using (a and g) ANOSIM analysis and (b
and h) Partial Least Squares Discriminant Analysis (PLS-DA) at phylum taxonomic level. The (c and i) PCA, (d and j) hierarchical cluster analysis and (e and k)
distance heatmap were used for the analysis of the bacteria and eukaryotic beta diversity of operational taxonomic units (OTUs based on Bray-Curtis distance al­
gorithm. While (f and i) shows the abundance of OTUs in the samples. TA and CA; TB and CB; and TC and CC denote the susceptible, intermediate, and resistant
dosage of tetracycline and ciprofloxacin that were administered into the wetland mesocosm. The control group (CO) received no antibiotics treatment.

6
O.E. Ohore et al. Water Research 267 (2024) 122484

Fig. 4. Composition of the bacterial community and the changes in the epiphytic microbial community structure following a series of antibiotics exposure at
susceptible (TA and CA), intermediate (TB and CB) and resistant (TC and CC) concentrations. The samples were compared with the (a) control groups (CO) and
among (c-d) antibiotic exposure groups using Fisher’s exact test. The bacteria with a significant difference (P < 0.005) demonstrated that the antibiotics caused a
significant shift in the bacterial composition. (e) and (f) shows the composition of the bacteria at the phylum and genus taxonomic level respectively.

7
O.E. Ohore et al. Water Research 267 (2024) 122484

Fig. 4. (continued).

managing antibiotic pollution in natural environments. shows that the changes in the microbial community composition were
Notably, inferential comparative analysis of the bacteria community clearly sensitive to antibiotics concentration.
composition showed that susceptible concentrations of TC (TA group) Finally, in the resistant group (TC and CC), TC and CIP generally
mostly decreased the bacteria phyla like Bacteroidota, Cyanobacteriota, decreased the abundance except for Bacillota, Bacteroidota, Armati­
and Acidobacteriota, while increasing the abundance of Bacillota, monadota, and Candidatus Dependentiae in TC treatment and Actino­
Actinomycetota, and Gemmatimonadota (P < 0.05) (Fig. 4). In contrast, mycetota, Planctomycetota, Verrucomicrobiota, Armatimonadota,
susceptible CIP (CA) concentrations generally increased the abundance Chlamydiota, and Candidatus Dependentiae in CIP treatment. These
of several bacterial phyla, with nine of 15 phyla, including Bacillota, results demonstrated that TC significantly inhibited bacterial abundance
Planctomycetota, and Actinomycetota, showing significant increases (P at all concentrations, while the epiphytic microbial community exhibi­
< 0.05) in the CIP treatment groups. This suggests that susceptible doses ted some resistance to CIP, although increased CIP concentrations had a
of CIP stimulated these bacterial phyla, indicating resistance to CIP, stronger inhibitory effect.
while Bacteroidota, Cyanobacteriota, Gemmatimonadota, Patescibac­ The summary of the mechanisms of the antibiotics on the epiphytic
teria, and Chloroflexota were significantly inhibited (P < 0.05). These biofilm eukaryotic community is provided in Fig. 6a. Based on the evi­
findings show that while susceptible doses of TC significantly inhibited dence provided in this study, the results demonstrated that TC generally
the bacterial community, CIP had a stimulatory effect. caused a significant reduction in the abundance of Cyanobacteriota,
In the intermediate treatment groups (TB and CB), 13 and 15 phyla, Acidobacteriota, Patescibacteria group, Planctomycetota, Chloro­
respectively, showed significant differences in response to TC and CIP flexota, Verrucomicrobiota, and Chlamydiota, but stimulated the
exposure. Intermediate concentrations of TC caused a drastic reduction abundance of phylum Bacillota, Bacteroidota, and Armatimonadota (P
in the abundance of most bacterial phyla, except for Bacillota, Bacter­ < 0.05). Conversely, CIP significantly decreased the relative abundance
oidota, and Armatimonadota. In the CIP group, an increase in the CIP of the phylum Bacteroidota, Cyanobacteriota, Gemmatimonadota,
concentration resulted in a stronger inhibitory effect compared to sus­ Bacillota, Patescibacteria group, and Chloroflexota, but stimulated the
ceptible concentrations, though it still increased the abundance of relative abundance of Bacillota, Planctomycetota, Actinomycetota,
Bacillota, Actinomycetota, Planctomycetota, Verrucomicrobiota, etc. Acidobacteriota, Verrucomicrobiota, Armatimonadota, Chlamydiota,
Except for Bacteroidota, Cyanobacteriota, Acidobacteriota, Gemmati­ and Candidatus Dependentiae (P < 0.05). Our findings revealed that TC
monadota, Bacillota, Patescibacteria group, and Deinococcota. This exhibited a more pronounced bacteriostatic effect on epiphytic microbes

8
O.E. Ohore et al. Water Research 267 (2024) 122484

Fig. 5. Composition of the eukaryotic community and the changes in the epiphytic microbial community structure following a series of antibiotics exposure at
susceptible (TA and CA), intermediate (TB and CB) and resistant (TC and CC) concentrations. The samples were compared with the (a) control groups and among (c-
d) antibiotic exposure groups using Fisher’s exact test. The eukaryotes with a significant difference (P < 0.005) demonstrated that the antibiotics caused a significant
shift in the eukaryotic composition. (e) and (f) shows the composition of the eukaryotes at the kingdom and phylum taxonomic level respectively.

