4. Discussion
The relative abundance of free-living
Listeria species found during this study and across all sampled sites is consistent with reports elsewhere [
13,
35]. There are no recommended standards specific for
Listeria pathogens in water and wastewater samples in South Africa for obvious reasons; thus the fecal coliforms standard (0 cfu/100 ml) for domestic water uses [
30] was applied in this report. Based on this standard, the water quality across the studied water system and throughout the year (
Table 2) fell short of acceptable target limits for domestic applications, thus disqualifying the waters for use in drinking and other domestic purposes.
Listeria abundance did not vary significantly with season, either as free-living or plankton-associated species, consistent with the observation of Murrel
et al. [
36], but contrary to our previous report [
13]. The significant positive correlation observed between
Listeria species attached to large (180 μm) planktons and those attached to small (20 μm) planktons suggests that the two groups of
Listeria species may occupy the same niche in the ecosystem; this is contrary to our previous report [
13], where
Listeria species attached to large (180 μm) planktons negatively correlated with those attached to small (20 μm) planktons. The lack of significant correlations between and among other treatments in this study suggests that free-living
Listeria species and
Listeria species attached to medium-sized (60 μm) planktons occupy separate niches in the water system, different from those occupied by
Listeria species attached to large (180 μm) and small (20 μm) planktons. The observation is consistent with those of Maugeri
et al. [
24] who reported lack of significant correlation between free-living bacteria and plankton associated bacterial populations in a marine coastal zone in Italy. However, another study [
37] reported a negative correlation between planktonic
Vibrio cells and sessile populations.
Listeria species were isolated from all sampled sites and throughout the year during this study, suggesting a 100% prevalence of the pathogen in the water system. Consistent with observations in a previous study [
13], free-living
Listeria species were most prevalent (84%) both in treated effluent and the receiving watershed; followed by
Listeria cells associated with planktons of sizes 180 μm (75%), 20 μm (68%), and 60 μm (59%), respectively. Corroborating this observation, Maugeri
et al. [
24] reported higher prevalence for free-living bacteria compared to their plankton-associated counterparts.
Listeria species were generally more prevalent in the treated effluents (FE), both as free-living and/or plankton-associated cells, compared to the receiving watershed (
Table 2). The observation could be as a result of higher nutrient levels in the wastewater effluents compared to the receiving watershed, in agreement with previous reports [
10,
11,
38]. Consistent with the observation of this study, high prevalence of
Listeria species has been reported in water systems impacted by wastewater effluents in Iraq [
8,
9], Poland [
10] France [
11], the United Kingdom [
12] and rural South Africa [
13]. Watkins and Sleath [
12] reported 100% prevalence of
Listeria species in sewage, river water, and trade effluent at densities (7.0 × 10
2 to >1.8 × 10
4 Most Probable Number (MPN)/mL), slightly higher than those observed in this study. The sewage effluent reported by Watkins and colleague, however, only underwent primary treatment unlike ours that was disinfected by chlorination, which could account for the differences. Al-Ghazali and Al-Azawi [
8,
9] also reported 100% prevalence in treated wastewater effluent in Iraq but at lower densities of <3 to 28 MPN/mL, and Paillard
et al. [
11] reported 84.4% prevalence of
Listeria species in treated wastewater in France at densities ranging from <0.3 to 21 MPN/ml, while Odjadjare and Okoh [
13] recorded 100% prevalence in a rural water system in South Africa at densities ranging from 1.0 × 10
1 to 1.1 × 10
4 cfu/mL. Contrary to the observation of this study, lower prevalence has been reported for
Listeria species in a variety of surface water systems. Frances
et al. [
39] reported the isolation of
Listeria species from 21% of freshwater samples collected from sites in Cheshire and North Wales; while Lyautey
et al. [
40] reported 64% for surface waters of the South Nation River Watershed in Ontario, Canada. These observations were consistent with expectations for surface waters that were not impacted by wastewater effluent in agreement with a report elsewhere [
38].
The significant variation observed between raw and treated sewage for most physicochemical parameters (
Table 3) is an indication that the wastewater treatment process remarkably improved the quality of the raw wastewater influent. However, despite the improvement on raw sewage quality, the treated effluent did not measure up to the desired target quality for turbidity, DO, COD, and NO
2 with respect to domestic applications [
30] and PO
4 with reference to preserving the integrity of the aquatic ecosystem [
34]. This suggests that the wastewater effluent has a potential negative impact on the environment and public health. The effluent quality was, however, acceptable in terms of pH, temperature, TDS, and NO
3 (
Table 3).
The chlorine residual (
Figure 2) generally fell within acceptable target limits (0.3–0.6 mg/L) for domestic water at the point of use [
41], except in September and November 2007, and indicates that the water is safe for domestic applications with reference to chlorine residual. The scatter plot (
Figure 3) indicates that the relationship between chlorine residual and listerial density did not follow any particular trend. This observation suggests that factors other than chlorine disinfection affected the abundance of
Listeria species during this study; some of these factors may also be responsible for the inability of chlorine to adequately eliminate the pathogens from the wastewater even at relatively high doses. LeChevallier
et al. [
42] observed attachment of bacteria to planktons and/or other suspended particles as a factor which enhanced resistance of bacteria to chlorine disinfection while Obi
et al. [
41] reported other factors to include contact time, temperature, and pH. This suggests that turbidity (which is a measure of suspended particles including planktons) could be a factor in the ineffectiveness of chlorine disinfection during this study; turbidity fell short of recommended target limits throughout the study (
Table 3). Attachment of
Listeria species to plankton may, however, not be a significant factor in the bacterial survival of chlorine disinfection in this study, as free-living
Listeria species were more abundant compared to their plankton attached counterparts even after chlorine disinfection in agreement with the observation of our study elsewhere [
13]. The reason for this observation is not clear.
