Detailed information of bacterial strains (Supplementary Table 2), plasmids (Supplementary Table 3), and primers (Supplementary Table 4) used in this study are listed in Supplementary Material.

The microorganism research and animal subject research (for preparation of anti-NagZ antibody) were approved by the Ethics Committee of the Clinical Medical College and the First Affiliated Hospital of Chengdu Medical College. After clearly explaining the nature and purposes of this scientific research to all participants, sufficient time was provided for questions and answers, written consents were acquired from all participants.

Antibiotic susceptibility test was performed by using broth microdilution and Kirby-Bauer method according to protocols recommended by Clinical Laboratory Standard Institute (CLSI, 2018). E. cloacae subsp. cloacae ATCC 13047 and E. coli ATCC 25922 were used for quality control. All antibiotics and culture medium used in antibiotic susceptibility test were purchased from Wenzhou Kangtai company (Bio-kont Co., Ltd., Wenzhou, China). Each assay was performed independently at least three times.

Anti-NagZ antibody was generated through rabbit immunization by an “antigen intersection” strategy immunization and purification (Arora et al., 2014; Zhou et al., 2016). Briefly, nagZ coding sequence (CDS) from EC was obtained by polymerase chain reaction (PCR), cloned into a pET28a vector with a 6His-label. Then, pET28a-nagZ-6His vector was transformed into E. coli B21 for expression of NagZ recombinant protein, which was purified by Ni-NAT and identified by electrophoresis. Next, purified NagZ-6His recombinant protein was used to immunize rabbit. Enzyme linked immunosorbent assay (ELISA) was applied to evaluate titer of antiserum (over 1:8000) after immunization of NagZ-6His. Finally, antiserum was purified by affinity of antibody to NagZ-6His-coupled antigen. Western blot showed an excellent specificity of the antibody (Supplementary Figure 1). Reagents and materials used in generation of anti-NagZ antibody were purchased Shenggong Biological Company (Sangon Biotech Co., Ltd., Shanghai, China), primers for obtaining nagZ CDS are listed in Supplementary Table 4.

AmpC β-lactamase activity was determined by a nitrocefin hydrolysis assay as previously described (Cavallari et al., 2013; Guerin et al., 2015). EC isolates were inoculated into LB medium and incubated at 37°C with 250 rpm overnight. It was sub-cultured in LB medium with a concentration of 1:100. When absorbance of OD600 reached 0.8, bacteria were collected and washed once with 1 ml of phosphate buffer (pH 7.0), and resuspended in 1 ml of protein lysate (Sangon Biotech Co., Ltd., Shanghai, China). Samples were placed on ice and lysed by sonication with a microprobe by using a 10-s pulse three times with a 10-s interval during each pulse. The samples were centrifuged at 10,000g for 10 min and supernatant was collected. The concentration of protein in supernatant was determined by a protein quantitative kit (Beyotime, Biotechnology, Shanghai, China). The nitrocefin hydrolysis assay was performed in 250 μl of phosphate buffer (pH 7.0) containing 5 μg of total protein and 50 μg/ml nitrocefin (Sigma-Aldrich; Merck KGaA, St. Louis, MO, United States). The hydrolysis rate of nitrocefin was determined at 486 nm at room temperature every 5 min. AmpC β-lactamase activity was calculated by extinction coefficient of nitrocefin 20, 500 M^–1 cm^–1, each assay was performed independently at least three times.

The total RNA was extracted from cellular lysates by using RNA extraction kit (Sangon Biotech Co., Ltd., Shanghai, China) according to the manufacturer’s instructions. Briefly, genus (EC isolates) were inoculated into LB medium and incubated at 37°C with 250 rpm overnight. It was sub-cultured in LB medium with a concentration of 1:100. When absorbance of OD600 reached 0.8, bacteria were collected by centrifuge at 12,000g for 2 min and the supernatant was discarded, the precipitate was washed once with 1 ml of phosphate buffer (pH 7.0), bacterial pellet was resuspended in 100 μl of TE buffer containing 400 μg/ml lysozyme, and incubated for 5 min at room temperature. Next, 900 μl of lysis solution was added and mixed at room temperature for 3 min, 200 μl of chloroform (Sangon Biotech Co., Ltd., Shanghai, China) was added, mixed, and centrifuged at 12,000g at 4°C for 5 min. Consequently, 600 μl of supernatant (aqueous liquid) was acquired and 200 μl anhydrous ethanol was added, the mixture was incubated at room temperature for 3 min, centrifuged at 12,000g at 4°C for 5 min. The supernatant was discarded, and precipitate was washed with 70% ethanol twice, dried naturally, dissolved in ddH₂O, the concentration of RNA was determined by NanoDrop^TM8000 spectro-photometer (Thermo Fisher Scientific, Waltham, MA, United States) and stored at −70°C. For detecting the expression of AmpR target genes, LB medium containing 5 mg/L 1,6-anhydromuropeptides (anhMurNAc, Medicilon, Co., Ltd., Shanghai, China) was used at the stage of sub-culture.

