Outer membrane protein of OmpF contributes to swimming motility, biofilm formation, osmotic response as well as the transcription of maltose metabolic genes in Citrobacter werkmanii

Bacterial outer membrane proteins (Omps) are essential for environmental sensing, stress responses, and substance transport. Our previous study discovered that OmpA contributes to planktonic growth, biocide resistance, biofilm formation, and swimming motility in Citrobacter werkmanii, whereas the molecular functions of OmpF in this strain are largely unknown. Thus, in this study, the ompF gene was firstly knocked out from the genome of C. werkmanii using a homologous recombination method, and its phenotypical alternations of ∆ompF were then thoroughly characterized using biochemical and molecular approaches with the parental wild type (WT) and complementary (∆ompF-com) strains. The results demonstrated that the swimming ability of ∆ompF on semi-solid plates was reduced compared to WT due to the down-regulation of flgC, flgH, fliK, and fliF. Meanwhile, ompF deletion reduces biofilm formation on both glass and polystyrene surfaces due to decreased cell aggregation. Furthermore, ompF inactivation induced different osmotic stress (carbon sources and metal ions) responses in its biofilms when compared to WT and ∆ompF-com. Finally, a total of 6 maltose metabolic genes of lamB, malE, malK, malG, malM, and malF were all up-regulated in ∆ompF. The gene knockout and HPLC results revealed that the MalEFGK2 cluster was primarily responsible for maltose transport in C. werkmanii. Furthermore, we discovered for the first time that the upstream promoter of OmpF and its transcription can be combined with and negatively regulated by MalT. Overall, OmpF plays a role in a variety of biochemical processes and molecular functions in C. werkmanii, and it may even act as a targeted site to inhibit biofilm formation.


Introduction
Citrobacter belongs to Enterobacteriaceae family, and a great number of Citrobacter sp. frequently cause serious opportunistic infections, particularly in the urinary and respiratory tracts, as well as in enteric diseases (Hodges et al. 1978;Katzenellenbogen et al. 2008). Meanwhile, some Citrobacter sp. has been used for heavy metal bioremediation, such as uranyl ion (Finlay et al. 1999;Jeong et al. 1997;Macaskie et al. 1995), and the production of bio-based products, such as 1, 3-propanediol (Maervoet et al. 2012(Maervoet et al. , 2014(Maervoet et al. , 2016. Undeniably, only by fully mastering the response mechanisms of Citrobacter sp. to external nutritional and environmental stimuli can the strains be more fully controlled and utilized. Bacterial cell membranes are well known for their importance in providing bacteria with a protective and functional barrier, as well as surface specificity (Salton 1967). Gramnegative bacteria capture double membranes, with the outermost one containing outer membrane proteins (Omps) and other components (Enjalbert et al. 2006). In particular, an outer membrane porin protein of OmpF is widely distributed in Gram-negative bacteria, and its expression is usually regulated by a variety of environmental factors (Song and Leff 2006). It has been reported that nutrient limitation controls ompF expression in Escherichia coli, and that a glucose limitation stimulated more OmpF activities than an ammonia limitation (Long et al. 2020).
Furthermore, OmpF was involved in a number of bacterial responses to antibiotic exposure, acidic resistance, hyperosmotic adaptation, and bacterial virulence. The decreased expression of ompF in E. coli did not affect bacterial survival, but it was responsible for the spread of quinolone resistance (Cruz et al. 2012). Similarly, the absence of OmpF in Serratia marcescens reduced antibiotic permeability and resulted in significantly higher antibiotic minimum inhibitory concentration (MIC) values for beta-lactam drugs such as ampicillin and cefoxitin, as well as nitrofurantoin (Otto and Hermansson 2004). The OmpF-deficient mutant was resistant to a variety of antibiotics, implying that OmpF serves as the primary route of outer membrane penetration for many antibiotics (Stoorvogel et al. 1991). OmpF also collaborates with the TolA or TonB systems to translocate some bacteriocin (Bootman and Berridge 1995). Meanwhile, when ompF was deleted from E. coli genome, its acidic resistance of ΔompF decreased, and the addition of glutamate but not arginine or lysine intensified or elevated this resistance, indicating that OmpF is required for E. coli cell survival under extremely acidic conditions, and the transportation of arginine, lysine, and their decarboxylated products was primarily via OmpF (Briones et al. 2022). Moreover, OmpF in E. coli or Yersinia pestis has been shown to participate in the hyperosmotic adaptation by altering its protein synthesis and the RNA expression levels of micF (Brown et al. 1995;Calderón et al. 2011;Smith 1995;Wang et al. 2019). ΔompF of E. coli significantly reduced its adherence and invasion capabilities to mouse brain microvascular endothelial cells, and this mutant also reduced the bacterial virulence in both ducklings and mouse models when compared to the wild type (WT) strain, which could be attributed to lower expression levels of ompA, fimC, and iBeA (Dupont et al. 2004). Collectively, there have been extensive studies that have characterized OmpF biochemical and molecular features, but its regulatory networks, additional functions, and regulated mechanisms, particularly in C. werkmanii, remain elusive.
In this study, ompF was deleted from C. werkmanii, which was isolated from industrial putrefaction in our laboratory . The phenotypic and genetic properties of ΔompF were studied in conjunction with its parental WT. The results demonstrated that ompF contributes to swimming motility, biofilm formation, and osmotic responses. In addition, we also discovered that werkmanii transports maltose primarily via the MalEFGK2 cluster and that MalT can bind to the promoter of OmpF and influence its transcription. Our findings shed light on the molecular functions of ompF and pave the way for more effective use of OmpF as a potential target site for biofilm controls.

