*2.5. Effects of Ampicillin*−*Resistant (AR) Gene Mutagenesis on Genetic Transformation Efficiency*

The transformation efficiencies of both ZM4∆0103 and ZM4∆ARs were significantly reduced using the plasmids of pEZ15A (~3 kb) and pE39−MVA (~10 kb). The electroporation efficiency of plasmid pEZ15A in ZM4 was (1.50 <sup>±</sup> 0.08) <sup>×</sup> <sup>10</sup><sup>5</sup> CFU/µg DNA, which was decreased to (1.66 <sup>±</sup> 1.98) <sup>×</sup> <sup>10</sup><sup>4</sup> and (2.66 <sup>±</sup> 1.11) <sup>×</sup> <sup>10</sup><sup>3</sup> CFU/µg DNA in ZM4∆0103 and ZM4∆ARs, respectively (Figure 4, Supplementary Materials Figure S5), with the transformation efficiencies reduced ca 10~100 fold. Few colonies grew on the plate after the electroporation of

pE39−MVA to ZM4∆0103 and ZM4∆ARs, while the efficiency of (1.35 <sup>±</sup> 0.01) <sup>×</sup> <sup>10</sup><sup>4</sup> CFU/µ<sup>g</sup> DNA in ZM4 was observed (Figure 4, Supplementary Materials Figure S6). The transformation efficiency of ZM4∆0103 and ZM4∆ARs cannot be calculated when the pE39−MVA plasmid with a large size was used. These results suggest that the predicted β−lactamase genes influenced the genetic transformation efficiency and significantly reduced the electroporation efficiency especially with the larger plasmid.

**Figure 4.** Electroporation efficiency of ZM4, ZM4∆0103, and ZM4∆ARs using plasmids of pEZ15A (~3 kb) and pE39−MVA (~10 kb). Three replicates were performed for the experiment. The error bar represents standard deviation (SD). When transformation of a plasmid was below the limit of detection (0.00001), the sample is marked "ND" (not detected). \*\*\* represents a very significant difference (*p*−value < 0.001).

#### **3. Discussion**

In this study, we first predicted six functional β−lactamase genes that may cause ampicillin antibiotic resistance in *Z. mobilis* ZM4 by bioinformatics analysis. ZMO0103, ZMO0893, and ZMO1650 belong to the AmpC superfamily containing similar conserved structures, indicating that they may have similar functions in cellular processes. ZMO1967 belongs to the PenP superfamily, which is associated with β−lactamase class A. ZMO1094 and ZMO1866 are not closely related to the function of β−lactams. We attempted to knock out all six of them in ZM4; unfortunately, only 5 of them were deleted individually using the CRISPR−Cas12a genome−editing system. The ZMO1967 protein has a transmembrane domain at the N−terminus with a probability of 0.98 by TMHMM−2.0 (https://dtu.biolib. com/DeepTMHMM, accessed on 24 August 2022). In addition, *ZMO1967* is probably an essential gene according to the results of the genome−wide CRISPRi (unpublished data), which may be the reason that it failed to be knocked out in this study. We successfully deleted five β−lactamase genes individually in ZM4, but the editing plasmid for *ZMO1866* deletion was not able to be eliminated.

Our study demonstrated that in addition to ampicillin resistance, these genes annotated as ampicillin−resistant (AR) genes had other effects on cell morphology, cell growth, and transformation efficiency. The knockouts *ZMO0893*, *ZMO1094*, and *ZMO1866* slightly inhibited the growth under the addition of 150 µg/mL. Especially, when *ZMO0103* was deleted, the growth of the strain was inhibited mostly with different concentrations of ampicillin. The biomass of ZM4∆0103 hardly increased under ampicillin ≥ 150 µg/mL. In addition, similar growth inhibition was also observed in four ampicillin−resistant (AR) genes of *ZMO0103*, *ZMO0893*, *ZMO1094* and the *ZMO1650* knockout strain ZM4∆ARs. The previous study reported that *ZMO0103* was a β−lactamase gene, which contains a 55%

amino acid sequence identity with class C β−lactamase genes [20]. A higher expression level [30,34–36] of *ZMO0103* in ZM4 may be ascribed to the absence of the AmpR that can inhibit the expression of *ZMO0103*. The result of multiple sequence alignment shows that the homologous protein of β−lactamase ZMO0103 was only found in *Sphingomonas* (Supplementary Materials Figures S7 and S8). Combining the result of the improved sensitivity of ZM4∆0103 to ampicillin in this study, we speculated that ZMO0103 is a unique protein and the most important β−lactamase in *Z. mobilis* ZM4, resulting in the high resistance of ZM4 to β−lactam antibiotics, such as ampicillin [17].

