Pseudomonas aeruginosa from river water: antimicrobial resistance, virulence and molecular typing

Abstract Pseudomonas aeruginosa isolates were recovered from surface river water samples in La Rioja region (Spain) to characterise their antibiotic resistance, molecular typing and virulence mechanisms. Fifty-two P. aeruginosa isolates were isolated from 15 different water samples (45.4%) and belonged to 23 different pulsed-field electrophoresis (PFGE) patterns. All isolates were susceptible to all antibiotics tested, except one carbapenem-resistant P. aeruginosa that showed a premature stop codon in OprD porin. Twenty-two sequence types (STs) (six new ones) were detected among 29 selected P. aeruginosa (one strain with a different PFGE pattern per sample), with ST274 (14%) being the most frequent one. O:6 and O:3 were the predominant serotypes (31%). Seven virulotypes were detected, being 59% exoS-exoY-exoT-exoA-lasA-lasB-lasI-lasR-rhlAB-rhlI-rhlR-aprA-positive P. aeruginosa. It is noteworthy that the exlA gene was identified in three strains (10.3%), and the exoU gene in seven (24.1%), exoS in 18 (62.1%), and both exoS and exoU genes in one strain. High motility ranges were found in these strains. Twenty-seven per cent of strains produced more biofilm biomass, 90% more pyorubin, 83% more pyocyanin and 65.5% more than twice the elastase activity compared with the PAO1 strain. These results highlight the importance of rivers as temporary reservoirs and sources of P. aeruginosa transmission, and show the importance of their epidemiological surveillance in the environment.


Introduction
Pseudomonas aeruginosa is a ubiquitous bacterium that is widespread in natural en vironments , survives on minimal nutritional r equir ements and toler ates a v ariety of physical conditions .T hese c har acteristics allow this species to colonise soil, vegetables, animals and a wide range of water sources, but also hospitals and community settings (Pirnay et al. 2005, Lister et al. 2009, Moradali et al. 2017 ).Pseudomonas aeruginosa is an opportunistic pathogen of great clinical importance, because it is one of the most frequent and severe agents causing nosocomial infections, particularly affecting immunocompromised, chronically infected and intensive care unit (ICU) patients.Along with the wide variety of infections and issues caused by P. aeruginosa , it can also be very difficult to tr eat.This or ganism is r esistant to many antibiotics and has a high capacity to express virulence factors and to form biofilm. Indeed, the World Health Organization identified P. aeruginosa as one of the top three priority pathogens for whic h ne w antibiotics ar e ur gentl y needed (Tacconelli et al. 2018 ).
The success of P. aeruginosa in infecting the host cell and e v ading the host immune system is due to a broad arsenal of pathogenicity factors such as the secretion of adhesins , toxins , proteases and pigments, as well as biofilm production.The Type 3 Secretion System (T3SS) is the major virulence weapon of this mi-cr oor ganism that contributes to cytotoxicity and acute infections, injecting potent exotoxins (ExoU, ExoS, ExoY and ExoT) into cytoplasm of the host cell (Moradali et al. 2017, Pena et al. 2019 ).The ExoU effector is e v en associated with an incr eased risk of earl y clinical mortality (Hauser 2009, Tümmler and Klockgether 2017, Foulkes et al. 2019 ).On the other hand, the absence of T3SS in P. aeruginosa strains has also been associated with the two-partner secretion system, ExlAB (Reboud et al. 2016, Ruiz-Roldán et al. 2020, Huber 2022 ).The secretion of the exolysin ExlA is responsible for the hypervirulent behaviour of some clinical strains (Elsen et al. 2014, Reboud et al. 2016 ).Furthermore, the production of biofilm, defined as organized bacterial communities embedded in an extracellular polymeric matrix attached to living or abiotic surfaces, is recognised as one of the major determinants of P. aeruginosa to favour its occurrence and persistence at different niches.
Humans use river water mainly for water supply, agriculture, industry, ener gy pr oduction and r ecr eational pur poses.Aquatic envir onments ar e r ecognised as one of the r eservoirs and tr ansmission routes for the dissemination of antimicr obial-r esistant and virulent pathogens, and consequently, water-borne bacterial diseases can be acquired (Amarasiri et al. 2020 ).High microbiological quality of rivers is r equir ed to r educe human infections.La Rioja is a region located in northern Spain with an estimated population of 323,377 inhabitants (in 2023), and the Iregua river is the main drinking water supplier (60% of the population).T hus , the present work aims to study the occurrence of P. aeruginosa in the Ir egua riv er, and to anal yse their antimicr obial r esistance, virulence factors and molecular typing.

