Acid phosphatase test proves superior to standard phenotypic identification procedure for Clostridium perfringens strains isolated from water.

Clostridium perfringens is used as an indicator for persistent faecal pollution as well as to monitor the efficacy of water treatment processes. For these purposes, differentiation between C. perfringens and other Clostridia is essential and is routinely carried out by phenotypic standard tests as proposed in the ISO/CD 6461-2:2002 (ISO_LGMN: lactose fermentation, gelatine liquidation, motility and nitrate reduction). Because the ISO_LGMN procedure is time consuming and labour intensive, the acid phosphatase test was investigated as a possible and much more rapid alternative method for confirmation. The aim of our study was to evaluate and compare confirmation results obtained by these two phenotypic methods using genotypically identified strains, what to our knowledge has not been accomplished before. For this purpose, a species specific PCR method was selected based on the results received for type strains and genotypically characterised environmental strains. For the comparative investigation type strains as well as presumptive C. perfringens isolates from water and faeces samples were used. The acid phosphatase test revealed higher percentage (92%) of correctly identified environmental strains (n=127) than the ISO_LGMN procedure (83%) and proved to be a sensitive and reliable confirmation method.

Introduction

Clostridium perfringens has been successfully used to detect faecal pollution in the aquatic environment in various geographic areas (Byamukama et al., 2005; Fujioka and Shizumura, 1985; Lisle et al., 2004; Mueller-Spitz et al., 2010a; Reischer et al., 2008). C. perfringens is able to form spores that allow prolonged detection of faecal contamination in the environment. In contrast, the faecal indicators Escherichia coli and enterococci are considered to be short-lived in aquatic systems. Monitoring of C. perfringens has shown to be useful to monitor water quality during the stages of natural and technical water treatment to evaluate efficacy (Payment, 1991; Schijven et al., 2003). Thus, C. perfringens (including spores) has been specified as an indicator parameter in the Directive of the European Union on the quality of water for human consumption. This parameter has to be investigated if drinking water originates from or is influenced by surface water (EU, 1998). Since no ISO method for C. perfringens was available at the time when the European Directive was enacted, mCP medium (Bisson and Cabelli, 1979) has been given as guidance, pending the possible future adoption. However, the most commonly used cultivation method for C. perfringens in water, food and faeces is based on tryptose sulphite cycloserine (TSC) agar (Hauschild and Hilsheimer, 1974). Cultivation on TSC agar has shown higher recovery rates compared to mCP agar in numerous studies on type strains (Byrne et al., 2008), water samples (Araujo et al., 2004; Burger et al., 1984; Sartory et al., 1998), and food samples (de Jong et al., 2003). Therefore this agar has been proposed in the ISO/CD 6461-2:2002.

If the assessment of faecal pollution is a matter of interest, the differentiation between mesophilic sulphite-reducing Clostridia and C. perfringens is essential. In the ISO/CD 6461-2:2002 (ISO, 2002) confirmation of presumptive C. perfringens is proposed by a procedure consisting of four phenotypic tests (ISO_LGMN: lactose fermentation, gelatine liquidation, motility and nitrate reduction). This procedure is time consuming (up to 72 h) and labour intensive. Ueno et al. (1970) described a more rapid and simple test for the identification of C. perfringens based on acid phosphatase activity. The phosphatase releases naphtyl from 1-naphtyl phosphate which forms an azo dye with diazonium o-dianisidine (Eisgruber et al., 2003). Sartory et al. (2006) stated, that the test for acid phosphatase is a reliable alternative method for confirming presumptive C. perfringens and also much shorter and easier to handle. Studies with methylumbelliferyl phosphate, a substrate enabling detection of acid and alkaline phosphatase, suggest being not suitable due to lack of selectivity (Adcock and Saint, 2001; Eisgruber et al., 2003). So far the reliability of the acid phosphatase test was compared to other phenotypical methods by means of biochemical reactions, solely (Sartory et al., 2006; Eisgruber et al., 2003).