9
O.E. Ohore et al. Water Research 267 (2024) 122484

Fig. 5. (continued).

compared to CIP, while CIP contributed to an increased abundance of _k_Alveolata. Similarly, CIP exposure significantly increased the abun­
bacterial phyla within the biofilm. dance of Mollusca, Cercozoa, Chytridiomycota, Choanoflagellida,
The responses of the eukaryotic community to susceptible, inter­ Cryptomycota, Neocallismastigomycota, Protalveolata, LKM15, and
mediate and resistant concentrations of the antibiotics are provided in Peronosporomycetes, while significantly decreasing the abundance of
Figs. 5 and 6b. This study showed that exposure to TC increased the norank_Chloroplastida, Phragmoplastophyta, Ochrophyta, Rotifera,
abundance of phyla such as Arthropoda, Phragmoplastophyta, Gastro­ Annelida, unclassified_d_Eukaryota, unclassified_k_Metazoa_Animalia,
tricha, Rotifera, Annelida, Cercozoa, Chytridiomycota, Protalveolata, Platyhelminthes, Porifera, unclassified_k_Alveolata, and Ciliophora.
Coanoflagellida, Neocallismastigomycota, LKM15. While decreasing the These results showed the distinct effects of TC and CIP on the eukaryotic
abundance of norank_Chloroplastida, Mollusca, Ochrophyta, uncl. community. TC primarily increased the abundance of various phyla,
_d_Eukaryota, uncl._k_Metazoa_Animalia, Porifera, Ciliophora, uncl. including Arthropoda and Rotifera, while CIP tends to increase the

10
O.E. Ohore et al. Water Research 267 (2024) 122484

(caption on next page)

11
O.E. Ohore et al. Water Research 267 (2024) 122484

Fig. 6. Illustration of the mechanisms of action of TC and CIP on the epiphytic biofilm (a) bacterial and (b) eukaryotic phyla. The data regarding the impact of the
antibiotics at various concentrations were pools and were used to perform a directed network analysis based on Fisher’s exacts test analysis at P < 0.001. The edges
indicated the mechanisms of action of the antibiotics, with the red lines (edges) indicating an inhibitory effect, and the black lines indicating a stimulatory effect on
the bacterial phyla and eukaryotic kingdoms represented as nodes. C-f shows the Network analysis of co-occurrence relations in microbial communities demon­
strating the impact of tetracycline (c and e) and ciprofloxacin (d and f) on the (c and d) bacterial and (e and f) eukaryotic and the interspecies interaction using
operational taxonomic units (OTUs) subclassified based on the bacterial and eukaryotic phyla. The size of each node is determined by the average degree, which
represents the number of connections and the node’s level of influence. Nodes of the same colour belong to the same phylum within the bacterial and eukary­
otic community.

abundance of Mollusca and Cryptomycota. This suggests that these an­ more conducive environment for eukaryotes than TC.
tibiotics may stimulate growth or enhance the resilience of certain taxa The network analysis identified specific keystone species: 12 and 15
in response to their presence (Tang et al., 2024b). phyla were keystones in the bacterial communities under TC and CIP
To understand the differences in the ecosystem structural changes treatments, respectively, while 25 keystone phyla were present in the
due to TC and CIP, we conducted a network analysis following the eukaryotic communities for both treatments. The top six keystone phyla
methods described in our previous study (Ohore et al., 2022a). The to­ in the TC treatment for bacteria were Pseudomonadota, Bacteroidia,
pological characteristics of the network are provided in Tables S4–S6. Bacillota, Actinomycetota, Gemmatimonadia, and Acidobacteriota
For bacteria community, surprisingly, the network nodes, edges, and (Fig. 6c). In contrast, in the CIP treatment, the most influential keystone
average weighted degree, were higher in the TC treatment than CIP phyla were Pseudomonadota, Bacteroidia, Actinomycetota, Plancto­
treatment, but the modularity and connected components were in mycetia, Verrucomicrobiota, and Bacillota (Fig. 6d). For eukaryotes, the
reverse. This suggested that TC increased the selection pressure, acti­ top six keystone phyla in the TC treatment were norank_k__­
vating diverse species as a survival mechanism. The increase in modu­ Chloroplastida, Phragmoplastophyta, Cercozoa, Ochrophyta, Rotifera,
larity and connected components in the CIP treatment indicates the and Chytridiomycota (Fig. 6e). In the CIP treatment, the trends were
stronger influence of environmental factors on the bacterial community similar but with a slight variation in order: norank_k__Chloroplastida,
under CIP exposure (Scheffer et al., 2012), demonstrating the impact of Cercozoa, Ochrophyta, Phragmoplastophyta, Chytridiomycota, and
the antibiotics on the bacterial communities. In contrast, the microbial Rotifera (Fig. 6f). Likewise, the phylogenetic tree showed differences in
community structure and interactions were stronger in the CIP treat­ the microbial assemblages due to antibiotics (Fig. S2). These findings
ment for the eukaryotic community compared to TC. Previous reports illustrate how TC and CIP disrupt the microbial communities within
have shown that antibiotic exposure can shift ecosystems toward a epiphytic biofilms, each through different mechanisms.
predatory food web (Li et al., 2024, 2023), indicating that CIP created a

Fig. 7. shows the consequence of the TC and CIP exposures on the relative abundance of efflux pump genes and resistance genes (ARGs copies per copy of 16S rRNA
gene). The genes were classified and the top 10 occurring resistance genes in each antibiotic resistance gene class were provided.