cDNA was synthesized from 500 ng of total RNA with a FastKing gDNA Dispelling RT SuperMix kit (Tiangen Biotech Co., Ltd., Beijing, China) according to the manufacturer’s instructions. Real-time fluorescence quantitative PCR (qPCR) was performed with a SuperReal PreMix Color (SYBR Green) kit (Tiangen Biotech Co., Ltd., Beijing, China) according to the manufacturer’s instructions (volume: 20 μL. PCR program: pre-denaturation: 95°C/10 min. Denaturation: 95°C/30 s, Annealing:58°C/30 s, Elongation:72°C/30 s, and 30 cycles), with 16S as an internal control. Sequences of primers used in RT-qPCR assays are listed in the Supplementary Table 4. Each assay was performed independently at least three times.

Total protein was extracted from EC by a bacterial protein extraction kit (Sangon Biotech Co., Ltd., Shanghai, China) according to the manufacturer’s instructions. Briefly, strains were inoculated into LB medium and incubated at 37°C with 250 rpm overnight. It was sub-cultured in LB medium with a dilution concentration of 1:100 and continue incubated at 37°C with 250 rpm. When absorbance of OD600 reached 0.8, bacteria were collected by centrifuge at 12,000g for 2 min and the supernatant was discarded, the precipitate was washed once with 1 ml of phosphate buffer (pH 7.0), bacterial pellet was resuspended 1 ml of protein lysate (Sangon Biotech Co., Ltd., Shanghai, China). Samples were placed on ice and lysed by sonication with a microprobe by using a 10-s pulse three times with a 10-s interval during each pulse. The samples were centrifuged at 10,000g for 10 min and supernatant was collected. The concentration of protein in supernatant was determined by a protein quantitative kit (Beyotime, Biotechnology, Shanghai, China), and 30 μg total protein was used to western blot assay. Western blot analysis was performed with a standard method as previously described (Yang et al., 2017). Information of antibodies used are as followings: rabbit anti-AmpC (Abnova Taipei, Taiwan, China), mouse anti-DnaK (Abcam, Cambridge, MA, United States), rabbit anti-NagZ (preparation by ourselves), goat anti-rabbit IgG-HRP (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, United States), goat anti-mouse IgG-HRP (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, United States). Images were taken with a SPOT-CCD camera. For quantitative analysis of western blot, intensities of protein bands were quantified by application of ImageJ, and DnaK was applied as an internal control. Each assay was performed independently at least three times.

nagZ-knockout EC was obtained by homologous recombination method with application of a suicide vector (Luo et al., 2015). Briefly, two homologous arms of DNA fragments (A: 522 bp-upstream fragment of initiator codon, and B: 544 bp-downstream fragment of termination codon) of nagZ gene were obtained by PCR. The fusion DNA fragment (AB fragment: 1066 bp) was obtained by the fusion PCR. The fused DNA fragment of AB was cloned into the suicide plasmid pLP12 and verified by PCR and sequencing. The recombinant plasmid was transformed into E. coliβ2163. Finally, nagZ-knockout EC strain was obtained by co-culture E. coliβ2163 with DNA fragment AB and wild-type E, cloacae. The strains and reagents used in this experiment were purchased from Nuojing Biological Company (Knogen Biotech Co., Ltd., Guangzhou, China).

The CDS of nagZ was obtained by PCR, then cloned into a plasmid of pBAD33cm-rp4 (Knogen Biotech Co., Ltd., Guangzhou, China), and verified by sequencing. The recombinant plasmid (pBAD33-nagZ) was transformed into competent E. coliβ2163 (Knogen Biotech Co., Ltd., Guangzhou, China). Finally, the recombinant plasmid from E. coli β2163 was transformed into E. cloacae by a conjugation assay, 0.05% L-Arabinose (Sangon Biotech Co., Ltd., Shanghai, China) was used to induce gene expressions of the recombinant plasmids. For antibiotic susceptibility test, L-Arabinose was added at the initial stage of antibiotic susceptibility test. For western blot, RNA Extraction and AmpC β-lactamase Activity Assay, L-Arabinose was added at the stage of sub-culture. Primers for obtaining CDS of nagZ are listed in the Supplementary Table 4.

All data were presented as mean ± standard deviation. Two-tailed t-test was used to determine the significant difference between two groups by GraphPad Prism 5. ^∗P < 0.05 and ^∗∗P < 0.01 were applied to be statistically significant and statistically highly significant, respectively. All experiments were performed independently at least three times.

