Escherichia coli DH5α was cultured in Lysogeny Broth (LB, 10 g/L NaCl, 10 g/L tryptone, 5 g/L yeast extract) at 37°C with shaking at 220 rpm. Z. mobilis ZM4 (ATCC 31821) was cultured in Rich Medium (RM, 50 g/L glucose, 10 g/L yeast extract, 2 g/L KH₂PO₄) and Minimal Medium (MM, 50 g/L glucose, 1 g/L KH₂PO₄, 1 g/L K₂HPO₄, 1 g/L (NH₄)₂SO₄, 0.5 g/L NaCl, 0.42 g/L MgSO₄·7H₂O, 0.001 g/L calcium pantothenate) at 30°C with shaking at 100 rpm. When required, 50 μg/ml of chloramphenicol was added to the LB and RM medium, respectively. The final concentration of furfural or ethanol added to the RM medium was 2.60 g/L (RMF) and 47.36 g/L (RME), respectively. The final concentration of furfural or ethanol added to the MM medium was 1.25 g/L (MMF) and 20 g/L (MME), respectively.

Xylose mother liquor used in this study was provided by Zhejiang Huakang Pharmaceutical Co., Ltd. (Kaihua, Zhejiang, China). The original hydrolysate contained 216.87 g/L glucose, 103.98 g/L xylose, 4.44 g/L furfural, and 1.18 g/L acetic acid. The hydrolysate was diluted with ×10 RM⁻ medium (100 g/L yeast extract, 20 g/L KH₂PO₄) for ethanol fermentation.

Plasmid pL2R (Zheng et al., 2019) was used to knock out cas12 gene in the chromosome. Meanwhile, gene ZMO0038 was complemented at this position using the native Type I-F CRISPR-Cas system of Z. mobilis (Zheng et al., 2019). The spacer was 32 bp sequence which was immediately after a 5-NCC-3’ PAM in cas12a. The oligonucleotides (TsingKe, Beijing, China) of spacer were 36 bp with 4 bp protruding sequences in 5’-end. The editing plasmid was constructed following the previous description(Zheng et al., 2019). Briefly, the targeting gRNA sequence was constructed by annealed two single-stranded oligonucleotides (Cas12a-gR-F: 5’-gaa​atg​cgt​ttt​gaa​ctg​att​ccg​cag​ggt​aaa​acc, Cas12a-gR-R: 5’-gaa​cgg​ttt​tac​cct​gcg​gaa​tca​gtt​caa​aac​gca). Specifically, two single-stranded oligonucleotides were first heating at 95°C for 5 min and subsequently cooling down gradually to room temperature. Then the annealed spacer was ligated into BsaI-linearized pL2R by T4 ligase at 22°C for 3 h. The resulting plasmid was named as pL2R-cas12a.

Gibson assembly method as described before (Li et al., 2022) was utilized for donor construction. Donor sequences including extra ∼800 bp upstream sequence and downstream sequence of the candidate gene were amplified using Primer STAR polymerase (Takara, Japan) from the genomic DNA of Z. mobilis ZM4. The upstream sequence was amplified using the oligonucleotide primer pair 0038-US-F (5’-ggt​cac​cag​ctc​acc​gtc​tgt​tag​gcg​aga​agg​gaa​agg​g) and 0038-US-R (5’-gtt​ggg​ttg​agc​cgc​gat​agt​cgt​taa​ata​ttc​aga​tag​acg​gag​ata​ata​aac​g), and the downstream sequence was amplified using the oligonucleotide primer pair 0038-DS-F (tcacgcccgacgccag) and 0038-DS-R (gct​cga​gat​ctg​ata​tca​ctc​acc​ctc​tgg​tga​ttg​tcg​at). The oligonucleotide primer pairs pL2R-FK-F (agt​gat​atc​aga​tct​cga​gct​cgg​tac​ccg​g) and pL2R-FK-R (aga​cgg​tga​gct​ggt​gac​ct) were used to amplify the pL2R-cas12a vector. The upstream sequence and downstream sequence were then cloned into pL2R-cas12a vector by T5 exonuclease (NEB, WA, United States). The resulting plasmids were named as pL2R-cas12aD. The editing plasmid pL2R-cas12aD verified by colony PCR and Sanger sequencing were transformed into Z. mobilis ZMNP-Cas12a to construct the final strain ZMNP.