Beside phenotypic methods, PCR protocols have been developed for the specific identification of C. perfringens (Baums et al., 2004; Fach and Popoff, 1997; Kikuchi et al., 2002; Songer and Meer, 1996; Wang et al., 1994). PCR techniques provide more reliable results, because they directly detect the presence of species specific genes independent of the state of the microorganism (Petit et al., 1999; Ruengwilysup et al., 2009). C. perfringens is known to form a monophyletic well characterised branch within the genus of Clostridia — making it an attractive target for PCR identification (Roenner and Stackebrandt, 1994). Mueller-Spitz et al. (2010b) characterised presumptive isolates obtained from mCP media by molecular tools. All these strains turned out to be C. perfringens. However, sensitivity and selectivity of the mCP-method were not taken into consideration.

The aim of our study was to evaluate and compare the phenotypic confirmation methods – ISO_LGMN procedure and acid phosphatase test – in terms of their accuracy to discriminate between C. perfringens and non C. perfringens using genotypically identified isolates.

Materials and methods

Strategy to evaluate ISO_LGMN procedure and acid phosphatase test for the confirmation of presumptive C. perfringens

Two PCR methods for specific detection of C. perfringens (CP_PCR and PLC_PCR) were chosen. These PCR methods are based on the amplification of species-specific targets as there were C. perfringens specific segment of the 16S rRNA gene (CP_PCR) and the phospholipase C gene of C. perfringens (PLC_PCR) (Fach and Popoff, 1997; Wang et al., 1994). Both target domains are chromosomally encoded and therefore should be available in all strains of C. perfringens, regardless of toxin type (Canard et al., 1992).

The two PCR methods were tested on type strains and on a subset of environmental strains, which had been characterised by sequencing of 16S rRNA genes and aligning the sequences to the sequences contained in the ribosomal database project (http://rdp.cme.msu.edu). Sequence analysis of all strains was considered as too time consuming and cost intensive.

The PCR method with the higher accuracy was chosen to assess the results received from the ISO_LGMN procedure and the acid phosphatase test applied on type strains as well as on presumptive C. perfringens isolates from water and faeces samples. The statistical measures sensitivity, specificity and percentage of correct identification were used to compare the performance characteristics of the phenotypic methods.

Bacterial strains

14 type strains of C. perfringens and other Clostridia were obtained from DSMZ (Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH, Braunschweig, Germany) (Table 2a).

Table 2a

Results of testing PLC_PCR and CP_PCR on C. perfringens and other Clostridia type strains.

Type strains	DSMZ no.	PLC_PCR	CP_PCR
Clostridium perfringens	756	+	+
Clostridium perfringens	11783	+	+
Clostridium perfringens	2755	+	+
Clostridium acetabutylicuma	792	−	−
Clostridium bifermentans	14991	−	−
Clostridium butyricum	10702	−	−
Clostridium clostridioforme	933	−	−
Clostridium histolyticum	2158	−	−
Clostridium novyi	14992	−	−
Clostridium pasteurianuma	525	−	−
Clostridium puniceuma	2619	−	−
Clostridium sardiniense	2632	−	−
Clostridium sordelli	2141	−	−
Clostridium sporiformea	1552	−	−

Strains not able to grow on TSC agar at 44 °C.

127 presumptive C. perfringens environmental strains were isolated from mountainous spring water (49) and from faecal samples (78) obtained from potential source groups located in the corresponding area (human and animal origin).

Culture preparation of the type strains

Type strains were incubated anaerobically at (36 ± 2) °C for 1 to 7 days in thioglycolate broth (Oxoid, Basingstoke, UK). Subsequently, the strains were subcultured anaerobically onto Tryptose Sulphite Cycloserine (TSC) agar (No 1–178, Scharlau, Sentmenat, Spain and supplement cycloserine, Oxoid, UK) at (44 ± 1) °C for (21 ± 3) h and on blood agar (BioMerieux, Craponne, France) at (36 ± 2) °C for (21 ± 3) h. The anaerobic conditions were provided in anaerobic jars with a gas generating system (No AN0025A, Oxoid, UK). Indicator strips (BR42, Oxoid, UK) were used to control if sufficient anaerobic conditions were achieved.

