A Real-Time Polymerase Chain Reaction Assay for Quantification Of

Journal of Aquatic Animal Health 23:178–188, 2011 C American Fisheries Society 2011 ISSN: 0899-7659 print / 1548-8667 online DOI: 10.1080/08997659.2011.637006

ARTICLE

A Real-Time Polymerase Chain Reaction Assay for Quantiﬁcation of Edwardsiella ictaluri in Catﬁsh Pond Water and Genetic Homogeneity of Diagnostic Case Isolates from Mississippi

Matt J. Grifﬁn,* Michael J. Mauel, Terrence E. Greenway, Lester H. Khoo, and David J. Wise Thad Cochran National Warmwater Aquaculture Center, Mississippi State University, 127 Experiment Station Road, Stoneville, Mississippi 38776, USA

Abstract A quantitative polymerase chain reaction (qPCR) assay was developed for the detection and quantification of Edwardsiella ictaluri in channel catfish Ictalurus punctatus pond water using modifications to a published E. ictaluri– specific qPCR assay and previously established protocols for the molecular detection of myxozoan parasites in catfish ponds. Genomic DNA equivalents indicative of the number of bacteria in a sample were determined and standard curves correlating to bacterial numbers were established. The assay was found to be highly repeatable and reproducible, with a linear dynamic range of five orders of magnitude. There was no interference of the assay from the presence of large quantities of nontarget DNA. Known quantities of bacteria were added to sample volumes of 40 or 500 mL of pond water collected from several different ponds. The minimum level of detection was approximately 100 cell equivalents (CE) in 40 (2.5 CE/mL) or 500 mL of pond water (0.2 CE/mL). Sample volumes of 40 mL yielded the most consistent results, which were not significantly different from those obtained from broth culture alone. Cell equivalents determined by qPCR in 40-mL pond water samples spiked with known quantities of bacteria were within one order of magnitude of the actual number of cells added. Repetitive element-based polymerase chain reaction analysis of archived isolates demonstrated the genetic homogeneity of E. ictaluri, and consistent amplification of these isolates by qPCR analysis demonstrated the stability of the PCR target. The assay described here provides a reliable method for the detection and quantification of E. ictaluri in pond water and will be an invaluable tool in epidemiological studies. Additionally, the assay provides a way to evaluate the effects that vaccination, antibiotic treatments, and restricted feeding practices have on E. ictaluri populations during an outbreak. Information obtained with these tools will aid in optimizing disease management practices designed to maximize productivity while minimizing losses.

Enteric septicemia of catfish (ESC), which is caused by the Quantitative polymerase chain reaction (qPCR) has been bacterium Edwardsiella ictaluri (Hawke et al. 1981), is consid- demonstrated to be useful for the quantitative analysis of mi- ered one of the most significant bacterial diseases affecting the croorganisms from a wide variety of environmental samples, culture of channel catfish Ictalurus punctatus in the southeast- including channel catfish ponds (Roe et al. 2001; Stults et al. ern United States. The pathogen is endemic at most commercial 2001; Brinkman et al. 2003; Hallett and Bartholomew 2006; operations and epizootics occur in the spring and fall, when Griffin et al. 2009). Bilodeau et al. (2003) developed qPCR water temperatures are between 20◦C and 28◦C (Francis-Floyd primers and probes for the detection of E. ictaluri in channel et al. 1987). The year-round prevalence of the bacteria in pond catfish, but to our knowledge this approach has not been applied water remains unknown. to the detection and analysis of E. ictaluri in catfish pond water.

*Corresponding author: grifﬁ[email protected] Received October 20, 2010; accepted June 13, 2011 Published online November 28, 2011

