lable at ScienceDirect

Toxicon 97 (2015) 23e31

Contents lists avai

Toxicon

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

Inflammatory responses and potencies of various lipopolysaccharidesfrom bacteria and cyanobacteria in aquatic environments and watersupply systems

Yumiko Ohkouchi a, *, Satoshi Tajima b, 1, Masahiro Nomura b, 2, Sadahiko Itoh b

a Department of Global Ecology, Graduate School of Global Environmental Studies, Kyoto University, Kyoto Daigaku-Katsura CI-2-233, Nishikyo-ku, Kyoto,615-8540, Japanb Department of Environmental Engineering, Graduate School of Engineering, Kyoto University, Kyoto Daigaku-Katsura CI-2-233, Nishikyo-ku, Kyoto,615-8540, Japan

a r t i c l e i n f o

Article history:Received 5 September 2014Received in revised form5 January 2015Accepted 5 February 2015Available online 7 February 2015

Keywords:Lipopolysaccharide (LPS)PurificationEndotoxinIndigenous bacteria/cyanobacteriaHuman monocytic cellsCytokine secretion

* Corresponding author. Present address: School ofence, Azabu University, 1-17-71 Fuchinobe, Chuo-ku252-5201, Japan.

E-mail address: okochi@azabu-u.ac.jp (Y. Ohkouch1 Present address: West Japan Railway Company, S

Marunouchi 3-chome, Chiyoda-ku, Tokyo, 100-0005,2 Present address: Earthtech Toyo Co., Ltd., 44-32

ku, Kyoto, 601-1374, Japan.

http://dx.doi.org/10.1016/j.toxicon.2015.02.0030041-0101/© 2015 Elsevier Ltd. All rights reserved.

a b s t r a c t

Inflammatory substances derived from indigenous bacteria in aquatic environments or water systems areof great concern. Lipopolysaccharides (LPSs), one of the major inflammatory substances in water, areusually identified using Limurus amoebocyte lysate (LAL) assay on the basis of their endotoxic activity,but endotoxin levels do not accurately represent their inflammatory potency in humans. In this inves-tigation, the cellular endotoxin contents of pure-cultured bacteria/cyanobacteria, which are frequentlydetected in water sources and distribution systems, and of indigenous bacteria in a river and in bio-logically activated carbon (BAC) effluent, were investigated. The indigenous bacteria showed the highestendotoxin contents exceeding 10�3 EU/cell. The LPSs were then purified from those samples, and theirinflammatory potencies were examined using a human monocytic cell line. The LPSs from Acinetobacterlwoffii culture, the river water, and the BAC effluent sample revealed a unique cytokine secretion pattern;they induced both IL-8 and TNF-a more strongly than the other tested bacterial LPSs. These resultssuggest that natural bacterial/cyanobacterial flora in aquatic environments and water distribution sys-tems have the potential to induce relatively strong inflammatory responses in humans; therefore, furtheraccumulation of data on water quality from the perspective of not just endotoxins but inflammatorypotency is needed.

© 2015 Elsevier Ltd. All rights reserved.

1. Introduction

There are various naturally occurring substances that are toxicto humans in aquatic environments. Cyanobacterial toxins, such asmicrocystins, are one of the major toxic compounds in aquaticenvironments (Carmichael and Bent, 1981). Lipopolysaccharides(LPSs), which are the outer membranes of gram-negative bacterialor cyanobacterial cells, are also widely known to cause stronginnate immune reactions in humans via Toll-like receptors on the

Life and Environmental Sci-, Sagamihara-shi, Kanagawa,

i).hin-Kokusai Building 9F, 4e1,Japan.Nishioji-cho, Daigo, Fushimi-

cell surface. LPSs are also called endotoxins because of their bio-logical activity. Information on endotoxin levels in various watersamples has been accumulated. Just as an example, drinking waterin Japan, China, and Australia ranges 1e10, 4e10, and 72e186 EU/mL, respectively, while the endotoxin levels in source water variedlargely, but normally are in the range of 10e1000 EU/mL (Ohkouchiet al., 2007, 2009; Can et al., 2013; O'Toole et al., 2008). Rapala et al.(2002) also reported that water bloom sample with dense cyano-bacteria reached at 3.8 � 104 EU/mL in Finland. Limurus amoebo-cyte lysate (LAL) assay is widely used to determine endotoxin levelsin aquatic environments. This assay, based on coagulating reactionsof LAL with endotoxins, is very sensitive and can detect smallamounts of endotoxins. In this assay, the endotoxic activity of eachbacterial LPS was expressed in terms of its endotoxic equivalencewith the LPS of Escherichia coli-type strains. However, severalstudies have implied that there are big differences between theendotoxic activities of various bacteria determined by LAL assay

Y. Ohkouchi et al. / Toxicon 97 (2015) 23e3124

and their toxic effects on human health (Hansen et al., 1999; Dehuset al., 2006). Our previous study indicated that the cellular assaysystem using a monocytic cell line, THP-1, could be useful fordetecting the changes in inflammatory potencies caused by bac-terial communities in aquatic environments (Ohkouchi et al., 2012).However, the information on their inflammatory potencies ascaused by indigenous bacteria in various water samples is stillinsufficient, as is that on their endotoxic activities.

In water supply systems, there are three sites where endotoxinscan increase: at water sources as previously described (Rapala et al.,2002), in a treatment process using biologically activated carbon(BAC) (Ohkouchi et al., 2007; Can et al., 2013), and in distributionsystems (Ohkouchi et al., 2009). In general, cyanobacteria arerecognized as major contributors to the increase of endotoxins inwater sources, while gram-negative bacteria could be a dominantfactor in treatment processes and distribution systems. Ohkouchiet al. (2007) compared the endotoxic activities derived fromMicrocystis aeruginosa, Synechococcus sp. and E. coli cultured in thelaboratory. The results showed that the Synechococcus sp. could be amajor contributor to endotoxin increase in the Lake Biwa - YodoRiver Basin due to their abundance in the natural aquatic envi-ronment, especially in summer. The survey results also suggestedthat effluents from sewage treatment plants could be sources ofendotoxin contamination other than cyanobacterial bloom in thatbasin. The second site for the increase of endotoxins could be BACprocesses, which provide habitats for bacteria. Bacteria canmultiply on granular activated carbon (GAC) by using nutrients inwater. The proteobacteria, i.e. gram-negative bacteria, were found tobe dominant inhabitants in the bacterial community created on thesurface of BAC (Poitelon et al., 2010). In contrast, it is known thatthe bacterial populations in water distribution systems vary fromsystem to system depending on their conditions. In many cases, thebacteria in biofilm accumulated inside the pipes or in tap watersamples have been detected using culture techniques, but someresearchers have applied DNA-basedmolecular techniques. Variousspecies of gram-negative bacteria that belong to alpha-, beta-, andgamma-proteobacteria were frequently detected. For example,Hirata et al. (1993) detected Alcaligenes sp., Bacillus sp., Methyl-obacterium sp., Pseudomonas sp., and Flavobacterium sp. in distrib-uted water samples. Furuhata et al. (Furuhata, 2004) also reportedthat Methylobacterium sp. and Pseudomonas sp. were dominantspecies inwater samples taken at a hospital via a storage tank. Scottand Pepper (2010) identified frequently detected bacterial speciesin tap water samples in the United States as Sphingomonas sp.,Acidovorax sp., Aquabacterium sp., and Acinetobacter sp. based onthe homology of 16S rRNA sequences. The bacterial regrowth ofthese various species, except for Sphingomonas sp., whose mem-brane structure differs from typical LPSs (Kawasaki et al., 1994;Kawahara et al., 1999), could cause increases in endotoxin and in-flammatory potency during water distribution. Therefore, identi-fying and prioritizing the major contributors to increases inendotoxin and inflammatory potency in water supply systems

