Prevalence, antibiotic resistance and virulence of Enterococcus spp. from wastewater treatment plant effluent and receiving waters in South Africa

In spite of advances and improvements in wastewater treatments, numerous wastewater treatment plants (WWTPs) in South Africa have been reported to be functioning inefficiently (DWA 2012; Iweriebor et al. 2015; Hamiwe et al. 2019). According to Dąbrowska et al. (2017), municipal wastewater effluents contain various toxic substances, such as pharmaceuticals, hormones and pathogens. The standard microbiological indicator parameter for disinfection efficiency in South Africa is Escherichia coli (DWA 2013). However, studies have illustrated the presence of other microorganisms, in effluent, some of which are potential pathogens (Cai et al. 2014; Do et al. 2019; Numberger et al. 2019). Among these potentially pathogenic microorganisms are the Enterococcus group which have been frequently reported in effluents (Iweriebor et al. 2015; Taučer-Kapteijn et al. 2016; Hamiwe et al. 2019; Fouz et al. 2020).

Enterococcus are a diverse, complex and important group of bacteria with regard to their interaction with humans. They are commensal organisms in the mammalian gastrointestinal tract but are also found in soil, water and plants (Byappanahalli et al. 2012). Despite being harmless in vivo, their spread from the intestinal tract to other mucosal and skin surfaces can cause infections (Vu & Carvalho 2011). Enterococci are opportunistic pathogens and target patients with underlying diseases, as well as those treated with a broad spectrum of antibiotics, the elderly and immune-compromised persons (Vu & Carvalho 2011). For Enterococcus spp. to cause infection, they must possess virulence factors that allow them to colonize and invade the host's tissue, translocate through epithelial cells and evade the host's immune response (Hammad & Shimamoto 2014). Additionally, the emergence of infections in a susceptible host is determined by the presence of multiple virulence factors, in pathogenic bacteria, functioning collectively or singularly at various stages of infection (Wu et al. 2008). Virulence factors are gene products, traits or effector molecules produced by pathogens in order to enhance the ability of a microorganism to cause tissue damage, disease and systematic inflammation (Mundy et al. 2000; Vu & Carvalho 2011). For this reason, the study of these organisms’ virulence mechanisms has become an important aspect (Semedo et al. 2003). Studying these organisms’ virulence mechanisms allows for determining their ability to establish infection in a susceptible host (Wu et al. 2008).

Additionally, enterococci form part of the ESKAPE (Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa and Enterobacter spp.) pathogens on the World Health Organisation's priority list (Ramsamy et al. 2018). These pathogens have been recognized as common causes of nosocomial infections due to their ability to escape the biocidal action of readily prescribed antibiotics (Hrenović et al. 2019). This is because Enterococcus spp. have been exposed to multiple antibiotics in the hospital setting, such as (but not limited to) amoxicillin, ampicillin and vancomycin, that have provided them with an evolutionary pressure for selection (SAASP 2014). The resistance of Enterococcus spp. to various antibiotics allows them to survive the hospital environment which provides an opportunity for their dissemination (Murray 1990). This dissemination can be via wastewater from domestic, industrial, veterinary and hospital settings; inevitably ending up in municipal WWTPs. Consequently, the WWTPs do not remove pollutants adequately allowing for bacteria, as well as antimicrobial residues, to enter receiving environmental water bodies (Michael et al. 2013).

Several studies have reported the presence, emergence, as well as outbreak of antibiotic-resistant clinical Enterococcus spp., in South Africa (McCarthy et al. 2000; von Gottberg et al. 2000), while in recent times, studies have reported the presence of antibiotic-resistant Enterococcus pathogens in hospital and domestic wastewater effluents in other parts of the country (Iweriebor et al. 2015; Hamiwe et al. 2019). However, there is a paucity of data on susceptibility patterns and epidemiology of antibiotic-resistant Enterococcus isolated from WWTPs harbouring virulence genes in the North West Province. The presence of Enterococcus spp. exhibiting antibiotic resistance and harbouring clinically relevant virulence genes is of particular interest due to the possible link of community-acquired enterococcal infections and recreational activities. The data generated in this study will be relevant and of value to surveillance programmes, such as the WHO Global Antimicrobial Resistance Surveillance System.