To clarify mechanism of developing resistance in EC, 12 clinically isolated EC were randomly collected. Minimum inhibitory concentrations (MICs) of piperacillin (PIP), piperacillin-tazobactam (TZP), aztreonam (ATM), ceftriaxone (CRO), cefotaxime (CTX), cefoperazone (CFP), ceftazidime (CAZ), cefepime (FEP), imipenem (IMP), meropenem (MEM), levofloxacin (LVX), ciprofloxacin (CIP), amikacin (AMK) and gentamicin (GEN) against the 12 clinically isolated EC were determined according to protocols recommended by Clinical Laboratory Standard Institute (Supplementary Table 1; CLSI, 2018). Based on the MICs, 12 clinical isolates were divided into two groups (six susceptible and six resistant isolates, abbreviated as S1, S2, S3, S4, S5, S6, and R1, R2, R3, R4, R5, R6, respectively. Susceptible isolate: susceptible to all types of the third and fourth generation cephalosporins; resistant isolate: resistant to at least one type of the third or fourth generation cephalosporins). To determine whether NagZ was involved in developing resistance in EC, nagZ mRNA expression was examined by reverse transcription-quantitative polymerase chain reaction (RT-qPCR), as indicated in Figure 1A. nagZ mRNA expression was significantly enhanced in the resistant isolates compared with susceptible ones. To further detect the different protein expressions of nagZ between susceptible and resistant strains, we prepared anti-NagZ antibody for the first time, and its specificity was verified by nagZ-knockout EC model (Supplementary Figure 1). nagZ protein expressions were detected by western blot in six resistant and six susceptible strains, as indicated in Figures 1B,C, protein expressions of nagZ were dramatically up-regulated in six resistant EC isolates compared to susceptible ones.

NagZ is a cytosolic glucosaminidase and acts a crucial role in peptidoglycan recycling pathway, some publications reported there also exists a correlation between peptidoglycan recycling and ampC expression in P. aeruginosa (Reith and Mayer, 2011; Mayer, 2019). Therefore, mRNA expression level of ampC in 12 clinically isolated EC was determined by RT-qPCR (Figure 1D), the results indicated ampC mRNA expression was up-regulated in the resistant EC compared to the susceptible ones. Furthermore, protein expressions of ampC were enhanced in resistant EC isolates, which is shown in Figures 1E,F. Additionally, to investigate whether a highly expressed AmpC β-lactamase was associated with a higher β-lactamase activity, nitrocefin hydrolysis assay was used to determine the β-lactamase activity of AmpC. Our results confirmed that increasing protein level of AmpC had an excellent ability to hydrolyze nitrocefin (Figure 1G). All these data confirmed that expression of NagZ and β-lactamase activity of AmpC were enhanced in resistant EC isolates.

As indicated in Figure 1, our results demonstrated that nagZ expression was up-regulated in resistant EC isolates. It was further determined whether increased NagZ was significantly functional in developing resistance in EC. nagZ CDS was cloned into the pBAD33cm-rp4 vector (pBAD33-nagZ), and then pBAD33-nagZ (NagZ complementation vector) and a pBAD33cm-rp4 vector (pBAD33, control vector) were transformed into S1 and S2, respectively. RT-qPCR and western blot were used to detect whether pBAD33-nagZ vector was effective (Figures 2A,B), the results indicated that mRNA and protein expressions of nagZ were significantly increased complemented with the pBAD33-nagZ vector compared with pBAD33 vector. To further identify the role of NagZ in developing resistance, inhibition zones and MICs of PIP, TZP, ATM, CRO, cefoperazone-sulbactam (SCF), CAZ against S1 and S2 complemented with or without pBAD33-nagZ vector were determined by Kirby-Bauer method and broth microdilution according to Clinical Laboratory Standard Institute guideline (CLSI, 2018). As shown in Supplementary Figure 2A, inhibition zones of S1 and S2 complemented with pBAD33-nagZ were severely reduced compared with pBAD33. Furthermore, MICs of PIP, TZP, ATM, CRO, SCF, and CAZ were significantly increased in the EC complemented with pBAD33-nagZ compared to EC complemented with pBAD33 (Table 1). These results indicated that increased expression of NagZ enhanced resistance of EC to β-lactam antibiotics.