Bacterial strain, culture conditions, and chemicals
Citrobacter werkmanii BF-6 (WT) was originally isolated from an industrial putrefaction sample by ourselves and deposited in the Guangdong Culture Collection Center (Guangzhou, Guangdong, China) under the accession number GDMCC 1.1242 . Meanwhile, its genome has been completely sequenced and is available at the NCBI under the accession number NZ_CP019986.1. Both BF-6 and E. coli S17-1 were routinely cultured in a liquid Luria Bertani (LB) medium at 30 °C and 37 °C under static or shaking conditions, respectively. Unless otherwise specified, all chemicals used in this study were reagent grade and obtained from Sigma (St Louis, MO, USA). All strains and plasmids used in this study, as well as their properties, were listed in additional file 1.

Genetic management
The gene deletion and complementation of ompF in C. werkmanii were carried out using previously described methods with minor modifications (Zhou et al. 2018). Briefly, the upstream and downstream of ompF (locus_tag = B2G73_ RS09925) were amplified with the primer pairs of ompFup-F/R and ompF-down-F/R using PrimeSTAR® Max DNA Polymerase (TaKaRa, Dalian, China) and the C. werkmanii genome as a template, respectively. To create a knockout plasmid of pYG4-ompF, the confirmed upstream and downstream of ompF were mixed with the linearized plasmid of pYG4 with BglII (TaKaRa) and In-Fusion® HD Cloning Kit (TaKaRa) according to the manufacturer's instructions. Following that, the pYG4-ompF was used to transform competent E. coli S17-1 cells, which were then mobilized into WT via biparental mating. The exconjugants and deletion mutant candidates of ΔompF were screened on solid LB medium supplemented with kanamycin (50 mg/L), rifampicin (100 mg/L), or sucrose (5%), and then confirmed using the ompF-QJ-F/R. Furthermore, the full length of the ompF gene with its upstream promoter was amplified with ompFcom-F/R and then cloned into the shuttle vector pSRK-GM linearized with XbaI using the In-Fusion® HD Cloning Kit (TaKaRa). The recombinant plasmid pSRK-GM-ompF-com was then mobilized into ∆ompF with the assistance of E. coli S17-1. The complemented strain was confirmed with ompFcom-F/R and named ∆ompF-com. Moreover, deletions of six genes from the maltose metabolic pathway, including lamb, malE, malK, malG, malM, and malF, as well as their regulated gene malT, were conducted according to the methods mentioned above. While for the constructions of double gene mutants of both ompF and maltose-related genes, each constructed knockout plasmid of maltose-related genes was transformed into ∆ompF as the recipient objects. In addition, all primers used in this study are listed in additional file 1.