12
O.E. Ohore et al. Water Research 267 (2024) 122484

3.2.2. Impact of antibiotics exposures on the resistome profiles efflux pump genes (e.g., ceoB, tetG) were significantly increased (p <
In this study, the metagenomic data were searched against the CARD 0.05). Meanwhile, glycopeptides (e.g., vanRO), polymyxins (e.g., PmrE),
database to screen the ARGs in the epiphytic biofilm samples, and a total elfamycins (e.g., Escherichia coli EF-Tu mutants), aminocoumarins (e.g.,
of 459 ARGs and 184 efflux pump genes. A total of 18 ARGs types were Staphylococcus aureus gyrB), isoniazid (e.g., Mycobacterium tuberculosis
obtained, corresponding to 127 bata lactam genes, 48 aminoglycoside, katG mutations), and fluoroquinolones (e.g., Salmonella serovars parE)
48 glycopeptide, 33 fluoroquinolone, 12 tetracycline, 12 rifampin, 11 were significantly decreased.
trimethoprim, 8 aminocoumarin, 8 MDR, 7 fosfomycin, 7 polymyxin, 6 In the TC group, ARGs for macrolides (e.g., Mrx, EreB), chloram­
peptide, 6 chloramphenicol, 5 elfamycin, 4 streptogramin, 4 macrolide, phenicols (e.g., cmlv, catB9), trimethoprim (e.g., dfrG), sulfonamides (e.
4 isoniazid, 3 sulfonamide. There were significant differences in the g., sul1), aminoglycosides (e.g., APH(3′)-Ib), tetracyclines (e.g., tetS,
relative abundance of ARGs between the treatment groups and the tetQ), fosfomycin (e.g., Mycobacterium tuberculosis murA), and multidrug
control (p < 0.05) and the total abundance of ARGs was significantly resistance (e.g., clbC) were significantly increased (p < 0.05).
increased with increasing antibiotics concentration (Fig. 7). Conversely, glycopeptides (e.g., vanRM), elfamycins (e.g., Escherichia
Interestingly, the antibiotic treatment had both inhibitory and coli EF-Tu mutants), rifampin (e.g., Mycobacterium leprae rpoB mutants),
stimulatory effects on certain ARGs. Susceptible concentration of CIP aminocoumarins (e.g., Staphylococcus aureus gyrB), beta-lactams (e.g.,
(CA) significantly influenced the relative abundance of Efflux pump GES-13), isoniazid (e.g., Mycobacterium tuberculosis katG mutations),
genes and 12 ARG classes. The relative abundance of glycopeptide (e.g., fluoroquinolones (e.g., Salmonella serovars parE), streptogramins (e.g.,
vanRF, vanRM, vanD), aminocoumarin (e.g., cysB, parY, kdpE), poly­ vatB), peptides (e.g., Bacillus subtilis mprF), and polymyxins (e.g., PmrB)
myxin (e.g., PmrC, pmrE, pmrA), and fluoroquinolone (e.g., Propioni­ were significantly decreased.
bacterium acnes gyrA, Bartonella bacilliformis gyrA) ARGs significantly The data reveal that TC had the most significant effect on ARG
decreased. In contrast, multidrug resistance (e.g., Erm(35), clba, clbB), profiles compared to CIP, likely due to its stronger impact on the bac­
tetracycline (e.g., tetT, tet32, tet36), trimethoprim (e.g., dfrE, dfrA3, terial community. Both antibiotics increased ARG levels with higher
dfrA14), fosfomycin (e.g., Chlamydia trachomatis murA), sulfonamide (e. concentrations. ARGs related to multidrug resistance, tetracyclines,
g., sul1), elfamycin (e.g., Clostridium difficile EF-Tu mutants), and efflux trimethoprim, sulfonamides, aminoglycosides, and fosfomycin consis­
pump (e.g., ceoB, tetV, mtrD) genes were significantly increased with TA tently increased across all treatment groups, highlighting their role in
treatment. Additionally, the abundance of chloramphenicol (e.g., cmlv, antimicrobial resistance. Efflux pump genes also increased in both sus­
catB7, catB10), trimethoprim (e.g., dfrE, dfrA3), aminoglycoside (e.g., ceptible and resistant groups. This suggests that antibiotics can either
AAC(3)-VIIIa, AAC(3)-Ib), tetracycline (e.g., tet44, tetS), macrolide (e.g., stimulate or inhibit ARGs depending on their effect on the bacterial
mphG, EreB), fosfomycin (e.g., Escherichia coli mutant murA), and poly­ community (Cunningham et al., 2020). The selection pressure from
myxin (e.g., arnA, pmrC) ARGs, along with efflux pump genes (e.g., antibiotics may cause bacteria to either die off or acquire new ARGs,
opmH, mtrC), increased. Conversely, glycopeptide (e.g., vanSO, vanD), affecting ARG abundance. Overall, the study indicates that TC has a
rifampin (e.g., Mycobacterium leprae rpoB mutants), beta-lactam (e.g., greater potential than CIP to drive ARG proliferation in the microbial
GES-13, CTX-M-59), peptide (e.g., Bacillus subtilis mprF), and isoniazid communities of wetland epiphytic biofilms.
