USING GENOMICS TO UNDERSTAND DISINFECTION WITH SILVER

1 downloads 0 Views 290KB Size Report
Disinfection with silver increasingly is being considered as an alternative to ... than environmental challenges (Beloin et al., 2004; Gabig and Wegrzyn, 2001; ...
WEFTEC®.06

USING GENOMICS TO UNDERSTAND DISINFECTION WITH SILVER Mau-Yi Wu* and Daniel. B. Oerther** *Department of Civil and Environment Engineering, University of Cincinnati, 765 Baldwin Hall (ML0071), Cincinnati, OH 45221-0071, USA (E-mail: [email protected]) **Department of Civil and Environment Engineering, University of Cincinnati, 765 Baldwin Hall (ML0071), Cincinnati, OH 45221-0071, USA (E-mail: [email protected]) ABSTRACT Disinfection with silver increasingly is being considered as an alternative to disinfection with oxidizing agents in certain water quality protection scenarios. The objective of this study was to identify the genes involved in the toxic response of Escherichia coliT when challenged with silver. Whole genome Affymetrix microarray results demonstrated that silver resistance was related to zinc resistance genes as well as copper resistance genes. Specifically, the transcription of periplasmic zinc-binding protein gene cluster including the regulatory genes (zraRS) and protein gene (zraP) were upregulated during early exponential phase in E. coliT. The microarray result also suggested the energy used to resist silver toxicity would come from repressing flagellin and chemotaxis gene expressions. KEYWORDS Chemotaxis, Escherichia coli, flagellum, microarray, silver, silver resistance mechanism, zraP, zraRS. INTRODUCTION Controlling harmful microorganisms in aquatic environments is an important aspect of the water quality profession. Often, oxidizing agents such as ozone, chlorine, monochloramine, and peroxide are used as disinfectants to control microbial growth. Alternative disinfectants also exist. For example, silver has been used for many years as a disinfectant for hot water systems as well as a topical disinfectant in medical applications (e.g., burn salve) (Lin et al., 2002; Sreekumari et al., 2005). Ag(I) as a disinfectant agent has several advantage. Ag(I) does not produce harmful disinfection byproducts; has a higher heat resistance compared to other disinfectants; discharges of Ag(I) to the environment do not pose significant threats to environmental nor human health (Russel and Hugo, 1994); and in recent years silver has become more affordable as the use of digital photography has greatly reduced the use of silver-based photographic film. Because of the mild disinfectant characteristic of Ag(I), it can suffer from significant silver resistance rendering it useless as a disinfectant. For example, Rohr et al. (1999) reported that the long term efficacy of silver ion decreased in hospital hot water systems, presumably due to the build up of silver resistant organisms. For improving the use of Ag(I) as the disinfectant, it is important to understand the genetic response of bacteria when challenged with silver. Gupta et al. (1999) first reported silver resistance mechanisms in bacteria. They observed one silver specific binding protein and two Copyright ©2006 Water Environment Foundation. All Rights Reserved

