Disinfection of sulfate-reducing clostridia, including Clostridium perfringens, in sewage effluents using peracetic acid

Clostridium perfringens is a bacterium that produces a variety of exotoxins and enzymes that cause disease, necessitating public health control measures (Ohtani & Shimizu 2016). The major toxins (α, β, ε, and ι) produced by bacteria are used to classify them as types A–E. α toxin causes gas necrosis and is the major toxin produced by type A bacteria. In contrast, β toxin causes necrotizing enterocolitis and is produced by type C bacteria, and ε toxin is neurotoxic to animals (Rood et al. 2018). C. perfringens enterotoxin (cpe) is pathogenic, and C. perfringens that express the cpe gene can cause food poisoning (Johansson et al. 2006); this can involve mass food poisoning affecting large numbers of patients, ranking second only to verotoxin-producing Escherichia coli in the number of bacterial food poisoning cases in Japan in 2020 (1,288 cases) (Ministry of Health, Labour and Welfare 2020). Similarly, in the USA, it ranked second only to Salmonella enterica (7,834 cases) from 2009 to 2015 (Mattia & Manikonda 2018). Since this bacterium is anaerobic (Iacumin & Comi 2021), curries (Kitadokoro et al. 2013) and stews (Wahl et al. 2013) prepared in large quantities and served warm over long periods are a potential source of C. perfringens food poisoning because of the slightly anaerobic conditions in the center of the food.

As complex dishes comprising a variety of ingredients are often the source of bacterial food poisoning, it is difficult to identify the causative ingredient. Meat and meat products have long been regarded as causative agents because C. perfringens is ubiquitous in animal intestinal tracts and can be detected at high rates therein (Grass et al. 2013). However, while recent studies based on genetic testing have detected C. perfringens in commercial meat, this did not include cpe gene-carrying C. perfringens (Miki et al. 2008), implying that meats may not be the only causative ingredient. Although there have been some reports of C. perfringens food poisoning attributed to vegetables, research indicates that they accounted for 36% of all C. perfringens food poisoning cases in France from 2013 to 2017 (Mahamat Abdelrahim et al. 2019). Furthermore, recent studies found C. perfringens in cultivated potato soils (Voidarou et al. 2011), as well as a high frequency (83%) of C. perfringens in commercial potato surface soils, 30% of which harbored the cpe gene (Hashimoto et al. 2023).

There have also been various reports on the isolation of cpe gene-carrying C. perfringens from the environment, as well as soil (Li et al. 2007), river water (La Sala et al. 2015), and marine areas (Yanagimoto et al. 2020); these represent a possible source of food contamination (Hashimoto et al. 2023). Of these, sewage effluent after chlorination is a major source of C. perfringens contamination in aquatic environments (Suzuki et al. 2021). Sulfite-reducing clostridia (SRC), including C. perfringens originating from human feces, are abundant in sewage effluent (3.1 × 10¹ cfu/mL) and include a high frequency (30.2%) of cpe gene-positive C. perfringens (Suzuki et al. 2021). Reports suggest that cpe gene-positive C. perfringens have reached the sea through sewage effluent and contaminated marine products (Yanagimoto et al. 2020). Alternatively, cpe gene-carrying C. perfringens could have contaminated root crops via soil through sewage effluent (Hashimoto et al. 2023). Overall, this shows that treated human sewage effluent may play a role in C. perfringens pollution. The average pollution load of C. perfringens in treated human sewage effluent, especially after chlorine disinfection, is 9.9 × 10¹⁰ cfu/m³/day, of which 2.5 × 10¹⁰ cfu/m³/day of C. perfringens carrying the cpe gene is released into the environment, levels that cannot be ignored (Suzuki et al. 2021).

A major cause of this environmental load may be the high environmental viability and disinfection resistance of C. perfringens due to spore formation (Bisson & Cabelli 1980). This bacterium is resistant to chlorine, the most widely used disinfectant for sewage. The time required for 5 log inactivation when exposed to 8 mg/L chlorine is approximately 5 min for E. coli and 10 min for Enterococci (Tree et al. 2003), while C. perfringens spores were only 1.4 log inactivated even when exposed to 5 mg/L chlorine for 4 h (Venczel et al. 1997). In other words, SRC-containing cpe gene-positive C. perfringens are highly resistant to chlorine and may be contaminating soil, rivers, and marine areas through sewage effluent. As a result, these bacteria could be causing food poisoning through root crops and marine products. This demonstrates that chlorine, a conventional disinfectant, might not ensure the microbiological safety of the environment into which sewage effluent is discharged.