Sample collection and processing, P. aeruginosa identification
Thirty-three surface water samples were collected from different points along the Ir egua riv er in La Rioja region during November 2015 (Fig. 1 ).The Iregua river is born in Sierra Cebollera at 2100 m, and descends over 64 km to flow into the Ebro river at 360 m.The Ir egua riv er basin pr esents a land occupation dominated by the area of forest (55% of the entire basin), scrubland (20%) and irrigated land (11%).The head zone is dominated by forest and scrub, and the low section for irrigation.The Ir egua riv er basin has tw o reserv oirs, which supply drinking water to 60% of the La Rioja population.The temper atur e and pH of the riv er w ater w ere measured during sampling.Supplementary Table S1 shows specific data from each collected sample.
A total of 500 ml of river water were collected with sterile bottles pre-dosed with sodium thiosulfate (10 mg) (Gosselin).The samples were transported under refrigeration conditions and dir ectl y pr ocessed.
A volume of 250 ml of each sample was filtered through a 0.45μm nitrocellulose filter that was dir ectl y deposited on the surface of cetrimide agar plates and incubated at 37ºC o vernight.T hen the filter was placed and washed in a tube with 1 ml of sterile water, and after that, serial dilutions were done.Aliquots of 100 μL were seeded on cetrimide agar plates, and incubated at 37 • C during 24 h.Colonies with Pseudomonas morphology (up to five per sample) were selected, identified by classical biochemical methods (Triple Sugar Iron and oxidase reactions), then confirmed by PCR and subsequent sequencing of the 16S rRNA (Rojo-Bezares et al. 2014 ).

Char acterisa tion of the porin coding oprD gene
Mutations in the oprD gene wer e anal ysed in the carbapenemresistant P. aeruginosa isolates by PCR, sequencing and comparison with the sequence of the P. aeruginosa PAO1 r efer ence str ain (GenBank accession number AE004091) (Ruiz-Roldán et al. 2018 ).

Molecular typing
The clonal diversity of the recovered P. aeruginosa isolates was analysed by pulsed-field electrophoresis (PFGE) with the SpeI enzyme (Rojo-Bezares et al. 2014 ).DNA profiles were analysed by the GelJ software 2.3 (UPGMA algorithm; Dice coefficient) (Heras et al. 2015 ).Multilocus sequence typing (MLST) was performed by PCR and sequencing (Ruiz-Roldán et al. 2018 ).The nucleotide sequences of alleles wer e compar ed with those of the PubMLST database ( http:// pubmlst.org/paeruginosa/ ) to obtain the specific sequence type (ST).New STs were submitted to the PubMLST website.

Biofilm production
Crystal violet (CV) staining assay was performed to analyse total biofilm biomass, and fluorescein diacetate (FDA) assay to study the bacterial metabolic activity inside the biofilm structure.Both methods were performed in 96-well microtiter plates using an initial inoculum 10 6 CFU/mL, and measured after 24 h of incubation, as pr e viousl y r ecommended (Peeters et al. 2008 ).Measur es wer e performed using a POLARstar Omega microplate reader (BMG Labtec h).All assays wer e performed in triplicate .T he percentage of biofilm biomass and bacterial metabolic activity was calculated in comparison with the biofilm production of the r efer ence P. aeruginosa PAO1 strain.

Pigments and elastase production
The c hlor oform-extr act method w as used for quantification of p yoc y anin and p y orubin pigments, b y measuring the absorbance at 520 and 525 nm, r espectiv el y (Ananthar ajah et al. 2017 ).
Elastase activity was tested by the Elastin-Congo-Red assay and the absorbance was measured at 450 and 600 nm as pr e viousl y described (Pearson et al. 1997 ).
All assays were performed in triplicate.The percentage of pigments and elastase production was calculated in comparison with the production of the reference P. aeruginosa PAO1 strain.

