Survival of emerging pathogen Acinetobacter baumannii in water environment exposed to different oxygen conditions

Acinetobacter baumannii is a Gram-negative coccobacillus, non-motile, non-spore forming, non-fermentative, strict aerobe (Towner 2009). Although not an obligate pathogen, A. baumannii is nowadays a leading cause of nosocomial infections and hospital outbreaks worldwide. A. baumannii, primary sensitive to antibiotics, during the last 30 years has developed resistance to commonly used antimicrobial agents (Roca et al. 2012). Today, clinical isolates resistant to last-resort antibiotics such as carbapenems and colistin are being reported worldwide (Towner 2009; Roca et al. 2012; Gottig et al. 2014). Due to the expression of the virulence factors, A. baumannii can survive for prolonged periods on abiotic surfaces in the hospital environment (Espinal et al. 2012).

Over the last decade, the occurrence of A. baumannii in natural environments outside of hospital settings has been reported. Viable A. baumannii of clinical significance were recovered from untreated hospital wastewaters (Ferreira et al. 2011; Zhang et al. 2013; Kovacic et al. 2017) urban wastewaters and wastewater treatment plants (Hrenovic et al. 2016; Higgins et al. 2018), as well as from rivers (Girlich et al. 2010; Seruga Music et al. 2017).

However, very little is known about the environmental factors that influence the survival of A. baumannii in waters from which it was recovered. A recent study by Dekic et al. (2018) demonstrated the survival of A. baumannii under different temperature (−20 to 44 °C) and pH (5–10) regimes as well as nutrient availability. Optimal conditions for the long-term survival of A. baumannii were temperatures of 4 and 22 °C regardless of nutrient availability (Dekic et al. 2018). Additionally, Hrenovic et al. (2016) investigated influent and effluent at the wastewater treatment plant from which A. baumannii isolates were recovered. From these data, it could be concluded that A. baumannii existed in waters with different nutrient availability at almost no dissolved oxygen (0.09 mg O₂/L, 18.6 °C) in influent to the very high oxygen concentration in aerated effluent. The multiplication and survival of A. baumannii in the aerated system of autoclaved effluent wastewater was also reported (Hrenovic et al. 2016). The composition of effluent wastewater varies from time to time and from place to place, which hinders the deeper insight into the survival of this pathogen in the water environment. Recently, the survival of A. baumannii through the technological process of anaerobic mesophilic sludge digestion at the wastewater treatment plant was reported (Higgins et al. 2018). These isolates were able to survive on nutrient agar in laboratory-controlled anaerobic conditions during 30 days, after which they grew normally in aerobic conditions. However, no growth of colonies on the inoculated nutrient agar was detected after direct exposure to anaerobic conditions. This finding by Higgins et al. (2018) indicated the anaerobic environment as an ecological niche important in the epidemiology of clinically relevant A. baumannii. However, the anaerobic conditions in digested sludge and nutrient agar are not much applicable to the conditions in the natural environment. The prediction of the behaviour of this emerging pathogen in waters still could not be established due to the lack of data.

Therefore, the aim of this study was to examine the survival of environmental as well as clinical isolates of A. baumannii in different water media exposed to different ranges of oxygen supply.

Five isolates of A. baumannii deposited at the University of Zagreb Faculty of Science were chosen for the experiments (Table 1). Four environmental isolates were recovered from Zagreb wastewater treatment plant (WWTP), one from influent (IN39) and three from effluent (EF7, EF8, EF11) wastewater (Goic-Barisic et al. 2017; Dekic et al. 2018; Higgins et al. 2018). One clinical isolate (OB4138) was recovered from a patient suffering from hospital-acquired pneumonia at the Special Hospital for Pulmonary Diseases in Zagreb, Croatia (Seruga Music et al. 2017; Dekic et al. 2018). According to the antibiotic susceptibility profile (Magiorakos et al. 2012), isolates were grouped into several categories: susceptible to all tested antibiotics (S) – EF11; multidrug-resistant (MDR) – IN39; extensively drug-resistant (XDR) – EF8 and OB4138; pandrug-resistant (PDR) – EF7. MDR isolates are resistant to at least one antibiotic in three or more antibiotic categories, XDR are resistant to at least one antibiotic in all but one or two antibiotic categories, while PDR isolates are resistant to all antibiotics in all tested categories.