The editing plasmid pL2R-cas12aD was then transformed into Z. mobilis ZMNP-Cas12a competent cells (100 ng DNA with 50 μl competent cells) via electroporation using a Bio-Rad Gene Pulser (Bio-Rad, CA, United States). Immediately, the electroporated cells were transferred to 1 ml mating medium (50 g/L glucose, 10 g/L yeast extract, 5 g/L tryptone, 2.5 g/L (NH₄)₂SO₄, 0.2 g/L K₂HPO₄, 1 mM MgSO₄), and recovered at 30°C for 3 h. The cells were then spread on RM agar plates containing 50 μg/ml of chloramphenicol at 30°C for 2 days to isolate single colonies.

For the seed culture preparation, glycerol stock solution of Z. mobilis strains ZM4 and ZMNP were inoculated into 5-ml RM medium and then cultured at 30°C. The cultures were then transformed into 200 ml RM medium which was in 250-ml flasks, and then cultured at 30 C without shaking. The seed culture of mid-log phase was then inoculated into 50-ml shake flasks containing 40 ml RM, MM, RMF, MMF, RME, MME medium, 1/3 xylose mother liquor with an initial OD₆₀₀ value of 0.1. During the different time points of fermentation, the OD₆₀₀ value of cell culture was determined by an ultraviolet spectrophotometer UV 1800 (AOE, Shanghai, China). Simultaneously, samples collected at different time points were centrifuged at ×12,000 g for 2 min, and then the supernatants were filtered through a 0.22-μm filters and stored at −80°C for measuring the concentrations of glucose and ethanol by HPLC analysis.

For HPLC analysis, HPLC (LC-20AD, Shimadzu, Japan) with an Aminex HPX-87H column (Bio-Rad, CA, United States) were used to measure the glucose and ethanol concentrations at 65°C. At the same time, 5 mM H₂SO₄ was as mobile phase at a flow rate of 0.5 ml/min.

The sample ZMNP for whole-genome resequencing was collected and entrusted to GENEWIZ (Suzhou, China). The paired-end sequencing technology according to standard Illumina protocols by IgeneCode, Inc. (Beijing, China) was used in this work. The paired-end reads quality was checked using FastQC program (http://www.bioinformatics.babra.ham.ac.uk/projects/fastqc/). Data that passed the quality control were then mapped to the reference genome sequences of Z. mobilis ZM4 ATCC 31821 (GenBank accession No. of chromosome: NZ_CP023715, and plasmids: NZ_CP023716, NZ_CP023717, NZ_ CP023718, and NZ_CP023719) using the CLC Genomics Workbench (version 11.0) to identify the genomic variations. The objective mutations of the mutant strain ZMNP were obtained with the parental wild-type strain ZM4 as control. The mutation frequency which was more than 30% would be filtered.

The method of transcriptomic study is the same as reported previously (Yang et al., 2018b; He et al., 2018). Briefly, cell cultured under RM and MM medium were collected at the mid-log phase. Then total RNA extraction was using TRIzol reagent (Invitrogen, CA, United States). Then rRNA was removed from total RNA by using Ribo-off rRNA Depletion Kit (Bacteria NR407). Subsequently, mRNA was interrupted to short fragments by adding the fragmentation buffer. After synthesizing the first strand cDNA using random hexamer-primers, the second-strand cDNA was synthesized using buffer, dNTPs, RNase H and DNA polymerase I, respectively. The sequencing library were constructed by connecting fragments and sequencing adapters. And the transcriptome data were sequenced based on the Illumina NovaSeq 6000 System.

RNA-Seq fastq data which passed the quality control by FastQC program were imported into CLC Genomics Workbench (version 14.0) for reads trimming and RNA-Seq analysis to get the RPKM values of each gene with Z. mobilis ZM4 ATCC 31821 (GenBank accession No. of chromosome: NZ_CP023715, and plasmids: NZ_CP023716, NZ_CP023717, NZ_CP023718, NZ_CP023719) and four native plasmids as the reference genome. JMP Genomics (version 9.0) was used to normalize gene expression, analyze variance (ANOVA) and hierarchical clustering to identify differentially expressed genes at different conditions. Differentially expressed genes were determined with a selection threshold of p-value ≤ 0.01 and log₂-fold change≥ ± 1 (significant induction).