Isolation of presumptive C. perfringens from water and faeces samples (environmental strains)

The isolation of presumptive C. perfringens was performed according to ISO/CD 6461-2:2002 (ISO, 2002). Water samples were processed directly; faeces samples were suspended in sterile peptone saline solution (No 64544, Bio-Rad, Hercules, USA). The samples and dilutions thereof were filtered through a membrane with a pore size of 0.45 μm (Cellulose Nitrate, No 11406, Sartorius, Goettingen, Germany). The membranes were incubated anaerobically at (44 ± 1) °C for (21 ± 3) h on Tryptose Sulphite Cycloserine (TSC) agar (No 1–178, Scharlau, Spain, and supplement cycloserine, No SR0088E, Oxoid, UK). Black, grey or yellow brown colonies were considered as presumptive C. perfringens and subcultured onto blood agar (No 43041, BioMerieux, France) and incubated at (36 ± 2) °C for (21 ± 3) h anaerobically. Anaerobic conditions were provided in anaerobic jars with a gas generating system (No AN0025A, Oxoid). Indicator strips (BR42, Oxoid, UK) were used to control if sufficient anaerobic conditions were achieved.

Confirmation of presumptive C. perfringens

ISO_LGMN procedure

The confirmation procedure consisting of lactose fermentation, gelatine liquefaction, motility and nitrate reduction (LGMN) is described in detail in ISO/CD 6461-2:2002 (ISO, 2002). Bacteria that produced characteristic colonies on TSC agar under anaerobic conditions and were non-motile, reduced nitrate, fermented lactose and liquefied gelatine were considered as C. perfringens.

Acid phosphatase test

The acid phosphatase test was performed according to Ueno et al. (1970). Typical colonies from TSC agar grown anaerobically on blood agar plates (36 ± 2) °C for (21 ± 3) h were spread on filter paper and 2 to 3 drops of the acid phosphatase reagent were placed onto the cell material. A purplish colour developed within 3 min was considered as positive for the presence of acid phosphatase. The test was performed in duplicate.

The acid phosphatase reagent was produced as follows: 0.4 g 1-naphtyl phosphate disodium salt (No N 7255-5 G Sigma Aldrich, St. Louis, USA) and 0.8 g Fast Blue B Salt (o-Dianisidine bis(diazotized) zinc double salt; No D 9805-15 G Sigma Aldrich, USA) were dissolved in 20 ml acetate buffer (pH 4.6; No T182.1, Roth, Karlsruhe, Germany). After a storage time of (60 ± 5) min at (5 ± 3) °C the solution was filtered through a fluted filter. The prepared reagent was stored at (5 ± 3) °C and used within two weeks.

Bacteria that produced characteristic colonies on TSC agar, grew anaerobically on blood agar and possessed acid phosphatase were considered as C. perfringens.

Molecular characterisation

DNA extraction

Cell pellets of the cultivated strains were subjected to phenol/chloroform DNA extraction procedure according to Ausubel et al. (1994), omitting cultivation in liquid culture (steps 1 to 2 of protocol). Supplemental 0.3 w/v lysozyme was added for predigestion to the cell pellet suspended in TE-buffer and incubated for 45 min at 37 °C before adding 10% SDS. Furthermore, as a last step 3 μl RNAse was added to the DNA extract and incubated at 37 °C for 1 h to remove residual RNA. Extracted DNA was examined by agarose gel electrophoresis before subsequent PCR analysis. DNA concentration was estimated using the Quant-IT dsDNA BR Assay Kit with the Qubit™ fluorometer (Invitrogen, Paisley, UK) according to the manufacturer's instructions.

Full length 16S-rRNA gene PCR

A subset of 30 environmental strains (water and faeces) was analysed by sequencing of 16S rRNA genes (700–1300 bp) and aligning the full and partial sequences to the characterised sequences contained in the ribosomal database (http://rdp.cme.msu.edu).