The aim of this study was to determine the antimicrobial resistance patterns and presence of virulence genes in Enterococcus spp. isolated from three municipal WWTPs' final effluent, as well as receiving environmental water systems in the North West Province, South Africa.

Three WWTPs in the North West Province which receive wastewater from urban households, industries, farms and hospitals were sampled during May 2015. All samples were collected at the final effluent points downstream and a point between two of the WWTPs. The GPS co-ordinates illustrating the position of the three WWTPs are displayed in Table 1. Three maps indicating the various sampling sites can be found in the Supplementary Appendix. Plant 1 uses a trickling filter system, plants 2 and 3 are activated sludge systems.

The dip sampling technique was employed at each sampling site (US EPA 1994). Briefly, a rope was tied to the neck of the bottle and the bottle was dipped into the surface water system. All water samples were collected aseptically in a sterile glass bottle. Sterilized bottles were labelled before sampling and gloves were worn when handling the bottles. All samples were kept on ice in a sealed cooler box and analysed within 6 h of collection. The membrane filtration technique was employed for Enterococcus isolation and enumeration (Greenburg et al. 1998). Triplicates of 100 ml water samples were filtered through 0.45 μm (47 mm grid) PALL Corporation sterilized filter membranes [(PALL Life Sciences, Mexico) (cat. no. GN-6 Metricel Membrane 66191)] and placed on KF-Streptococcus agar containing 1 ml of 2,3,5-Triphenyltetrazolium chloride per 100 ml (Sigma Aldrich, South Africa). The KF-Streptococcus agar plates were incubated at 37 °C for 48 h. Single well-isolated pink colonies were aseptically sub-cultured three times on nutrient agar using the streak plate technique and incubated for 24 h at 37 °C. A total of nine isolates were chosen for each sampling site.

For each Enterococcus isolate screened, genomic DNA was extracted using a Macherey-Nagel Nucelospin Tissue kit using the manufacturer's instructions (Macherey-Nagel, Germany). The quantity and quality of the isolated total genomic DNA were determined using a NanoDrop^TM 1000 Spectrophotometer (Thermo Fischer Scientific, USA) and agarose electrophoresis. The genomic DNA was then amplified using an ICycler thermal cycler (Bio-Rad, UK). A polymerase chain reaction (PCR) reaction mixture with a final volume of 25 μl was used and it contained RNase/DNase free water (Fermentas Life Sciences, USA), 12.5 μl 2 × PCR Master Mix (0.05 U/μl Taq DNA Polymerase in reaction buffer, 0.4 mM of each dNTP, 4 mM MgCl₂) (Fermentas Life Sciences, USA), 50 μM primer and 1 μl bacterial DNA template (50–100 ng/μl). A portion of 16S small-subunit ribosomal genes positioned 341–907 (V3-V5) was amplified using the primer set described in Muyzer et al. (1993). Furthermore, the PCR reaction was set such that initial denaturation occurred at 95 °C for 30 s. Thereafter, the PCR mixture was subjected to 35 cycles of 30 s at 95 °C, 30 s at 52 °C and 60 s at 72 °C, followed by a final extension step of 180 s at 72 °C. All PCR amplifications were confirmed using electrophoresis.

PCR products were purified using ZR DNA Sequencing Clean-up Kit (Zymo Research, USA) and as described by Li et al. (2010). The partial 16S rRNA gene was sequenced and analysed as described in Jordaan & Bezuidenhout (2013). A total of 63 Enterococcus spp. nucleotide sequences were submitted to the GenBank database under accession numbers KT598397-KT598459.

Five oligonucleotide primer pairs (Applied Biosystems, UK) used for the detection of virulence genes were obtained from Vankerckhoven et al. (2004). Identification of virulence genes for each isolate was performed by separate PCR for the virulence genes asa1, hyl and gelE while a duplex reaction was used for cylA and esp. PCR amplifications were performed in a Techne Prime Elite thermocycler (Cambridge, UK), in 0.2 ml reaction tubes. PCR assays were performed in a total volume of 25 μl containing 1 μl bacterial DNA template (50–100 ng/μl) for the single assays and 2 μl bacterial DNA template (100–150 ng/μl) for the duplex reactions. Additionally, the reaction contained RNase/DNase free water (Fermentas Life Sciences, USA), 2× DreamTaq PCR Master Mix (0.05 U/μL Taq DNA polymerase in reaction buffer, 0.4 mM of each dNTP and 4 mM MgCl₂), 0.2 μM of primers asa1 and gelE and 0.4 μM of primers cylA, esp and hyl. The PCR cycling conditions consisted of denaturing at 95 °C for 180 s, followed by 30 cycles of 95 °C for 30 s, annealing at 56 °C for 30 s and 72 °C for 60 s. This was followed by a final step of 72 °C for 600 s. PCR products were confirmed by gel electrophoresis.