Behaviour of five isolates of A. baumannii was tested in two different water media exposed to three types of different oxygen supply. The nutrient-depleted commercially available spring water (SW) was used as well as nutrient broth (Biolife) diluted with distilled water to 1:100 (DNB). In comparison to SW, DNB can be considered nutrient-rich, while it has fewer nutrients than regular nutrient broth. Chemical properties of the tested water media are given in Table 2. A full 10 μL loop of overnight culture of each A. baumannii isolate was suspended in a tube containing 10 mL of autoclaved SW or DNB. One mL of suspension was inoculated in duplicate in the 50 mL tubes containing 30 mL of autoclaved SW or DNB. Aerated conditions were achieved by forced aeration of water media with sterile air filtered through the membrane filters and serological pipette. Intermediate oxygen supply was achieved in the stationary loosely capped tubes. Anaerobic conditions were maintained by placing the loosely capped tubes in the commercial Anaerocult A system (Merck Millipore). Prepared inoculated tubes were incubated at 22 ± 2 °C during 50 days of monitoring.

Measurement of oxygen concentrations was done with a WTW Oxi 330i electrode, in duplicate series of tubes to prevent cross contamination. In aerated and intermediate conditions, the dissolved oxygen concentration was constant during 50 days of monitoring: 8.4 mg/L (oxygen saturation of 96%) and 4.6 mg/L (oxygen saturation of 56%), respectively. To generate anaerobic conditions, sachets made of absorptive and highly gas-permeable paper filled with an oxygen-binding mixture were put inside an Anaerocult jar. The addition of water started the reaction, which consisted of oxidation of iron inside the sachets and release of CO₂. An oxygen strip indicator in the Anaerocult A system indicated no dissolved oxygen 3 h after the closing of the jar.

In another series of tubes, the number of viable bacteria was monitored at the beginning of the experiment, after 1, 2 and further every 7 days of incubation. After a specified period of time, tubes were shaken, sub-samples were diluted in sterile saline solution, inoculated in technical triplicate onto Mueller-Hinton agar (Biolife) plates, and bacterial colonies were counted after incubation at 42 °C/24 h. The number of viable bacteria was determined as colony forming units (CFU), logarithmically transformed, and expressed as log CFU per one mL of water medium.

In order to check the purity of bacterial cultures during 50 days of monitoring, Gram staining was routinely preformed each time when the number of CFU was determined in all systems. The original subsamples at the start and at the end of aerobic and anaerobic experiments with DNB were fixed in 2.5% glutaraldehyde in phosphate buffered saline, and prepared for scanning (SEM) and transmission (TEM) electron microscopy using standard techniques. Briefly, after removal of the fixative, the samples were rinsed in phosphate buffer, post-fixed in 1% osmium tetroxide in phosphate buffer, again rinsed in phosphate buffer and then dehydrated in a graded ethanol series up to absolute ethanol. Cells for SEM were dried with hexamethyldisilazane and carbon coated before examination at low voltage (0.5 kV) with a Zeiss Ultra PLUS FEG SEM. The samples for TEM were infiltrated with an epoxy resin and ultrathin sections were examined at 200 kV with a Jeol 2100F TEM.

Statistical analyses were carried out using Statistica 13.3 software (TIBCO Software Inc.). For pairwise comparisons, factorial analysis of variance (ANOVA) and Duncan post hoc test were used. For making correlations, Spearman's rank correlation coefficient was used. Decisions regarding statistical significance were made at p < 0.05. For graphical presentation, the percentage of survival was calculated as follows: ((logCFU/mL_time:logCFU/mL_start)*100), where log CFU/mL_time represents the number of bacteria on a day of measurement and log CFU/mL_start the initial number of bacteria. In some cases the survival percent exceeds 100%, which signifies bacterial growth.

All five tested isolates of A. baumannii successfully survived the aerated, intermediate and anaerobic conditions in both SW and DNB during 50 days of monitoring (Figures 1–3). In nutrient-depleted SW there was no evident multiplication of bacteria. The multiplication of all isolates occurred in nutrient-rich DNB during the first two days of contact in aerated and intermediate conditions, and to a lesser extent during the first day in anaerobic conditions.