178 A REAL-TIME POLYMERASE CHAIN REACTION ASSAY 179

The facilitation of molecular methods of pathogen detection sequences. No alterations were made to the 25-bp TaqMan probe in channel catfish ponds has been encumbered by the presence designed by Bilodeau et al. (2003), which was labeled with a of compounds within the ponds that are inhibitory to the am- fluorescent reporter dye, 6-carboxyfluorescein (FAM), on the 5 plification of nucleic acids by PCR. Heavy metals, humic and end and the quencher dye, black hole quencher-1 (BHQ-1), on phenolic compounds, and constituents of bacterial cells have all the 3 end. been identified as potential inhibitors commonly associated with Sensitivity, specificity, and cross-reactivity. —Five mL were environmental samples (Wilson 1997). Protocols have been de- taken from the cultures of E. ictaluri (S97-773), E. tarda, F. veloped for the detection of the myxozoan parasite Henneguya columnare, and Aeromonas sp. and pelleted by centrifugation at ictaluri in channel catfish ponds using a series of centrifugation 16,000 × gravity (g) for 3 min. DNA isolation was performed steps and a commercial kit designed for the isolation of genomic with the QIAGEN DNeasy Blood and Tissue Kit (QIAGEN, DNA from soil (Griffin et al. 2009). These commercially avail- Hilden, Germany) protocol for gram-negative bacteria. The iso- able kits specifically target and remove environmental inhibitors lated DNA was quantified using a nanodrop spectrophotometer during sample processing. The objective of this study was to and diluted to the desired concentrations with Puregene DNA apply the E. ictaluri–specific primers and probe combinations hydration solution (QIAGEN). Quantitative PCR analysis was developed by Bilodeau et al. (2003) and the water sampling pro- performed on a 10-fold serial dilution starting with 500 ng of tocols described by Griffin et al. (2009) to establish protocols E. ictaluri genomic DNA on three separate occasions to estab- for the molecular quantification of E. ictaluri DNA from catfish lish the linear dynamic range of the assay. Separate analysis was pond water. also performed using 300, 30, 3, and 0.3 ng of genomic DNA from each nontarget bacterial species. Analysis of the potential inhibitory effects of nontarget DNA was conducted using three METHODS separate dilution series consisting of 150, 1.5, and 0.015 ng of E. Bacterial cultures.—The primary E. ictaluri strain used in ictaluri genomic DNA combined with 150 ng of genomic DNA the development of this assay was a case isolate (S97-773) from E. tarda, F. columnare,orAeromonas sp. or nuclease-free originally obtained during a natural ESC outbreak at a com- water. To ensure that there was no variation in amplification mercial catfish operation. The isolate was confirmed to be efficiency among the 19 different E. ictaluri isolates, separate E. ictaluri by PCR (Bilodeau et al. 2003) and expanded in analyses were conducted with 10 ng of genomic DNA from brain–heart infusion (BHI) broth (Becton Dickinson and Co., each isolate. The analyses were run on three separate plates on Sparks, Maryland) prior to storage at –60◦C in 20% glycerol. separate days. Each plate included a freshly diluted aliquot of Frozen cultures were streaked onto BHI agar plates and allowed genomic DNA from each isolate. to incubate for 48 h at 27◦C, at which time an individual colony Quantitative PCR.—The 15-µL PCR reactions contained was isolated and cultured in 10 mL of BHI broth at 27◦Cfor8h. 7.75 µL of Applied Biosystems Environmental Mastermix 2.0 In addition to S97-773, 18 different E. ictaluri isolates collected (Applied Biosystems, Foster City, California), 10 pM of each from 12 different catfish operations over a period of 20 years primer, 0.25 pM of probe, and 5 µL of template DNA. Ampli- (1991–2010) were obtained from the archival bacterial collec- fications were performed on a BioRad Icycler version 3.1 real- tion of the Thad Cochran National Warmwater Aquaculture time PCR system (BioRad, Hercules, California) programmed Center’s Aquatic Research and Diagnostic Laboratory (ARDL) for1cycleof50◦C for 2 min followed by 1 cycle of 95◦Cfor in Stoneville, Mississippi. Moreover, three species of bacteria 10 min and 40 cycles of 95◦C for 15 s and 60◦Cfor1min. commonly isolated from channel catfish (Edwardsiella tarda [4 Data collection was carried out following the 60◦C annealing– isolates], Flavobacterium columnare, and Aeromonas sp.) were extension step at the end of each cycle. For each plate, samples obtained from the ARDL archival collection and cryostocks and no-template negative controls were run in triplicate along- were expanded similarly in 10 mL of BHI broth overnight at side standard–positive controls consisting of 50, 5, 0.5, and 0.05 either 27◦C(E. ictaluri, F. columnare, and Aeromonas sp.) or ng of E. ictaluri genomic DNA, which were run in duplicate. The 37◦C(E. tarda). amplification efficiency (E) was estimated by the formula E = Design of TaqMan primer and probe sets.—The qPCR assay 10−1/s – 1, where s is the slope of the standard curve (Wong and targets the 129-base-pair (bp) fragment of a putative transposon Medrano 2005; Bustin et al. 2009). Efficiency estimates between located next to the phosphoserine transaminase (serC; enzyme 90% and 110% were considered acceptable (Griffin et al. 2009). number 2.6.1.52; IUBMB 1992) gene of E. ictaluri (GenBank Standard curves. —A 10-fold dilution (starting with 100 µL accession number AF110153; Thune et al. 1999; Bilodeau et of broth culture in 900 µL of nuclease-free H2O) was performed al. 2003). The original primer sequences from Bilodeau et al. for the E. ictaluri culture, and standard plate counts were per- (2003) were modified to reduce the differences in primer denatu- formed in duplicate using 50 µL from each aliquot. Additional ration temperatures (Tm; Table 1). A subsequent BLASTn search 50-µL aliquots from each dilution were added to a 1.5-mL mi- for somewhat similar sequences of the National Center for crocentrifuge tube and stored at –80◦C until processing. Biotechnology Information nonredundant nucleotide database Water samples (20 L) were collected from 7 different com- confirmed the primer combinations to be specific to their target mercial catfish ponds and processed within 24 h of collection. 180 GRIFFIN ET AL.

TABLE 1. Primers and dual-labeled probe (ESCTMP) for the speciﬁc detection of Edwardsiella ictaluri; Tm is the denaturing temperature. Primer or Sequence a ◦ b probe Sequence (5 –3 ) location (bp) Tm ( C) Reference Primers ESCF ACTTATCGCCCTCGCAACTC 481–500 65.3 Bilodeau et al. (2003) ESCR CCTCTGATAAGTGGTTCTCG 658–639 58.9 Bilodeau et al. (2003) ESCF2 ACTTATCGCCCTCGCAAC 481–498 62.8 This paper ESCR2 GCCTCTGATAAGTGGTTCTCG 659–639 62.7 This paper ESCTMP CCTCACATATTGCTTCAGCGTCGAC 537–561 77.7 Bilodeau et al. (2003)

aBased on GenBank accession number AF110153. bCalculated using the nearest-neighbor method assuming 50 mm NaCl and 0.5 µm oligonucleotide (http://www.sigma-genosys.com/calc/DNACalc.asp).