Antimicrobial susceptibility patterns of Enterococcus isolates were determined using the disk diffusion method (Bauer et al. 1966). Assays were performed on Mueller Hinton agar (Merck, Germany) using ampicillin (10 μg), amoxicillin (10 μg), penicillin G (10 μg), neomycin (30 μg), streptomycin (300 μg), vancomycin (30 μg), chloramphenicol (30 μg), ciprofloxacin (5 μg), tetracycline (30 μg), trimethoprim (2.5 μg) and erythromycin (15 μg). All antibiotics were obtained from Mast Diagnostics (UK). Enterococcus isolates were classified resistant, susceptible or intermediate according to the criteria from the Clinical and Laboratory Standards Institute (CLSI 2017).

Pure Enterococcus isolates showing resistance to two or more antibiotics were grown on nutrient agar, cultured overnight at 37°C in 20 ml Brain Heart Infusion broth (BHI, Merck, Germany) and harvested by centrifugation. A plasmid isolation kit was used to extract total plasmids according to the manufacturer's instructions (Macherey-Nagel, Germany). All plasmid products were electrophoresed for 1 h at 60 V and agarose gel images were captured under a UV light using a Bio-Rad ChemiDoc imaging system (Hercules, CA). Enterococcus isolates resistant to two or more antibiotics were subjected to plasmid curing according to Molina-Aja et al. (2002) and Carvalho et al. (2014). The isolates were once again subjected to antibiotic susceptibility testing and resistance classified as plasmid dependent when affected by curing.

The inhibition zone diameter data of Enterococcus spp. were subject to cluster analysis using Ward's method and Euclidean distances in Statistica 12.0 (StatSoft, USA).

In Table 2, the triplicate average values of presumptive Enterococcus spp. And ±standard deviation values are presented. The average presumptive Enterococcus levels ranged from 52.00 to >315.00 cfu/100 ml. The highest levels were observed at plant 2 final effluent (>315.00 ± 0.00 cfu/100 ml), whereas the lowest values were observed at plant 3 point downstream (52.00 ± 1.73 cfu/100 ml).

The most frequently detected Enterococcus spp. were E. hirae (21%), E. faecalis (21%) and E. faecium (19%). The remaining detected species were E. gallinarum (13%), E. casseliflavus (16%), E. mundtii (8%) and E. durans (3%). All Enterococcus species were identified to a sequence similarity threshold of >97%.

The antibiotic inhibition zone diameter data from the antimicrobial susceptibility test performed on all 63 Enterococcus spp. were subjected to cluster analysis which is presented in Figure 1. Two main clusters, A and B, were generated and cluster A sub-divided into minor clusters A1 and A2. Furthermore, the species composition of both clusters was determined and is presented in Table 3. Differences in species diversity and prevalence levels at the various sites were observed. Furthermore, as depicted in Figure 1, the cluster analysis indicated grouping of the antibiotic inhibition zone results of Enterococcus spp. isolated at the final effluent and points downstream, respectively. Cluster A1 comprised Enterococcus spp. from all sampled effluent. This cluster was predominated by E. faecalis isolates followed by E. faecium and E. hirae isolates (Table 3). Furthermore, as depicted in Figure 1, cluster A2 comprised Enterococcus spp. isolated from the points downstream the effluents.

The predominant species in this cluster (A2) were Enterococcus gallinarum, E. casseliflavus and E. mundtii (Table 3). Lastly, cluster B represented Enterococcus isolates from the point located between plants 1 and 2. In this cluster, the predominant species, as presented in Table 3, were E. faecium and E. hirae.

Presented in Table 4 are the characteristics of Enterococcus spp. isolated from three WWTPs' final effluent and points downstream with reference to the presence of virulence genes, as well as the MAR phenotypes prior to and post-plasmid curing.