Results and discussion
Fifty-two P. aeruginosa isolates were isolated from 15 different water samples that were recovered along the Iregua river in La Rioja r egion (Fig. 1 ).The pr esence of P. aeruginosa was detected in 45.4% of the tested samples .T he temper atur e of the water r anged fr om 7.5 to 14ºC, and the pH from 7.9 to 10.8 ( Supplementary Table S1 ).There was no evidence regarding the presence of w astew ater outflows near to any positive sampling points, whereas P. aeruginosa was consistently isolated from water samples collected further downstream along the river and at the sites with a larger population.In this part of the river, water quality is expected to be influenced by run-off from agricultural land and an anthropogenic tr end, e v en r eceiving waste water fr om both households and industry.Indeed, it should be pointed out that se v en samples that harboured P. aeruginosa were recollected in industrial zones (47% of total samples), in accordance with other studies that obtained a high pr e v alence of P. aeruginosa in samples from environments with intense human contact (Crone et al. 2020 ).
These 52 isolates were susceptible to all the antibiotics tested, except for P. aeruginosa MW131b isolated from an urban zone sample that was resistant to carbapenems (imipenem, meropenem and doripenem).No carbapenemases were found in strain MW131b; ho w ever, the follo wing amino acid changes w ere detected in its OprD porin: D43N, S57E, S59R, E202Q, I210A, E230K, S240T , N262T , A267S, A281G , K296Q, Q301E, R310G , V359L, Loop 7 short, and a pr ematur e stop codon at position 415.All these substitutions, with the exception of the stop codon, have been pr e viousl y detected in carba penem-r esistant and susceptible P. aeruginosa str ains r ecov er ed fr om differ ent origins (Estepa et  T he o v er all antibiotic r esistance le v el was v ery low in comparison with P. aeruginosa isolates r ecov er ed fr om clinical samples (Lister et al 2009, Peña et al. 2015, Recio et al. 2020, Rojo-Bezares et al. 2014 ) or w astew ater effluents (Okafor and Nwodo, 2023 ;Wu et al. 2023 ), but similar low antimicr obial r esistance le vels were detected in other studies performed with environmental P. aeruginosa str ains (Cr one et al. 2020 , Kittinger et al. 2016, Pirnay et al. 2005, Suzuki et al. 2013 ).Considering that carbapenems are last-resort antibiotics for treating infections caused by m ultidrug-r esistant Gr am-negativ e bacteria, it is concerning that we detected a carba penem-r esistant P. aeruginosa strain in this natur al envir onment, suggesting that riv ers could be consider ed a potential risk for human health.
Regarding the clonal relationship among the 52 P. aeruginosa isolates, 23 different PFGE patterns were observed (Fig. 2 ).Most of the strains with indistinguishable PFGE patterns were isolated from the same sample, although there were also strains with an equal PFGE pattern that were isolated from different samples (i.e.those strains with the PFGE pattern P1, P6, P16 and P18) (Fig. 2 ).Considering the strains from the same sample, the highest diversity was found among the five P. aeruginosa strains recovered from the I33 sample (collected in the Ebro river mouth) as they sho w ed four different PFGE patterns (P16, P18, P20 and P21), two of which (P16 and P18) were also detected in strains from the I30 sample.Sampling points of the two samples I30 and I33 were located close to each other ( Supplementary Table S1 ).One strain with a different PFGE pattern per sample was included in further anal ysis.Additionall y, str ains MW131b and MW133 were both included, e v en although they sho w ed an indistinguishable PFGE pattern, but differ ent r esistance phenotypes .T hus , a total of 29 P. aeruginosa strains were finally included in this study for further c har acterisation.