Table 1Bacterial strains tested in this investigation.

Tested bacteria Strain Class Proliferation

Escherichia coli NBRC 3301 g-Proteobacteria e

Microcystis aeruginosa NIES-44 Cyanophyceae Water resourSynechococcus sp. NIES-946 Cyanophyceae Water resourPseudomonas fluorescens ATCC 49642 g-Proteobacteria DistributionPseudomonas aeruginosa NBRC 12689 g-Proteobacteria DistributionAquabacterium commune ATCC BAA-209 b-Proteobacteria DistributionAcidovorax delafieldii ATCC 17606 b-Proteobacteria DistributionAcinetobacter lwoffii JCM 6840 g-Proteobacteria DistributionMethylobacterium fujisawaense NBRC 16843 a-Proteobacteria Distribution

would help in the development of a novel strategy of water qualitymanagement.

This investigation focused on several bacterial or cyanobacterialstrains which ubiquitously inhabit water sources or distributionsystems to determine what kind of bacteria could contribute to anincrease in inflammatory potencies as well as endotoxins in watersupply systems. Firstly, the cellular LPS contents of each purecultured bacteria/cyanobacteria or indigenous bacteria in riverwater or BAC effluent were compared. Secondly, their endotoxicactivities and the inflammatory potencies of LPSs after purificationwere also compared using a monocytic cell line based on the levelsand patterns of cytokine secretion to narrow down which strainscan impact human health. Also, a possibility that contaminantsubstances contribute to inflammatory potencies by the purifiedLPSs pretreated with/without polymixin B was examined. Lastly,the importance of water quality in terms of inflammatory potencywas discussed based on the data obtained.

2. Materials and methods

2.1. Bacterial/cyanobacterial culture conditions

The bacteria/cyanobacteria used for the LPS preparation andtheir cultivation conditions are listed in Table 1. These bacteria/cyanobacteria were chosen based on published data regarding theiroccurrence rates in water sources or distribution systems.Microcystis aeruginosa and Synechococcus sp., which are frequentlydetected in water sources (Rapala et al., 2002; Hoson et al., 2002),were cultured at 20 �C for 20e30 days with illumination of1500e2500 lux at 12-h intervals using CB and C media, respec-tively. Gram-negative bacteria, except Escherichia coli NBRC 3301,which are frequently detected in tap water samples or in biofilmaccumulated in distribution systems, were cultured at 20 �C for4e12 days using a medium recommended by each culture collec-tion. E. coli NBRC 3301 was cultured at 37 �C for 30 h using LBmedium.

2.2. Water sampling

A 21.8 L volume of river water was taken from the Yodo Riverdownstream of Osaka (N34.72.48.47, E135.51.30.37) in December2010. The sampling site was adjacent to the intake area of adrinking water purification plant. A 20 L volume of effluent waterfrom the BAC adsorption process was taken at the water purifica-tion plant on the Yodo River. The purification plant has coagulation-sedimentation, mid-ozonation, rapid sand filtration, post-ozonation, BAC, and chlorination processes. The containers usedfor water samplings were soaked with PyroCLEAN (Alerchek, Inc.,Portland, ME, USA) for more than one hour to degrade and solu-bilize endotoxins attached on the inner surface and rinsed thor-oughly with endotoxin-free Milli-Q water. The endotoxin-freeMilli-Q water was produced using a Milli-Q Academic equipped

site Culture condition Reference

LB broth, 37 �C, 30 hces CB medium, 20 �C, 30 days (Rapala et al., 2002)ces C medium, 20 �C, 20 days (Hoson et al., 2002)systems R2A broth, 20 �C, 5 days (Yano et al., 2009)systems NBRC 802, 20 �C, 12 days (Yano et al., 2009) (Szita et al., 2007)systems ATTC 2261, 20 �C, 7 days (Kalmbach et al., 1999)systems Nutrient broth, 20 �C, 4 days (Lee et al., 2010)systems Nutrient broth, 20 �C, 7 days (Scott and Pepper, 2010)systems NBRC 352, 20 �C, 7 days (Furuhata et al., 2006)

Y. Ohkouchi et al. / Toxicon 97 (2015) 23e31 25

with a BioPak cartridge as a final filter. The water sample wastransported to our laboratory under refrigerated conditions at 4 �C.

2.3. Enumeration of bacterial cells

The bacterial cell counts cultured in the laboratory were deter-mined by pour-plating using agar plates under the conditionsrecommended by each culture collection. TheMicrocystis cell countwas determined bymicroscopic examination at x400magnificationafter pressurization. The Synechococcus cell count was determinedby epifluorescence microscopy observation of intrinsic red fluo-rescence under G-excitation. The culturable heterotrophic bacteriain the river and BAC effluent water samples were enumerated usingR2A agar plates at 20 �C with 7 days incubation. The total bacterialcell counts were determined by epifluorescence microscopicenumeration under UV excitation after 40,6-diamidino-2-phenylindole (DAPI) staining.