1285

WEFTEC®.06

parallel silver efflux pump systems in the silver resistance plasmid, pMG101, of Samonella. These gene clusters are similar to the copper resistance gene clusters on the Escherichia coli genome. Frank et al. (2001) mentioned the ybdE(cusA), one of the copper resistance gene, is responsible for detoxification of Ag(I) for Escherichia coli K38. Although these early studies point to some of the mechanisms involved in the resistance to silver, they have not been performed using modern genome-based analysis tools such as whole-genome microarrays. Thus, the mechanisms of silver resistance for E. coli and other organisms remain to be studied to answer the currently unexplained phenomenon of ‘silver-resistance’ where the efficacy of silver to control harmful microorganisms fails unexpectedly (Gupta and Silver, 1998; Rohr et al., 1999). In this study, the central objective was to identify changes in transcription levels when when Escherichia coli was challenged with silver ions. Commercially available, whole-genome microarrays were used to analyze every single open reading frame (ORF) in E. coli (Campbell and Ghazal, 2004). This approach is novel because most whole genome studies have focused primarily upon the identification of changes in transcription levels related to pathogenicity rather than environmental challenges (Beloin et al., 2004; Gabig and Wegrzyn, 2001; Methé et al., 2005). In this research, Affymetrix Gene Chip® E. coli Antisense Genome array were used to identify the genes involved when E. coli responded to silver toxicity. It was believed that this improved information would provide a more comprehensive background to understand silver disinfect efficacy. METHODOLOGY Silver toxicity test Escherichia coliT (American Type Culture Collection 11775) was cultivated in 100 ml of Tryptone-Glucose-Yeast (TGY) extract medium (per liter of water: 5 g tryptone, 5 g yeast extract, 1 g glucose, 1 g K2HPO4) in 250 ml erlenmeyer flask at 37oC under aerobic conditions on a New Brunswick model C24 rotary shaker at 200rpm. Conventional batch growth curves were constructed by diluting exponential-phase pure cultures 2000-fold with fresh media and measuring increases in turbidity overtime using a Geneys Spectrophotometer at a wavelength of 600 nm. To test the toxicity of silver on E. coliT, batches of TGY media were prepared with various levels of Ag(I) ranging from 0 parts per million (ppm) to 6 ppm of Ag(I) ion (prepared from AgNO3(s)) in the TGY media. This experiment was repeated for three different pH values (6.5, 7.0, and 7.5) to evaluate pH has influence on the silver toxicity. The Ag(I) concentration of the TGY medium was confirmed using an atomic absorption spectrometer (AAS) (Perkin Elmer AAnalyst 300). E. coliT Cultivation for the microarray test Triplicate cultures of E. coliT samples were prepared using pH 6.5 TGY medium with either 0 or 2.9 ppm Ag(I) (measured by the AAS) at 37oC under aerobic conditions on a New Brunswick model C24 rotary shaker at 200rpm. The batch growth was started by diluting exponential-phase pure cultures 2000-fold with fresh media and measured the turbidity overtime by using a Geneys Spectrophotometer at a wavelength of 600 nm. At early exponential phase (optical density at

Copyright ©2006 Water Environment Foundation. All Rights Reserved

1286

WEFTEC®.06

600 nm of 0.1) during batch cultivation, 80 ml samples were taken, centrifuged at 4,000xg for 2 min, the spent media was decanted, and cell pellets were preserved at -80oC. RNA extraction Total RNA was isolated from the pellets by the low pH, hot phenol: chloroform method (Stahl et al., 1988). The concentration of total RNA was determined by measuring absorbance at 260nm on the spectrophotometer (1 absorbance unit = 40 µg/mL RNA). The concentration of total RNA in each triplicate subsample from each treatment was adjusted to 2 µg/µl. The triplicate subsamples were pooled to reduce variation among gene expression levels. The remaining triplicate samples were preserved at -80oC for the following real-time PCR experiment. cDNA preparation and microarray hybridization The commercial gene expression microarray, Affymetrix E. coli Antisense Genome Array, was used to evaluate changes in transcription level. The cDNA synthesis and labeling, hybridizing the target mix to the E. coli genome array, washing staining, and scanning of hybridized the E. coli genome array were performed by Genome Explorations (Memphis, TN) according to “Prokaryotic Sample and Array Processing” ( http://www.affymetrix.com/support/downloads/manuals/expression_s3_manual.pdf [Accessed 05/25/06]). The gene expression data was subsequently analyzed to identify genes that were impacted by Ag(I). Statistical analysis of gene expression Statistical analysis of the microarray data was performed using Affymetrix Micro Array Suite 5.0 (MAS5.0). MAS5.0 takes a robust average on each array of gene expression using one step Tukey’s biweight estimate, where outliers are penalized with low weights. Then, it normalizes arrays by scaling each array that all arrays have the same mean. After the normalization, MAS5.0 used the non-parametric Wilcoxon rank test to look for if the gene expressions of two arrays with different treatments have significant changes. (Affymetrix technical note: Statistical algorithms guide [http://www.uic.edu/depts/rrc/cgf/Recommended_Reading/statistical%20algorithm%20descripti on%20document.pdf, Accessed 05/25/06]). The function of candidate genes was determined using the pathway tools software v9.5 (Keseler et al., 2005) to identify operons that were influenced by Ag(I) exposure. Validation of microarray data by using real-time PCR After the statistic calculation, the gene clusters determined to be significantly up-regulated or down-regulated were identified. At least one gene of each significantly changed cluster (Table 1) was selected for further analysis using relative quantitative real-time polymerase chain reaction (rq-RT-PCR) using an Applied Biosystems Prism 5700 Sequence Detector to validate if the gene cluster truly demonstrated a significant change in expression level. rq-RT-PCR was performed following the procedures outlined in the Chemistry Guide of Sequence Detection Systems for the ABI Prism 5700 Sequence Detector (Applied Biosystems, Inc, CA).