In addition to the problem of chlorine-resistant microorganisms, chlorine added for disinfection synthesizes with dissolved organic matter, such as humic acid, in the water to produce disinfection byproducts (DBPs), including trihalomethanes (THMs) (World Health Organization 2011). DBPs have many disadvantages, such as genotoxicity, epigenetic effects (Shanks et al. 2013), carcinogenicity (Richardson et al. 2007), mutagenicity, and genetic degeneration (Guzzella et al. 2004; Monarco et al. 2005; Krasner 2009).

In Italy, which has recently set strict environmental standards for THM concentrations in sewage effluent, several studies have been conducted on new disinfectants to replace chlorine (Stampi et al. 2001; Mezzanotte et al. 2007; Bonetta et al. 2017). Among these, peracetic acid (PAA) is a promising alternative or complementary disinfectant for sewage effluent. Its advantages include a proven track record as a disinfectant in the medical and food hygiene fields (Block 2001), its highly degradable nature, and its lack of operationally toxic byproducts (only water, oxygen, and hydrogen peroxide are produced) (Wagner et al. 2002). Studies have begun to examine the application of PAA to sewage treatment from various angles, one research showing that pretreatment combined with UV contributed to rapid disinfection of sewage (Bai et al. 2024), and another showing that treatment with PAA during sewage sludge recycling can efficiently separate medium-chain fatty acids, a product with high added value (Wang et al. 2024). Furthermore, PAA induces protein denaturation in microorganisms (Baldry et al. 1991) or inhibition of metabolic enzymes (Tutumi et al. 1974), and it has an inactivation mechanism different from chlorine (Lefevre et al. 1992).

However, while there are reports that PAA is effective against C. perfringens spores (Briancesco et al. 2005), there are also reports that it has no inactivating effect on C. perfringens in sewage (Gehr et al. 2003). Thus, there are few studies evaluating the inactivation of C. perfringens by PAA and their results are inconsistent. Particular care must be taken when quantifying the inactivating effect of PAA on C. perfringens in sewage treatment where multiple wild strains are present as C. perfringens culture requires selective media, such as m-CP or Handford's, which take advantage of the sulfite-reducing ability of SRC. In this case, all SRC-containing C. perfringens would be quantified, making it a controversial method for accurate C. perfringens quantification. All SRC colonies grown on these selection media have generally been presumed to be C. perfringens; however, additional identification testing is required to definitively identify them as C. perfringens.

This study attempted to quantify the percentage of C. perfringens in SRC by detecting the cpa gene. The log inactivation rate was calculated based on the SRC concentration corrected for the proportion of C. perfringens in the SRC, and the log inactivation rate of C. perfringens by PAA was estimated. In addition, PAA-resistant SRC remaining after PAA exposure was identified using biochemical characterization tests.

Ten liters of the final sedimentation tank effluent (before chlorine addition) from the Shobara (SB) sewage treatment plant, in Hiroshima, Japan (oxidation ditch process, average daily effluent flow = 3.2 × 10³ m³/day), were collected and tested on the same day.

The sodium hypochlorite solution (Fujifilm Wako, Tokyo, Japan) was used as the chlorinating agent, and the PAA solution used was Perasan^®MP2-J (Kanto Chemical, Tokyo, Japan). The effective concentration of each disinfectant was measured with a spectrophotometer MP-100 (Lovibond, Dortmund, Germany) by using a common colorimetric reagent, N,N-diethyl-p-phenylenediamine (DPD), and PAA was measured with a PA-300 device (Hiranuma Sangyo, Ibaraki, Japan) using the amperometric titration method.

One liter of sewage effluent was poured into a 3-L Erlenmeyer flask. PAA or sodium hypochlorite was injected to reach the prescribed concentration for each experiment and allowed to react at room temperature (approximately 20 °C) with gentle stirring. Sampling was performed at every prescribed time point for each experiment before contact with the disinfecting solution began to measure the concentration of PAA or free residual chlorine. After the reaction was stopped by adding sodium thiosulfate, the concentration of E. coli and/or SRC was determined.