From each pond, 40- and 500-mL subsamples were collected. six different catfish ponds. All samples were processed as de- The 500-mL subsamples were transferred to a 1-L centrifuge scribed above, and qPCR analysis of each sample was performed bottle and spun for 10 min at 9,000 × g at room temperature in in triplicate. Standard plate counts were conducted to estimate a Sorvall RC-6 centrifuge (Thermo Fisher Scientific, Waltham, the number of bacteria added to the pond water samples. The Massachusetts). The supernatant from each subsample was re- approximate recovery of bacteria was calculated based on the moved and the pellet resuspended in 40 mL of distilled water standard curve generated from the dilution series aliquots added prior to transfer to a 40-mL round-bottom centrifuge tube. From directly to the pellets of 40-mL of pond water. Similar analysis this point on, the 40- and 500-mL samples were processed in the in 500-mL samples was not performed because of the incon- same manner. The samples were first centrifuged at 20,000 × sistent results obtained from spiked 500-mL samples collected g for 10 min. The supernatant was again removed and the pellet from eight separate ponds. was resuspended in 1.5 mL of nuclease-free H2O and transferred Genomic fingerprinting.—All E. ictaluri isolates (n = 19) to a 1.8-mL microcentrifuge tube. A 50-µL aliquot from each were analyzed by PCR-mediated genomic fingerprinting (rep- broth culture dilution was added directly to each pellet. DNA PCR) to determine the genetic homogeneity of the E. ictaluri isolation was then carried out using the UltraClean Soil DNA and to ensure the stability of the PCR target among isolates isolation kit (Mo Bio Laboratories, Carlsbad, California) fol- collected in different years from different locations. Four iso- lowing the manufacturer’s suggested protocol for wet samples. lates of E. tarda were included as outliers–negative controls. All centrifuge bottles were washed vigorously three times with Bacterial isolates were expanded and DNA isolation and quan- distilled water between samples. tification was carried out as described previously. Repetitive Assay validation.—To determine the accuracy of the water element-based polymerase chain reaction analysis was carried sampling protocol, a separate E. ictaluri broth culture was grown out using the ERIC-1 (5–ATG TAA GCT CCT GGG GAT TCA overnight and 100-µL aliquots of undiluted culture were added C–3), ERIC-2 (5–AAG TAA GTG ACT GGG GTG AGC G– directly to 40- (n = 8) or 500-mL (n = 8) water samples collected 3), or BOX (5–CTA CGG CAA GGC GAC GCT GAC G–3) from the same catfish pond or a 1.5-mL microcentrifuge tube primers (Versalovic et al. 1991, 1994). The 25-µL reaction mix- (n = 8). All samples were processed as described above, and tures comprised of 13 µL of IQ Supermix (Bio-Rad), 20 pmol qPCR analysis was performed in triplicate on three separate (ERIC) or 40 pmol (BOX) of primer, 100 ng of template DNA, days. Standard plate counts were performed to estimate the and nuclease-free water to volume. Amplifications were per- number of bacteria added to the pond water or microcentrifuge formed on a PTC-200 gradient cycler (MJ Research, Waltham, tube. The approximate recovery of bacteria was calculated based Massachusetts) with the following temperature profile: 1 cycle on the standard curve generated from the dilution series aliquots at 95◦C for 10 min; 5 cycles of 95◦Cfor1min,40◦Cfor1 added directly to the pellets of 40- or 500-mL of pond water. min, and 72◦C for 5 min; and 35 cycles of 95◦Cfor1min, This procedure was then repeated with a fresh E. ictaluri culture, 55◦C for 1 min, and 72◦C for 5 min. Aliquots of each ampli- this time adding aliquots to 40- (n = 8) and 500-mL (n = fication reaction (10 µL each) were electrophoresed through a 8) subsamples from eight different ponds as well as 1.5-mL 2% (weight : volume) agarose gel in the presence of ethidium microcentrifuge tubes (n = 8). bromide and visualized under ultraviolet light. Digital analy- To test the accuracy of the assay against a range of cell equiv- sis was performed with BioRad’s Quantity One software, and alents across several orders of magnitude, a separate E. ictaluri band sizes were assigned by direct comparison with concur- broth culture was grown overnight and serially diluted 10-fold. rently run 200-bp DNA ladder standards. Genomic fingerprints Aliquots (50 µL) from the undiluted culture as well as the 1:100, were evaluated based on the presence or absence of the am- 1:10,000, and 1:1,000,000 broth culture dilutions were added di- plified polymorphic fragments generated by the ERIC or BOX rectly to individual 40-mL pond water samples collected from primer. A REAL-TIME POLYMERASE CHAIN REACTION ASSAY 181

TABLE 2. Interference of Edwardsiella ictaluri assay by nontarget genomic DNA from bacteria commonly associated with commercial catfish ponds. Genomic DNA from E. ictaluri was combined with 150 ng of genomic DNA from nontarget bacteria and quantitative PCR analysis performed in triplicate within the same plate. Data are the mean quantification cycles for the different combinations, with SDs in parentheses; N/A = no amplification. Within columns, values with different letters are significantly different (ANOVA; p < 0.001)

Edwardsiella ictaluri genomic DNA quantity 150 ng 1.5 ng 15 pg 150 fg 0 fg Edwardsiella tarda 14.5 (0.08) 20.4 (0.23) 27.07 (0.15) 31.9 y (0.31) NA Flavobacterium columnare 14.6 (0.09) 21.2 (0.45) 27.5 (0.25) 33.2 z (0.23) NA Aeromonas sp. 14.2 (0.01) 21.0 (0.10) 27.7 (0.15) 33.8 z (0.72) NA

Statistical analysis.—Multiple comparisons of the treatments the presence of large amounts of nontarget DNA (even when were made using the General Linear Model and Least Signifi- the quantities of E. ictaluri DNA were six orders of magnitude cant Difference utilities of SAS version 9.2 (SAS, Cary, North lower than those of the nontarget DNA) (Table 2). Additionally, Carolina). there were no significant differences in the amplification of the 19 different E. ictaluri isolates (Table 3).

RESULTS Sensitivity Specificity Ten-fold serial dilutions of genomic DNA were linear over The modifications made to the original primers described by nine orders of magnitude, reaching a plateau at 10−5 ng of E. Bilodeau et al. (2003) did not change the specificity of the assay, ictaluri genomic DNA (∼20 cells). Edwardsiella ictaluri DNA and there were no significant differences in amplification of was amplified from at least one replicate of each pond water similar quantities of DNA from different isolates of E. ictaluri. sample (both 40 and 500 mL) with 10 and 1 colony-forming units There was no amplification of genomic DNA from F.columnare, (CFU) added directly to the pellet from several samples (but Aeromonas sp., or the closely related E. tarda, nor was there never in all three replicates from a given sample). This suggests inhibition of the amplification of E. ictaluri genomic DNA in that these samples were below the quantifiable limits of the

TABLE 3. Comparison of the ampliﬁcation of genomic DNA isolated from 19 different E. ictaluri isolates from the archived collection of the Aquatic Diagnostic and Research Laboratory in Stoneville, Mississippi. The values are the mean quantiﬁcation cycles and calculated DNA equivalents (SDs in parentheses) for triplicate reactions run on two separate days using 10 ng of genomic DNA per reaction. For each day, the isolates were run concurrently on the same PCR plate.