Full length 16S-rRNA gene PCR amplification was conducted using the primer set 63f and 1387r designed by Marchesi et al. (1998). The sequences are listed in Table 1.

Table 1

Primer sequences used for the applied PCR protocols.

PCR	Primer sequence	References
16S-rRNA gene PCR	63f (5′-CAGGCCTAACACATGCAAGTC-3′)1387r (5′-GGGCGGWGTGTACAAGGC-3′)	Marchesi et al. (1998)
	T7_63f (5′-TAATACGACTCACTATAGGGCAGGCCTAACACATGCAAGTC-3′)aSP6_1387r (5′-CATTTAGGTGACACTATAGGGGCGGWGTGTACAAGGC-3′)b
PLC_PCR	PL3 (5′-AAGTTACCTTTGCTGCATAATCCC-3′)PL7 (5′-ATAGATACTCCATATCATCCTGCT-3′)	Fach and Popoff (1997)
CP_PCR	CP1f (5′-AAAGATGGCATCATCATTCAAC-3′)CP2r (5′-TACCGTCATTATCTTCCCCAAA-3′)	Wang et al. (1994)

Primer 63f was combined with sequencing primer T7.

Primer 1387r was combined with sequencing primer SP6.

The 50 μl PCR reaction mixture consisted of 1× Taq buffer, 1 U Taq polymerase native, 2 mM of MgCl2, 0.2 mM dNTPs (all of them Fermentas GmbH, St. Leon-Rot, Germany) and 0.2 μM of each primer (Eurofins MWG GmbH, Ebersberg, Germany) and 100–1000 pg DNA sample. The PCR was performed using iCycler (Bio-Rad, USA) with the following optimised thermal profile: initial denaturation at 95 °C for 120 s followed by 30 cycles with 60 s of denaturation at 95 °C, 60 s of primer annealing at 55 °C and 90 s of primer extension at 72 °C and as last step final elongation at 72 °C for 600 s. Amplicon size of about 1300 bp was checked by electrophoresis on agarose gel (1.5%) containing ethidium bromide with GenRulter 100 bp plus (Fermentas, Germany) as size standard. Finally the PCR products were purified with QIAquick PCR Purification Kit (Quiagen GmbH, Hilden, Germany).

Cloning and sequencing

16S-rRNA gene fragments were cloned using the pGMET Vector system and E. coli JM109 High Efficiency Competent Cells (Promega, Madison, WI, USA) according to the manufacturer's instructions. White colonies were selected and grown in 3 ml LB broth (MERCK, Darmstadt, Germany) with ampicillin (100 μg/ml) overnight at 37 °C. Plasmid DNA was isolated with FastPlasmid Miniprep Kit (Eppendorf AG, Hamburg, Germany). Isolated plasmid DNA was diluted 10− 2 and subsequently the cloned 16S-rRNA gene fragment was amplified at slightly modified PCR conditions as described above: 0.8 U Taq polymerase; primer set T7_63f and SP6_1387r (Table 1); primer annealing temperature 52 °C. Purification of the PCR product was carried out using QIAquick PCR Purification Kit (Quiagen, Germany). Sequencing was carried out by Eurofins MWG, Germany.

Identification of the sequences was performed with the Sequence Match Tool (SeqMatch, version 3) provided by the Ribosomal Database Project II (RDP, Release 10) using a naïve Bayesian classifier trained on sequences from known type strains to assign sequences (Wang et al., 2007). Each match result is characterised by the seqmatch score (S_ab). S_ab describes the number of (unique) 7-base oligomers shared between the investigated sequence and a given RDP sequence divided by the lowest number of unique oligos in either of the two sequences (http://rdp.cme.msu.edu).