In the sterile distilled water system, the standard E. coli strain NBRC3301 was used after growth in the trypto-casein soy agar (TSA) medium for 24 h at 36 °C until the logarithmic growth phase. Surviving colonies were resuspended in phosphate buffer, and the absorbance was measured to check that the bacterial solution was at a concentration of 10⁸–10⁹ cfu/mL. The bacterial solution was added to 100 mL sterile distilled water to reach a concentration of 10⁸ cfu/mL before sodium hypochlorite was injected to reach the target concentration; the reaction was carried out at room temperature with gentle stirring. Sampling was performed at every given time before contact began to measure the concentration of PAA or free residual chlorine. After the reaction was stopped by adding 1 mL of 1 mol/L of sodium thiosulfate, the number of E. coli and/or SRC was determined.

E. coli concentration was calculated using pour plate culture or the membrane filter method with X-mg agar. Colonies were identified and quantified based on the number of blue to blue-purple colonies formed.

For SRC, sample water was dispensed into 100 mL polycentrifuge tubes, heated in a water bath at 75 °C for 20 min, and then immediately cooled on ice. The heated samples were inoculated onto a Handford-modified agar medium by the mixed culture or the membrane filter method, triple-layered, and placed into anaerobic jars for 24 h at 45 °C. After incubation, black colonies exhibiting sulfite-reducing properties were counted as SRC. These SRC colonies were randomly selected. Each selected single colony was spread onto a Colombia 5% sheep blood agar medium (Sysmex Biomérieux, Tokyo, Japan) with an inoculation loop, placed in an anaerobic jar, and incubated anaerobically for 24 h at 36 °C for enrichment culture. After multiplication, SRC colonies were collected with an inoculating loop, placed in screw-capped tubes containing 1 mL of Tris EDTA buffer solution, and frozen at −30 °C.

First, a 2.5 μL cryopreserved SRC bacterial solution was placed into microtubes before 0.5 μL of 10% TritonX and 2 μL of double distilled water were added and heated in a thermal cycler at 95 °C for 5 min to lyse the bacteria. The multiplex PCR was performed as described by van Asten et al. (2009) to detect cpa, cpb, cpb2, etx, iap, and cpe toxin genes (Table 1). The strain carrying the cpa gene encoding the alpha toxin was identified as C. perfringens as this gene is always retained by C. perfringens.

Each isolated SRC colony was grown on a Columbia 5% sheep blood agar medium and subjected to the biochemical bacterial identification test using Rapid ID 32A (Sysmex BioMérieux) and the dedicated database website (apiweb; Sysmex BioMérieux, https://apiweb.biomerieux.com) following the manufacturer's instructions.

The inactivation of E. coli in sterile distilled water by PAA was similar at all initial PAA concentrations (1.0, 2.5, and 5.0 mg/L). There was a period during which inactivation did not progress at CT values up to 10 mg·min/L.

There was a lag stage during which inactivation did not progress until the CT value reached 10 mg·min/L.

However, inactivation progressed rapidly above a CT value of 10 mg·min/L, reaching −4.83 and −8 log at CT values of 13.44 and 20.48 mg·min/L, respectively.

Using chlorine and these same conditions, an inactivation rate of −4 log was obtained immediately after contact at an initial concentration of 0.1 mg/L (data not shown). PAA inactivates E. coli more slowly than free chlorine, indicating that sufficient contact time is required to achieve inactivation equivalent to that of free chlorine.

Experiments were conducted with PAA and chlorine in sewage effluent to evaluate their inactivation effects on E. coli in conditions where various organic substances consume them. Free chlorine achieved an inactivation rate of −4 log at a CT value of 0.1 mg·min/L for E. coli in sterile distilled water, while, in sewage effluent, this decreased to approximately −3 log at a CT value of 3.4 mg·min/L. The inactivation effect of PAA was −4.83 log at a CT value of 13.44 mg·min/L for E. coli in sterile distilled water and −2 log at a CT value of 16.64 mg·min/L in sewage effluent. The CT values required for 2.5 log inactivation in sewage effluent were 11.2 and 18.9 mg·min/L for free chlorine and PAA, respectively. Unlike in sterile distilled water, the inactivation efficiency of free chlorine for E. coli in sewage effluent was significantly reduced, whereas that of PAA was less than free chlorine. PAA could maintain its concentration in the presence of organic matter and showed little attenuation of its inactivation effect on E. coli. The progression of inactivation was dependent on the CT value, and contact concentration had no significant effect.