Copyright ©2006 Water Environment Foundation. All Rights Reserved

1287

WEFTEC®.06

Table 1 - Gene whose expression levels were investigated by relative quantitative realtimenee PCR. LocusTag no. b1779 b1881 b1923 b1924 b3036 b3037 b4002

Common name gapA, (Housekeeping gene), GAPDH cheZ, CheY protein phophatase fliC, flagellin, filament structural protein fliD, filament capping protein ygiA, hypothetical protein ygiB, hypothetical protein zraP, Zn-binding periplasmic protein

Primer sequence Forward Reverse TCGTCTGGAAAAAGCTGCAACT TACGTCATCTTCGGTGTAGCCC CATGATGGCGCAGGATTTTC

CCAACAGCACCATCAGCAACT

GTATTCAGGACGCCGACTATGC

CCTGGTTAGCTTTTGCCAACAC

AAACTGGAACTGGATGCCGAC

GCCGGTTTTTTTACCATCGC

TTCTCGCTTCAATTTCGACCAG TGATGGCCGGTTACATGATG GACATCTGCATTTGCTCACGG

TGGCGTATGGATTTTGTCCG ATTTACCGTAAGCCGGACTGG GGATTTTCTGCCACGCTGTCT

The template for the rq-RT-PCR analysis was the same subsamples that had been used for microarray analysis. Two replicate subsamples were analyzed for each condition. Before performing rq-RT-PCR, the commercially available RNAqueous kit (Ambion®) was used to purify total RNA, and potential contamination with DNA was eliminated using DNaseI treatment (DNA-free kit, Ambion®). The 2.5 µg of purified total RNA samples were converted to complementary DNA (cDNA) using the EndoFree Reverse Transcription™ Kit (Ambion®) with random decamer primers (Ambion®). A subsample (1.5 µl) of cDNA was used as the template to perform rq-RT-PCR. The glyceraldehydes-3-phosphate dehydrogenase (GAPDH) gene was selected as a housekeeping gene for the internal control. Primers for target genes and the endogenous control were designed by using Primer Express® Software with an optimal annealing temperature of 60˚C and a PCR product length of 90-110 nts (Table 1). SYBR green I dye was used as the reagent to detect PCR products as they accumulated during PCR cycles. The results of rq-RT-PCR were compared to the results of the microarray analysis. RESULTS Silver toxicity on the Escherichia coliT. Fig. 1 presents the results of batch growth curves for E. coliT treated with various levels of Ag(I) at different pH. The rate of growth of E. coliT during the exponential phase were not changed significantly by the silver challenge. The primary change observed with silver challenge was the delay of exponential growth through an extension of the lag phase. For example, E. coliT treated with 1.5ppm of silver behaved similar to untreated E. coliT whereas the lag phase increased significantly at doses of 2.4, 3.0, or 6.0 ppm of silver. As the pH was increased, the effect of the dose of Ag(I) was reduced suggesting that bioavailability of silver ion was influenced by pH and complexation with hydroxide. According to the result of the initial batch studies, pH 6.5 TGY medium with a silver challenge of 2.9 ppm was used as the bases for the molecular assays.