Previous studies on the antimicrobial susceptibility profiles of
Listeria species focused mainly on clinical and/or food isolates with little information in the literature on antibiotic susceptibility profiles for
Listeria strains isolated from chlorinated municipal wastewater effluent. All 23
Listeria species tested in this study were sensitive to three (15%) of the 20 test antibiotics including amikacin (aminoglycosides), meropenem, and ertapenem (carbapenems) (
Table 4); suggesting that these antibiotics may be the best therapy in the event of listeriosis outbreak in South Africa. Consistent with the observation of this study, Hansen
et al. [
43] reported complete sensitivity of 106
Listeria species isolated from humans to meropenem, while Safdar and Armstrong [
44] observed 100% sensitivity to amikacin and Odjadjare and Okoh [
13] reported complete sensitivity to the three antibiotics by all 14
Listeria species isolated from chlorinated wastewater effluent in a previous study.
Listeria strains in this study showed resistance to at least one of 17 antibiotics at percentages ranging from 4.5%–91% (
Table 4), and particularly high levels for penicillin G (91%), ampicillin (87%), erythromycin (83%), and sulphamethoxazole (65%). Contrary to the observation of this study,
Listeria species were generally reported to be susceptible to penicillin G [
45], ampicillin [
46], erythromycin [
28,
44], and sulphamethoxazole [
13,
43]. Conversely, considerable resistance has been reported in the literature for
Listeria species against the penicillins (penicillin G and ampicillin) [
21], erythromycin [
13,
47], and sulphamethoxazole [
46]. The high resistance observed for penicillin G, ampicillin and sulphamethoxazole could be of serious public health concern as penicillin G and ampicillin are reported to be the antibiotics of choice in the treatment of listeriosis [
28,
43]; while sulphamethoxazole, usually in combination with trimethoprim, is considered second choice therapy, especially for patients who are allergic to the penicillins [
46]. The observation generally indicated that municipal wastewater effluent could be a significant source of highly resistant strains of
Listeria pathogens in the South African aquatic milieu.
The physicochemical quality of the wastewater effluent may be a factor in the level of resistance observed in this study, as it is widely reported [
48–
50] that conventional wastewater treatment plants lack the capacity to effectively remove antibiotics and a number of other chemicals from wastewater, thereby increasing the chances of bacterial pathogens resident in such wastewater effluent to develop resistance to common antibiotics due to selective pressure. Although we did not attempt to assay for residual antibiotics in the treated effluents in the course of this study, lack of capacity to remove some chemicals from the wastewater during the treatment process is evident in
Table 3. The table shows that the treated effluent fell short of recommended standard quality for critical parameters such as turbidity, DO, COD, NO
2, and PO
4 and suggests a possible influence on the listerial resistance.
Twenty-two (95.7%) of the 23 test isolates in this study showed multiple antibiotic resistance in combinations ranging from four to 10 antibiotics (
Table 5). Similar observations have been reported elsewhere [
13,
21]. On the contrary Conter
et al. [
28] reported more resistance to single antibiotics than multiple resistance amongst 120
Listeria isolates tested against 19 antibiotics; while Arslan and Ozdemir [
51] reported resistance to single antibiotics with no record of multiple antibiotic resistance amongst 47 strains of
Listeria species isolated from white cheese and tested against 13 antibiotics. Multiple drug resistance in
Listeria species have been attributed to antimicrobial selective pressure and gene transfer mechanism between and among
Listeria species and close relatives of the bacteria such as
Enterococcus,
Streptococcus and
Staphylococcus species [
44]. Donlan and Costerton [
52] also reported the acquisition of inherent resistance to antimicrobial agents due to bacterial attachment to surfaces; suggesting that attachment to plankton at one point or the other may have enhanced the multiple resistances of our listerial strains to several test antibiotics.
Although the penicillins (penicillin G and ampicillin) and erythromycin showed the highest phenotypic resistance during this study, the genes responsible for resistance to these antibiotics were not detected in our
Listeria isolates (
Table 6). In a similar report, Srinivasan
et al. [
21] observed high level (92%) of phenotypic resistance to ampicillin but failed to detect the genes responsible for ampicillin resistance in all of their 38
Listeria isolates. Consistent with the observation of this study, Davis and Jackson [
20] could not detect
penA genes (responsible for penicillin resistance) in
Listeria strains isolated from various sources; while Srinivasan
et al. [
21] reported their inability to detect genes responsible for erythromycin resistance in 38
Listeria isolates from dairy farms in spite of observed phenotypic resistance to the antibiotic. Contrary to the observation of this study, Srinivasan
et al. [
21] reported the detection of
penA genes in 37% of their
Listeria isolates while Roberts
et al. [
53] reported the detection of erythromycin resistance genes in
Listeria species isolated from food samples. To the best of our knowledge, this is the first report on the detection of dihydropteroate synthetase type II (
sulII) resistance gene markers in
Listeria species (
Table 6). Previous attempt by other workers [
20,
21] did not detect the genes in
Listeria species. The percentage of
Listeria isolates that harbored this gene was, however, relatively small (22%) compared to the high (65%) level of phenotypic resistance observed for the antibiotic (sulphametoxazole) in this study. The observations generally suggests that the presence of antimicrobial resistance genes in bacterial isolates do not always correlate with phenotypic antibiotic resistance, and indicates that other mechanisms such as decreased outer membrane permeability, activation of efflux pump, or mutation in a ribosomal protein may have contributed to the antimicrobial resistance phenotypes observed in this study [
21].