2.4. LPS purification

The bacterial cells and Synechococcus sp. cultured in our labo-ratorywere harvested by centrifugation at 8000 rpm for 10e20minand then washed three times with sterilized 50 mM phosphatebuffer (pH 7.2). Only Microcystis aeruginosa was harvested bycentrifugation at 6000 g for 10 min and then washed three timeswith sterilized 15 mg/L NaHCO3 solution. The Yodo River sampleand the BAC effluent samplewere filtered through a glass fiber filter(Advantec Toyo Kaisha, Ltd., GA-100, Tokyo, Japan) and thenenriched using an ultrafiltration device with membranes of MWCO10,000 (Advantec Toyo Kaisha, Ltd., Q0100, Tokyo, Japan). The de-vices and membranes were washed with 5 N and 1 N NaOH,respectively, by soaking for more than three hours to degradecontaminated organics, and then rinsed thoroughly withendotoxin-free Milli-Q water.

After the cell pellets or the enriched water sample was lyophi-lized, LPSs were extracted from the lyophilized cells according tothe method reported by Bernardov�a et al. (2008). A brief descrip-tion of the procedure follows. An equal volume of hot 90% phenol-water mixture was added to the suspension of lyophilized cells inendotoxin-free water, and then stirred at 65e70 �C for 20 min. Thisextraction procedure was repeated twice. After centrifugation, thesupernatant aqueous phase was collected and dialyzed usingSpectra/Por 7 (1000 Da; Spectrum Laboratories, Inc., RanchoDominguez, CA, USA) against endotoxin-free Milli-Q water for 3e4days. The dialyzed sample was lyophilized again, and dissolved in0.1 M TriseHCl buffer containing 25 mg/mL RNase A (from bovinepancreas, Waco Pure Chemicals, Osaka, Japan). After incubation for16 h at 37 �C to degrade RNA in the samples, an equal volume of 90%phenol-water solution was added again and stirred vigorously. Theaqueous phase containing LPS was separated by centrifugation at9600 � g and dialyzed sufficiently against endotoxin-free waterusing Amicon Ultra-15 (Millipore Japan, Tokyo). Finally, the dia-lyzed sample was lyophilized and stored at �80 �C until furtherassay.

2.5. Endotoxin determination

The endotoxins in pure bacterial/cyanobacterial culture, thewater samples, and the purified LPSs were determined by kinetic-chromogenic LAL assay using Pyrochrome with Glucashield buffer(Seikagaku Biobusiness Corporation, Tokyo, Japan). The uninten-tional reactionwith b-glucan remaining in the samples was blockedusing Glucashield buffer. LPS purified from Escherichia coli strainO113:H10 (Seikagaku Biobusiness Corporation, Tokyo, Japan) wasused for calibration, and the results were given in endotoxin units

(EU). Each sample was diluted with endotoxin-free Milli-Q water.Pipette chips and 96-well microplates guaranteed to be endotoxin-free were used for the assay.

For determination of endotoxin in pure culture, the culturemedia samples were taken at late logarithmic growth phases. Theendotoxins of media themselves for cultivation in Table 1 were alsodetermined before inoculation. The known weight of the purifiedLPS samples were dissolved in sterilized endotoxin-free Milli Qwater, and then the endotoxins were also determined in the sameway.

2.6. Cell culture

A monocytic cell line, THP-1 (JCRB0112.1), purchased from theJapanese Collection of Research Bioresources Cell Bank, wascultured in RPMI1640 medium with 5% heat-inactivated FBS. Allculture flasks were incubated at 37 �C in humidified 5% CO2.

2.7. Endotoxin stimulation

The non-adherent THP-1 cells were suspended in each mediumat 1.0 � 106 cells/mL. Aliquots of 100 mL cell suspensions wereseeded in flat-bottomed 96-well culture plates and were incubatedfor three hours prior to LPS stimulation. The commercial LPS ofE. coli O55:B5 (Sigma Aldrich Japan, Tokyo, japan) and the extractedLPS samples were dissolved in endotoxin-free Milli-Q water andfiltered through a sterilized 0.2-mm syringe filter (DISMIC-25CS,Advantec Toyo Kaisha, Ltd., Tokyo, Japan). After preincubation, theculture medium in each well was removed and 100 mL of freshmedia was added. The THP-1 cells were stimulated with LPS sam-ples for six hours by adding 10 mL LPS suspension (Ohkouchi et al.,2012).

2.8. Contaminant assay by blocking endotoxin activity

Polymixin B, cationic polypeptide, can bind to mono ordiphosphate group of the lipid A (Coyne and Fenwick, 1993), and soit is often used to neutralize biological activity of LPSs. The inhibi-tion of cytokine secretions by addition of polymixin B can excludethe possibility that contaminants remaining in the purified samplescontribute to cytokine production. The E. coli O55:B5 LPS (Sigma-eAldrich Japan), the LPSs extracted from Acinetobacter lwoffii JCM6840, the river water, and the BAC effluent samples (5000 EU/mL)were pretreated with 50 mg/mL polymixin B sulphate (SigmaAldrich Japan) for 1 h at room temperature. Then, THP-1 wasstimulated with the pretreated LPSs at 500 EU/mL in the same wayas described above.

2.9. Determination of cytokine levels and cell counts

After stimulation with LPS samples, the culture supernatantswere collected and stored at �80 �C until cytokine determination.An aliquot of 100 mL fresh medium was added to each well again,and the cells were incubated for 30 min. Then, viable cell countswere determined using Cell counting Kit-8 (Dojin Laboratories,Tokyo, Japan). The cytokines, TNF-a and IL-8, were measured byindirect sandwich ELISA assays (Diaclone, Gen-Probe Inc., SanDiego, CA, USA) following the manufacturer's instructions.

2.10. Statistical analysis

Data were compared by t-test using GraphPad Prism version 5.0(GraphPad Prism Inc., San Diego, CA, USA), and significant differ-ences were determined with a level of p < 0.05. Principal compo-nent analysis was also applied to data set of 3 variables (cellular

Y. Ohkouchi et al. / Toxicon 97 (2015) 23e3126

endotoxin contents, two types of cytokine secretions at LPS doses of1000 EU/mL) obtained from 11 different LPS samples using theMicrosoft Excel add-in software, Mac Tahen ver. 1.0a (ESUMI Co.,Ltd., Tokyo, Japan).

3. Results and discussion

3.1. Comparison of endotoxic activities derived from differentbacterial strains and water samples

The endotoxic activities of the pure bacteria/cyanobacteriastrains cultured in our laboratory were determined before LPSextraction. The results in terms of endotoxins per cell in each cul-ture medium are compared in Fig. 1. The endotoxin in each culti-vation mediumwas negligible less than 0.23 EU/mL. The endotoxicactivities of indigenous bacteria in riverwater/BAC effluent samplesare also shown in Fig. 1. The endotoxic activities per cell werecalculated by dividing the endotoxins determined by LAL assay bythe total cell counts in the samples determined by DAPI staining,because LPSs on cell surfaces exhibit endotoxic activity regardlessof the viability of bacterial cells. Aquabacterium commune andPseudomonas aeruginosa showed the highest endotoxic activities,greater than 10�3 EU/cell. The endotoxic activities per cell ofAcinetobacter lwoffii, Synechococcus sp., Pseudomonas fluorescens,Microcystis aeruginosa, Acidovorax delafieldii, and Escherichia colifollowed in descending order. In contrast, significantly lowerendotoxic activities were detected in Methylobacterium fujisa-waense, less than 10�9 EU/cell.