Figure 2 is an agarose gel with representative isolates confirming the presence of virulence genes in some Enterococcus spp. screened in this study. All five virulence genes screened were detected and Enterococcus isolates harbouring virulence genes are presented in Table 4. As presented in Table 4, the most frequently detected virulence determinants were esp (25%, 10/39), hyl (25%, 10/39), gelE (23%, 9/39) and cylA (17%, 7/39). Nonetheless, asa1 (7%, 3/39) was also detected though at lower prevalence levels. Enterococcus faecium spp. predominantly harboured the gelE (55%, 5/9) and hyl (3%, 3/10) genes (Table 4). The esp gene was predominantly harboured by E. faecalis, E. casseliflavus and E. hirae spp. at prevalence levels of 3% (3/10) each. Furthermore, virulence gene cylA was predominantly detected in E. faecalis and E. gallinarum spp. at prevalence levels of 29% (2/7) each (Table 4). In addition, six multi-virulence profiles were observed with cylA–esp being the most prominent.

Analysis of the antimicrobial susceptibility of the 63 Enterococcus isolates revealed that resistance to ampicillin (67%), vancomycin (62%), tetracycline (58%), penicillin (52%) and erythromycin (51%) was most frequent. Furthermore, resistance to chloramphenicol (22%) and streptomycin (17%) was also observed, although at slightly lower levels. Additionally, low frequencies of resistance to ciprofloxacin (8%) and gentamycin (6%) were detected.

Sixty-eight percent of the screened Enterococcus spp. were resistant to three or more antibiotics (Table 4), while resistance to multiple antibiotics was more common among the E. faecium isolates followed by E. faecalis and E. casseliflavus spp. As depicted in Table 4, a total of 19 multiple antibiotic resistance (MAR) phenotypes were identified with the most frequently identified MAR phenotype being ERY–STREP–PEN–TET. The latter MAR phenotype was observed in nine Enterococcus isolates. In addition, AMP–PEN–TET–VA was the second-most prevalent MAR phenotype and was identified in six Enterococcus isolates screened. Similarities in species composition of these MAR phenotypes were observed as they were composed of E. faecalis, E. faecium, E. casseliflavus and E. gallinarum spp. (Table 4).

Seventy-six percent (35/46) of the Enterococcus isolates resistant to two or more antibiotics had plasmids. The plasmids were of high-molecular weight. Furthermore, in Table 4, MAR phenotypes of the screened Enterococcus spp. pre- and post-plasmid curing are presented. Differences in MAR phenotypes were observed for all profiles pre- and post-plasmid curing. Thirty-three percent (10/30) of the Enterococcus spp. resistant to vancomycin were still resistant to this antibiotic after plasmid curing, while resistance to penicillin, ampicillin, erythromycin and chloramphenicol was observed in few isolates post-plasmid curing (Table 4).

Although antibiotic resistance profiles have often been detected in recreational and wastewaters (Soge et al. 2009; Araújo et al. 2010), the current knowledge on the prevalence of antibiotic-resistant bacteria and types of antibiotic resistance in the environment is insufficient (Rizzo et al. 2013). Urban WWTPs provide environmental waters, used for various agricultural, religious, cultural and recreational activities and effluent for reuse. However, ineffectively functioning plants are among the main sources of antibiotics released into the environment. This contamination of environmental surface water systems with antibiotics consequently facilitates the selection of antibiotic-resistant genes and antibiotic-resistant bacteria (Martinez 2009) which is a concern when the health and well-being of both humans and animals are considered.

In this study, 63 Enterococcus spp. isolated from the final effluent of three municipal WWTPs and receiving environmental water systems were screened for antimicrobial resistance and clinically relevant virulence genes. Differences in species composition and diversity at varying prevalence levels were observed for the various sites. Graves & Weaver (2010) explain that the observed differences are a result of influences exerted by environmental factors, as well as the sources of a sewer system. Carvalho et al. (2014) reported higher species diversity in raw sewage compared to sites more distant from the raw sewage. The latter findings support the findings of this study where species diversity was higher at the final effluent sites as compared to the site between plants 1 and 2. Furthermore, similar to the findings of other studies, E. hirae, E. faecalis and E. faecium were the most frequently identified Enterococcus spp. (da Silva et al. 2006; Łuczkiewicz et al. 2010).