(e.g., Mycobacterium tuberculosis katG) ARGs significantly decreased (p <
0.05). 3.2.3. Dynamics of electrical conductivity of the wetland’s surface water
In the CB groups, ARGs for multidrug resistance (e.g., clbA, Erm(36), during antibiotics exposure
clbB), macrolides (e.g., Mrx, EreB, mphG), tetracyclines (e.g., tetT, tetS, The impact of antibiotics on wetlands’ water quality and nutrient
tetQ), chloramphenicols (e.g., cmlv, catB3), trimethoprim (e.g., dfrE, removal capacity was investigated. The results showed that EC was
dfrA14, dfrA5), sulfonamides (e.g., sul1), aminoglycosides (e.g., AAC(3)- significantly (p < 0.05) increased in the antibiotics treatment groups,
Ic, AAC(3)-IIIc, AAC(6′)-If), and fosfomycin (e.g., Chlamydia trachomatis especially for the tetracycline group (TCs; including tetracycline (TC)
murA, Escherichia coli mutant GlpT) were significantly increased (p < and oxytetracycline (OTC)) which showed the highest increase
0.05). In contrast, ARGs for glycopeptides (e.g., vanRO, vanRM), compared to the quinolone groups (QNs; including ciprofloxacin (CIP)
rifampin (e.g., Mycobacterium leprae rpoB mutants), aminocoumarins (e. and levofloxacin (LVX) at high (H) and low (L) dosage) (Fig. S3). The
g., Staphylococcus aureus gyrB), beta-lactams (e.g., GES-13), peptides (e. highest increase in the electrical conductivity (EC) was recorded in the
g., Brucella suis mprF), fluoroquinolones (e.g., Salmonella serovars parE), H-TC group which demonstrated a 2 % increase followed by H-OTC>L-
isoniazid (e.g., Mycobacterium tuberculosis katG mutations), polymyxins OTC>H-LVX>L-CIP>L-TC>L-LVX>CT and having EC values (percent­
(e.g., PmrE), elfamycins (e.g., Escherichia coli EF-Tu mutants), and age increase) of 865.25 (1.6 %), 858.44 (1.5 %), 850.62 (1.4 %), 838.75
streptogramins (e.g., vatB) were significantly decreased (p < 0.05) (1.2 %), 830.83 (1.1 %), 822.33 (1.0 %) μs/cm, respectively compared
compared to the control group. with control CT (745.33 μs/cm). Noteworthy the dynamics of the EC
In the TB group, ARGs for tetracyclines (e.g., tetT, tetS, tetQ), chlor­ showed a concentration-specific response in all antibiotics treatment
amphenicols (e.g., catB9, catB10), trimethoprim (e.g., dfrA26, dfrG), groups when comparing the high concentration and the low concen­
aminoglycosides (e.g., PmrE, PmrA), macrolides (e.g., EreB, Mrx), fos­ tration group of the respective antibiotics. The increase in the EC of the
fomycin (e.g., Escherichia coli mutant murA), sulfonamides (e.g., sul1), water column may be ascribed to the presence of a relatively higher
and efflux pump genes (e.g., mdtf, coB) were significantly increased (p < amount of nutrient and organic matter in the antibiotics treatment
0.05). Conversely, the abundance of ARGs for glycopeptides (e.g., groups likely due to nutrient retention caused by diminished nutrient
vanRD, vanSO), polymyxins (e.g., PmrE, PmrA), rifampin (e.g., rgt1438), degradation capacity by microbes (Ramalingam et al., 2020; Si et al.,
aminocoumarins (e.g., Staphylococcus aureus gyrB), beta-lactams (e.g., 2020).
CTX-M-59), streptogramins (e.g., vatF), peptides (e.g., Bacillus subtilis
mprF), fluoroquinolones (e.g., Salmonella serovars parE), and isoniazid 3.2.4. The impact of antibiotics on ammonia removal by wetlands was
(e.g., Mycobacterium tuberculosis katG mutations) were significantly dependent on the antibiotic concentration
decreased. For ammonia removal, the highest ammonia removal rate was
Finally, in the resistant groups, the relative abundance of macrolide recorded in the control group (CT) at 73.7 % while the lowest was
(e.g., Mrx, EreB), rifampin (e.g., Mycobacterium leprae rpoB mutants), recorded in the LVX treatment (Fig. S4). On the contrary, the cipro­
tetracycline (e.g., tetO, tetQ), chloramphenicol (e.g., cmlv, catB10), beta- floxacin class (CIP) did not significantly impact on the ammonia removal
lactam (e.g., SPG-1), trimethoprim (e.g., dfrE, dfrA14), sulfonamide (e. rate. A removal rate of 71.3 % and 44.3 % was recorded for L-CIP and H-
g., sul1), peptide (e.g., tsnr), aminoglycoside (e.g., AAC(3)-Ib), strep­ CIP, respectively. The trend of the ammonia concentration (mg/L) and
togramin (e.g., vatF), and multidrug resistance (e.g., CIPa) ARGs, and removal rates (%) includes CT (10.5 mg/L; 73.7 %)> L-CIP (11.5 mg/L;