CP_PCR

This PCR approach is based on the amplification of a C. perfringens specific part of the 16S rRNA gene. Primers CP1f and CP2r according to Wang et al. (1994) have been used for the amplification of the 279 bp product (Table 1). The optimum of the amplification temperature was investigated in the range between 50 °C and 60 °C and was finally found to be at 55 °C. The 25 μl PCR reaction mixture consisted of 1× Taq buffer, 0.9 U Taq polymerase native, 3 mM of MgCl2, 0.25 mM dNTPs (all Fermentas, Germany), 0.05% w/v bovine serum albumin (Boehringer Mannheim, Mannheim, Germany), 0.25 μM of each primer (Eurofins MWG; Germany) and DNA sample. The PCR was performed using iCycler (Bio-Rad, USA) with the following optimised thermal profile: initial denaturation at 95 °C for 60 s followed by 35 cycles with 30 s of denaturation at 95 °C, 30 s of primer annealing at 55 °C and 60 s of primer extension at 74 °C and as last step final elongation at 74 °C for 120 s. This PCR approach showed to be very susceptible to false positive reaction as a function of DNA starting concentration. The optimal starting concentration of sample DNA was determined as approximately 1 pg sample DNA per 25 μl reaction mixture. For higher concentrations false positive results were obtained. The PCR products were separated by electrophoresis on agarose gel (2%) containing ethidium bromide. Fragment size was referred to the size standard GeneRuler Low Range DNA Ladder (Fermentas, Germany).

PLC_PCR

The second PCR approach was for the detection of a 283 bp fragment of the C. perfringens specific plc-Gene using the primers PL3 and PL7 (Table 1) according to Fach and Popoff (1997). The PCR was carried out on iCycler (Bio-Rad, USA) using the following thermal profile: initial denaturation at 94 °C for 60 s followed by 30 cycles with 30 s of denaturation at 94 °C, 30 s of primer annealing at 55 °C and 30 s of primer extension at 72 °C and as last step final elongation at 72 °C for 600 s. The 25 μl PCR reaction mixture consisted of 1× Taq buffer, 2.5 U Taq polymerase native, 1.5 mM of MgCl2, 0.2 mM dNTPs, (all Fermentas, Germany) and 0.5 μM of each primer (Eurofins MWG, Germany) and DNA sample (starting concentration 100–1000 pg sample DNA/25 μl reaction mixture). The PCR products were separated by electrophoresis on agarose gel (2%) containing ethidium bromide. Fragment size was referred to the size standard GeneRuler Low Range DNA Ladder (Fermentas, Germany).

PCR quality and inhibition control

The reliability of the PCR methods was tested in a pre-trial during the adaptation of the methods on a representative subset of strains in at least duplicate. The type strains and the strains which underwent sequencing were determined by the two PCR methods in duplicate. To monitor possible amplification inhibition of CP_PCR and PLC_PCR each sample was run twice. One of the samples was spiked with DNA extracted from type strain C. perfringens DSMZ No. 756 as internal control. For sensitive inhibition control, spiking was performed with the lowest amplificable concentration of DNA, which was estimated as 1 pg DNA for CP_PCR and 100 pg DNA for PLC_PCR per 25 μl PCR reaction mixture, respectively.

Statistics

Accuracy was described by the statistical measures sensitivity and specificity. Sensitivity expresses the proportion of correctly identified positives and specificity reveals the proportion of correctly identified negatives:

Sensitivity %: (true positive results ∗ 100) / (true positive + false negative results)

Specificity %: (true negative results ∗ 100) / (true negative + false positive results)

Percentage of correct identification ((true positive results + true negative results) ∗ 100 / all results) was used as a summative parameter for overall comparison.

These measures were used to compare and describe firstly the accuracy of the selected PCR methods and secondly the performance characteristics of the phenotypic methods.

Results and discussions

Evaluation and selection of a PCR based identification method for C. perfringens

All type strains (n = 14) exhibited a perfect match for both PCR reactions (sensitivity and specificity 100% each) (Table 2a).