In SW, there was no statistically significant difference between the survival of isolates in aerated, intermediate and anaerobic conditions (Figures 1–3). The survival of four isolates (IN39, EF7, EF8, OB4138) in SW after 50 days of contact ranged from 84–105%. The isolate EF11 sensitive to all tested antibiotics had the lowest survival in SW in all tested oxygen conditions as compared to other isolates. After 50 days of contact in aerobic, intermediate and anaerobic conditions, the survival of isolate EF11 in SW averaged 61, 58 and 64%, respectively.

In DNB, there was no difference between the survival of EF11 and other isolates. The survival of all isolates in DNB after 50 days of contact ranged from 73 to 117% in different oxygen conditions. The highest survival was observed in DNB in aerated and intermediate conditions (76–117%) with no statistically significant difference between these conditions (p = 0.075). However, the survival in anaerobic DNB (73–93%) was statistically lower than in both aerated (p = 0.007) and intermediate (p = 0.000) conditions. The survival in DNB was statistically positively correlated with dissolved oxygen concentration (R = 0.448, p = 0.013), whereas the survival in SW showed no statistically significant correlation. In the anaerobic DNB, after the sharp decrease in CFUs during the first nine days, A. baumannii behaved erratically, having irregular survival curve with sudden decreases and increases in concentration of viable bacteria (Figure 3).

Furthermore, A. baumannii from anaerobic conditions formed smaller translucent colonies together with normal opaque ones (Figure 4). Small colony variants irregularly but constantly appeared on the plates of Mueller-Hinton agar inoculated with all isolates either from SW and DNB after only one day of incubation in anaerobic conditions. The identity of translucent colonies as A. baumannii was confirmed by matrix-assisted laser desorption ionisation-time of flight mass spectrometry – MALDI-TOF MS based on the analysis of ribosomal proteins. No translucent colonies were observed in the aerated or intermediate experiments.

Due to the unusual survival curve, cells from the anaerobic DNB experiment were further analysed and compared to cells from aerobic conditions in DNB. After 14 days and further every 7 days, the subsamples were Gram stained and checked under the optical microscope. Among the normal red-stained A. baumannii cells of coccobacillus shape, the unusual bloated poorly stained cells were observed (data not shown). The SEM analysis of A. baumannii after the exposure to anaerobic DNB conditions showed no surface damage of cells as compared to those at the start of the experiment, but cells were stacked together with the extracellular substances (Figure 5). The TEM showed normal structure of A. baumannii cells after the exposure to aerobic conditions in DNB (Figure 6(a)). After the exposure to anaerobic conditions in DNB, the majority of A. baumannii cells showed damage with signs of lysis, with only spheroplasts remaining in some cases (Figure 6(b)). A minority of cells retained a dark protoplast with an intact membrane. Cells with thick envelopes, although in the minority, were easily distinguished among the damaged and undamaged cells (Figure 6(c) and 6(d)). These cells of coccobacillus shape remained osmotically stable. We conclude that the cells containing the thick envelopes are dormant cells. Dormant cells were not evidently bigger (1,317 × 1,029 nm) as compared to normal cells (1,181 × 996 nm), but there was an apparent difference in the thickness of the outer membrane. The thickness of the outer membrane of the control cells from aerobic conditions averaged 30 ± 4 nm (from 22 to 41 nm, n = 93). The outer membrane of dormant cells from anaerobic conditions was on average 3.3 times thicker, averaging 98 ± 11 nm (from 83 to 111 nm, n = 17). TEM images (Figure 6(d)) showed that the thickness of the outer membrane was due to an increase in the electron-dense, most likely, peptidoglycan layer.

The results presented here show that one clinical and four environmental, but clinically relevant, isolates of A. baumannii were able to survive both in nutrient-depleted and nutrient-rich water media in the aerated, intermediate and anaerobic conditions during 50 days. Moreover, this emergent human pathogen multiplied in the nutrient-rich water medium. The monitoring of survival in aerated DNB is consistent with the previous report (Hrenovic et al. 2016) of multiplication and good survival of XDR isolates of A. baumannii in sterilised effluent wastewater (of chemical composition close to the DNB) within 50 days. This implies that once in the environment, clinically relevant A. baumannii are able to persist in a wide range of water environments: from oligotrophic spring water to eutrophic lakes, at full oxygen saturation to complete anaerobic conditions. Although not commonly associated with anaerobic sediments, A. baumannii could possibly survive in such environments.