13
O.E. Ohore et al. Water Research 267 (2024) 122484

71.30 %)> L-OTC (15.3 mg/L; 61.63 %)> H-CIP (22.37 mg/L; 44 %)> respectively. The increased in the chlorophyl content may be ascribed to
H-TC (25.2 mg/L; 37 %)> H-OTC (27.8 mg/L; 35.86 %)> L-TC (29.2 the impact of the antibiotics on the eukaryotic community. This
mg/L; 30.58 %)> H-LVX (30.1 mg/L; 26.99 %)> L-LVX (32 mg/L; 24.76 demonstrated that the antibiotics promoted the photosynthesis and
%). Therefore, this data demonstrated that antibiotics inhibited the respiration rate of algae, especially for the TCs group compared to QNs.
nitrification process in the water column with the strongest impact
recorded for LVX than CIP. The mechanism for the diminished ammo­ 3.2.7. Antibiotics increased the DOM concentration in wetlands by
nium removal efficiency may be attributed to the inhibition of ammonia- destruction of plant tissue
oxidizing bacteria growth (AOB) and our findings are in accordance with This study showed that the DOM significantly increased with an in­
Tu et al. (2019). This study also demonstrated a crease in the antibiotic’s concentration (Figs. S7–S13). The content of
concentration-dependent response since higher removal rates were the DOM was mainly humic acid-like (V) substance and fulvic acid-like
mostly recorded in the low-concentration group compared to the substance (III), with humic-like substance having the highest intensity,
high-concentration groups. with higher intensity in the QNs groups compared with the TCs. Humic
materials are primarily formed through microbial degradation of plant
3.2.5. Quinolones demonstrated stronger impact on COD contents than residues (Dong et al., 2020). This demonstrated that the antibiotics
tetracyclines caused apoptosis of some plant cells, accumulation, and transformation
For COD removal, the results show that the concentration of COD of plant residues in the water column thereby increasing the proportion
was significantly (p < 0.05) higher in the antibiotics treatment group of humic substances in the DOM.
compared with control (CT) (Fig. S5). Noteworthy, the highest COD To further understand the impact of antibiotics on the dynamics of
concentration was recorded in the QNs group compared with the TCs DOM in wetlands, we adopted the PARAFAC model. This study is the
group. Among the QNs group, there was no significant difference in the first to integrate antibiotic pollution analyses in wetlands to characterize
COD concentration between the high-concentration groups. An average the distribution and transformation of DOM. According to Fellman et al.
concentration of 45.75 and 45.12 mg/L was recorded in H-LVX and H- (2010), C1 corresponds to humic-like or fulvic acid-like substance
CIP respectively. Astonishingly, the low-concentration group of the QNs (Ex:290–325 Em: 370–430) with low molecular weight, common in
antibiotics also showed higher COD concentration compared with the wetland environments associated with biological activity, C2 and C3
TCs groups except for H-OTC (32.4 mg/L) which was slightly higher correspond to humic-like substance (Ex 320–360, Em 420–460) with
than L-LVX (31.9 mg/L). Among the TCs groups, H-TC showed the high-molecular-weight humic, widespread, but highest in wetlands
highest increase in the COD concentration, followed by H-OTC. This (Fig. S11c). Noteworthy the C1 component had the highest fluorescent
trend was observed between the L-TC and L-OTC, thus indicating that TC values followed by component C2 in all samples. While component C3
increased the COD concentration of the water column more than OTC. was mostly detected in the QNs group. Some uncommon com­
These results showed that antibiotics increased the COD of the water pounds/residues were found in H-TC and H-OTC groups and were not
column and the impact is relative to the antibiotics type and concen­ taken into account in the PARAFAC model (Fig. S12). The fluorescence
tration of the antibiotics. The difference in the COD content between the was higher in the QNs group, especially the high-concentration groups,