a Isolate ID Year isolated Location Cq DNA equivalent Log10 DNA equivalent S91-160 1991 Schlater 19.2 (0.9) 8.6 (4.5) 0.9 (0.3) S94-709 1994 Moorheadx 18.7 (0.5) 11.3 (4.9) 1.0 (0.2) S94-711 1994 Moorheadx 18.9 (0.5) 10.5 (3.6) 1.0 (0.1) S97-773 1997 Elizabeth 20.5 (0.7) 3.4 (1.8) 0.5 (0.2) S99-1140 1999 Moorheadx 19.3 (0.7) 15.0 (3.9) 0.8 (0.3) S03-699 2003 Indianolax 18.7 (0.7) 11.6 (5.6) 1.0 (0.2) S03-707 2003 Inverness 19.7 (0.7) 8.9 (3.0) 0.7 (0.3) S03-740 2003 Stoneville 19.0 (0.7) 9.6 (5.0) 0.9 (0.3) S03-830 2003 Inverness 18.2 (1.4) 21.6 (19.0) 1.2 (0.4) S05-233 2005 Stoneville 19.4 (0.8) 7.5 (4.1) 0.8 (0.3) S05-329 2005 Moorheadx 19.1 (0.6) 8.9 (4.0) 0.9 (0.2) S05-518 2005 Indianolay 17.9 (0.5) 19.0 (7.7) 1.2 (0.2) S06-428 2006 Moorheady 18.1 (0.7) 17.5 (7.9) 1.2 (0.2) S06-437 2006 Tunica 19.2 (0.8) 8.6 (4.2) 0.9 (0.3) S07-696 2007 Stoneville 19.1 (0.8) 8.7 (4.2) 0.9 (0.3) S07-698 2007 Moorheady 19.0 (0.6) 9.6 (4.2) 0.9 (0.2) S07-794 2007 Indianolaz 18.9 (1.3) 12.1 (8.6) 0.9 (0.4) S07-885 2007 Leland 18.6 (0.6) 12.7 (6.9) 1.1 (0.2) S10-95 2010 Doddsville 19.6 (0.5) 6.6 (2.7) 0.8 (0.2)

aAll locations are in Mississippi. When there is more than one farm at the same location, the different farms are distinguished by different superscripts. 182 GRIFFIN ET AL.

FIGURE 1. Linear dynamic ranges for the detection of E. ictaluri DNA in the real-time PCR assay. A 10-fold dilution series of known quantities of target DNA covering 10 orders of magnitude was analyzed in triplicate on three separate occasions. The mean quantification cycle (Cq) of the three runs is presented for each DNA quantity. The equation for the linear range of the assay ( + 2.699 to –5.301) is also given; the error bars denote SDs. assay. However, the assay was able to reliably detect E. ictaluri Assay Repeatability DNA in all replicates from all water samples with 102 CFU Repeated determinations of cell equivalents by qPCR analy- added directly to the pellets (Figure 1; Table 4), establishing the sis from straight broth culture or 40 or 500 mL of pond water cutoff for quantification of the assay at 102 CFU. Amplification (collected from either the same or eight different ponds) spiked of the genomic DNA isolated from serial dilutions of bacterial with 100 µL of broth culture were performed on separate days. cell culture was linear over five orders of magnitude, with the The inter-run coefficient of variation for data collected on three curve reaching a plateau at cell equivalents less than 102 (>1 separate days ranged from 3.9% to 21.1%, indicating an ac- CFU/mL for 500-mL samples; ∼5 CFU/mL in 40-mL samples) ceptable degree of precision, with all but one of the treatments − or standard equivalents less than 1 × 10 5, similar to what was showing less than 15% variability between analyses. The great- observed in the linear dynamic range established by the 10-fold est variability was observed among the 500-mL spiked pond dilution of genomic DNA. water samples collected from the eight different ponds (Table 5). Additionally, when cell equivalents spanning several orders of magnitude was added to 40-mL pond water samples col- TABLE 4. Interassay repeatability of E. ictaluri qPCR assay. A 10-fold dilu- lected from several different ponds, the among-pond variability tion series of known quantities of target DNA covering 10 orders of magnitude was analyzed in triplicate on three separate occasions. The mean (SD in paren- was greatest at lower cell equivalents, with the most variability theses) quantification cycle (Cq) of the three separate runs and the coefficient of and poorest accuracy being observed beyond the quantifiable variation (CV) between runs are presented for each DNA quantity. range of the assay (<102 CFU).

Genomic DNA Log10 genomic Assay Validation quantity (ng) DNA quantity (ng) Cq CV (%) Standard curves generated from the aforementioned dilu- 500 2.699 13.5 (0.3) 2.5 tion series were used to calculate approximate cell equivalents, 50 1.699 16.4 (0.3) 2.0 where y is the standard equivalent and x is the log10 transformed 5 0.699 19.6 (0.2) 1.2 E. ictaluri CFU (Figure 2). When replicate pond water samples 0.5 –0.301 22.9 (0.6) 2.7 were all collected from the same pond, there were no significant µ 0.05 –1.301 26.3 (0.3) 1.0 differences in cell equivalents between the 100 L of straight 0.005 –2.301 29.2 (0.4) 1.3 broth culture and the 40 or 500 mL of pond water to which µ 0.0005 –3.301 31.8 (0.4) 1.1 100 L of broth culture was added. However, when replicates 0.00005 –4.301 36.0 (0.8) 2.2 were performed using water from eight different ponds, the cell 0.000005 –5.301 38.6 (0.7) 1.9 equivalents for 500 mL of spiked pond water were significantly 0.0000005 –6.301 39.1 (0.6) 2.7 lower and there was significant variability among ponds (Table 5). The cell equivalents determined by the assay in straight broth A REAL-TIME POLYMERASE CHAIN REACTION ASSAY 183

FIGURE 2. Standard curves generated from DNA isolated from known quantities of E. ictaluri bacteria in straight broth culture or broth culture added to 40 or 500 mL of pond water from seven different catﬁsh ponds. Values are presented in terms of the mean standard equivalent determined by qPCR for each quantity, which was carried out in triplicate; the error bars denote SEs. For each curve, the equation for the linear range of the assay covers ﬁve orders of magnitude (102–106 CFU). 184 GRIFFIN ET AL.