Antibiotic-resistant pathogens like A. baumannii are transferred from hospitals into the natural environment via the hospital wastewater (Ferreira et al. 2011; Zhang et al. 2013; Kovacic et al. 2017). In some countries, hospital wastewater is not disinfected, meaning that raw hospital wastewater is discharged directly into a river (Hrenovic et al. 2019). In other situations, the hospital wastewater is mixed with domestic wastewater and storm runoff, after which it goes to a wastewater treatment plant. There is evidence that antibiotic-resistant A. baumannii survive the wastewater treatment and are discharged into the natural environment (Hrenovic et al. 2016; Seruga Music et al. 2017; Higgins et al. 2018). The time duration of 50 days is sufficient for water masses to spread far from the source point. This fact is disturbing and represents a public health problem. Sporadic acute community-acquired human infections with A. baumannii (Dexter et al. 2015) are well known, but the source of infection has not been identified thus far. Even anaerobic natural water environments could present a source of A. baumannii, where this bacterium could survive for prolonged periods.

The antibiotic sensitive isolate EF11 showed the lowest survival in SW as compared to MDR, XDR or PDR isolates. In the nutrient-rich urban wastewater, the persistence of antibiotic-sensitive isolates among dominant antibiotic-resistant A. baumannii isolates was reported, suggesting the urban sewage is a natural habitat of susceptible A. baumannii (Higgins et al. 2018). The EF11 isolate recovered from wastewater is therefore probably accustomed to the high nutrient concentration, which explains its lower survival in nutrient-poor water. This may suggest better survival of the problematic antibiotic-resistant isolates in clean oligotrophic waters, but the association between the genetic determinants for the resistance to antibiotics and the resistance to environmental conditions needs to be examined further.

Small colony variants represent a subpopulation with distinctive phenotypic traits as compared to the parent bacterial population (Proctor et al. 2006). Besides the atypical colony morphology, small colony variants exhibit unusual biochemical characteristics, which often leads to the misidentification of the species by using the routine identification techniques. Since the small colony variants are resistant to antibiotics successful in treatment of the parent population, they represent a challenge in clinical microbiology. Small colony variants are described in different bacterial species, and most often in Staphylococcus aureus (Seifert et al. 2003), as a cause of recurrent human infections. It is hypothesized that small colony variants arise from the parent population as a consequence of the selective pressure of prolonged antimicrobial treatment of infected patients. Translucent colonies formed on LB plates incubated at 37 °C were recently reported in A. baumannii (Tipton et al. 2015). However, the described translucent colonies were not smaller than the opaque ones, and the translucency could be viewed only under a dissecting microscope with oblique indirect illumination at density of colonies greater than 200 per plate. Contrary to the increased resistance to aminoglycosides of S. aureus small colony variants (Seifert et al. 2003; Proctor et al. 2006), translucent A. baumannii colonies showed the increased susceptibility to aminoglycosides (Tipton et al. 2015). In our case, small colony variants of A. baumannii appeared after the exposure of isolates to anaerobic conditions without contact with antibiotics. Therefore, we presume that the appearance of small colony variants is a consequence of environmental stress caused by depletion of oxygen necessary for the normal metabolism of A. baumannii. In addition, Dekic et al. (2018) reported the occurrence of small colony variants in A. baumannii after the exposure to extreme temperatures and pH. These data imply that A. baumannii forms translucent colonies in unfavourable environmental conditions. Translucency of smaller colonies was visible under normal room lighting, differentiating them from the translucent A. baumannii colonies reported by Tipton et al. (2015). The biochemical characteristics of small translucent colonies were not examined, but the identification of A. baumannii was confirmed by MALDI-TOF MS. This implies the necessity of advanced techniques for the identification of A. baumannii in addition to routine techniques based on the biochemical characteristics, especially when working with the environmental samples where A. baumannii could be exposed to oxygen depleted conditions.