The high prevalence of E. hirae isolates identified in this study correlates with the findings of Bonilla et al. (2006) and da Silva et al. (2006), who reported high E. hirae levels in recreational and urban wastewater. In addition, studies have also reported the dominance of E. hirae species in animal faeces (Fisher & Phillips 2009; Peeters et al. 2013). Thus, the high levels of E. hirae spp. identified in this study suggest faecal contamination. The presence of E. hirae spp. in the WWTPs' final effluent and points downstream used for recreational, agricultural and cultural activities is a cause for concern since several studies have reported the ability of E. hirae to cause wound infections, gastritis and bacteraemia (Talarmin et al. 2011; Chan et al. 2012).

Enterococcus faecalis and E. faecium were detected in the WWTPs' final effluent. This is in accordance with the findings of several other studies (Cheng et al. 2012; Varela et al. 2013; Iweriebor et al. 2015). Finding E. faecalis and E. faecium at high levels in all three sampled effluents suggests poorly disinfected sewage effluent as these two species are the most commonly found Enterococcus spp. in the human gastrointestinal tract (Manero et al. 2002; Klein 2003). Cheng et al. (2012) suggested that frequently detecting E. faecalis and E. faecium in treated effluent may imply that these species are naturally present at higher concentrations in the influent or may be more resistant to treatment processes, particularly disinfection. Furthermore, their presence in effluent and points downstream may suggest the dissemination of these species from the WWTP into the environment. Their presence in the effluent is also a cause for concern seeing that they account for a majority of infections caused by Enterococcus spp. (Vu & Carvalho 2011).

Enterococcus gallinarum, E. casseliflavus and E. mundtii were also isolated in this study and were the most prevalent isolates in surface water bodies receiving the final effluent from the WWTPs. The presence of these species in the downstream sites is not unexpected due to the well-documented presence of these Enterococcus species in plants, soil and water non-human animal hosts (Byappanahalli et al. 2012; Ran et al. 2013). However, their presence in downstream sites of receiving water bodies that are used for various social activities is a cause for concern as E. gallinarum, E. casseliflavus and E. mundtii have been listed among Enterococcus spp. known to occasionally cause human infection (Perio et al. 2006). Thus, implications for the immuno-compromised could have far-reaching effects.

Enterococcus isolates screened in this study displayed the presence of multi-virulence profiles. This is in accordance with the findings of Poeta et al. (2005) and Lata et al. (2009). Furthermore, the findings of Lata et al. (2009) showed that Enterococcus isolates exhibiting two to three virulence-markers per isolate were mostly associated with polluted sites.

The esp gene was the predominant virulence gene identified and was harboured by E. faecalis, E. casseliflavus, E. hirae and E. faecium isolates. Enterococcal surface protein (esp) is considered a pathogenicity marker in E. faecium and E. faecalis clinical isolates (Said et al. 2015). Therefore, the predominance of esp in Enterococcus isolates obtained from the final effluent and receiving water bodies is not strange as the WWTPs receive wastewater from the hospitals. Furthermore, the presence of esp in Enterococcus spp. other than E. faecalis and E. faecium has not been extensively reported. However, it should be noted that its presence in other Enterococcus spp. may possibly be a result of genetic exchange by conjugation (Oancea et al. 2004).

The cylA, gelE and hyl virulence genes were also detected in the various Enterococcus spp. This observation is supported by various studies, conducted in most parts of the world, that reported the presence of these virulence genes in various Enterococcus spp. isolated from different sources (Semedo et al. 2003; Gonçalves et al. 2011; Sidhu et al. 2014). However, a recent study in South Africa reported the absence of virulence genes cylA, esp and hyl in Enterococcus spp. isolated from WWTP in the Eastern Cape (Iweriebor et al. 2015). However, a similar study reported on the presence of asa and esp in Enterococcus spp. isolated from WWTP in Gauteng. Thus, in the present study, we report on virulence patterns previously observed. However, here we also report on some virulence patterns not previously reported, particularly for South African Enterococcus spp.