The cells cultured in MM medium were collected and suspended into a 1 × RIPA buffer (×10: 150 mM NaCl, 10% NP40, 10% sodium deoxycholate, 10% SDS, 250 mM Tris–HCl, pH 7.6) at mid-log phase. Samples were grinded under low temperature for 5 min, and then ultrasonication on ice for 5 min. Subsequently, samples were kept at 4°C for 2 h and then centrifuged for 15 min at 4°C and ×12,000 g. 20 μl supernatant liquor containing extracted proteins of each sample was added into 96-well plate containing BCA buffer (Thermo Scientific, Rockford, United States). Then 96-well plate was shock at 37°C for 30 min, and the absorbance was detected at 562 nm (Thermo Scientific, CA, United States). The standard curve was then fitted, and the protein concentration of the corresponding sample is calculated. 100 μg of each sample was diluted to ∼1 mg/ml with 1 × RIPA lysis buffer.

Then each sample was precipitated by acetone overnight at −20°C. Subsequently, the protein precipitation was collected by centrifugation for 10 min at 4°C and ×12,000 g and then rewashed with pre-chilled 80% acetone two times. The protein precipitation was re-dissolved, reduced, and alkylated. The sequencing grade trypsin was then added at the ratio of 1:50 (wt:wt), and the digestion was run overnight at 37°C. The final protein precipitation was treated by sodium deoxycholate (SDC) cleanup and peptide desalting.

1 μg total peptides of each sample were separated and analyzed with a nano-UPLC (EASY- nLC1200) coupled to a Q Exactive HFX Orbitrap instrument (Thermo Fisher Scientific, CA, United States) with a nano-electrospray ion source. A reversed phase column (100 μm ID ×15 cm, Reprosil Pur 120 C18-AQ, 1.9 μm, Dr. Maisch) was used to separate different sizes of peptides. Mobile phases were H₂O with 0.1% FA, 2% ACN (phase A) and 80% ACN, 0.1% FA (phase B). Each ample was separated by executed with a 120 min gradient at 300 nL/min flow rate. Gradient B: 2%–5% for 2 min, 5%–22% for 88 min, 22%–45% for 26 min, 45%–95% for 2 min, 95% for 2 min. Data dependent acquisition (DDA) was performed in profile and positive mode with Orbitrap analyzer at a resolution of 120,000 (@200 m/z) and m/z range of 350–1,600 for MS1; For MS2, the resolution was set to 15,000 with a dynamic first mass. The automatic gain control (AGC) target for MS1 was set to 3E6 with max IT 50 ms, and 1E5 for MS2 with max IT 110 ms. The top 20 of most intense ions were fragmented by HCD with normalized collision energy (NCE) of 27%, and isolation window of 1.2 m/z. The dynamic exclusion time window was 45 s, single charged peaks and peaks with charge exceeding six were excluded from the DDA procedure.

Proteome Discoverer (PD) software (Version 2.4.0.305) and the built-in Sequest HT search engine were used to process Vendor’s raw MS files. MS spectra lists were searched against their species-level UniProt FASTA databases (uniprot-Mus + musculus-10090-2020-10.fasta). Here, carbamidomethyl (C) as a fixed modification, oxidation (M) and acetyl (N-term) as variable modifications, and trypsin as proteases. A maximum of two missed cleavage(s) was allowed. The threshold of false discovery rate (FDR) was 0.01 for both PSM and peptide levels. Peptide identification was performed with an initial precursor mass deviation of up to 10 ppm and a fragment mass deviation of 0.02 Da. Unique peptide and Razor peptide were used for protein quantification and total peptide amount for normalization. All other parameters were reserved as default.

Here, the measurement of intracellular ROS levels were using 2', 7'-dichlorodihydrofluorescein diacetate (H₂DCF-DA) (Beyotime Biotechnology, Hangzhou, China). When cells are cultured to mid-log phase, 0.6 OD₆₀₀ cells were collected, and then washed with 1×phosphate-buffered saline (PBS) once. The pellets were re-suspended with 500 μl PBS, then added H₂DCF-DA at a final concentration of 100 μM. The mixture then incubated in darkness for 1 h at 30°C, 100 rpm. After that, cells were collected and washed three times with 1 × PBS. 300 μl 1 × PBS were used to re-suspended the pellets. The mixture then detected by the DCF fluorescence using CytoFLEX FCM flow cytometry (Beckman coulter, CA, United States). The excitation and emission wavelength were set up at 488 nm and 525 nm, respectively. Cells with fluorescence intensities ranging 10³–10⁵ were selected and counted, and at least 20,000 events were collected for each sample.

We previously constructed a mutant strain ZMNP-Cas12a that contains a cassette of TetR-Ptet-cas12a that replaced ZMO0038 locating between ZMO0037 and ZMO0039. Compared to wild-type ZM4, ZMNP-Cas12a also lacks four native plasmids that were cured by CRISPR-Cas system. Here, we further constructed the mutant ZMNP with the cassette TetR-Ptet-cas12a replaced with the native ZMO0038 by native type I-F CRISPR system (Figure 1). The final mutant strain ZMNP only lacks four native plasmids compared to ZM4.