Special concern rises from the fact that A. baumannii isolates formed dormant cells in anaerobic conditions, which enhances its survival in an unfavourable environment. Dormancy in non-sporulating bacteria has been recognised for decades (Kaprelyants et al. 1993). Exogenous dormancy refers to the reversible state of low metabolic activity, which helps the cells survive for a long time without multiplication. Dormant cells are unable to divide and form colonies on agar plates without a preceding resuscitation phase (Kaprelyants et al. 1993). In our case, the resuscitation of dormant cells was provided by incubation of inoculated agar plates in the air, where dormant together with undamaged cells were allowed to form CFUs. Therefore, the initial nine-day drops in CFUs of A. baumannii in anaerobic DNB actually did not reflect the formation of dormant cells. But, the sudden increases in CFUs after 14 days of contact could be connected to the formation of dormant cells. The erratic curve of CFUs in DNB exposed to anaerobic conditions could not be explained by the ‘bust and boom’ survival strategy where weak bacterial cells die in unfavourable conditions and the remaining persister cells live on the expense of dead cells (Bravo et al. 2016). Namely, in the oxygen depleted conditions multiplication of A. baumannii is unlikely to happen. As judged by relatively high CFUs from the anaerobic experiment, A. baumannii isolates did not form the viable but non-culturable cells, which is in agreement with previous records for A. baumannii ATCC 19606 T under nutrient-deprived conditions (Bravo et al. 2016). Dormancy per se is one of the possible consequences of starvation. However, until now it has not been clear what enables the fraction of the population to respond differently to the starvation stress (Pu et al. 2017). In this study, there was no difference in the survival curve of A. baumannii in nutrient-poor medium regardless of oxygen concentration. An unusual survival curve was registered only in the experiment with DNB subjected to anaerobic conditions. The nutrient availability in DNB together with exposure to anaerobic conditions governed the erratic behaviour of A. baumannii and the formation of dormant cells.

Here, observed dormant cells could be classified as those employing a ‘passive defence’; they are different from persister cells employing an ‘active defence’ to expel antibiotics (Barth et al. 2013; Pu et al. 2017). The thick envelope of dormant cells probably reduces the permeability of the outer membrane, which enables the minimum metabolism of cells in an unfavourable environment. It should be emphasized that small colony variants should not be equated with the dormant cells of A. baumannii. Although not quantified, small colony variants were much more abundant than dormant cells and did not match in either the same time (1 day of contact for small colony variants vs 14 days of contact for dormant cells) or trend of appearance. It seems that the small colony variants are a fast response to the adverse environmental conditions, while the formation of dormant cells requires prolonged exposure.

A. baumannii is considered a strict aerobe, producing energy by aerobic respiration with oxygen as the terminal electron acceptor, but is unable to grow with nitrate as the terminal electron acceptor (Garrity et al. 2005). In routine bacteriology, it is known that A. baumannii gives a characteristic weak orange reaction in the depth of Kligler Iron Agar. This suggest the formation of weak or low concentration of acid from glucose, and also could be an indication of fermentation potential. Pseudomonas aeruginosa is another strict aerobe confirmed to perform fermentation in the absence of oxygen (Eschbach et al. 2004; Schreiber et al. 2006). The mechanisms of P. aeruginosa fermentation are well investigated; however, the mechanism of A. baumannii anaerobic metabolism has not been investigated at all which opens new research opportunities.

To our knowledge, this work provides the first evidence of the survival of A. baumannii in anaerobic conditions. The findings call for the future usage of microscopic, biochemical and molecular techniques in order to elucidate the mechanisms of survival of non-spore forming, strictly aerobic, human pathogens in anaerobic environments.

A. baumannii isolates survived both in nutrient-depleted and nutrient-rich water media in aerated, intermediate and anaerobic conditions (oxygen saturation of 96, 56 and 0%, respectively) for 50 days. Clinical A. baumannii isolate behaved the same as environmental isolates of clinical significance. In addition, the antibiotic-sensitive isolate showed the lowest survival in nutrient-depleted water as compared to MDR, XDR or PDR isolates. The availability of dissolved oxygen did not influence the survival of A. baumannii in nutrient-depleted water, but enhanced its survival in nutrient-rich water. The formation of small colony variants as the fast response and dormant cells as the prolonged response to anaerobic conditions was observed. The results provide insight into the behaviour of A. baumannii and make an important contribution to raise awareness that this emerging pathogen persists in the natural environment, which should lead to appropriate monitoring and control measures to reduce the threat to public health.

This study was supported by Croatian Science Foundation (IP-2014-09-5656). Many thanks to Snježana Kazazić for enabling the use of the MALDI-TOF MS device.