The endotoxin levels of the river water and the BAC effluentwere found to be 440 and 13 EU/mL, respectively. The endotoxicactivity per cell of indigenous bacteria in the BAC effluent waterwas slightly higher than that in the river water (river water1.1 � 10�3; BAC effluent 1.5 � 10�3 EU/cell). Both water sampleswere taken in the winter season; therefore, gram-negative bacteriarather than cyanobacteria were considered to be major contribu-tors to the endotoxic activities. This result showed that the cellularendotoxin content was somewhat increased in the BAC effluent,whereas 97% of endotoxins in the river water regarded as sourcewater were removed through rapid sand filtration and ozonation-BAC processes. The increase can be attributed to the shift in thebacterial community during treatment processes, especially theBAC adsorption process, which is known to be a place for bacterialregrowth. Kasuga et al. (2011) reported that the bacterial commu-nity structures in the BAC can shift with the maturing process.

Fig. 1. Comparison of cellular endotoxin contents of pure-cultured bacteria/cyano-bacteria and water samples. Those values are calculated by dividing the endotoxicactivity in the culture medium or water samples by the total cell counts.

However, the water purification plant where we conducted sam-pling replaced a subset, not all, of the BAC columns with new GACevery 5 years in order to attain constant organic removal. Therefore,the effect of the operation period of BAC does not need to be takeninto account in this case. The result also indicates that bacteriaubiquitously present in thewater source or even after thewater hasundergone drinking-water treatment have relatively high endo-toxic activities.

Based on the comparison of the cellular endotoxin contents ofeach bacteria/cyanobacteria, there is a possibility thatAquabacterium commune, Pseudomonas spp., andAcinetobacter lwoffii, which exhibited endotoxin contents around10�3 EU/cell, may bemajor contributors to endotoxin increases thatoccur during water distribution. Two Pseudomonas spp. strainswere frequently detected in biofilm samples or tap water samplesinmany countries (Yano et al., 2009; Szita et al., 2007). Additionally,Kalmbach et al. (1999) reported that Aquabacterium commune is adominant and ubiquitous community member in biofilms accu-mulated in several distribution systems in Hamburg, Berlin, Mainz,and Stockholm, which had different water sources and treatmentprocesses. Acinetobacter lwoffii, which is known as an opportunisticpathogen, was also widely detected in water supply systems (Scottand Pepper, 2010). If these ubiquitously present bacteria, whichhave potent endotoxic activity, multiply in water distribution sys-tems, significant elevations of endotoxin levels in water couldoccur.

3.2. LPS purification from pure bacterial cultures, river water, andBAC effluent

Table 2 shows the changes in endotoxic activities and dryweights during the LPS purification steps from the pure bacterialcultures and the indigenous bacterial cells in the river water or theBAC effluent samples. The results showed that the endotoxic ac-tivities per dry weight were not always increased by purification inthe case of pure bacterial cultures. Theoretically, the total endotoxicactivities per dry weight should have increased as the purityincreased due to the removal of many of the coexisting cellularcomponents. In fact, the dry weights after purification were largelyreduced by 0.8e10%. However, the total endotoxins were increasedin some cases in pure culture samples after phenol extraction,which destroys cell membranes and removes proteinaceous com-pounds contained in bacterial cells by denaturing. In contrast, theendotoxic activities per unit weight of indigenous bacteria in watersamples were increased step-by-step during purification. Thedifferent behaviors of the endotoxic activities in the environmentalsamples from the cultured samples during purification might arisedue to changes in the association state of the LPSs and in thedamage to the cell membrane, or due to a non-specific LAL reactionwith some components contained in the bacterial cultures, but theexact reason is still unclear. An additional data accumulationwouldbe needed.

The endotoxic activities per dry weight in the purified LPSs arecompared in Fig. 2. Synechococcus sp. exhibited the highest endo-toxic activity of 50 EU/ng, and the activities of Aquabacteriumcommune and Escherichia coli followed at levels greater than 10 EU/ng. Generally, the endotoxic activities of purified LPSs fromEscherichia coli were reported in the range of 5e10 EU/ng(Anderson et al., 2003; Bernardov�a et al., 2008). The endotoxicactivity of the purified LPS from E. coli NBRC 3301 cultured in ourlaboratory was consistent with that value, and this consistencysuggests that the purification procedure was adequate.

The endotoxic activity of the LPS purified from the river watersample was comparable with the LPSs purified from two Pseudo-monas species. In contrast, the LPS purified from the BAC effluent

Table 2Changes of endotoxic activities and dry weight during purification steps.

Bacteria/watersamples

Harvested After phenol extraction Purified

Endotoxin(EU)

Dryweight (mg)

Endotoxin per dryweight (EU/ng)

Endotoxin(EU)

Dryweight (mg)

Endotoxin per dryweight (EU/ng)

Endotoxin (EU) Dry weight(mg)

Endotoxin per dryweight (EU/ng)

E. coli 4.1 � 107 280 1.48 � 10�1 7.4 � 108 29.6 2.49 � 101 3.8 � 107 2.5 1.51 � 101

M. aeruginosa 3.9 � 105 34.7 1.12 � 10�2 9.9 � 104 8.9 1.11 � 10�2 5.4 � 103 1.3 4.13 � 10�3

Synechococcus sp. 1.4 � 108 194 7.12 � 10�1 6.4 � 107 15.2 4.18 � 100 1.7 � 108 3.3 5.06 � 101

River water 9.6 � 106 20.3 4.72 � 10�1 4.8 � 106 4.4 1.09 � 100 2.2 � 106 1.7 1.29 � 100