The MLST was performed for the 29 selected P. aeruginosa and a high diversity of STs was found, detecting 22 different STs (Fig. 2 ).Six new STs were identified and registered in the PubMLST database as ST2540, ST2543, ST2544, ST2545, ST2547 and ST2548 (Table 1 ).None of the worldwide P. aeruginosa high-risk clones (including ST235, ST111 and ST175) were detected among our P. aeruginosa from river water.ST136, ST274, ST679, ST782 and ST2540 were found more than once among the studied strains.The clone ST274 was the most pr e v alent and was detected in four strains (14%).The ST274 has been previously described as an intercontinental P. aeruginosa clone disseminated worldwide and detected in strains from clinical samples (blood cultures , urine , ulcers , wounds , sputum and stool, etc.), healthy humans and raw foods (Estepa et al. 2014, 2015, 2017, López-Causapé et al. 2017, Ocampo-Sosa et al. 2015, Rojo-Bezares et al. 2016, Ruiz-Roldan et al. 2018 ).In the case of ST136, it was also found in clinical strains of the ICU of a hospital in Ho Chi Minh City (Tada et al. 2013 ).According to the MLST database, the following STs have also been reported among clinical strains: ST116, ST217, ST455, ST483, ST773, ST782 and ST792 in sputum, ST136, ST217 and ST252 in soft tissue infections and ST252 in br onc hial lav a ge.The P. aeruginosa MW131b strain was resistant to imipenem, meropenem and doripenem.NA, non-agglutinable.
F igure 3. P er centages of production of biofilm biomass (CV) and biofilm metabolic activity (FDA) of 29 P. aeruginosa strains compared with the reference P. aeruginosa PAO1 strain (the black dotted line shows 100%).Regarding virulence profiles, the 29 strains were classified in se v en virulotypes (Table 1 ).The virulotype 1, c har acterised by amplifying exoS, exoY, exoT, exoA, lasA, lasB, lasI, lasR, rhlAB , rhlI , rhlR and aprA genes, was the most common (58.6%).This virulotype was also the most frequent in P. aeruginosa from blood samples, animal and human faecal samples, and foods such as raw vegetables, observed in studies of our group (Rojo-Bezares et al. 2016, Ruiz-Roldán et al. 2018, 2021 ).The exoU, exoS, exoY and exoT genes were found in 27.6%, 65.5%, 79.3% and 89.6% of the strains, respectiv el y.The exoU gene was detected in se v en P. aeruginosa strains, exoS in 18 strains and both genes in one strain.This proportion of T3SS genotypes is in accordance with other Spanish studies carried out in the clinical setting (Peña et al. 2015, Recio et al. 2020 ).
ExoU is a potent cytotoxin with phospholipase A2 activity that has been clinically associated with early mortality, and ExoS plays an important role in the inv asiv e ca pability of P. aeruginosa (Hauser 2009, Moradali et al. 2017, Peña et al. 2015 ).Se v er al authors hav e reported that the exoS and exoU genes are mutually exclusive (Feltman et al. 2001, Pirnay et al. 2009 ), pr obabl y because of enhanced fitness in distinct ecological niches (Ozer et al. 2019, Rutherford et al. 2018 ).Ho w e v er, the number of reported works that described the concomitant presence of both genes is increasing (Horna et al. 2019, Morales-Espinosa et al. 2017, Ozer et al. 2019, Song et al. 2023, Yi et al. 2021 ).Considering that the exoU gene is located in an island structure, it has been suggested that exoU could be a horizontall y acquir ed virulence determinant, mobilised onto a tr ansmissible plasmid (Kulasekara et al. 2006 ).In our work, only strain MW200 co-carried the exoU and exoS genes (3.4%), a low percenta ge pr e viousl y observ ed in studies by our gr oup (Ruiz-Roldán et al. 2018 ), in contrast to other studies with higher percentages of exoS and exoU co-carriers (Horna et al. 2019, Morales-Espinosa et al. 2017 ).The presence of both exoS and exoU genes has been associated with acute infection in humans and with hypervirulent P. aeruginosa clones (Horna et al. 2019, Morales-Espinosa et al. 2017, Song et al. 2023 ).T hus , further r esearc h is r equir ed to determine the cytotoxicity and pathogenicity of this MW200 strain isolated from a surface river water sample.
On the other hand, in three of our 29 strains (10%), all T3SS genes ( exoU, exoS, exoT, exoY ) were absent, but the exlA gene was amplified (Table 1 ).The ExlA toxin induces plasma membrane rupture in host cells .Furthermore , exlA -harbouring strains ha ve been found in non-clinical samples (such as plants, soil, wildanimal and, as in this stud y, ri v er water), pr omoting their ca pacity F igure 5. P er centage of p y orubin and p y oc y anin production and elastase activity of the 29 P. aeruginosa strains compared with the P. aeruginosa PAO1 r efer ence str ain (the blac k dotted line shows 100%).The gr ey line indicates the mean v alue.