High levels of multiple antibiotic-resistant Enterococcus spp. were isolated in this study. This is in accordance with the findings of several other studies that also isolated large numbers of Enterococcus from water sources that were resistant to multiple antibiotics (Da Costa et al. 2006; Carvalho et al. 2014). Furthermore, the observed resistance to vancomycin, tetracycline, penicillin and erythromycin is in agreement with the findings of other studies (Moore et al. 2010; Carvalho et al. 2014; Iweriebor et al. 2015). On the other hand, resistance to ampicillin, chloramphenicol and streptomycin was also observed in the present study. Resistance to the latter antibiotics may not be surprising as these antibiotics are frequently prescribed at South African Hospitals (NDoH 2012).

Similar to the findings of Xu et al. (2007) and Carvalho et al. (2014), 21 distinct MAR profiles were observed. The presence of Enterococcus spp. in wastewater and receiving water systems that are resistant to multiple antibiotics is a cause for concern as these species are associated with both community and hospital-acquired infections. If these antibiotics are commonly used for the treatment of infections (Xu et al. 2007), then it could be a further health burden as treatment failure could arise.

However, 67% of the Enterococcus spp. with MAR was susceptible to the same antibiotics after plasmid curing, implying that the resistance is plasmid mediated. Similar results were observed by Amaechi (2015) and Carvalho et al. (2014). According to McBride et al. (2007), the presence of plasmids is common in the Enterococcus genus, as these mobile genetic elements constitute a substantial fraction of their genome and are responsible for horizontal gene transfer. Therefore, Enterococcus spp. may acquire antimicrobial-resistant genes through mutation or horizontal transfer of mobile genetic elements, such as plasmids and transposons (Hasmann et al. 2005). Furthermore, environmental water systems polluted by sewage may act as vehicles of bacteria carrying and disseminating antibiotic-resistant plasmids (Carvalho et al. 2014).

Although only one plasmid with a high-molecular weight was observed in all Enterococcus spp. screened, studies by Szczepanowski et al. (2004) and Schlüter et al. (2003) demonstrated that antibiotic resistance genes were commonly contained in high-molecular weight plasmids. Therefore, the presence of Enterococcus spp. harbouring plasmids in final effluent flowing into receiving water systems implies that these aquatic systems could potentially be acting as reservoirs of antibiotic-resistant Enterococcus harbouring high-molecular weight plasmids.

In this study, the antimicrobial resistance patterns and presence of virulence genes in Enterococcus spp. isolated from three municipal WWTPs' final effluent, as well as receiving environmental water systems in the North West Province, South Africa, were determined. Plant 2 had the highest levels of Enterococcus spp. while the point downstream Plant 3 had the lowest levels. Cluster analysis grouped Enterococcus spp. according to isolates detected in all three plants' effluent, the point between plants 1 and 2 and the points downstream, respectively. The predominant species identified in the sampled effluent sites were Enterococcus faecalis, E. faecium and E. hirae. The point between plants 1 and 2 was dominated by Enterococcus gallinarum, E. casseliflavus and E. mundtii, while the predominant species in the downstream sites were Enterococcus faecium and E. hirae. All five virulence factors screened were identified in this study. Over 60% of the screened Enterococcus spp. was resistant to three or more antibiotics with the most prevalent MAR phenotype being ERY–STREP–PEN–TET. Additionally, differences were observed in MAR phenotypes pre- and post-plasmid curing. The findings of this study have demonstrated that Enterococcus spp. harbouring plasmids that mediate MAR are not effectively removed during treatment consequently entering environmental water systems that support a variety of social needs. Furthermore, these Enterococcus spp. harbour virulence factors, which play a significant role in the rupture of a variety of target membranes, colonization and biofilm production, as well as the catabolism of various target molecules, such as gelatine, collagen, fibrinogen, casein and insulin. Therefore, the data presented here serve as an alert for public health authorities. It is recommended that interventions, such as advanced treatment technologies and disinfection processes, be implemented in WWTPs while environmental water systems used by humans and animals be tested regularly for Enterococcus levels.

The authors wish to thank Dr C Mienie for assistance with sequencing, Ms Daleen Du Plessis for the maps, the National Research Foundation Deutscher Akademischer Austauchdienst (NRF-DAAD) (SFH1208219435) for a bursary to LG Molale-Tom and Water Research Commission of South Africa (K5/1966 & K5/2347) for financial support. The views expressed are those of the authors and not of the funding agencies.

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