P. fluorescens 1.9 � 109 516 3.67 � 100 2.6 � 108 104 2.49 � 100 3.0 � 107 13.3 2.27 � 100

P. aeruginosa 9.3 � 108 339 2.74 � 100 4.3 � 108 70.2 6.13 � 100 5.5 � 106 3.9 1.41 � 100

A. commune 6.2 � 108 106 5.84 � 100 6.9 � 108 16.3 4.23 � 101 3.2 � 107 1.4 2.29 � 101

A. delafieldii 2.0 � 107 266 7.51 � 10�2 1.6 � 108 37.0 4.32 � 100 1.0 � 105 2.3 4.35 � 10�2

A. lwoffii 1.4 � 108 111 1.26 � 100 5.2 � 108 14.3 3.64 � 101 9.5 � 105 1.5 6.32 � 10�1

M. fujisawaense 7.3 � 101 140 5.20 � 10�4 2.7 � 101 12.5 2.16 � 10�3 5.9 � 10�1 2.8 2.09 � 10�7

BAC effluent 4.3 � 105 6.0 7.13 � 10�2 4.5 � 105 1.4 3.24 � 10�2 3.4 � 104 0.6 5.73 � 10�2

Y. Ohkouchi et al. / Toxicon 97 (2015) 23e31 27

sample showed much lower endotoxic activity, less than 0.1 EU/ng.Also Microcystis aeruginosa LPS revealed much lower endotoxinactivity than E. coli LPS as reported by Bl�ahov�a et al. (Bl�ahov�a et al.,2013).

3.3. Comparison of inflammatory responses caused by the purifiedLPSs

The IL-8 and TNF-a secretions from THP-1 after six hoursstimulationwith each purified LPS are presented in Fig. 3 and Fig. 4,respectively. The cell counts of THP-1 after six hours stimulationshowed no significant change in the examined range of LPSs puri-fied from different lab-cultured bacteria/cyanobacteria. Each bac-terial/cyanobacterial LPS induced different potencies and patternsof inflammatory responses in THP-1. Most LPSs induced higher IL-8secretions than TNF-a, whereas only Acinetobacter lwoffii LPSinduced TNF-a secretions that were twice as high as IL-8. Methyl-obacterium fujisawaense LPS, which showed very little endotoxicactivity, induced IL-8 at 36 pg/106 cells, even at 0.2 EU/mL. Becauseof its extremely low endotoxic activities, the LPS from Methyl-obacterium fujisawaensewas excluded from the following analyses.The IL-8 secretion levels caused by the other LPSs were categorizedinto three groups: the first group including Acinetobacter lwoffii andMicrocystis aeruginosa showed high levels of IL-8 secretion over1000 pg/106 THP-1 cells at a dose of 1000 EU/mL. The second groupincluding Escherichia coli and Synechococcus sp. showed moderateIL-8 secretions (approximately 100e400 pg/106 cells) at a 1000 EU/mL dose. The last group including two Pseudomonas sp.,

Fig. 2. Comparison of endotoxin contents based on dry weight in purified LPS samplesfrom pure-cultured bacteria/cyanobacteria and water samples.

Aquabacterium commune, and Acidovorax delafieldii exhibited fewIL-8 secretions over the entire endotoxin range. Regarding the twoPseudomonas spp. and Aquabacterium commune, very slight in-creases in IL-8 secretion were observed at the extremely highendotoxin dose of 10000 EU/mL, but no significant doseeresponserelationship was observed.

As described previously, the TNF-a secretions were much lowerthan the IL-8 secretions in most bacterial samples other than Aci-netobacter lwoffii LPS. These results are consistent with previousreports concluded that gram-negative bacteria more stronglyinduced IL-10-, IL-8-, and IL-6-type cytokines than gram-positivebacteria, while gram-positive bacteria induced more of othertypes of cytokines such as IFN-g, TNF-a, and IL-1b (Skovbjerg et al.,2010). However, only Acinetobacter lwoffii showed greater TNF-asecretion in the examined endotoxin range. Erridge et al. (2007)reported that Acinetobacter baumannii and Acinetobacter‘genomespecies 9’ LPSs showed high potency to induce IL-8 andTNF-a in differentiated THP-1 cells. They indicated that stimulationwith Acinetobacter LPSs induced significantly higher TNF-a secre-tion even at 1 ng/mL LPS. They did not mention the levels ofendotoxic activity of the Acinetobacter LPSs in that research, butthey explained that Acinetobacter baumannii LPS exhibited endo-toxic activity and cytokine secretions equal to those of Escherichiacoli LPS. In our investigation, Figs. 3 and 4 clearly show that thepurified A. lwoffii LPS induced 8- and 20-times higher secretions ofIL-8 and TNF-a, respectively, than did E. coli at a dose of 1000 EU/mL, although A. lwoffii LPS had one-tenth the endotoxic activity perdry weight of E. coli LPS as shown in Fig. 2. Acinetobacter spp.including A. lwoffii are known not just as opportunistic pathogensbut also ubiquitous habitant in biofilm in water distribution sys-tems. (Scott and Pepper, 2010; Vaz-Moreira et al., 2013; Kelly et al.,2014). This result suggests that inflammatory potency can increasewith the regrowth of Acinetobacter sp. in distribution systems eventhough the endotoxin activity remains low.

3.4. Comparison of inflammatory potencies of river water and BACeffluent

The IL-8 and TNF-a secretions from THP-1 stimulated with theLPSs purified from the river water and the BAC effluent samples arealso presented in Figs. 3 and 4, respectively. Recently data onendotoxin levels in various aquatic environments has beencollected all over the world. For example, several studies reportedthat endotoxin levels in drinking water derived from treated sur-face water were in the range of 1e10 EU/mL (Ohkouchi et al., 2009;Can et al., 2013). Usually the endotoxin levels in surface water werea few hundred EU/mL, but they rose above 1000 EU/mL if the sur-facewater was impacted by effluents from sewage treatment plants

Fig. 3. The effects of purified LPSs from various pure-cultured bacteria/cyanobacteria and water samples on the IL-8 secretion of the THP-1 cell-line. The LPSs purified from (a)M. fujisawaense ( � ), (b) A. lwoffii (C), M. aeruginosa (-), A. delafieldii (B), river water (A), BAC effluent (◊), and (c) E. coli (▫), Synechococcus sp. (-), P. fluorescens (:),P. aeruginosa (▵), A. commune (C) were used for 6 h stimulation. The cytokine data represent means of two independent assays.

Y. Ohkouchi et al. / Toxicon 97 (2015) 23e3128

(Ohkouchi et al., 2007) or cyanobacterial bloom (Rapala et al.,2002). Based on our result, relatively strong cytokine secretionscaused by stimulation with river water and BAC effluent wereobserved at normal endotoxin levels in an aquatic environment,around 100 EU/mL (104e105 cells/mL as total cells). Of special noteis that the BAC effluent LPS induced a moderate inflammatoryresponse (approx. 250 pg/106 THP-1 cells) even at 10 EU/mL, whichis the same endotoxin level as in drinking water. At LPS doses above500 EU/mL, cytokine secretions were increased in a dose-dependent manner by stimulation with the river water LPS,whereas cytokine levels did not increase significantly by stimula-tion with the BAC effluent LPS.