TABLE 5. Comparison of methods of detecting E. ictaluri from pond water. Genomic DNA was isolated from known numbers of bacteria in straight broth culture or broth culture added to either 40 or 500 mL of pond water collected from the same pond or eight different ponds. Log10 cell equivalents were calculated using standard curves established for each treatment (see Figure 2). The values are the mean log10 cell equivalents (SDs in parentheses) determined by qPCR for the different runs. The analysis was performed in triplicate in three separate runs. The coefﬁcient of variation (CV) between samples from all runs is listed for each treatment. Within groups and columns, values with different letters are signiﬁcantly different (ANOVA; p < 0.01)

Log10 CFU Treatment added Run 1 Run 2 Run 3 All runs CV (%) Same pond 40 mL (n = 8) 8.04 7.39 (0.31) 7.65 (0.30) 7.64 (0.23) 7.56 (0.30) 4.0 500 mL (n = 8) 8.04 7.42 (0.32) 7.31 (0.57) 7.40 (0.74) 7.37 (0.56) 7.7 Straight bacteria (n = 8) 8.04 7.36 (0.48) 7.57 (0.32) 7.62 (0.36) 7.52 (0.41) 5.4

Different ponds 40 mL (n = 8) 6.30 5.42 (0.76) 5.67 (0.62) 5.57 (0.78) 5.56 z (0.72) 12.9 500 mL (n = 8) 6.30 4.63 (0.71) 4.95 (0.65) 4.15 (1.27) 4.58 y (0.96) 21.1 Straight bacteria (n = 8) 6.30 5.44 (0.55) 5.67 (0.61) 5.61 (0.51) 5.57 z (0.56) 10.0

culture or in straight broth culture added to 40 mL of pond water contrast, the ERIC primers generated 23 different bands for the from either the same or different ponds were consistently within E. tarda isolates, some of which were shared by all four isolates, one order of magnitude of the actual number of bacterial cells some of which were shared by different isolates, and some of added to the samples (Tables 5, 6), demonstrating a reasonable which were unique to individual isolates (Figure 3; Table 7). level of accuracy. Alternatively, a considerable amount of bacte- In similar fashion, the BOX primers generated electrophoretic rial DNA was lost during the processing of the 500-mL samples profiles for the E. ictaluri isolates that consisted of 9 distinct collected from different ponds, resulting in significantly lower bands all of which were shared by the 19 isolates, while the cell equivalents being detected than were actually added to the four profiles generated for E. tarda shared some homology but samples (Table 5). Reaction efficiencies ranged from 90% to had significantly more variability than was seen with E. ictaluri 104% throughout the study, with a mean of 94%. (Figure 3; Table 8).

Genomic Fingerprinting For both the ERIC and BOX primer sets, electrophoretic profiles demonstrated high levels of homogeneity for the DISCUSSION E. ictaluri isolates yet generated unique profiles for all four The protocols described here provide an accurate, sensitive, E. tarda isolates. The E. ictaluri profiles generated by the ERIC and relatively precise method of quantifying E. ictaluri from primers consisted of 13 distinct bands ranging in size from 150 commercial catfish ponds and demonstrate that 40 mL is the to 2,000 bp, all of which were shared by the 19 isolates. By optimal volume for analysis. There were no significant differences between the qPCR-determined cell equivalents obtained from straight broth culture and those obtained from a mixture TABLE 6. Accuracy of the assay for the detection of E. ictaluri in pond water. of broth culture and 40 mL of pond water, whether the wa- Genomic DNA was isolated from known numbers of bacteria added directly to ter was collected from the same pond or eight different ponds. 40 mL of pond water collected from six different ponds. Log10 cell equivalents Similarly, when the eight replicate pond water samples were were calculated using established standard curves (see Figure 2). The values are collected from the same pond, there were no differences in the mean log10 cell equivalents (SDs in parentheses) determined by qPCR. All cell equivalents between straight broth culture and broth culture analyses were performed in triplicate. The coefficient of variation (CV) between ponds for each CFU quantity is also listed. added directly to 40 mL or 500 mL of pond water. However, when replicate pond water samples were collected from differ- Log10 CFU ent ponds, significantly lower amounts of bacterial DNA were Log10 CFU determined detected from the 500-mL samples than the straight broth cul- CFU added added (SD) CV (%) ture or the 40-mL samples. This suggests that greater volumes of water decrease the efficiency of the DNA isolation process. This 4(n = 6) 0.60 2.03 (0.64) 31.6 can be attributed to an increase in the influence that pond turbid- 400 (n = 6) 2.60 2.87 (0.29) 10.2 ity and chemistry have on the DNA isolation process as well as 40,000 (n = 6) 4.60 4.86 (0.21) 4.4 the greater physical manipulation required for processing larger 4,000,000 (n = 6) 6.60 7.02 (0.24) 3.4 sample volumes. By contrast, pond water characteristics did not A REAL-TIME POLYMERASE CHAIN REACTION ASSAY 185

TABLE 7. Presence ( + ) or absence (–) of speciﬁc bands generated by the ERIC primers for E. ictaluri isolates used to validate the qPCR assay. Edwardsiella tarda isolates were included as outliers–negative controls.

Isolate

E. ictaluri E. tarda Band (∼molecular weight [bp]) 1991 1994a 1994b 1997 1999 2003a 2003b 2003c 2003d 2005a 2005b 2005c 2006a 2006b 2007a 2007b 2007c 2007d 2010 1998 2004 2007 2011

