Quinolone antibiotics are chemically synthesized drugs with advantages of high blood concentrations, excellent broad-spectrum antibacterial activity, and strong tissue penetration (Ahmed and Daneshtalab, 2012; Ball et al., 1998), making them a preferred agent for empirical treatment of infections (Gupta et al., 2011; Pilatz et al., 2020; Taplitz et al., 2018). They were the second most frequently prescribed antimicrobial drugs after β-lactam (Mercuro et al., 2018; Sanchez et al., 2016; Shapiro et al., 2014), but followed by serious bacterial resistance (Logan et al., 2019; Pitout et al., 2022; Zhao et al., 2024). Researches have shown that bacterial resistance to quinolones is mainly mediated by chromosomal mechanisms, including alterations in drug action sites (DNA gyrase and topoisomerase IV gene mutations) and decreased drug accumulation capacity (loss of membrane porins and overexpression of efflux pumps) (Darby et al., 2023; Pham et al., 2019). Of greater concern were plasmid-mediated quinolone resistance (PMQR) due to their higher transmissibility (Ellabaan et al., 2021), including: Qnr protein families, which protect quinolone targets; the AAC(6′)-Ib-cr enzyme, which acetylates ciprofloxacin and norfloxacin; and efflux pumps mediated by qepA and oqxAB plasmid genes (Ruiz et al., 2012). Among them, the aac(6′)-Ib-cr gene is the most threatening as it provides a selective advantage in the presence of ciprofloxacin (Phan et al., 2022). Since its first report in 2006 (Robicsek et al., 2006), the prevalence of strains carrying the aac(6′)-Ib-cr gene has been increasing in both clinical and environmental settings. According to recent studies, this gene has been detected in various countries of Asia (Al-Khafaji et al., 2024; Kuo et al., 2024; Sohrabi et al., 2024; Zhan et al., 2021), Europe (Hrovat et al., 2019; Mounsey et al., 2021; Piccirilli et al., 2021), Africa (Al-Gallas et al., 2024; Masoud et al., 2021; Swedan et al., 2023), and the Americas (de Oliveira Alves et al., 2022; Golden et al., 2021; Logan et al., 2023) across multiple Gram-negative bacteria. The detection rates of the gene vary significantly by region and source, typically ranging from 10% to 50%. Furthermore, the distribution of the aac(6′)-Ib-cr gene is widespread, with its presence detected not only in clinical strains but also in animals, food, and water sources (Al-Gallas et al., 2024; Mounsey et al., 2021; Patel et al., 2024; Zahra et al., 2023).

However, unlike the qnr, qepA, and oqxAB screening, which can be accomplished by PCR amplification of the target genes, the detection of aac(6′)-Ib-cr is more complicated. This is because AAC(6′)-Ib-cr belongs to a variant of aminoglycoside acetyltransferases, and the differences between its encoding gene aac(6′)-Ib-cr and the wild-type aac(6′)-Ib exist only on just two nucleotides (Trp102Arg and Asp179Tyr) (Robicsek et al., 2006). These specific point mutations enable the acetylation of norfloxacin and ciprofloxacin at the amino nitrogen on position 7-C of the piperazine ring, thereby reducing drug activity (Vetting et al., 2008; Maurice et al., 2008). Therefore, the screening for aac(6′)-Ib-cr traditionally involves amplification of the target gene, followed by sequencing or restriction analysis (Pitout et al., 2008). Obviously, this is a time-consuming and expensive work. Although several studies have proposed alternative methods for aac(6′)-Ib-cr sequencing (Hidalgo-Grass and Strahilevitz, 2010; Wareham et al., 2010), these methods still relied on the use of large equipment not typically available in ordinary laboratories. To date, no study has proposed a method for detecting the AAC(6′)-Ib-cr enzyme based on common antimicrobial susceptibility testing consumables.