Principal component analysis (PAC) was then applied to groupthe purified LPS samples, except Methylobacterium fujisawaense,based on their cytokine profiles at a dose of 1000 EU/mL and theircellular endotoxin contents. The endotoxin dose of 1000 EU/mLwas chosen by considering actual endotoxin levels in aquatic en-vironments. The results of the analysis are presented in Fig. 5. ThePCA score plot showed that the river water LPS, the BAC effluentLPS, and Acinetobacter lwoffii LPS formed a group distinct from theother bacterial/cyanobacterial LPSs. This group represented bacte-ria that induced strong cytokine secretions in human monocytesalthough their endotoxic activities per cell were moderate (approx.10�3 EU/cell). In contrast, Aquabacterium commune LPS exhibitedvery high endotoxin activity but very low induction of inflamma-tory responses and was distant from the other LPSs. These results

Fig. 4. The effects of purified LPSs from various pure-cultured bacteria/cyanobacteria and wM. fujisawaense ( � ), (b) A. lwoffii (C), M. aeruginosa (-), A. delafieldii (B), river waterP. aeruginosa (▵), A. commune (C) were used for 6 h stimulation. The cytokine data repre

prove that Acinetobacter lwoffii and indigenous bacteria in aquaticenvironments have relatively high cellular endotoxin contents andreveal a common pattern by which significant cytokine secretionscan be induced in human cells.

Thus, LPSs purified from both the river water and the BACeffluent inducedmuch stronger inflammatory responses at doses of10e5000 EU/mL than did the LPSs of most lab-cultured bacteria/cyanobacteria. At the same time, both water samples induced veryhigh TNF-a secretions as well as IL-8 secretions like AcinetobacterLPS rather than the LPSs purified from other lab-cultured bacteria/cyanobacteria. The reasons for such stronger responses by stimu-lation with the indigenous bacterial LPSs have not yet been iden-tified, but there are several possibilities including the following:

1) In aquatic environments, bacteria like Acinetobacter lwoffii,which have LPSs that cause strong inflammatory responses in hu-man monocytes, are present in abundance; 2) low water temper-ature or low nutritional/osmotic conditions in aquaticenvironments can increase the potency of indigenous bacterialLPSs; and 3) some substances that enhance inflammatory re-sponses exist in those purified samples. The third possibility will bediscussed in a later section. Kawahara et al. (2002) and Dentovskayaet al. (2008) reported that the changes in the structure and in-flammatory potency of Yersinia pestis LPS depend on the growthtemperature. According to their results, the purified LPS fromY. pestis grown at 25 or 27 �C could induce stronger TNF-a secretionin mouse and human macrophages than that grown at 37 �C. They

ater samples on the TNF-a secretion of the THP-1 cell-line. The LPSs purified from (a)(A), BAC effluent (◊), and (c) E. coli (▫), Synechococcus sp. (-), P. fluorescens (:),sent means of two independent assays.

Fig. 5. Principal component analysis of various LPS samples purified in this investi-gation. The first and second principal components explained 63.8% and 33.4% of thetotal variability, respectively.

Y. Ohkouchi et al. / Toxicon 97 (2015) 23e31 29

also proved that hexaacylated lipid-A, which is synthesized inbacteria only grown at low temperatures, contributes to thepotentiation of inflammatory responses. In addition, Ernst et al.(2006) indicated that the LPS structure of Pseudomonas aerugi-nosa change depending on temperature for cultivation. Especially,hexa-acylated LPS was dominant at 15 �C while after growth atelevated temperature penta-acylated LPS was dominant. Theysuggest that the dominance of hexa-avylated LPS at lower tem-perature could be attributed to lower deacylase activity, which isdeficient in one-third of clinical isolates. In our investigation, boththe water samples from the river and from the BAC process weretaken in winter when the water temperature was below 10 �C,whereas all bacteria/cyanobacteria except for Escherichia coli werecultured at 20 �C in the laboratory in order to make conditionscloser to the natural aquatic environment. Based on the above in-formation, as a consequence of a decrease in deacylase activity

Fig. 6. Effects of polymixin B addition as an inhibitor of endotoxin on cytokine secretion isamples and with LPS samples pretreated with polymixin B, respectively. The cytokine data ra significant difference in comparison to the samples without the polymixin B pretreatmen

caused by water temperature or other environmental conditions inthe aquatic environment, LPS structures of indigenous bacteria mayshift toward the more potent.

3.5. Effect of contaminants in purified LPSs on TNF-a secretions

As discussed above, there is a possibility that contaminatedsubstances that remained in the purified LPSs of the river water andthe BAC effluent contribute to the elevation of TNF-a secretions. Inorder to exclude this possibility, the cytokine secretions from THP-1stimulated with each LPS with/without polymixin B as an inhibitorof endotoxins were evaluated. The results are presented in Fig. 6.

Except for the LPS samples from the BAC effluent, the addition ofpolymixin B significantly inhibited both cytokine secretions. Thisresult proved that the LPS components in the purified sample fromriverwater greatly contributed to the cytokine elevations, includingTNF-a, as with the Acinetobacter LPS. In contrast, the cytokine levelsstimulated with the BAC effluent LPS were not affected by theaddition of polymixin B. The contribution of contaminant sub-stances such as peptidoglycan or lipoteichoic acid to the strongcytokine secretions was suspected. However, the contribution ofcontaminant substances in the BAC effluent LPS was unlikelybecause several researches reported that more than 100-foldamounts of contaminated substances such as peptidoglycan orlipoteichoic acid were required to induce the same level of cytokinesecretions with LPSs. Another reason that polymixin B failed toinhibit the secretions of both types of cytokines, IL-8 and TNF-a, inthat sample may be the lower affinity of polymixin B with the LPSsin the BAC effluent. Kasuga et al. (2011) reported that differentbacterial communities from those in the water sources weredeveloped on the BAC surface installed after the ozonation process.Such a difference could have resulted in the diversity in the LPSchemical structures in the BAC effluent sample. At the same time,several researchers also suggested that the surrounding conditionssuch as thewater temperature or osmosis can also contribute to thediversity of the chemical structures of LPSs in the environmentalsamples (Merino et al., 1998; Suomalainen et al., 2010). On theother hand, Coyne and Fenwick (1993) reported that the ability ofpolymixin B to inhibit TNF-a synthesis in macrophages hadconsiderable variability. For example, they showed that polymixinB of 1 IU/mL completely inhibited TNF-a synthesis in macrophagesin the presence of LPS from Escherichia coli B4:O111. The relative

n the THP-1 cell line. IL-8 and TNF-a were determined after 6 h stimulation with LPSepresent means with standard deviations of four independent assays. Asterisks indicatet.