1 (2,050) +++++++++++++++++++–––– 2 (1,730) +++++++++++++++++++–––– 3 (1,600) – – – – – – – – – – – – – – – – – – – ++ –– 4 (1,360) +++++++++++++++++++–––– 5 (1,210) +++++++++++++++++++––+ – 6 (1,110) – – – – – – – – – – – – – – – – – – – + ––– 7 (1,030) +++++++++++++++++++––++ 8 (1,000) – – – – – – – – – – – – – – – – – – – + ––– 9 (870) – – – – – – – – – – – – – – – – – – – ++ –– 10 (840) – – – – – – – – – – – – – – – – – – – – +++ 11 (760) – – – – – – – – – – – – – – – – – – – – + –– 12 (730) ++++++++++++++++++++– ++ 13 (670) + + + ++ + + + + + + + + + + + + + +++++ 14 (620) +++++++++++++++++++++–– 15 (530) – – – – – – – – – – – – – – – – – – – ++ –– 16 (490) +++++++++++++++++++––++ 17 (450) – – – – – – – – – – – – – – – – – – – ++++ 18 (420) – – – – – – – – – – – – – – – – – – – – ++ – 19 (390) – – – – – – – – – – – – – – – – – – – – + –– 20 (360) – – – – – – – – – – – – – – – – – – – – + –– 21 (350) – – – – – – – – – – – – – – – – – – – – – – + 22 (320) – – – – – – – – – – – – – – – – – – – ++ –– 23 (290) +++++++++++++++++++–––– 24 (240) +++++++++++++++++++–––+ 25 (200) +++++++++++++++++++––+ – 26 (170) + + + ++ + + + + + + + + + + + + + +++++ 27 (130) – – – – – – – – – – – – – – – – – – – – – ++ have a significant effect on smaller volumes of water (40 mL) The accuracy of the assay did not appear to be affected by as cell equivalents and variability between samples (even those increasing or decreasing the number of cells added to the sample collected from different ponds) were similar to those observed unless the bacterial numbers were below the quantifiable range 2 with straight broth culture. oftheassay(<10 CFU). Within the quantifiable range, the log10

TABLE 8. Presence ( + ) or absence (–) of speciﬁc bands generated by the BOX primers for E. ictaluri isolates used to validate the qPCR assay. Edwardsiella tarda isolates were included as outliers–negative controls.

Isolate

E. ictaluri E. tarda Band (∼molecular weight [bp]) 1991 1994a 1994b 1997 1999 2003a 2003b 2003c 2003d 2005a 2005b 2005c 2006a 2006b 2007a 2007b 2007c 2007d 2010 1998 2004 2007 2011

1 (1,550) +++++++++++++++++++–––– 2 (1,360) – – – – – – – – – – – – – – – – – – – ++++ 3 (1,310) +++++++++++++++++++–––– 4 (1,120) – – – – – – – – – – – – – – – – – – – ++ –– 5 (1,050) – – – – – – – – – – – – – – – – – – – – – ++ 6 (1,000) – – – – – – – – – – – – – – – – – – – – – ++ 7 (840) +++++++++++++++++++–––– 8 (800) – – – – – – – – – – – – – – – – – – – – + –– 9 (700) – – – – – – – – – – – – – – – – – – – – – + – 10 (640) – – – – – – – – – – – – – – – – – – – ++ – + 11 (530) +++++++++++++++++++++–– 12 (390) +++++++++++++++++++++++ 13 (360) +++++++++++++++++++–––– 14 (290) +++++++++++++++++++++–– 15 (250) – – – – – – – – – – – – – – – – – – – – – ++ 16 (210) – – – – – – – – – – – – – – – – – – – ++ –– 17 (200) +++++++++++++++++++––++ 18 (190) – – – – – – – – – – – – – – – – – – – ++ –– 19 (160) +++++++++++++++++++–––– 20 (140) – – – – – – – – – – – – – – – – – – – ++ –– 21 (130) – – – – – – – – – – – – – – – – – – – – – + – 186 GRIFFIN ET AL.

magnitude, from 2.5 cells/mL to 0.2 cells/mL. However, when the volume of water is increased the speed at which the ini- tial centrifugation step can be carried out is reduced, resulting in a looser pellet that may become dislodged during the de- canting of the supernatant. Concurrent analysis of spiked pond water samples of 40 and 500 mL collected from different ponds showed a decrease in the precision of the assay as sample volume increased. In short, water volumes of 500 mL collected from different ponds resulted in unacceptable levels of precision (>15%; Huang and Pan 2004; Wang et al. 2005; Scheurer et al. 2005). Additionally, larger volumes of water contain a greater amount of organic particulates and other substances that make the samples more difficult to work with, as transferring the pellet from the 40-mL centrifuge tube to the 1.5-mL microcentrifuge tube becomes cumbersome. Lastly, the addition of more organic-solid particulates can increase the number of environmental compounds that inhibit downstream PCR appli- cations (Wilson 1997). The low variability between replicates for the 500-mL samples collected from the same pond suggests that large volumes can be used in some ponds. However, the variability observed between replicate samples collected from different ponds suggests that the sampling protocol for larger volumes of water is inconsistent and the efficiency of DNA isolation and amplification is more sensitive to variations in pond turbidity and chemistry than for smaller volumes. Thus, given the high levels of variability in turbidity, water chemistry, zoo- plankton communities, and algal blooms between catfish ponds, 40-mL volumes are the most reliable, resulting in an assay sensitivity of approximately 2.5 cell equivalents/mL. Although the challenge dose required to cause significant mortality within a pond population is unknown, the level of detection is significantly lower than the dose required to induce more than 50% mortality under controlled conditions in our laboratory (Wise et al. 2000, 2008; Wise and Terhune 2001). FIGURE 3. Genomic fingerprints of the E. ictaluri and E. tarda isolates used to validate the qPCR assay. The electrophoretic profiles were generated from Polymerase chain reaction–mediated genomic fingerprint- either (A) ERIC or (B) BOX primer sets (see text for additional details). Lanes ing using the ERIC and BOX primers is a well-recognized, 1 and 26 are the molecular weight marker (200-bp ladder), lane 2 shows the reliable method for taxonomic identification and discrimina- results for the no template control, lanes 3–21 show the results for the E. ictaluri tion of bacteria (Hulton et al. 1991; Louws et al. 1994, 1999; isolates, and lanes 22–25 show the results for the E. tarda isolates. Rodriguez-Barradas et al. 1995; Dombek et al. 2000; Saxena et al. 2002; Tacao˜ et al. 2005; Maiti et al. 2008; Valdeben- ito and Avendano-Herrara˜ 2009; Yuan et al. 2010). Given cell equivalents determined by qPCR from pond water samples the success that these primer sets have had in differentiat- spiked with a range of bacterial cell quantities (102–106)were ing between closely related isolates of other species of bacte- within one standard deviation of the log10 of the actual number ria, the similarities demonstrated by the isolates used in this of bacteria added to the sample. The accuracy of the assay was study make a strong argument for the genetic homogeneity significantly reduced when the number of cells added was below of E. ictaluri. Similarly, the results of the repetitive element- the quantifiable range, and between-pond variability increased based polymerase chain reaction analysis described here sup- by nearly 20%. The variability between ponds increased as the port previous work suggesting that there is limited genetic cell equivalents decreased, which is consistent with other qPCR variation between E. ictaluri isolates from different years and assays, in which inter- and intra = assay variability increased as geographic locations (Panangala et al. 2006). Moreover, the DNA template quantity decreased (Cankar et al. 2006). limited variation in the amplification of genomic DNA from Given the sensitivity of the qPCR assay (102 CFU), increas- isolates collected from several different catfish farms over a pe- ing the volume of water from 40 to 500 mL should theoretically riod of 20 years suggests a high degree of stability in the PCR increase the overall sensitivity of the protocol by one order of target. A REAL-TIME POLYMERASE CHAIN REACTION ASSAY 187