The inactivation rate by PAA was evaluated 16 times, and linear inactivation was observed up to approximately 2 log with a CT value of 2,000 mg·min/L. However, it was observed that the trend of increasing inactivation rate became slower when the CT value exceeded 3,000 mg·min/L. Compared to E. coli, SRC was resistant to both free chlorine and PAA and required a higher contact time for adequate inactivation. Furthermore, inactivation by PAA was delayed up to approximately 3 log, even with a high CT value.

The SRC detected before and after PAA contact and the prevalence of cpa genes are shown in Table 2. Before PAA contact, strains carrying genes other than cpb and iap were detected in 16.7–96.7% of strains, respectively. Before PAA inactivation, the cpa-positive rate of SRC was very high, reaching 80% (24/30) to 96.7% (29/30); this decreased to 82.5% (33/40) and 0% (0/30) for CT values of 360 and 4,772 mg·min/L, respectively.

The cpa-positive rate of SRC was not detected under high CT conditions, indicating that PAA is effective in inactivating C. perfringens and that non-C. perfringens PAA-resistant SRCs are present in sewage effluents and remain after PAA inactivation.

At CT values of 89 and 192 mg·min/L, the estimated log inactivation rates of C. perfringens and the log inactivation rate of SRC were almost identical: 0.097, 0.014, 1.25, and 1.31, respectively.

At CT values of 89 mg·min/L, the estimated log inactivation rates of C. perfringens and SRC were almost identical: −0.097 and −0.014, respectively (Figure 5). A similar trend was observed with CT values of 192 mg·min/L, with inactivation rates of −0.58 and −0.59, respectively (Figure 5).

In contrast, at a CT value of 3,025 mg·min/L, the SRC inactivation rate was 2.41 log, while that of C. perfringens was 3.17 log. The slope for the former was 0.0007, while that of the latter was 0.001, indicating that PAA inactivates C. perfringens 10 times more rapidly than SRC.

In this study, we investigated the effect of PAA, which has been used experimentally in Italy and other countries for sewage effluent disinfection (Stampi et al. 2001; Mezzanotte et al. 2007; Bonetta et al. 2017), on C. perfringens spores, as previous evaluations of its disinfectant effects are inconsistent (Gehr et al. 2003; Briancesco et al. 2005).

When PAA was added to sewage effluent and its persistence evaluated, 80–90% remained after 120 min of contact. The disinfection effect in the presence of organic matter was compared between chlorine and PAA for E. coli in sewage effluent. The results showed that in proportion to the persistence of each disinfectant, the inactivation effect of chlorine was greatly reduced, while PAA was not affected by organic matter as much as chlorine. The results indicate that PAA can persist for a long time in sewage effluent with organic matter and continue to disinfect therein. The persistence of PAA even in the presence of organic matter is a significant advantage over chlorine, which rapidly oxidizes and decomposes in the presence of organic matter, losing its effectiveness (Teng et al. 2018). This characteristic may allow for disinfection with a lower environmental impact by ensuring contact time at lower concentrations, such as injection immediately after biological treatment.

The inactivation curve of E. coli in pure water by PAA was gentler than that of free chlorine, and with a delay in the low CT region, a multi-hit aspect was observed, i.e., a model in which more than one disinfection factor needs to reach for inactivation. Although the mechanism underlying inactivation by chlorine and PAA is not fully understood, the difference in the inactivation curve of E. coli by chlorine and PAA is potentially influenced by differences in their inactivation mechanism. The mechanism of bacterial inactivation by chlorine generally involves significant damage to cell walls and biological membranes by oxidation (Jefri et al. 2022). The inactivation mechanism of PAA not only involves a weak oxidation process compared to chlorine but also multiple other mechanisms, such as metabolic enzyme inhibition (Tutumi et al. 1974) or inhibition of cellular protein denaturation and related transport (Fraise et al. 2013). This mild oxidation and the characteristics of the inactivation mechanism of PAA, which are not solely dependent on oxidative action, may be responsible for its long persistence and continued inactivation effect in the presence of organic matter. In addition, its persistence and inactivation mechanisms might be an advantage for sewage effluent disinfection that could replace or supplement chlorine disinfection.