Y. Ohkouchi et al. / Toxicon 97 (2015) 23e3130

TNF-a synthesis with polymixin B to the one without polymixin Bwas 25.3% for Salmonella minnesota, whereas much higher TNF-asyntheses were observed in the presence of LPSs fromS. Typhimurium (98.2%), K, pneumoniae (96.5%), Pseudomonas aer-uginosa (73.4%). They also proposed that the efficiency of inhibitionwas affected by the saturated carbon chains of LPSs and the phos-phate groups of the lipid A core moiety. Based on the above infor-mation, it is more likely that the LPSs with various chemicalstructures in the BAC effluent led to a decrease in the inhibitionability of polymixin B as a consequence of a lower capability to formcomplexes with LPSs.

4. Conclusions

Our results indicated that the LPSs of some bacteria that areubiquitous in aquatic environments or distribution systems caninduce stronger inflammatory responses in human cells than E. coliLPS, which is commonly used as a reference standard for endotoxindetermination. Some of these LPSs, such as Acinetobacter lwoffii,indigenous bacteria in river water and BAC effluent, also exhibited aunique response pattern to induce TNF-a secretions comparable toor higher than the IL-8 secretions, whereas most other LPSsinduced more IL-8 than TNF-a. These results suggest that naturalbacterial/cyanobacterial flora in aquatic environments and waterdistribution systems have the potential to induce relatively stronginflammatory responses in humans and that water quality moni-toring based on endotoxin levels is not sufficient for evaluatinghealth impacts. The discussion of the inflammatory potency ofwater samples is in its early stages (Wichmann et al., 2004; Bl�ahov�aet al., 2013; Marghani et al., 2014), whereas the health impacts ofcyanobacteria/bacteria have been discussed for a long time fromthe perspectives of their pathogenicity, exotoxins, and drug resis-tance. A current climate change we are facing can affect bacterial/cyanobacterial flora in aquatic environments and thus their in-flammatory potency in response to the flora shifts. An effort toaccumulate data on the inflammatory potency of various watersamples in aquatic environments, especially those used for recre-ation or tap water, should be continued.

Acknowledgments

This research was supported by a GSGES Research Grant forYoung Researchers 2011, Kyoto University.

Transparency document

Transparency document related to this article can be foundonline at http://dx.doi.org/10.1016/j.toxicon.2015.02.003.

References

Anderson, W.B., Mayfield, C.I., Dixon, D.G., Huck, P.M., 2003. Endotoxin inactivationby selected drinking water treatment oxidants. Water Res. 37, 4553e4560.

Bernardov�a, K., Babica, P., Mar�s�alek, B., Bl�aha, L., 2008. Isolation and endotoxinactivities of lipopolysaccharides from cyanobacterial cultures and complexwater blooms and comparison with the effects of heterotrophic bacteria andgreen alga. J. Appl. Toxicol. 28, 72e77.

Bl�ahov�a, L., Adamovský, O., Kubala, L., �Svih�alkov�a �Sindlerov�a, L., Zounkov�a, R.,Bl�aha, L., 2013. The isolation and characterization of lipopolysaccharides fromMicrocystis aeruginosa, a prominent toxic water bloom forming cyanobacteria.Toxicon 76, 187e196.

Can, Z., Wenjun, L., Wen, S., Minglu, Z., Lingjia, Q., Cuiping, L., Fang, T., 2013.Endotoxin contamination and control in surface water sources and a drinkingwater treatment plant in Beijing, China. Water Res. 47, 3591e3599.

Carmichael, W.W., Bent, P.E., 1981. Hemagglutination method for detection offreshwater cyanobacteria (blue-green algae) toxins. Appl. Environ. Microbiol.41, 1383e1388.

Coyne, C.P., Fenwick, B.W., 1993. Inhibition of lipopolysaccharide-induced macro-phage tumor necrosis factor-alpha synthesis by polymyxin B sulfate. Am. J. Vet.

Res. 54, 305e314.Dehus, O., Hartung, T., Hermann, C., 2006. Endotoxin evaluation of eleven lipo-

polysaccharides by whole blood assay does not correlate with Limurusamebocyte lysate assay. J. Endotoxin Res. 12, 171e180.

Dentovskaya, S.V., Bakhteeva, I.V., Titareva, G.M., Shaikhutdinova, R.Z.,Kondakova, A.N., Bystrova, O.V., Lindner, B., Knirel, Y.A., Anisimov, A.P., 2008.Structural diversity and endotoxic activity of the lipopolysaccharide of Yersiniapestis. Biochem. (Mosc.) 73, 192e199.

El Marghani, A., Pradhan, A., Seyoum, A., Khalaf, H., Ros, T., Forsberg, L.H.,Nermark, T., Osterman, L., Wiklund, U., Ivarsson, P., Jass, J., Olsson, P.E., 2014.Contribution of pharmaceuticals, fecal bacteria and endotoxin to the inflam-matory responses to inland waters. Sci. Total Environ. 488e489, 228e235.

Ernst, R.K., Adams, K.N., Moskowitz, S.M., Kraig, G.M., Kawasaki, K., Stead, C.M.,Trent, M.S., Miller, S.I., 2006. The Pseudomonas aeruginosa lipid A deacylase:selection for expression and loss within the cystic fibrosis airway. J. Bacteriol.188, 191e201.

Erridge, C., Moncayo-Nieto, O.L., Morgan, R., Young, M., Poxton, I.R., 2007. Acineto-bacter baumannii lipopolysaccharides are potent stimulators of human mono-cyte activation via Toll-like receptor 4 signalling. J. Med. Microbiol. 56, 165e171.

Furuhata, K., 2004. Heterotrophic bacteria in tap water. Shokuhin Eiseigaku Zasshi45, J1eJ5 (in Japanese).

Furuhata, K., Kato, Y., Goto, K., Hara, M., Yoshida, S., Fukuyama, M., 2006. Isolationand identification of Methylobacterium species from the tap water in hospitalsin Japan and their antibiotic susceptibility. Microbiol. Immunol. 50, 11e17.

Hansen, L.A., Poulsen, O.M., Wurtz, H., 1999. Endotoxin potency in the A549 lungepithelial cell bioassay and the Limulus amebocyte lysate assay. J. Immunol.Methods 226, 49e58.