Admittedly, a significant limitation of qPCR analysis is that REFERENCES the technique only detects target DNA and does not provide Bilodeau, A. L., G. C. Waldbieser, J. S. Terhune, D. J. Wise, and W. R. Wolters. any information regarding the viability of the target organism. 2003. A real-time polymerase chain reaction assay of the bacterium Edward- In controlled laboratory settings, Plumb and Quinlan (1986) siella ictaluri in channel catfish. Journal of Aquatic Animal Health 15:80–86. Brinkman, N. E., R. A. Haugland, L. J. Wymer, M. Byappanahalli, R. L. Whit- demonstrated that E. ictaluri is fairly short lived in pond water, man, and S. J. Vesper. 2003. Evaluation of a rapid, quantitative real-time PCR ◦ surviving for less than 10 d at 25 C. They also demonstrated method for cellular enumeration of pathogenic Candida species in water. that E. ictaluri can survive for extended periods in the bottom Applied and Environmental Microbiology 69:1775–1782. mud of catfish ponds, but as water samples are taken from the Bustin, S. A., V. Benes, J. A. Garson, J. Hellemans, J. Huggett, M. Kubista, R. upper portions of the water column bacteria persisting in bottom Mueller, T. Nolan, M. W. Pfaffl, G. L. Shipley, J. Vandesompele, and C. T. Wittwer. 2009. The MIQE guidelines: minimum information for publication mud would have little effect on the assay. As it seems unlikely of quantitative real-time PCR experiments. Clinical Chemistry 55:611–622. that dead bacteria would persist in the pond environment for Cankar, K., D. Stebih,ˇ T. Dreo, J. Zel,ˇ and K. Gruden. 2006. Critical points significant periods of time, the bacterial DNA detected by this of DNA quantification by real-time PCR: effects of DNA extraction method assay will most likely be attributed to organisms that have only and sample matrix on quantification of genetically modified organisms. BMC recently (i.e., within the past 10–15 d) been shed by the fish Biotechnology [online serial] 6:37. Dombek, P. E., L. K. Johnson, S. T. Zimmerley, and M. J. Sadowsky. 2000. host. Use of repetitive DNA sequences and the PCR to differentiate Escherichia The determination of E. ictaluri concentrations in pond water coli isolates from human and animal sources. Applied and Environmental samples has been hindered by the normal microflora of the Microbiology 66:2572–2577. pond environment, which makes conventional microbiological Francis-Floyd, R., M. H. Beleau, P. R. Waterstrat, and P. R. Bowser. 1987. techniques such as plate counts obsolete. Additionally, the use Effect of water temperature on the clinical outcome of infection with Ed- wardsiella ictaluri in channel catfish. Journal of the American Veterinary of molecular techniques to determine microbial concentrations Medical Association 191:1413–1416. in catfish ponds has been prevented by compounds present in the Griffin, M. J., L. M. Pote, A. C. Camus, M. J. Mauel, T. E. Greenway, and ponds that inhibit the amplification of nucleic acids by PCR. The D. J. Wise. 2009. Application of a real-time PCR assay for the detection of techniques described here successfully address these issues and Henneguya ictaluri in commercial channel catfish ponds. Diseases of Aquatic provide a rapid, accurate means for detecting and quantifying Organisms 86:223–233. Hallett, S. L., and J. L. Bartholomew. 2006. Application of a real-time PCR E. ictaluri in the pond environment. assay to detect and quantify the myxozoan parasite Ceratomyxa shasta in Currently, the year-round prevalence of E. ictaluri in the cul- river water samples. Diseases of Aquatic Organisms 71:109–118. ture system is unknown, along with the dynamics of the bacterial Hawke, J. P., A. C. McWhorter, A. G. Steigerwalt, and D. J. Brenner. 1981. Ed- populations in the days leading up to and during an outbreak. wardsiella ictaluri sp. nov., the causative agent of enteric septicemia of catfish. The ability to quantify microbial numbers in pond water and International Journal of Systematic Evolutionary Microbiology 31:396–400. Huang, H., and T. Pan. 2004. Detection of genetically modified maize MON810 identify fluctuations within the population will provide a means and NK603 by multiplex and real-time polymerase chain reaction methods. of evaluating the year-round prevalence of the organism within Journal of Agricultural and Food Chemistry 52:3264–3268. the system. This information could enable us to identify the Hulton, C. S., C. F. Higgins, and P. M. Sharp. 1991. ERIC sequences: a novel conditions in early spring or midsummer when water tempera- family of repetitive elements in the genomes of Esherichia coli, Salmonella tures are nonpermissive for the organism, which may predispose typhimurium, and other enterobacteria. Molecular Microbiology 5:825–834. IUBMB (International Union of Biochemistry and Molecular Biology). 1992. ponds to outbreaks of ESC when water temperatures return to Enzyme nomenclature 1992. Academic Press, San Diego, California. the permissive range in the late spring and early fall. Similarly, Louws, F. J., D. W. Fulbright, C. T. Stephens, and F. J. de Bruijn. 1994. Specific the effects that vaccination, antibiotic therapies, and restricted genomic fingerprints of phytopathogenic Xanthomonas and Pseudomonas feeding regimes have on the E. ictaluri population within ponds pathovars and strains generated with repetitive sequences and PCR. Applied are unknown. This assay will provide an invaluable tool for re- and Environmental Microbiology 60:2286–2295. Louws, F. J., J. L. W. Rademaker, and F. J. de Bruijn. 1999. The three Ds searchers conducting epidemiological studies as well as a means of PCR-based genomic analysis of phytobacteria: diversity, detection, and to evaluate the effects that various management practices have disease diagnosis. Annual Review of Phytopathology 37:81–125. on an E. ictaluri population during an outbreak. This informa- Maiti, N. K., A. Mandal, S. Mohanty, and M. Samanta. 2008. Comparative tion could prove invaluable in optimizing disease management analysis of genome of Edwardsiella tarda by BOX-PCR and PCR-ribotyping. practices, enabling practitioners to maximize productivity while Aquaculture 280:60–63. Panangala, V.S., C. A. Shoemaker, S. T. McNulty, C. R. Arias, and P.H. Klesius. reducing disease-related losses. 2006. Intra- and interspecific phenotypic characteristics of fish-pathogenic Edwardsiella ictaluri and E. tarda. Aquaculture Research 37:49–60. Plumb, J. A., and E. E. Quinlan. 1986. Survival of Edwardsiella ic- ACKNOWLEDGMENTS taluri in pond water and bottom mud. Progressive Fish-Culturist 48:212– The authors would like to thank Cyndi Ware and Megan 214. Barnes for their technical contributions to this project. This re- Rodriguez-Barradas, M. C., R. J. Hamill, E. D. Houston, P. R. Georghiou, J. E. search was supported by the College of Veterinary Medicine and Clarridge, R. L. Regnery, and J. E. Koehler. 1995. Genomic fingerprinting of the Mississippi Agricultural and Forestry Experiment Station, Bartonella species by repetitive-element PCR for distinguishing species and isolates. Journal of Clinical Microbiology 33:1089–1093. both of Mississippi State University. This is MAFES publication Roe, J. D., R. A. Haugland, S. J. Vesper, and L. J. Wymer. 2001. Quantifi- J-12046. cation of Stachybotrys chartarum conidia in indoor dust using real time, 188 GRIFFIN ET AL.