antibiotic treatment groups may be ascribed to the degradation of the and the fluorescence increased with increasing days (Fig. S13b). In
antibiotics. Pollutant influx and persistent organic pollutants in the addition, C3 was negatively related with C1 (r2 = -0.977) and C2 (r2 =
water column can drastically increase the COD content. Therefore, we -0.898), while C1 (r2 = 0.787) was positively related to C2 (Fig. S13c).
can extrapolate that the QNs antibiotics persist in the environment This further demonstrated that antibiotics toxicity resulted in the
compared to TCs which are easily biodegraded (Kandi et al., 2020) and destruction of plant tissues and the QNs have stronger toxicity effects in
may be the reason for their higher impact on COD with increasing days, wetlands.
compared to their TCs counterparts.
4. Conclusion
3.2.6. Quinolone and tetracycline antibiotics stimulated chlorophyll content
in the water column This study comprehensively investigated the fate and ecological
In this study, to further understand the impact of antibiotic pollution impact of tetracyclines and quinolones on the epiphytic microbial
on vegetation of wetland we investigated the impact of the QNs and TCs community in mesocosm wetlands. The research tracked the antibiotics’
on the chlorophyll content of the water column. Generally, the result behaviour in water, sediment, and plants, assessed their effects on mi­
indicated that the average “total chlorophyll” concentration was higher crobial biofilms and ARGs, and evaluated their influence on nutrient
in the antibiotics treatment groups (0.67 μg/L) than in control (0.57 μg/ removal and plant health in wetlands. In conclusion, this study reveals
L), except for H-LVX (0.53 μg/L) and H-OTC (0.45 μg/L) (Fig. S6). the far-reaching consequences of tetracyclines and quinolones on the
Likewise, the chlorophyll-a (Chl-a) content was increased in the anti­ epiphytic microbial community of mesocosm wetlands. The wetlands
biotic’s treatment groups compared with the control, especially for the demonstrated a remarkable capacity for degrading antibiotics. Howev­
TCs group, which recorded the highest increase in the total chlorophyll er, the presence of antibiotics not only impaired the wetland’s ability to
content compared with the QNs group. All the TCs group showed higher remove nutrients it also disrupted the delicate balance of microbial
Chl-a concentration than the quinolone group of antibiotics. A pareto community assemblages and spawned hotspots of ARGs. Importantly,
chart was used to show the distribution of Chl-a content in descending this study elucidated the ecological mechanisms through which TC and
order of frequency with a cumulative line on the secondary axis as a CIP affected the wetland ecosystems. While the research was designed to
percentage of the total. The decreasing trend of the chlorophyll a con­ simulate the natural conditions of mesocosm wetlands, it did not ac­
tent includes H-OTC>L-TC>L-OTC>H-TC>L-LVX>L-CIP>H-CIP>H- count for seasonal and temporal variations that could influence micro­
LVX>CT. Antibiotics exposure also increased the Chl-b and Chl-c content bial community responses. These findings demonstrated the urgent need
of the water column. Both Chl-b and Chl-c showed the highest chloro­ for critical attention to the environmental implications of antibiotic
phyll concentration in L-OTC while the lowest was recorded in the CT pollution on wetland ecosystems, highlighting the importance of miti­
group. The decreasing trend of chlorophyll concentration includes L- gating their devastating effects on these fragile ecosystems.
OTC>LVX>LTC>H-OTC>L-CIP>H-CIP> H-TC>CT and L-OTC> L-TC>
H-TC> HOTC> L-LVX> L-CIP> H-CIP> H-LVX>CT for Chl-b and Chl-c, CRediT authorship contribution statement
respectively. Noteworthy Chl-c was also higher in the TCs group
compared with the QNs group. The antibiotics increased the chlorophyll Okugbe Ebiotubo Ohore: Writing – review & editing, Writing –
content by 30 %, 61 %, and 40 % for Chl-a, Chl-b, and Chl-c, original draft, Visualization, Supervision, Methodology, Investigation,

