*2.4. Determination of β-Lactamase Genes*

Determination of resistance genes was carried out for all *Acinetobacter* spp. isolates that revealed a resistance to at least one tested carbapenem. PCR detection and gene identification were performed for five different β-lactamases gene families, *bla*CTX-M-1group, *bla*CTX-M-2group, *bla*CTX-M-9group, *bla*GES, *bla*KPC, *bla*OXA-23, *bla*OXA-24, *bla*OXA-51, *bla*OXA-48, *bla*OXA-58, *bla*NDM, *bla*SHV, *bla*TEM, and *bla*VIM., PCR and sequencing procedures were performed as described previously [22–27].


**Table 2.** List of tested antibiotics, concentration on the disc (Sensi-DiscTM paper discs, BD, Vienna, Austria) or Etest® (Biomerieux) and antibiotic classes.

#### **3. Results**

In total, 262 *Acinetobacter* were isolated. *Acinetobacter baumannii* was the most common species with 135 isolates. *Acinetobacter johnsonii* was second most with 62 isolates; all other species were represented by less than 20 isolates; *Acinetobacter haemolyticus* 19 isolates, *Acinetobacter junii* 17 isolates, *Acinetobacter lwoffii* 16 isolates, *Acinetobacter radioresistens* four, *Acinetobacter ursingii* two and seven isolates where no distinct species identification was possible. Non-*baumannii Acinetobacter* spp. were subsumed for further analyses.

Susceptibility testing revealed that resistance to the most tested antibiotics was rare in *Acinetobacter baumannii* and non*-baumannii Acinetobacter* spp. river water isolates. The only resistance present in both groups in more than 10% of the isolates was to cefotaxime with 95.6% (129/135) of *Acinetobacter baumannii* and 43.3% (55/127) non*-baumannii Acinetobacter* spp. In addition, resistances to ceftazidime in non*-baumannii* (12.6%, 16/127) and to piperacillin/tazobactam in *Acinetobacter baumannii* (12.6%, 17/125) were present in more than 10% in one of the sample subgroups (Figure 1).

**Figure 1.** Percentage of resistance to tested antibiotics of isolated *Acinetobacter baumannii* (blue) and non*-baumannii Acinetobacter* spp. (red).

Resistance to fluoroquinolones showed one notable detail: all five *Acinetobacter baumannii* were resistant to ciprofloxacin and levofloxacin, whereas non-*baumannii Acinetobacter* spp. remained susceptible to the second tested fluoroquinolone. Less than 10 *Acinetobacter baumannii* isolates revealed resistance to all other tested antibiotics, all isolates were susceptible to colistin. In contrast to this, colistin resistance could be detected in six (4.7%) non*-baumannii Acinetobacter* spp., and no resistance was found to levofloxacin, imipenem, meropenem, amikacine, and tigecycline (Figure 1).

Only six (4.4%) *Acinetobacter baumannii* revealed susceptibility to all tested antibiotics, but only 16 (11.9%) were resistant to one or more tested antibiotics additionally to cefotaxime. Six (4.4%) isolates could be classified as multiresistant, with resistance to antibiotics from at least three different antibiotic classes, including one isolate only susceptible to colistin and four to colistin and tigecyline (Table 3).

**Table 3.** Detected resistance genes in carbapenem resistant *Acinetobacter baumannii*; us = upstream; ds = downstream.


Susceptibility to all antibiotics was 10 times higher (55 isolates, 43.3%) in the non*-baumannii Acinetobacter* spp. group compared to the *Acinetobacter baumannii group*. Only 10 isolates revealed multiresistance. Two isolates with different resistance profile showed resistance to four antibiotics (JDS10AC012 to CTX, piperacillin/tazobactam, FEP, and ceftazidime; JDS38AC048 to CTX, TZP SXT and CAZ).

Five *Acinetobacter baumannii* were resistant to carbapenems (meropenem and imipenem). These isolates were analyzed for the presence of several β-lactamases genes, resulting in four different gene patterns. Classic intrinsic OXA carbapenemases were present in all isolates, but no gene was present in all isolates. JDS59AC007 and JDS59AC001 were positive for OXA-23 and OXA-51, and JDS38AC020 was positive for OXA-24 and OXA-51. In addition to OXA-23 and OXA-51, two isolates revealed β-lactamases from another Ambler class: JDS38AC018 and JDS38AC017 harbored both the gene for carbapenemase VIM-2 and JDS38AC017 harbored additionally the gene for the broad spectrum β-lactamases TEM-1 (Table 3).

#### **4. Discussion**

The presence of *Acinetobacter* with human-induced multidrug resistance phenotypes in surface water has been reported from all over the world. Their origin seems to be influenced by (treated and untreated) hospital waste water. The impact (proportion and persistence) of these strains on the *Acinetobacter* water population is not well documented [9,14,15,28]. This study shows for the first time the susceptibility phenotypes of *Acinetobacter* of a total European river system. Furthermore, our study provides a first glimpse of the anthropogenic impact on the *Acinetobacter* river population. In the *Acinetobacter* population of the River Danube, even resistance to last line antibiotics (e.g., colistin and tigecycline) is detectable, and this without using selective media by screening only a relatively small volume of water. This screening led to the detection of multidrug-resistant *Acinetobacter baumannii* isolates, whose multidrug resistance would normally be found and related only to intensive care units. The isolation of multiresistant *Acinetobacter* was limited to the area of influence of megacities, but even there the great majority of isolates remained not or only slightly influenced on their susceptibility pattern.

Looking at the resistance data for invasive *Acinetobacter* isolates for the Danube neighboring countries reveals a rather gloomy picture: In clinical isolates, the ratio for carbapenem resistance spans from 5.5% in Germany to over 80% in Romania. In our study of the river water, however, only five isolates (less than 2% of all isolates) showed resistance to carbapenems, a very low proportion of resistant *Acinetobacter* spp. compared with the clinical settings [29]. This ratio corresponds with our findings in the *Pseudomonas* population in the river Danube, where we also found a ratio of around 2% [30,31].

In a study of the Jadro River, carried out by Maravic et al., only selected multi-drug-resistant *Acinetobacter* were isolated (using selective media with supplements). Comparing these isolates with the Danube isolates, there is a remediable difference as regards aminoglycoside resistance. Maravic et al. did not detect a single isolate that was resistant to the tested aminoglycoside in contrast to 5.7% (15 isolates) from the River Danube. Furthermore, only two multiresistant Danube isolates revealed no resistance to one of the tested aminoglycoside. Carbapenem resistance in Jadro River isolates was restricted to meropenem, but the number of isolates was too low to be noteworthy. Interestingly enough, the proportion of cefotaxime resistance is nearly identical with 68% in Jadro River and 66% in the river Danube [9].