Temperature is one of the common physical factors affecting cell growth and microbial fermentation. Moreover, high-temperature ethanol fermentation has advantages of reduced pollution risk as well as cooling costs making it suitable for large-scale bioethanol fermentation (Li et al., 2021). Substrate is another factor that restricts the cost of microbial fermentation. Corncob is abundant with ca 250 million tons produced each year, which is an excellent cellulosic material for commercial xylose and xylitol production. However, the industrial effluent of xylose mother liquor accompanying the xylose and xylitol production becomes a problem. We evaluated cell growth and fermentation performance of ZM4 and ZMNP under different temperatures and different media including xylose mother liquor.

Cell growth and fermentation performance of ZMNP using glucose under different temperatures or using the xylose mother liquor were compared with ZM4, and the results demonstrated that ZMNP performed similarly to ZM4 using glucose with a growth rate of 0.525 h⁻¹ vs 0.512 h⁻¹ as well as ethanol productivity of 2.44 g/L h⁻¹ vs 2.38 g/L h⁻¹ when cultured under 30°C (Figure 2A), and with a growth rate of 0.395 h^-1 vs 0.399 h⁻¹ as well as ethanol productivity of 2.66 g/L h⁻¹ vs 2.67 g/L h⁻¹ when cultured under 40°C (Figure 2B). However, ZMNP did exhibit better glucose consumption and ethanol production than ZM4 when the toxic xylose mother liquor was used. The glucose utilization in ZMNP was 24 h faster than that observed for ZM4, and the maximum ethanol titers of 39.13 g/L and 36.91 g/L were achieved by ZMNP and ZM4, respectively (Figure 2C).

The normal growth and fermentation performance of ZMNP at high temperature and its excellent fermentation performance using xylose mother liquor exhibited that ZMNP could be an ideal chassis for biochemical production from lignocellulosic materials.

To determine the potential genetic determinants of ZMNP for increased utilization of xylose mother liquor, next-generation sequencing (NGS) and third-generation sequencing (TGS) technology were applied to identify the potential genetic changes in ZMNP. Combining the data collected from NGS and TGS, we assembled the genome of ZMNP to explore the potential genetic determinants for native plasmid deletion through comparative genomic analysis using the genome of parental strain ZM4 (ATCC 31821) as the reference (Yang et al., 2018b). The genome of ZMNP contains one circular chromosome of 2,058,754 bp only without plasmids that exist in the wild-type ZM4.

The whole-genome resequencing (WGR) results identified a total of five single nucleotide polymorphisms (SNPs) that occurred in open reading frames (ORF) in ZMNP (Table 1 ). Among these mutations, two missense mutations were identified in ZMNP located in the coding sequence region of ZMO0651 (D172N) and ZMO1733 (S267P), respectively. ZMO0651 encodes a flagellar hook protein FliD, which is a flagellar cap at the distal end of the flagellar filament protecting the tip of the flagellum (Nedeljkovic et al., 2021). FliD also helps insert flagellin proteins repetitively to grow the flagellar filament (Cho et al., 2019). It was observed that the motility was defective in fliD mutants because the flagellin monomers were shed into the outside of cell causing the failure of flagellin polymerization and the form a functional flagellum, revealing that fliD was an essential factor in cell motility (Arora et al., 1998; Kim et al., 1999). We simulated the 3D structure of the FliD mutant, and it showed that this missense mutation didn’t cause changes in the 3D structure (data not shown). Therefore, we speculate that this point mutation D172N mayn’t affect its protein function.

The S267P mutation of OxyR is in LysR-substrate binding domain (IPR005119, 93-295 aa), which probably changes the binding affinity with its substrate (H₂O₂) due to the amino acid change from serine to proline, resulting in a protrusion at the 267 aa (Figure 3). Previous research reported that residue 266 aa was an important residue in reduced monomer conformation of OxyR (Anand et al., 2020), and beneficial to stabilize reduced tetramer via side-chain interactions (Anand et al., 2020). We speculated that the mutation at the 267 aa might affect the stabilization of the OxyR reduction state, making it more prone to transition to the oxidation state by altering the transcription initiation, and further respond to intracellular oxidative stress signals (Figure 1). This may be one of the reasons that ZMNP is better at utilizing xylose mother liquor.