14
O.E. Ohore et al. Water Research 267 (2024) 122484

Funding acquisition, Formal analysis, Data curation, Conceptualization. Li, J., Yu, Z., Warren, A., Lin, X., 2023. Predation risk affects the ecotoxicity evaluation of
antibiotics: population growth and antioxidase activity in the ciliate Paramecium
Jingli Zhang: Writing – review & editing, Writing – original draft,
jenningsi. Ecotoxicol. Environ. Saf. 251, 114536.
Formal analysis, Data curation. Sanji Zhou: Writing – review & editing, Li, J., Yu, Z., Zheng, Q., Chen, W., Lin, X., 2024. How antibiotic exposure affect predator-
Formal analysis, Data curation. Edmond Sanganyado: Writing – review prey interactions by population dynamics in ciliates? Aquat. Toxicol. 267, 106814.
& editing. Ji-Dong Gu: Writing – review & editing, Supervision. Guoj­ Li, S., Zhang, R., Hu, J., Shi, W., Kuang, Y., Guo, X., Sun, W., 2019. Occurrence and
removal of antibiotics and antibiotic resistance genes in natural and constructed
ing Yang: Writing – review & editing, Supervision. riverine wetlands in Beijing, China. Sci. Total Environ. 664, 546–553.
Liu, H., Li, S., Xie, X., Shi, Q., 2021. Pseudomonas putida actively forms biofilms to protect
Declaration of competing interest the population under antibiotic stress. Environ. Pollut. 270, 116261.
Michael, I., Rizzo, L., McArdell, C.S., Manaia, C.M., Merlin, C., Schwartz, T., Dagot, C.,
Fatta-Kassinos, D., 2013. Urban wastewater treatment plants as hotspots for the
The authors declare that they have no known competing financial release of antibiotics in the environment: a review. Water Res. 47, 957–995.
interests or personal relationships that could have appeared to influence Mu, X., Zhang, S., Lu, J., Huang, Y., Ji, J., 2024. Fate and removal of fluoroquinolone
antibiotics in mesocosmic wetlands: impact on wetland performance, resistance
the work reported in this paper. genes and microbial communities. J. Hazard. Mater. 470, 133740.
Nwankwegu, A.S., Zhang, L., Xie, D., Ohore, O.E., Li, Y., Yang, G., Yao, X., Song, Z.,
Data availability Yang, Q., 2023. Metabolites dynamics exacerbated by external nutrients inputs into a
Ceratium hirundinella-dominated bloom in the Pengxi River, Three Gorges
Reservoir, China. Aquat. Toxicol. 258, 106507.
Data will be made available on request. Ohore, O.E., Wei, Y., Wang, J., Wang, Y., Ifon, B.E., Liu, W., Wang, Z., 2022a. Vertical
characterisation of phylogenetic divergence of microbial community structures,
interaction, and sustainability in estuary and marine ecosystems. Sci. Total Environ.
851, 158369.
Acknowledgments Ohore, O.E., Wei, Y., Wang, Y., Nwankwegu, A.S., Wang, Z., 2022b. Tracking the
influence of antibiotics, antibiotic resistomes, and salinity gradient in modulating
We would like to thank Prof. Songhe Zhang for his support in this microbial community assemblage of surface water and the ecological consequences.
Chemosphere 305, 135428.
research. This work research was supported by Hainan Medical Uni­ Ohore, O.E., Zhang, S., Guo, S., Addo, F.G., Manirakiza, B., Zhang, W., 2021a.
versity Talent Research Launch Fund (RZ2300006042), the National Ciprofloxacin increased abundance of antibiotic resistance genes and shaped
Nature Science Foundation (grant no. 82260655) and Hainan Nature microbial community in epiphytic biofilm on Vallisneria spiralis in mesocosmic
wetland. Bioresour. Technol. 323, 124574.
Science Foundation (grant no. 821CXTD440), and National Key Ohore, O.E., Zhang, S., Guo, S., Manirakiza, B., Addo, F.G., Zhang, W., 2021b. The fate of
Research and Development Program of China (2021YFA0910300) tetracycline in vegetated mesocosmic wetlands and its impact on the water quality
(JDG). and epiphytic microbes. J. Hazard. Mater. 417, 126148.
Park, J., Gasparrini, A.J., Reck, M.R., Symister, C.T., Elliott, J.L., Vogel, J.P.,
Wencewicz, T.A., Dantas, G., Tolia, N.H., 2017. Plasticity, dynamics, and inhibition
Supplementary materials of emerging tetracycline resistance enzymes. Nat. Chem. Biol. 13, 730–736.
Ramalingam, G., Vignesh, R., Ragupathi, C., Magdalane, C.M., Kaviyarasu, K.,
Kennedy, J., 2020. Electrical and chemical stability of CuS nanofluids for
Supplementary material associated with this article can be found, in conductivity of water soluble based nanocomposites. Surf. Interfaces 19, 100475.
the online version, at doi:10.1016/j.watres.2024.122484. Scheffer, M., Carpenter, S.R., Lenton, T.M., Bascompte, J., Brock, W., Dakos, V., Van De
Koppel, J., Van De Leemput, I.A., Levin, S.A., Van Nes, E.H., Pascual, M.,
Vandermeer, J., 2012. Anticipating critical transitions. Science (1979) 338,
References
344–348.
Schierano, M.C., Panigatti, M.C., Maine, M.A., Griffa, C.A., Boglione, R., 2020.
Albero, B., Tadeo, J.L., Escario, M., Miguel, E., Pérez, R.A., 2018. Persistence and Horizontal subsurface flow constructed wetland for tertiary treatment of dairy
availability of veterinary antibiotics in soil and soil-manure systems. Sci. Total wastewater: Removal efficiencies and plant uptake. J. Environ. Manage 272,
Environ. 643, 1562–1570. 111094.
Austin, D., Vazquez-Burney, R., Dyke, G., King, T., 2019. Nitrification and total nitrogen Si, T., Chen, H., Qiu, Z., Zhang, L., Ohore, O.E., Zhang, S., 2020. Bacterial succession in
removal in a super-oxygenated wetland. Sci. Total Environ. 652, 307–313. epiphytic biofilms and deciduous layer sediments during Hydrilla verticillata decay: a