Sequence analysis showed that 13 of the subset of environmental strains (n = 30) were C. perfringens (S_ab score 1.00–0.99) and 17 non C. perfringens (S_ab score in general 1.00–0.95, except two strains of Clostridium tertium with S_ab scores 0.93 and 0.88). Obtained sequences were submitted to NCBI GenBank. For detailed GenBank accession numbers and S_ab scores see Table 2b. This subset of environmental strains revealed a sensitivity of 92% and a specificity of 88% for PLC_PCR due to one false negative and two false positive results (Table 3). PLC_PCR gave a false positive result for one Clostridium bifermentans strain and one out of three Clostridium sordelli. This might be explained by the fact, that phospholipase genes of C. bifermenans and C. sordelli show partial homology with that of C. perfringens (Karasawa et al., 2003). It should be emphasised that this discrepancy did not appear with the respective type strains but only with the isolates from the environment. The environmental strains revealed for CP_PCR a slightly lower accuracy with two false negative and three false positive results. Accordingly, the sensitivity decreased to 85% and the specificity to 82% (Table 3). The percentage of correct identification was 90% for PLC_PCR and 83% for CP_PCR. Furthermore PLC_PCR proved to be more robust against varying DNA template concentrations compared to CP_PCR, which gave false positive results for higher DNA concentrations (≥ 10 pg/25 μl PCR reaction mixture). Thus, PLC_PCR was selected for evaluation and comparison of the ISO_LGMN procedure and the acid phosphatase test (PLC_PCR sensitivity and specificity was set to 100%).

Table 2b

Results of testing PLC_PCR and CP_PCR on environmental strains (C. perfringens and other Clostridia).

Closest match RDP	S_ab score	NCBI GenBank accession no.	PLC_PCR	CP_PCR
Clostridium perfringens	1.00	JN048955	+	+
Clostridium perfringens	1.00	JN048944	+	+
Clostridium perfringens	1.00	JN048945	+	+
Clostridium perfringens	1.00	JN048948	+	+
Clostridium perfringens	1.00	JN048940	+	+
Clostridium perfringens	1.00	JN048934	+	−
Clostridium perfringens	0.99	JN048953	+	+
Clostridium perfringens	0.99	JN048947	−	−
Clostridium perfringens	0.99	JN048959	+	+
Clostridium perfringens	0.99	JN048950	+	+
Clostridium perfringens	0.99	JN048946	+	+
Clostridium perfringens	0.99	JN048958	+	+
Clostridium perfringens	0.98	JN048949	+	+
Clostridium baratii	0.98	JN048942	−	−
Clostridium bifermentans	1.00	JN048952	+	+
Clostridium butyricum	1.00	JN048941	−	−
Clostridium ghonii	0.99	JN048963	−	−
Clostridium ghonii	0.99	JN048961	−	−
Clostridium ghonii	0.97	JN048937	−	−
Clostridium ghonii	0.96	JN048936	−	−
Clostridium ghonii	0.96	JN048962	−	−
Clostridium sardiniense	0.98	JN048956	−	−
Clostridium sardiniense	0.95	JN048939	−	−
Clostridium sordellii	1.00	JN048938	−	−
Clostridium sordellii	0.99	JN048960	+	+
Clostridium sordellii	0.99	JN048957	−	−
Clostridium sporogenes subsp. tusciae	0.99	JN048943	−	−
Clostridium tertium	0.93	JN048951	−	−
Clostridium tertium	0.88	JN048954	−	+
Eubacterium tenue	0.98	JN048935	−	−

Table 3

Evaluation of the accuracy of PLC_PCR and CP_PCR used for identification of presumptive C. perfringens isolates from environmental samples.

	PLC_PCR	CP_PCR
True positive results	12	11
True negative results	15	14
False positive results	2	3
False negative results	1	2
Total number of strains	30	30
Sensitivity %	92	85
Specificity%	88	82
% of correct identification	90	83

Accuracy of the confirmation methods ISO_LGMN procedure and acid phosphatase test

Ten type strains gave equal results for the ISO_LGMN procedure and the acid phosphatase test resulting in sensitivity and specificity of 100% and 86%, respectively. Both methods showed a false positive result for a strain of Clostridium sardiniense. False positive results of the acid phosphatase test for C. sardiniense have been reported previously (Sartory et al., 2006). Four non C. perfringens type strains were not able to grow on TSC agar at 44 °C (Table 2a) and were therefore excluded from phenotypic investigation (ISO_LGMN procedure and acid phosphatase test).