Copyright ©2006 Water Environment Foundation. All Rights Reserved

1288

WEFTEC®.06

Figure 1 - Results of growth curves with E. coliT exposed to various concentrations of Ag(I), at a pH of 6.5, 7.0, and 7.5. pH 6.5

pH 7.0

pH 7.5

1.0

0 ppm 1.5 ppm 2.4ppm 3.0 ppm 6.0 ppm

0.5

0.0 0

5

10 15 Time (hr)

20

25

absorbance (600 nm)

1.0

1.5

absorbance (600 nm)

1.5

absorbance (600 nm)

1.5

1.0

0 ppm 1.5 ppm 2.4ppm 3.0 ppm 6.0 ppm

0.5

0.0 0

5

10 15 Time (hr)

20

25

0 ppm 1.5 ppm 2.4ppm 3.0 ppm 6.0 ppm

0.5

0.0 0

5

10 15 Time (hr)

20

25

The gene clusters influenced by the silver toxicity A comparison of the levels of gene expression for samples collected from the 0 ppm treatment and the 2.9 ppm treatment showed that 37 ORFs were up-regulated while 69 ORFs were downregulated out of a total possible 4344 predicted ORFs (Table 2). Table 2 - Expression ratio of functional groups. Functional groupc All ORF's Unknown function

Total 4344 1056

Number of genesa Expression ratiob (silver/control) Induced Repressed 37 6

69 23

1601 221

18 6

16 3

250 36

2 0

7 6

Information transfer RNA related protein related

966 526 367

10 5 4

13 3 8

Regulation

508

5

10

Transport

680

6

6

Cell process celll division motility, chemotaxis, energy taxis Adaptations Protection

454 70 55 161 164

8 3 0 4 2

28 0 17 5 5

Cell structure Membrane Flagella

951 702 43

6 6 0

17 9 7

Metabolism energy metabolism, carbon biosybthesis of macromolecules Flagellum

1719 21 Location of gene products Cytoplasm 938 12 Periplasm 133 3 inner membrane 599 8 outer membrane 74 1 a Number of genes showing significant (Wilcoxon Signed Rank Test). b Expression ratio (silver/control)- genes that their relative fluorescence in silver condition was at least 1 fold higher (induced) or 1 fold lower (repressed) of the control. c Partial list of functional group.

Copyright ©2006 Water Environment Foundation. All Rights Reserved

1289

19 9 1 7 2

WEFTEC®.06

Thus, the majority of genes (97.6%) were not impacted by the dose of Ag(I). An analysis of the gene clusters impacted by Ag(I) indicated that the most significant changes occurred in the gene clusters were flagellum and chemotaxis, both of them were down-regulated (Table 2). In the specific gene cluster regarding to the heavy metal resistance, the results showed that none of copper resistance gene clusters (cueO, copA, cusRS, and cusCFBA) had significant change between the silver and no-silver treatments (Table 3). In contrast, the gene clusters for zinc periplasmic binding protein (zraP and zraRS) were significant up-regulated. zraP, the structural gene of zinc-binding periplasmic protein, demonstrated the largest fold increase among the 4344 predicted ORFs. In addition, a previously undefined gene cluster (ygiA and ygiB) was also upregulated suggesting that this operon may play a role when responding to silver toxicity. Table 3 - The expression ratios of zinc and copper resistance proteins Zinc resistance proteins LocusTag no. b3469 b4002 b4003

Common namea

Functiona

zntA, Pb/Cd/Zn/Hg transporting ATPase zraP, Zn-binding periplasmic protein zraS, sensor kinase of Zn/Pb responsive two-component regulatory system

b4004

zraR, response regulator in twocomponent reguatory system with ZraS, regulates zraP expression

zinc-transporting ATPase periplasmic Zn-binding protein sensory histidine kinase in two-component regulatory system with ZraR, regulates zraP expression, senses Zn response regulator of Zn/Pb responsive twocomponent regulatory system

Mean log2 ratiobc +0.3 +6.5 +2.4

Fold Changec +1.2 +90.5 +5.3

+1.0

+2.0

-0.2

-1.12

Copper resistance proteins b0123 b0484 b0570

cueO, Probable periplasmic copper oxidase copA, Cu(I)-translocating P-type ATPase cusS, sensor kinase of copper sensing twocomponent system.

b0571

multicopper oxidase P-type ATPase, copper transporting sensory histidine kinase in two-component regulatory system with CusR, regulation of copper resistance, senses copper ions response regulator of copper sensing system

cusR, response regulator in twocomponent regulatory system with cusS b0572 cusC, outer membrane transport protein putative outer membrane protein involved involved in copper (silver) tolerance in copper transport b0573 ylcC (cusF), putative periplasmic copperputative periplasmic copper-binding protein binding protein b0574 cusB, outer membrane transport protein possible component of copper transport involved in copper (silver) tolerance system b0575 cusA, outer membrane copper (silver) and putative copper transport protein drug transport protein a. Cited from http://chase.ou.edu/oubcf/ [Accessed 05/26/06] b. The log2 ratio (silver/control) less than 1.0 is considered as no significant change; “A” means absent signal. c. “+” means induced by silver; “-“ means suppressed by silver.