Hirata, T., Tosa, K., Kawamura, K., Nakajima, J., Kaneko, M., Taguchi, K., 1993. Het-erotrophic bacteria in chlorinated drinking water distribution systems: detec-tion and identification. Water Sci. Technol. 27, 155e158.

Hoson, T., Kaji, S., Mizumachi, M., 2002. Occurrence of picophytoplankton in YodoRiver Basin and its effect on turbidity control in water treatment system.J. Water Waste 44, 755e762 (in Japanese).

Kalmbach, S., Manz, W., Wecke, J., Szewzyk, U., 1999. Aquabacterium gen. nov., withdescription of Aquabacterium citratiphilum sp. nov., Aquabacterium parvum sp.nov. and Aquabacterium commune sp. nov., three in situ dominant bacterialspecies from the Berlin drinking water system. Int. J. Syst. Bacteriol. 49 (Pt 2),769e777.

Kasuga, I., Kurisu, F., Furumai, H., Iwamoto, T., 2011. Development of microbialcommunity on granular activated carbon used for advanced drinking watertreatment. In: Proceedings of the 4th IWA-ASPIRE (Tokyo), pp. 141e148.

Kawahara, K., Kuraishi, H., Zahringer, U., 1999. Chemical structure and function ofglycosphingolipids of Sphingomonas spp. and their distribution among mem-bers of the alpha-4 subclass of Proteobacteria. J. Ind. Microbiol. Biotechnol. 23,408e413.

Kawahara, K., Tsukano, H., Watanabe, H., Lindner, B., Matsuura, M., 2002. Modifi-cation of the structure and activity of lipid A in Yersinia pestis lipopolysaccha-ride by growth temperature. Infect. Immun. 70, 4092e4098.

Kawasaki, S., Moriguchi, R., Sekiya, K., Nakai, T., Ono, E., Kume, K., Kawahara, K.,1994. The cell envelope structure of the lipopolysaccharide-lacking gram-negative bacterium Sphingomonas paucimobilis. J. Bacteriol. 176, 284e290.

Kelly, J.J., Minalt, N., Culotti, A., Pryor, M., Packman, A., 2014. Temporal variations inthe abundance and composition of biofilm communities colonizing drinkingwater distribution pipes. PLoS One 9, e98542.

Lee, J., Lee, C.S., Hugunin, K.M., Maute, C.J., Dysko, R.C., 2010. Bacteria from drinkingwater supply and their fate in gastrointestinal tracts of germ-free mice: aphylogenetic comparison study. Water Res. 44, 5050e5058.

Merino, S., Aguilar, A., Rubires, X., Tomas, J.M., 1998. Mesophilic Aeromonas strainsfrom different serogroups: the influence of growth temperature and osmolarityon lipopolysaccharide and virulence. Res. Microbiol. 149, 407e416.

O'Toole, J., Sinclair, M., Jeavons, T., Leder, K., 2008. Alternative water sources andendotoxin. Water Sci. Technol. 58, 603e607.

Ohkouchi, Y., Ishikawa, S., Takahashi, K., Itoh, S., 2007. Factors associated withendotoxin fluctuation in aquatic environment and characterization of endo-toxin removal in water treatment process. Environ. Eng. Res. 44, 247e254 (inJapanese).

Ohkouchi, Y., Ly, B.T., Itoh, S., 2009. Detection of bacterial regrowth in water dis-tribution system using endotoxin as an alternative indicator. Adv. Asian Envi-ron. Eng. 8, 13e19.

Ohkouchi, Y., Tajima, S., Nomura, M., Itoh, S., 2012. Comparison of inflammatoryresponses in human cells caused by lipopolysaccharides from Escherichia coliand from indigenous bacteria in aquatic environment. J. Environ. Sci. Health AToxic Hazard Subst. Environ. Eng. 47, 1966e1974.

Poitelon, J.B., Joyeux, M., Welte, B., Duguet, J.P., Prestel, E., DuBow, M.S., 2010. Var-iations of bacterial 16S rDNA phylotypes prior to and after chlorination fordrinking water production from two surface water treatment plants. J. Ind.Microbiol. Biotechnol. 37, 117e128.

Rapala, J., Lahti, K., Rasanen, L.A., Esala, A.L., Niemela, S.I., Sivonen, K., 2002. En-dotoxins associated with cyanobacteria and their removal during drinkingwater treatment. Water Res. 36, 2627e2635.

Scott, B.A., Pepper, I.L., 2010. Water distribution systems as living ecosystems:impact on taste and odor. J. Environ. Sci. Health A Toxic Hazard Subst. Environ.Eng. 45, 890e900.

Skovbjerg, S., Martner, A., Hynsjo, L., Hessle, C., Olsen, I., Dewhirst, F.E., Tham, W.,Wold, A.E., 2010. Gram-positive and gram-negative bacteria induce different

Y. Ohkouchi et al. / Toxicon 97 (2015) 23e31 31

patterns of cytokine production in human mononuclear cells irrespective oftaxonomic relatedness. J. Interferon Cytokine Res. 30, 23e32.

Suomalainen, M., Lobo, L.A., Brandenburg, K., Lindner, B., Virkola, R., Knirel, Y.A.,Anisimov, A.P., Holst, O., Korhonen, T.K., 2010. Temperature-induced changes inthe lipopolysaccharide of Yersinia pestis affect plasminogen activation by the plasurface protease. Infect. Immun. 78, 2644e2652.

Szita, G., Gyenes, M., Soos, L., Retfalvi, T., Bekesi, L., Csiko, G., Bernath, S., 2007.Detection of Pseudomonas aeruginosa in water samples using a novel syntheticmedium and impedimetric technology. Lett. Appl. Microbiol. 45, 42e46.

Vaz-Moreira, I., Egas, C., Nunes, O.C., Manaia, C.M., 2013. Bacterial diversity from thesource to the tap: a comparative study based on 16S rRNA gene-DGGE andculture-dependent methods. FEMS Microbiol. Ecol. 83, 361e374.

Wichmann, G., Daegelmann, C., Herbarth, O., Strauch, G., Schirmer, K.,Wostemeyer, J., Lehmann, I., 2004. Inflammatory activity in river-water sam-ples. Environ. Toxicol. 19, 594e602.

Yano, T., Hashimoto, K., Tashiro, N., Tanamachi, C., Horita, R., Itoyama, T., Ishii, K.,Sagawa, K., 2009. Bacterial contamination study for in-hospital water supplyinstallation. Jpn. J. Environ. Infect. 24, 303e311 (in Japanese).

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具