The rhlI/R and lasI/R genes, involved in QS regulation, were detected in all strains, with the exception of one strain belonging to ST217 that did not amplify the lasR gene .T he lasA, lasB , rhlAB and aprA genes were detected in all strains .T he exoA gene , encoding an exotoxin A that is an important virulence factor of type 2 secretion system (T2SS) (Yousefi-Avarvand et al. 2015 ), was found in all but two P. aeruginosa str ains.Curiousl y, two exlA -positiv e str ains lac ked the exoA gene and one str ain has both exlA and exoA genes.Further future studies invite the analysis of the cytotoxicity of different strains with a v ery div erse battery of virulence genes.
Biofilm pr oduction anal ysis sho w ed that 27.6% of str ains pr oduced more biofilm biomass, and 17.2% more bacterial metabolic activity than the PAO1 strain (Fig. 3 ).T he s wimming and swarming r esults r e v ealed that P. aeruginosa str ains fr om Ir egua riv er had great motility, because 72.4% (21 strains) sho w ed the highest swimming, and 58.6% (17 strains) the highest swarming capacity, covering the entire Petri dish surface (6400 mm 2 ) (Fig. 4 ).Moreover, 65.5% of the strains produced more than twice the elastase activity compared with the PAO1 strain (Fig. 5 ).Regarding pyorubin and p y oc y anin production, 90% and 83% of the str ains pr oduced, r espectiv el y, mor e than the PAO1 strain (Fig. 5 ).This high percenta ge of str ains with gr eat pigment pr oduction could be because the p y oc y anin is toxic to the majority of the population when cells are nutrient limited; a subset of cells is intrinsically p y oc y anin resistant.The effect of p y oc y anin on the producer population thus appears to be dynamic, and helping biofilm de v elopment (Meirelles and Newman 2018 ).Interestingly, the three exlAharbouring strains were the major producers of biofilm biomass and p y oc y anin production, but w ere lo w elastase producers.Our r esults ar e in accordance with Ruiz-Roldán et al. ( 2020 ), wher e six exlA strains sho w ed three to four times higher biofilm biomass values and p y oc y anin pr oduction than the contr ol P. aeruginosa PAO1 strain.
Pr e vious r eports suggested a relationship between different serotypes and the presence of the exoU and exoS virulence genes (del Barrio-Tofiño et al. 2019, Faure et al. 2003, Recio et al. 2020). Faure et al. ( 2003 ) observed that none of the strains with the O:1 ser otype harbour ed the exoU gene; by contr ast, 67% of O:1 str ains wer e exoU positiv e in our work.On the other hand, ser otype O:11 was associated with the presence of exoU (Recio et al. 2020 ), whereas in our work only the ST773 strain met this requirement.It is noted that strains belonging to serotype O:4 are frequently associated with high mortality rates (23.5%)(Faure et al. 2003 ), and additionall y, as a pr e vious Spanish nationwide study reported (del Barrio-Tofiño et al. 2019 ), the O:4 serotype is str ongl y linked to the m ultidrug/extensiv el y drug-r esistant pr ofile of the widespr ead ST175 high-risk clone.Ho w e v er, in our work the unique strain with serotype O:4 contained both the exoU and exoS genes and was ascribed to the new ST2547.

Conclusion
Despite their antimicrobial susceptibility , P .aeruginosa strains recov er ed fr om envir onmental samples hav e gr eat pathogenic potential, given the high presence of virulence factors, great motility, as well as high pigment production and elastase activity.
T his study pro vides significant data on the great diversity and pathogenicity of P. aeruginosa r ecov er ed fr om riv er water .These r esults highlight the importance of epidemiological surveillance of this species in the environment, and its implication in the clinical setting.

Figure 1 .
Figure 1.Location of sampling sites of the 33 surface water samples of Iregua river in La Rioja region.Those points where P. aeruginosa strains were isolated are marked with star symbols.

Figure 2 .
Figure 2. Dendr ogr am of PFGE patterns obtained from the 52 P. aeruginosa isolates studied.The names of strains and samples, as well as the PFGE patterns and STs, are also indicated.

Figure 4 .
Figure 4. Values of motility area (swimming and swarming) of 29 P. aeruginosa strains .T he entir e plate ar ea was 6400 mm 2 .

Table 1 .
Virulence profiles, serotypes and sequence types found in 29 P. aeruginosa strains from Iregua river.