A A A A A A A

rq-RT-PCR verse microarray Fig. 2 shows a comparison of the results of rq-RT-PCR and the microarray. As expected, a strong positive correlation (R2 = 0.94) was observed between the complementary molecular methods indicating the consistency of microarray data and rq-RT-PCR results. The signals of selected genes of cheZ, fliC, fliD, and zraP all suggested have been impacted by the Ag(I) (more than 1 fold changes). Unfortunately, the ygiA and ygiB did not appear to be significantly impacted by the silver dose.

Copyright ©2006 Water Environment Foundation. All Rights Reserved

1290

WEFTEC®.06

Figure 2 - Log2 expression ratio from genes determined by relative quantitative real-time PCR (rq-RT-PCR) versus microarray analysis. rq-RT-PCR results were produced by averaging triplicate measurements from two biological replicates. The linear relationship is given by y = 0.7014x - 0.5883 (R2 = 0.94) 5

rq-RT-PCR log 2 ratio

2

R = 0.94

zraP

3 1 -7

-5

-3

fliD -1-1

cheZ fliC

ygiA

ygiB

1

3

5

7

-3 -5

Microarray log2 ratio DISCUSSION In the silver toxicity test on E. coliT, it was demonstrated that no impact is observed until the level of silver is greater than 1.5 ppm. Increasingly dose resulted in increasing lag phase before exponential growth was observed. The results indicate that Ag(I) does not inhibit the growth of E. coliT until reaching a threshold concentration. This result also suggested the silver toxicity in the low ppm can only delay the lag phase of E. coliT; and the early stage of growth is crucial phase for the adaptation of silver toxicity. As the pH was increased, the effect of Ag(I) dose was reduced suggesting that E.coliT is more sensitive to Ag(I) at lower pH. The microarray results suggested the flagellin and chemotaxis were significant downregulated by the Ag(I). The same genetic response has been reported for other adverse conditions such as the presence of high temperature, high concentrations of salts, high concentrations of carbohydrates, high concentrations of low-molecular-weight alcohols (Li et al., 1993; Shi et al., 1993). The silver toxicity is also an adverse condition for E. coliT. The energy cost for synthesis and functioning of a flagellar mobility is high (Soutourina et al., 2001). Therefore, E. coliT would save energy from stopping synthesis these proteins for fighting against silver toxicity. The gene cluster related to silver resistance mechanism is the periplasmic zinc-binding protein, zraP, including its two component regulatory system (zraRS). zraRS is a two-component regulatory system found induced by high Zn(II) and Pb(II) which can activate ZraP to bind the excess of Zn(II) (Leonhartsberger et al., 2001). The connection between zinc and silver resistance mechanism has never been mentioned before. The rq-RT-PCR result also supported this hypothesis. This study showed that the silver resistance mechanism could be related to the