For environmental strains (n = 127), the sensitivity of the acid phosphatase test was 94% compared to 79% of the ISO_LGMN procedure (Table 4a). Specificity of the acid phosphatase test was with 87% lower than 95% for the ISO_LGMN procedure. However, percentage of correct identification was 92% for the acid phosphatase test and 83% for the ISO_LGMN procedure. Our results are comparable to those of a study by Eisgruber et al. (2000), in which 94% of the C. perfringens strains could be confirmed by the acid phosphatase test and 89% by the ISO_LGMN procedure. Another study revealed that 95% of the C. perfringens strains are positive for acid phosphatase (Eisgruber et al., 2003).

Table 4

Results of ISO_LGMN procedure and acid phosphatase test (APT).

	LGMN	APT
a.) Environmental isolates
True positive results	70	84
True negative results	36	33
False positive results	2	5
False negative results	19	5
Total number of strains	127	127
Sensitivity %	79	94
Specificity %	95	87
% of correct identification	83	92
b.) Isolates from water
True positive results	22	30
True negative results	16	13
False positive results	0	3
False negative results	11	3
Total number of strains	49	49
Sensitivity %	67	91
Specificity %	100	81
% of correct identification	78	88
c.) Isolates from faeces
True positive results	48	54
True negative results	20	20
False positive results	2	2
False negative results	8	2
Total number of strains	78	78
Sensitivity %	86	96
Specificity %	91	91
% of correct identification	87	95

Results of the phenotypic methods for strains isolated from water samples (n = 49) differed significantly from the results obtained with the isolates from faeces. The isolates from water revealed a sensitivity for the acid phosphatase test of 91% compared to 67% for the ISO_LGMN procedure whereas specificity of the acid phosphatase test was 81% compared to 100% for the ISO_LGMN procedure (Table 4b). The low sensitivity means that 33% (n = 11) of the C. perfringens water strains (identified with PLC_PCR) could not be confirmed with the ISO_LGMN procedure. Four of the false negative C. perfringens strains were motile and did not reduce nitrate, further five were motile and another two did not ferment lactose. The occurrence of motility in C. perfringens strains was surprising. However, recent studies showed, that C. perfringens, which is considered to be non-motile, is able to type IV pilus (TFP)-dependent gliding motility (Mendez et al., 2008; Varga et al., 2006). Sartory et al. (2006) reported that 10% or more of C. perfringens strains were negative with respect to the nitrate reduction test. Overall percentage of correct identification was with 88% significantly higher for the acid phosphatase test than for the ISO_LGMN procedure (78%).

Faecal isolates (n = 78) exhibited, compared to isolates from water, lower differences in sensitivity for the acid phosphatase test (96%) and the ISO_LGMN procedure (86%) and the same specificity of 91% for both methods (Table 4c). Again the percentage of correct identification was higher for the acid phosphatase test (95%) than for the ISO_LGMN procedure (87%). False negative results for ISO_ LGMN procedure were due to two gelatinase negative strains, one motile strain, three strains negative for nitrate reduction and two lactose fermentation negative strains. The negative results for gelatinase test are in accordance with Eisgruber et al. (2000) who observed strains of C. perfringens that are not able to liquefy gelatine.

Based on the selected species specific PLC_PCR method an evaluation and comparison of the two phenotypic confirmation methods was possible. The acid phosphatase test proved to be a reliable method for the confirmation of C. perfringens isolates from water and faeces. The advantages compared to the ISO_LGMN procedure are the much easier application and the shorter duration of analysis. This permits to investigate a much higher percentage of the presumptive C. perfringens isolates, leading to an increased quality of analysis. Furthermore the acid phosphatase test showed a higher sensitivity for the identification of C. perfringens isolates from water. From the results presented here, we conclude that confirmation of C. perfringens isolates by the acid phosphatase test is superior to the ISO_LGMN procedure.