Next, we aimed to investigate whether expression of ampC is regulated by NagZ in EC isolates. pBAD33-nagZ and pBAD33 were transformed into S1 and S2, respectively. RT-qPCR and western blot were adopted to detect the effect of NagZ on AmpC expression. It is shown in Figures 2C–E, NagZ promoted mRNA (Figure 2C) and protein (Figures 2D,E) expressions of ampC. Nitrocefin hydrolysis assay was used to determine the β-lactamase activity of AmpC, results showed that AmpC hydrolysis activity was significantly improved in EC complemented with pBAD33-nagZ compared with pBAD33 (Figure 2F). Therefore, NagZ enhanced AmpC expression and increased AmpC β-lactamase activity in susceptible EC isolates.

To investigate the regulating role of NagZ in resistance and AmpC β-lactamase expression, we constructed a R1 nagZ-knockout model (R1-ΔnagZ) by homologous recombination. RT-qPCR (Figure 3A) and western blot (Figure 3B) confirmed that nagZ gene was successfully knocked out in clinical isolate of R1. Firstly, mRNA and protein levels of AmpC were detected by RT-qPCR and western blot, which suggested loss of NagZ reduced expression of ampC (Figures 3C–E). Secondly, nitrocefin hydrolysis assay indicated that β-lactamase activity of R1-ΔnagZ was significantly decreased compared with wild-type R1 (Figure 3F). Finally, the effect of NagZ on resistance of R1 was evaluated by broth microdilution and Kirby-Bauer method. The results suggested deletion of NagZ increased inhibition zones of CRO, CAZ, ATM, SCF, PIP, and TZP in R1 (Supplementary Figure 2B), while MICs of CRO, CAZ, ATM, SCF, PIP, and TZP against R1-ΔnagZ were at least fourfold lower than wild-type R1 (Table 1).

To further explore whether complementation of NagZ could rescue ampC expression and resistance in R1-ΔnagZ model, pBAD33 and pBAD33-nagZ were transformed into R1-ΔnagZ strain, respectively. RT-qPCR analyses confirmed that mRNA level of ampC was significantly increased by complementation of NagZ (Figure 3C). Moreover, the protein expression of ampC was rescued by NagZ complementation (Figures 3D,E). Furthermore, decreased β- lactamase activity induced by deletion of nagZ was rescued by complementation of NagZ (Figure 3F). In addition, inhibition zones and MICs of CRO, CAZ, ATM, SCF, PIP, and TZP against R1-ΔnagZ complemented with pBAD33 or pBAD33-nagZ vector were measured, and the results showed that increased inhibition zones induced by knockout of nagZ were reversed by complementation of NagZ (Supplementary Figure 2C). Consistently, knockout of nagZ significantly reduced MICs, which was rescued by complementation of NagZ as well (Table 1). In summary, NagZ promoted expression of ampC and β-lactamase activity, and enhanced resistance in strain of R1-ΔnagZ.

In P. aeruginosa, overproduction of the chromosomally encoded AmpC β-lactamase is the major mechanism of β-lactam resistance (Lodge et al., 1990; Kong et al., 2005). During normal physiological growth, N-acetylglucosaminyl-1,6-anhydromuropeptides (GlcNAc-1,6-anhydroMurNAc) are been transport into the cytoplasm by permease AmpG (Park and Uehara, 2008), where the glucosaminidase NagZ removes the GlcNAc moiety and form 1,6-anhydromuropeptides (anhMurNAc) (Park and Uehara, 2008; Ho et al., 2018). It has been proposed that anhMurNAc induces a conformational change of AmpR and maintains AmpR in an active conformation that promote the expression of ampC in P. aeruginosa (Caille et al., 2014), AmpR is a global transcriptional factor that regulates expression of hundreds of genes (such as rsmA, oxyR, rpoS, grpE, and phoP) (Kong et al., 2005; Caille et al., 2014). To explore the detail mechanism of NagZ promoting AmpC expression in E. cloacae, NagZ and AmpR Sequence homology were analyzed between P. aeruginosa and E. cloacae, the results revealed that E. cloacae NagZ (66.9%) and AmpR (100%) bears a high degree of homology to its counterpart P. aeruginosa (Figures 4A,B). A high degree of homology is also seen in the upstream region (transcriptional factor binding zone) of ampC between P. aeruginosa and E. cloacae (Figure 4C). Here, to further verify whether the regulation of NagZ on AmpC is mediated by the activation of AmpR by anhMurNAc, we examined the effect of NagZ on the expression of target genes of AmpR. The results indicate that NagZ can promote the expression of AmpR target genes such as rsmA, oxyR, rpoS, grpE, phoP, etc. (Figure 5). To further verify that the activation of AmpR is initiated by the NagZ hydrolyzate anhMurNAc, the effect of anhMurNAc on expression of AmpR target genes were examined. The results showed, consistent with NagZ, anhMurNAc could enhance the expression of rsmA, oxyR, rpoS, grpE, and phoP genes (Figure 5).