Growth curve determination
Overnight cultures of WT, ∆ompF, and ∆ompF-com were diluted with fresh LB medium to prepare suspensions with an optical density (OD) of 0.1 at 600 nm. Then, an aliquot of 200 μl of each suspension was transferred into 96-well plates (Corning Incorporated, Corning, NY, USA) for growth curve determination using a plate reader (Tecan Spark, Tecan, Switzerland). The inoculated plates were incubated at 30 °C for 48 h, and the bacterial growths with eight replicates were recorded every 30 min by measuring their OD 600 .

Aggregation and biofilm formation in glass tubes
The aggregation and biofilm formation assays were performed according to the methods as described previously with slight modifications (Shanks et al. 2008;Stepanović et al. 2000). Briefly, transparent glass bottles with caps (Ø = 27.5 mm, height = 72.5 mm; Guangdong Huankai Microbial Science & Technology, Guangzhou, China) were filled with 18 ml LB medium and 2 ml of WT, ∆ompF or ∆ompF-com cells (initial OD 600 = 1.0). After that, all of the inoculated glass bottles were placed in an incubator and cultured vertically at 30 °C in a static condition. After 4 days of cultivation, the upper 1 ml of the cultures was carefully pipetted from each tube, and its OD 600 was measured and recorded as OD 600 pre-vortex. After thoroughly stirring the residues to resuspend the aggregated cells, their OD 600 was recorded as OD 600 post-vortex. Finally, the following formula: 100% × (OD 600 post-vortex − OD 600 pre-vortex)/ OD 600 post-vortex was used to calculate the aggregation percent. In addition, 25 ml of 0.1% crystal violet was then added to the glass bottles to stain the biofilms that had formed on the inner walls and the attached crystal violet was finally dissolved in 95% ethanol to measure their OD 590 for assaying their biofilm biomass (Zhou et al. 2013).

Swimming ability assessment
The swimming abilities of C. werkmanii WT, ∆ompF, and ∆ompF-com on semi-solid plates were determined according to the procedures reported previously with minor modifications (Rashid and Kornberg 2000). Briefly, 2 µL of bacterial suspensions (initial OD 600 = 1.0) were inoculated at the center of dried swimming plates containing 10 g/L tryptone, 5 g/L NaCl, and 0.3% agarose. Subsequently, the inoculated plates were gently placed in an incubator and cultured at 30 °C for 24 h and then taken out to measure the diameters of all colonies using a digital Vernier caliper (CD-20CP, Kawasaki, Japan). The swimming motility experiments described above were conducted in triplicate and repeated at least twice.

Effects of diverse environmental factors on bacterial growth and biofilm formation
To evaluate the effects of several nutritional and environmental factors on the planktonic growth and biofilm formation of WT, ∆ompF, and ∆ompF-com, the microtiter wells were inoculated with bacterial suspensions (OD 600 = 0.10) and the same volumes of LB medium supplemented with the following reagents: 400 mM sucrose (final concentration, the same as below) or xylitol or glucose, 40 mM or 400 mM CaCl 2 , 160 mM or 320 mM MgCl 2 . After 4 days of static culture at 30 °C, planktonic growth and biofilm formation were determined using a Multiskan GO reader at 600 nm (OD 600 ) and 590 nm (OD 590 ), respectively, according to the methods described previously (Zhou et al. 2013). Eight replicate wells were inoculated for each concentration, and cultures without any reagents served as controls.