Copyright ©2006 Water Environment Foundation. All Rights Reserved

1291

WEFTEC®.06

zinc resistance mechanism. The copper resistance gene clusters which were discussed in prior studies did not show significant changes in the levels of gene expression in the current study. mentioned in the other researches didn’t show significant signal. CONCLUSIONS This study demonstrated that Ag(I) has more toxicity at a lower pH and could only delay the early stage of grow curve in the low Ag(I) concentration on E. coliT. This study also showed for the first time that zinc resistance genes were related to the silver resistance mechanism. The periplasmic zinc-binding protein gene cluster including the regulatory genes (zraRS) and protein gene (zraP) significantly involved the silver adaptation mechanism in the early exponential phase of E. coliT. The microarray result also suggested the energy reserved for the silver resistance would come from repressing flagellin and chemotaxis gene express; and the microarray pairing with RT-PCR research would be a better way to interpret the gene expression. REFERENCES Beloin, C., Valle, J., Latour-Lambert, P., Faure, P., Kzreminski, M., Balestrino, D., Haagensen, J. A. H., Molin, S., Prensier, G., Arbeille, B., and Ghigo, J. M. (2004). Global impact of mature biofilm lifestyle on Escherichia coli K-12 gene expression. Mol. Microbiol., 51(3), 659-674. Campbell, C. J., and Ghazal, P. (2004). Molecular signatures for diagnosis of infection: application of microarray technology. Journal of Applied Microbiology, 96, 18-23. Franke, S., Grass, G., and Nies, D. H. (2001). The product of the ybdE gene of the Escherichia coli chromosome is involved in detoxification of silver ions. Microbiology, 147, 965-972. Gabig, M. and Wegrzyn, G. (2001). An introduction to DNA chips: principles, technology, applications and analysis. Acta biochimica polonica, 48(3), 615-622. Gupta, A., and Silver S. (1998) Silver as a biocide: will resistance become a problem? Nat. Biotechnol. 16(10) 888. Gupta, A., Matsui, K., Lo, J. F., and Silver, S. (1999). Molecular basis for resistance to silver cations in Salmonella. Nature Medicine, 5(2), 183-188. Keseler, I. M., Collado-Vides, J., Ingraham, J., Paley, S., Paulsen, I. T., Peralta-Gill, M., and Karp, P. D. (2005). EcoCyc: a comprehensive database resource for Escherichia coli. Nucleic Acids Res., 33, D334-D337. Leonhartsberger, S., Huber, A., Lottspeich, F., and Bock, A. (2001). The hydH/G genes from Escherichia coli code for a zinc and lead responsive two-component regulatory system. J. Mo. Biol., 307(1), 93-105. Li, C. Y., Louise, C. J., SHI, W. Y., and Adler J. (1993). Adverse conditions which cause lack of flagella in Escherichia coli. J. Bacteriol., 175(8), 2229-2235. Lin, Y. S. E., Vidic, R. D., Stout, J. E., and Yu, V. V. (2002). Negative effect of high pH on biocidal efficacy of copper and silver ion in controlling Legionella pneumophila. Appl. Environ. Microbiol., 68(6), 2711-2715. Methé, B. A., Webster, J., Nevin, K., Butler, J., and Lovely, D. R. (2005). DNA microarray analysis of nitrogen fixation and Fe(III) reduction in Geobacter sulfurreducens. Appl. Environ. Microbiol., 71(5), 2530-2538.

Copyright ©2006 Water Environment Foundation. All Rights Reserved

1292

WEFTEC®.06

Rohr, U., Senger, M., Selenka, F., Turley, R., and Wilhelm, M. (1999). Four years of experience with silver-copper ionization for control of Legionella in German Universirt hospital hot water plumbing system. Clin. Infect. Dis., 29, 1507-1511. Russel, A. D., and Hugo, W. B. (1994). Antibacterial activity and action of silver. Progress in Medicinal Chemistry, 31, 351-371. Shi, W. Y., Li, C. Y., Louise, C. J., and Alder, J. (1993). Mechanism of adverse conditions causing lack of flagella in Escherichia coli. J. Bacteriol., 175(8), 2236-2240. Soutourina, O. A., Semenova, E. A., Parfenova, V. V., Danchin, A., and Bertin, P. (2001). Control of bacterial motility by environmental factors in polarly flagellated and peritrichous bacteria isolated from Lake Baikal. Appl. Environ. Microbiol., 67(9), 38523859. Sreekumari, K. R., Sato, Y. and Kikuchi, Y. (2005) Antibacterial metals-a viable solution for bacterial attachment and microbiologically influenced corrosion. Materials Transactions 46(7), 1636-1645. Stahl, D. A., Flesher, B., Mansfield, H. R., and Montgomery, L. (1988). Use of phylogenetically based hybridization probes for studies of ruminal microbial ecology. Appl. Environ. Microbiol, 54, 1079-1084.

Copyright ©2006 Water Environment Foundation. All Rights Reserved

1293

Suggest Documents