In addition, we also found that the growth of ZM4∆0103 and ZM4∆ARs was inhibited even without the addition of ampicillin. Considering the existence of transmembrane helices and signal peptides at the N−terminus (https://services.healthtech.dtu.dk/service. php?SignalP, accessed on 28 August 2022) of ZMO0103, ZMO0103 located on the cell membrane may directly hydrolyze ampicillin in the periplasmic space, and the integrity of the membrane could be disrupted by knocking out *ZMO0103* leading to the defective growth of ZM4∆0103 and ZM4∆ARs. Furthermore, cell sizes of ZM4∆0103 and ZM4∆ARs became longer and further lengthened with the addition of ampicillin indicating that the AmpC family lactamase protein ZMO0103 is related to cell wall biosynthesis and deconstruction and crucial for cell morphology and growth. Microscopic observations under RMG5 showed that the longest cells in ZM4∆0103 and ZM4∆ARs increased in length by 13.7 µm and 13.1 µm, respectively, compared to the longest cells in the wild−type ZM4. In addition, when 100 µg/mL of ampicillin (RMA100) was added, the cell lengths of the longest cells in ZM4∆0103 and ZM4∆ARs could increase by 7.6 µm and 6.6 µm, respectively, compared to the lengths of ZM4∆0103 and ZM4∆ARs under RM. These results suggest that the deletion of ZMO0103 affected the cell wall structure and the cell membrane and therefore led to changes in the intracellular osmotic pressure, thus enlarging the cells similar to the result of the previous study [37].

Based on our speculation that the deletion of *ZMO0103* may affect cell membrane and cell wall structures, we expected that the exogenous DNA could be more easily transferred into the cells. So, we further tested the transform efficiency of ZM4∆0103 and ZM4∆ARs using two plasmids of pEZ15A and pE39−MVA with different sizes. However, lower transformation efficiencies were observed when pEZ15A and pE39−MVA were electroporated into ZM4∆0103 and ZM4∆ARs compared with ZM4. Particularly, in the case of pE39−MVA, almost no transformants were obtained, which suggested that the transformation efficiency of the knockout strains was affected by plasmids sizes. However, the mechanisms of how *ZMO0103* influences the transformation efficiency remain to be investigated including the constructing a truncated mutant of *ZMO0103* by deleting the catalytic domain (amino acids from position 52 to 405) or the C−terminal domain (amino acids from position 406 to 520) that may influence the transformation of exogenous DNA. In addition, the restriction–modification (R–M) system genes (*ZMO0028*, *ZMO1933*, and *ZMOp32x025\_028*) could be knocked out in ampicillin−resistant gene knockout strains to further improve the transformation efficiency [38–40].

#### **4. Materials and Methods**

#### *4.1. Strains and Cultural Conditions*

*E. coli* DH5α was stored in our laboratory and used for plasmid maintenance and construction. During culturing, 50 µg/mL of chloramphenicol was added to Luria−Bertani (LB) medium (10 g/L tryptone, 5 g/L yeast extract, and 10 g/L NaCl, pH 7.0) for *E. coli* and cultured at 37 ◦C. *Z. mobilis* ZM4 was used as the parental strain for the construction of derived mutants and cultured with Rich Medium (RMG5) (50 g/L glucose, 10 g/L yeast extract, and 2 g/L KH2PO4) at 30 ◦C. When required, 50 µg/mL of chloramphenicol and 100 µg/mL of spectinomycin were added to the LB and RMG5. All *E. coli*, *Z. mobilis,* and their derivative strains used in this study are listed in the Supplementary Materials Table S3.

#### *4.2. In Silico Analysis of the AR Genes of Z. mobilis ZM4*

A BLASTP analysis of the *Z. mobilis* ZM4 proteome was performed using two databases for resistance genes: CARD (Comprehensive antibiotic resistance database) and MEGARes database. In addition to using BLASTP to investigate the relatedness of the sequences to those contained within the CARD and MEGARes databases, we further aligned with the gene sequences of the β−lactamase class in the UniProt database with the *Z. mobilis* ZM4 protein sequences. We chose the common genes of candidate β−lactamase genes after BLASTP with CARD and MEGARes databases and genes annotated as β−lactamase as candidate gene list 1. Then, we chose the common genes of candidate β−lactamase genes after BLASTP with the UniProt database and genes annotated as β−lactamase as candidate gene list 2. The common genes within both candidate gene list 1 and list 2 were the final candidate β−lactamase genes used for investigation in this work. The β−lactamase genes with a significant E−value of 2 <sup>×</sup> <sup>10</sup>−<sup>10</sup> were selected.