The results of this study indicate that the inactivation of C. perfringens by PAA is approximately 1 log higher than that of SRC and that a CT value of approximately 3,000 mg·min/L can ensure the inactivation of approximately 3 log. Considering that C. perfringens is discharged into water systems at high loads from sewage effluent after chlorine disinfection (Suzuki et al. 2021) and its detection in river water and clams (Yanagimoto et al. 2020), as well as the possibility of root crop contamination through the contamination of agricultural land (Hashimoto et al. 2023), it is important for public health that chlorine disinfection is supplemented to control the release of C. perfringens into the environment

In this study, SRC-containing C. perfringens were counted using the Handford-modified agar plate counting method to evaluate inactivation by PAA against C. perfringens. In addition, the cpa gene of SRC strains at each inactivation stage was evaluated to determine the percentage of C. perfringens in the SRC and to estimate the percentage of C. perfringens inactivated by PAA. Furthermore, the biochemical identification of viable SRC bacteria species was performed using Rapid ID 32A. As a result, we succeeded in providing a reliable quantitative evaluation of the inactivation effect of PAA against C. perfringens, which had previously been evaluated differently (Gehr et al. 2003; Briancesco et al. 2005). A more accurate qualitative and resistance assessment of the surviving strains requires further investigation.

While PAA was found to have a certain inactivating effect on C. perfringens, non-C. perfringens PAA-resistant SRC was found to survive even at CT values above 3,000 mg·min/L. Further investigation of the toxicity and health effects of these resistant strains, such as C. fallax and C. baratii, is essential. Alternatively, tracking the distribution of disinfectant-resistant strains in the environment could be considered as a tool to investigate the extent of the impact of sewage treatment plant effluent in the destination river.

Chlorine is currently used as a disinfectant for sewage effluent due to its broad antimicrobial spectrum, accumulated knowledge, ease of use, and relatively low cost (Zheng et al. 2017). However, residual chlorine in discharged water has adverse effects on aquatic organisms in rivers (Brungs 1973). In particular, it can inhibit the growth of attached algae and reduce algal biomass (Aratani et al. 2007). Additional problems, such as carcinogens generated from DBPs, including THMs, have been identified (Guzzella et al. 2004; Monarca et al. 2005; Richardson et al. 2007; Krasner 2009; World Health Organization 2011; Shanks et al. 2013). Furthermore, the presence of chlorine-resistant pathogenic microorganisms, represented by spore-forming bacteria, remains a concern (Bisson & Cabelli 1980).

Chlorine dioxide and ozone are typical chemical alternative disinfectants, but these oxidizing disinfectants, like chlorine, produce DBPs (World Health Organization 2011). Ultraviolet disinfection, which is a typical physical treatment used in Japan for sewage and drinking water, is expensive (Vilhunen et al. 2009) and can produce issues if a series of light recovery enzymes correct the genetic defects caused by ultraviolet irradiation when visible light is present after treatment (Oguma et al. 2002). There have also been attempts to introduce membrane filtration using ultrafiltration or microfiltration membranes into sewage treatment, but the introduction cost and equipment difficulties remain significant (Fenu et al. 2010).

Previous data on the inactivation effect of PAA on C. perfringens have been inconsistent, with some studies finding it to be as effective as chlorine on C. perfringens in sewage effluent, while others found it to be ineffective. In this study, we evaluated the inactivation of SRC and C. perfringens by PAA in sewage effluent with large amounts of organic matter and found that a log inactivation rate of −2.4 log at the CT value of 3,000 mg·min/L.

PAA has little impact on the environment and ecosystem of the discharge site and was shown to be effective as a disinfectant against spore-forming bacteria such as C. perfringens that are resistant to chlorine disinfection even in the presence of organic matter such as in treated sewage water. Although certain issues remain, such as the existence of PAA-resistant strains and the need for contact with higher concentrations than chlorine, PAA has merit enough to be considered as an alternative sewage disinfectant to chlorine.

This work was supported by a Grant-in-Aid for Scientific Research(B), JSPS KAKENHI Grant Number 23K26234, and SATAKE Technical Foundation.

H.S. designed the study, the main conceptual ideas, and the proof outline. M.T. and T.S. collected the data. M.Y. and K.O. aided in interpreting the results and worked on the manuscript. A.H. supervised the project. All authors discussed the results and commented on the manuscript.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.