Clinical and Laboratory Standards Institute (CLSI), 2015. Performance standards for field investigation. J. Environ. Sci. 93, 193–201.
antimicrobial M100-S23, Vol. 35 No. 3. Sridhar, S., Forrest, S., Pickard, D., Cormie, C., Lees, E.A., Thomson, N.R., Dougan, G.,
Cunningham, C.J., Kuyukina, M.S., Ivshina, I.B., Konev, A.I., Peshkur, T.A., Knapp, C.W., Baker, S., 2021. Inhibitory Concentrations of Ciprofloxacin Induce an Adaptive
2020. Potential risks of antibiotic resistant bacteria and genes in bioremediation of Response Promoting the Intracellular Survival of Salmonella enterica Serovar
petroleum hydrocarbon contaminated soils. Environ. Sci. Process. Impacts 22, Typhimurium. mBio 12, 01093-21.
1110–1124. Sun, J., Zeng, Q., Tsang, D.C.W., Zhu, L.Z., Li, X.D., 2017. Antibiotics in the agricultural
Danner, M.C., Robertson, A., Behrends, V., Reiss, J., 2019. Antibiotic pollution in surface soils from the Yangtze River Delta, China. Chemosphere 189, 301–308. https://doi.
fresh waters: occurrence and effects. Sci. Total Environ. 664, 793–804. org/10.1016/j.chemosphere.2017.09.040.
Dong, Y., Li, Y., Kong, F., Zhang, J., Xi, M., 2020. Source, structural characteristics and Tang, J., Wang, S., Tai, Y., Tam, N.F., Su, L., Shi, Y., Luo, B., Tao, R., Yang, Y., Zhang, X.,
ecological indication of dissolved organic matter extracted from sediments in the 2020. Evaluation of factors influencing annual occurrence, bioaccumulation, and
primary tributaries of the Dagu River. Ecol. Indic. 109, 105776. biomagnification of antibiotics in planktonic food webs of a large subtropical river in
Fellman, J.B., Hood, E., Spencer, R.G.M., 2010. Fluorescence spectroscopy opens new South China. Water Res. 170, 115302.
windows into dissolved organic matter dynamics in freshwater ecosystems: a review. Tang, M., Chen, Q., Zhong, H., Liu, S., Sun, W., 2024a. CPR bacteria and DPANN archaea
Limnol. Oceanogr. 55, 2452–2462. play pivotal roles in response of microbial community to antibiotic stress in
Fu, L., Huang, T., Wang, S., Wang, X., Su, L., Li, C., Zhao, Y., 2017. Toxicity of 13 groundwater. Water Res. 251, 121137.
different antibiotics towards freshwater green algae Pseudokirchneriella subcapitata Tang, M., Chen, Q., Zhong, H., Liu, S., Sun, W., 2024b. CPR bacteria and DPANN archaea
and their modes of action. Chemosphere 168, 217–222. play pivotal roles in response of microbial community to antibiotic stress in
Han, Y., Yang, L., Chen, Xueming, Cai, Y., Zhang, X., Qian, M., Chen, Xingkui, Zhao, H., groundwater. Water Res. 251, 121137.
Sheng, M., Cao, G., Shen, G., 2020. Removal of veterinary antibiotics from swine Tu, R., Jin, W., Han, S.F., Zhou, X., Wang, T., Gao, S.H., Wang, Qing, Chen, C., Xie, G.J.,
wastewater using anaerobic and aerobic biodegradation. Sci. Total Environ. 709, Wang, Qilin, 2019. Rapid enrichment and ammonia oxidation performance of
136094. ammonia-oxidizing archaea from an urban polluted river of China. Environ. Pollut.
Hu, X., He, K., Zhou, Q., 2012. Occurrence, accumulation, attenuation and priority of 255, 113258.
typical antibiotics in sediments based on long-term field and modeling studies. Vymazal, J., 2020. Removal of nutrients in constructed wetlands for wastewater
J. Hazard. Mater. 225–226, 91–98. treatment through plant harvesting – biomass and load matter the most. Ecol. Eng.
Kandi, D., Behera, A., Sahoo, S., Parida, K., 2020. CdS QDs modified BiOI/Bi2MoO6 155, 105962.
nanocomposite for degradation of quinolone and tetracycline types of antibiotics Wang, Y., Hu, Y., Liu, F., Cao, J., Lv, N., Zhu, B., Zhang, G., Gao, G.F., 2020. Integrated
towards environmental remediation. Sep. Purif. Technol. 253, 117523. metagenomic and metatranscriptomic profiling reveals differentially expressed
Kiki, C., Rashid, A., Wang, Y., Li, Y., Zeng, Q., Yu, C.P., Sun, Q., 2020. Dissipation of resistomes in human, chicken, and pig gut microbiomes. Environ. Int. 138, 105649.
antibiotics by microalgae: kinetics, identification of transformation products and Yan, L., Herrmann, M., Kampe, B., Lehmann, R., Totsche, K.U., Küsel, K., 2020.
pathways. J. Hazard. Mater. 387, 121985. Environmental selection shapes the formation of near-surface groundwater
Lee, O., Kim, H.S., Kim, S., 2020. Hydrological simple water balance modeling for microbiomes. Water Res. 170, 115341.
increasing geographically isolated doline wetland functions and its application to Yan, L., Mu, X., Han, B., Zhang, S., Qiu, C., Ohore, O.E., 2019. Ammonium loading
climate change. Ecol. Eng. 149, 105812. disturbed the microbial food webs in biofilms attached to submersed macrophyte
Vallisneria natans. Sci. Total Environ. 659, 691–698.

15
O.E. Ohore et al. Water Research 267 (2024) 122484

Yang, Y., Du, W., Cui, Z., Zhao, T., Wang, X., Lv, J., 2020. Spectroscopic characteristics of Zhou, L.-J., Ying, G.-G., Liu, S., Zhao, J.-L., Chen, F., Zhang, R.-Q., Peng, F.-Q., Zhang, Q.-
dissolved organic matter during pig manure composting with bean dregs and biochar Q., 2012. Simultaneous determination of human and veterinary antibiotics in
amendments. Microchem. J. 158, 105226. various environmental matrices by rapid resolution liquid
Zhao, Q., Hu, Z., Zhang, J., Wang, Y., 2023. Determination of the fate of antibiotic chromatography–electrospray ionization tandem mass spectrometry. J. Chromatogr.
resistance genes and the response mechanism of plants during enhanced antibiotic A 1244, 123–138.
degradation in a bioelectrochemical-constructed wetland system. J. Hazard. Mater.
451, 131207.

16