Here, we proposed three low-cost phenotypic methods of detecting the AAC(6′)-Ib-cr enzyme in Enterobacteriaceae. They each possess distinct characteristics and are suitable for clinical microbiology laboratories under different conditions. We aim to (i) improve existing phenotypic detection methods for AAC(6′)-Ib-cr, (ii) develop and validate QIM and QHT as reliable alternatives, and (iii) compare their performance with mass spectrometry and PCR. These methods are based on the following principles: (i) Matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF MS) is able to measure the mass transfer of 42 Da resulting from acetylation of ciprofloxacin by the AAC(6′)-Ib-cr enzyme. Based on this, the appearance of spectrum peaks representing acetylated ciprofloxacin and its sodium or potassium adduct reveals the presence of the AAC(6′)-Ib-cr enzyme. This method was initially reported by Pardo et al. (2016), and we followed their steps with some modifications while enriching the species of the measured strains. (ii) The quinolone inactivation method (QIM), which is introduced and named for the first time in this study, is based on a principle derived from the Carbapenem Inactivation Method (CIM) (Pierce et al., 2017; van der Zwaluw et al., 2015): after co-incubation with test strains for some time, the ciprofloxacin adhered onto the disk will be inactivated by the enzyme-producing strains, thereby unable to inhibit the growth of the indicator organism. Furthermore, replacing the indicator organism E. coli ATCC 25922 with an isolate with only slight resistance to quinolones can greatly improve the observation of results. Because AAC(6′)-Ib-cr enzyme can easily reduce the quinolone concentration to a level where slightly resistant strains can grow, but it still cannot be lowered enough to support the growth of highly sensitive wild-type strains, as this would require nearly complete depletion of the quinolone. (iii) The last approach is based on the reduction of quinolone drug activity by the AAC(6′)-Ib-cr enzyme-producing Enterobacteriaceae (APE), allowing indicator organisms to grow in a curved manner along the inoculum of the tested strain toward the disk. This principle was first proposed by Hodge et al. (1978) to detect penicillinase and has subsequently been modified multiple times for detection of β-lactamases. The excellent performance of Modified Hodge Test (MHT) for screening carbapenase has been repeatedly demonstrated (Lee et al., 2001; Pasteran et al., 2016). In this study, we attempt to apply this principle to the detection of quinolone acetyltransferases, and therefore name it the quinolone Hodge Test (QHT).

The resistance phenotype to quinolones is typically insufficient to distinguish between PMQR and other resistance mechanisms (Rodríguez-Martínez et al., 2016; Strahilevitz et al., 2009). As the most prevalent PMQR (Guillard et al., 2014; Jomehzadeh et al., 2022; Shrestha et al., 2023; Tayh et al., 2024), AAC(6′)-Ib-cr confers a low level of quinolone resistance in host organisms, generally below the resistance breakpoints established by CLSI (Machuca et al., 2016; Robicsek et al., 2006; Ruiz, 2019), yet it poses a significantly grave clinical threat (Redgrave et al., 2014; Tacconelli et al., 2018). Several factors substantiate its significance in this process, including its increasing prevalence (Kuo et al., 2024; Zhan et al., 2021), association with other resistance elements (Dulyayangkul et al., 2024; Raherison et al., 2017), as well as the accelerated induction of mutations in QRDRs (Liang et al., 2023; Ortiz-Padilla et al., 2020; Recacha et al., 2019). Of concern is that clinical microbiology laboratories have not yet considered these genes in antimicrobial resistance screening, which may be attributed to the current lack of reliable screening assays for AAC(6′)-Ib-cr (Strahilevitz et al., 2009). In this study, we proposed three methods for detecting the AAC(6′)-Ib-cr enzyme in Enterobacteriaceae, all of which demonstrated an overall performance with 100% specificity and sensitivity. Additionally, all three methods cost less than $1 per test, and QIM and QHT did not require complex equipment and professional personnel to perform, as they were easy to operate and yield simple results. Furthermore, our experiments clearly showed that neither mutations in quinolone action target genes nor plasmid-mediated Qnr proteins would interfere with the determination of AAC(6′)-Ib-cr in these new methods. This is not difficult to explain. All tests are based on the acetylation modification activity of the AAC(6′)-Ib-cr enzyme toward ciprofloxacin, which is specific and independent among all mechanisms of resistance to quinolone(Robicsek et al., 2006; Ruiz et al., 2012).

In fact, the concept of using mass spectrometry for the detection of resistant enzymes was already commonplace and dated back to the early 2000s (Burckhardt and Zimmermann, 2011). Currently, it is mainly applied in the detection of carbapenemases (Alvarez-Buylla et al., 2013; Zhang et al., 2023). In contrast to carbapenemases, the detection of the AAC(6′)-Ib-cr enzyme by MS has an additional advantage, because the former result is explained by the complete disappearance of the spectral peaks representing ertapenem (Burckhardt and Zimmermann, 2011), which requires sufficient reaction time, while the latter result is explained by the appearance of new peaks representing the modified products, which are produced early in the reaction (Oviaño et al., 2017; Pardo et al., 2016). Therefore, although the activity of modifying enzymes is much lower than that of hydrolyzing enzymes, the incubation time required for AAC(6′)-Ib-cr detection is still shorter than that needed for carbapenemase detection. In addition, we found that MS was extremely sensitive to CIP detection, as clear ion peaks could be obtained with a 10 μg/ml CIP solution. This was why we believed that CIP reagents could be directly obtained by immersing CIP disks in normal saline. Such adjustments offered numerous benefits: simplifying reagent preparation, reducing costs, and the fact that antibiotic disks are often easy to store for a long time. Furthermore, preliminary research experience showed us that adding the matrix solution directly before the sample droplet dries did not affect the spectral peaks (allowing co-drying). Such a research was necessary because the drying time for sample droplets without surfactants was extremely long at room temperature. Adding the matrix solution could reduce the surface tension of the liquid, similar to the effect of surfactants, and greatly shorten the waiting time for drying (Miller et al., 2017; Wu et al., 2019).