fluorescent-probe-based detection of PCR products. Journal of Exposure Versalovic, J., M. Schneider, F. J. deBruijn, and J. R. Lupski. 1994. Genomic Analysis and Environmental Epidemiology 11:12–20. fingerprinting of bacteria using repetitive-sequence-based polymerase chain Saxena, M. K., V.P.Singh, B. D. Lakhcharua, G. Taj, and B. Sharma. 2002. Strain reaction. Methods in Molecular and Cellular Biology 5:25–40. differentiation of Indian isolates of Salmonella by ERIC-PCR. Research in Wang, Y., V. W. S. Liu, W. Xue, P. C. K. Tsang, A. N. Y. Cheung, and H. Y. S. Veterinary Science 73:313–314. Ngan. 2005. The increase of mitochondrial DNA content in endometrial ade- Scheurer, M. E., G. Tortolero-Luna, M. Guillaud, M. Follen, Z. Chen, L. M. nocarcinoma cells: a quantitative study using laser-captured microdissected Dillon, and K. Adler-Storthz. 2005. Correlation of human papillomavirus tissues. Gynecologic Oncology 98:104–110. type 16 and human papillomavirus type 18 e7 messenger RNA levels with Wilson, I. G. 1997. Inhibition and facilitation of nucleic acid amplification. degree of cervical dysplasia. Cancer Epidemiology, Biomarkers, and Preven- Applied and Environmental Microbiology 63:3741–3751. tion 14:1948–1952. Wise, D. J., T. E. Greenway, M. H. Li, A. C. Camus, and E. H. Robinson. 2008. Stults, J. R., O. Snoeyenbos-West, B. Methe, D. R. Lovley, and D. P. Chandler. Effects of variable periods of food deprivation on the development of enteric 2001. Application of the 5 fluorogenic exonuclease assay (TaqMan) for septicemia in channel catfish. Journal of Aquatic Animal Health 20:39–44. quantitative ribosomal DNA and rRNA analysis in sediments. Applied and Wise, D. J., P. H. Klesius, C. A. Shoemaker, and W. R. Wolters. 2000. Vacci- Environmental Microbiology 67:2781–2789. nation of mixed and full-sib families of channel catfish Ictalurus punctatus Tacao,˜ M., A. Alves, M. J. Saavedra, and A. Correia. 2005. BOX-PCR is an against enteric septicemia of catfish with a live attenuated Edwardsiella ic- adequate tool for typing Aeromonas spp. Antonie van Leeuwenhoek 88:173– taluri isolate (RE-33). Journal of the World Aquaculture Society 31:206– 179. 212. Thune, R. L., D. H. Fernandez, and J. R. Battista. 1999. An aroA mutant of Wise, D. J., and J. S. Terhune. 2001. The relationship between vaccine dose Edwardsiella ictaluri is safe and efficacious as a live, attenuated vaccine. and efficacy in channel catfish Ictalurus punctatus vaccinated as fry with a Journal of Aquatic Animal Health 11:358–372. live attenuated strain of Edwardsiella ictaluri (RE-33). Journal of the World Valdebenito, S., and R. Avendano-Herrara.˜ 2009. Phenotypic, serological, and Aquaculture Society 32:177–183. genetic characterization of Flavobacterium psychrophilum strains isolated Wong, M. L., and J. F. Medrano. 2005. Real-time PCR for mRNA quantitation. from salmonids in Chile. Journal of Fish Diseases 32:321–333. BioTechniques 39:75–85. Versalovic, J., T. Koeuth, and J. R. Lupski. 1991. Distribution of repetitive Yuan, W., T. J. Chai, and Z. M. Miao. 2010. ERIC-PCR identification of the DNA sequences in eubacteria and application to fingerprinting of bacterial spread of airborne Escherichia coli in pig houses. Science of the Total Envi- genomes. Nucleic Acids Research 19:6823–6831. ronment 408:1446–1450.