The QIM and QHT methods were proposed for the first time in this study, and our objective was to design a low-cost, as minimally instrument-free method for the detection of the plasmid-mediated quinolone resistance determinant AAC(6′)-Ib-cr in Enterobacteriaceae. However, in our previous study (Yunxiang et al., 2015), we found that when E. coli ATCC 25922 was used as an indicator organism, it was only suitable for enzymes with strong hydrolytic activity, such as carbapenemases. Its performance was unsatisfactory in detecting aac(6′)-Ib-cr-mediated acetyltransferase. Therefore, through the comparative experiments involving four strains of E. coli with different degrees of QRDR mutations, we attempted to propose new insights into indicator organisms that played a critical role in the test.

The selection of an ideal indicator organism must meet two simple criteria: (i) its minimum inhibitory concentration (MIC) of ciprofloxacin should range between 0.25 to 1 μg/ml. (ii) in the Kirby-Bauer disk diffusion susceptibility test, the indicator organism which must be non-mucoid, should exhibit a clear boundary with an inhibition zone diameter of 28–30 mm. Such organisms, which are readily obtainable from clinical samples, share a common feature of a single-point mutation (S83L or D87N) in the gyrA gene, resulting in low-level resistance to quinolones. However, they still exhibit good susceptibility to ciprofloxacin in the Kirby-Bauer test (Machuca et al., 2016; Pham et al., 2019). Our approaches are based on the fact that the enzyme AAC(6′)-Ib-cr produced by the test strains modifies ciprofloxacin, reducing the drug’s effective concentration to such a level that the indicator organism with weak quinolone resistance is sufficient to protect itself, while the highly sensitive wild-type strain (E. coli ATCC 25922) remains difficult to survive (Robicsek et al., 2006). This principle is similar to that when a bacterium possesses both a gyrA mutation and plasmid-mediated AAC(6′)-Ib-cr enzyme, it always exhibits a higher level of resistance to quinolones (Machuca et al., 2016). We redesigned the indicator organism aiming to combine these two low-level resistance mechanisms to exhibit enhanced drug resistance effects in both QIM and QHT methods. Indeed, The combination of resistance mechanisms is also common to other classes of antibiotics. For example, the resistance of Pseudomonas aeruginosa to imipenem is due to a decrease in antibiotic uptake (via downregulation of OMPD2) and the production of AmpC-type β-lactamase (Livermore, 1992). Either mechanism alone is not sufficient to cause clinically significant levels of resistance, but only their combination results in high-level resistance. Therefore, as our experimental results showed, when E. coli ATCC25922 was used as the indicator organism, its indication effect for AAC(6′)-Ib-cr production was not significant. However, when it was replaced by strain E. coli 28B220, there would be an “illusion” of high level of resistance at the junction of the two bacteria (quinolone Hodge test), similar to that shown when the two resistance mechanisms worked together. When another resistance mechanisms is added, the inherently small inhibition zone may dilute the changes caused by the AAC(6′)-Ib-cr enzyme.

In addition, it was worth noting that the aac(6′)-Ib-cr genes carried by all 46 strains of APE in this study exhibited identical mutations at codons 102 (Trp-Arg) and 179 (Asp-Tyr). The remaining codons showed high similarity to the wild-type aac(6′)-Ib gene, and no independent occurrences of the W102R or D179Y mutations were observed. This was consistent with previous studies (Robicsek et al., 2006). We believed that the high sequence homology ensures relatively consistent enzyme activities of the encoded products (Vetting et al., 2008), which was the primary reason for the high performance of these enzyme activity methods. Therefore, compared to labor-intensive, time-consuming and expensive genotype testing, direct detection of AAC(6′)-Ib-cr enzyme activity represents a more cost-effective laboratory strategy. On the other hand, we observed that CIP was completely modified after incubation with APE for up to 18 h (no 322 Da peak was shown; Figure 2g). This compelled us to reassess the AAC(6′)-Ib-cr enzyme. To our knowledge, the concept of AAC(6′)-Ib-cr as a low-level resistance mechanism is based on in vitro experiments (Robicsek et al., 2006), which are inherently short-term studies. Thus, its prolonged presence in the human body and its sustained impact on antibiotic efficacy warrant re-evaluation.

There are some limitations to our study. Although all three methods demonstrated 100% specificity and sensitivity, real-world variability should be acknowledged. For example, we are still unable to explain whether a single mutation of the aac(6′)-Ib gene (W102R or D179Y) in strains might challenge these methods, whether our methods are applicable to non-fermenting Gram-negative bacilli, whether severe porin protein loss that prevents ciprofloxacin from entering the periplasmic space could lead to false negatives, and the variability between different laboratories.

In summary, this study presents three simple, efficient, and cost-effective methods for the detection of AAC(6′)-Ib-cr in Enterobacteriaceae. Each method has its own advantages: MSM is a rapid and highly accurate method for AAC(6′)-Ib-cr detection, but it relies on MALDI-TOF MS. QIM and QHT offer cost-effective alternatives but are unsuitable for P. mirabilis. Overall, these methods can enhance routine clinical screening for PMQR mechanisms.