Transcriptome sequencing and data analysis
Citrobacter werkmanii WT and ∆ompF were cultured in 96-well microtiter plates at 30 °C for 4 days under a static condition. And then, the cultured planktonic cells of these two strains were pipetted out from the plates and their pellets were harvested by centrifugation, before sending all samples to the Guangzhou Meige Biotechnology Company for RNA isolation and library preparation (Zhou et al. 2019(Zhou et al. , 2016a. Following cluster generation, the library preparations were sequenced on an Illumina Novaseq platform. Differential gene expression analyses of WT and ∆ompF were performed using the DESeq R package (1.18.0). The obtained genes identified by DESeq with an adjusted p-value < 0.05 were assigned as differentially expressed genes (DEGs). Meanwhile, the GOseq R package was used to perform GO enrichment analysis on DEGs, and GO terms with corrected p-value less than 0.05 were considered significant. Moreover, the KOBAS software was used to test the statistical enrichment of DEGs in the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways.

qRT-PCR
Both WT and ∆ompF were cultured at 30 °C for 4 days in fresh LB liquid medium and then their planktonic cells were collected independently. Their total RNAs were extracted and translated into cDNA using TRNzol Reagent (Tiangen, Beijing, China) and PrimeScript RT Master Mix (TaKaRa, Dalian, China) according to the manufacturer's instructions. The transcriptional levels of flagellar assembly-related genes were determined via qRT-PCR with paired primers and tenfold dilutions of the samples' cDNA as templates (Additional file 1). All qRT-PCRs were performed on a Mastercycler ep realplex (Eppendorf, Hamburg, Germany) using TB Green Premix Ex Taq II (TaKaRa) following the user's guides. The transcription level of each gene in the cDNAs was calculated using the 2 −∆∆Ct method with rpsL serving as an internal standard (Livak and Schmittgen 2001).

Disk diffusion tests for resistance to H 2 O 2 and polymyxin B
The obtained RNA-Seq results show that ompF deletion altered the expression of some genes involved in the cationic antimicrobial peptide (CAMP) resistance, benzoate degradation, and quorum sensing. We hypothesized that deleting ompF would result in altered susceptibility to certain biocides. Therefore, the antimicrobial sensitivity of WT, ∆ompF, and ∆ompF-com to H 2 O 2 and polymyxin B was measured using disk diffusion methods (Hall 2013). Briefly, the entire surface of LB plates (90 mm in diameter) was first covered with the suspensions of WT, ∆ompF, or ∆ompF-com, and the inoculated plates were air-dried at room temperature for at least 2 h before use. The disks containing different concentrations of H 2 O 2 and polymyxin B (6 mm in diameter) were gently and carefully placed on the dried plates. After 48 h of static cultivation at 30 °C, the diameters of the inhibition zones for each strain were measured and recorded using a digital Vernier caliper (Kawasaki).

Assay of maltose consumption rates
The RNA-Seq results also revealed that 6 genes in the maltose metabolic pathway were promoted in ∆ompF when compared to WT. Therefore, the maltose metabolic rates of C. werkmanii WT, ∆ompF, ∆lamB, ∆malE, ∆malK, ∆malG, ∆malM, ∆malF, ∆ompF∆lamB, ∆ompF∆malE, ∆ompF∆malK, ∆ompF∆malG, ∆ompF∆malM, and ∆ompF∆malF, as well as ∆malT and ∆ompF∆malT, were determined using HPLC according to the methods described previously (Whitman 2015) with some modifications. Briefly, all of the above strains were cultivated in glass bottles at 30 °C for 4 days before being harvested separately by centrifugation at 16,000×g for 15 min to collect the supernatants. After filtration with a 0.22 μm filter, the supernatants were then analyzed using an HPLC system on Thermo U3000 equipped with RefractoMax 520 RI detector. Maltose was separated and analyzed on a Phenomenex Rezex™ ROA-Organic Acid column (7.5 × 300 mm) that was heated to 65 °C with a flow rate of 0.6 ml/min and 0.005 N sulfuric acid in water as a mobile phase. Meanwhile, a standard curve was constructed using synthetic maltose purchased from Biolog (San Diego, CA), and metabolite concentrations are expressed as g/L of dough.

MalT expression and electrophoretic mobility shift assay (EMSA)
MalT protein expression and purification were performed according to the previous methods with slight modifications (Long et al. 2020;Zhao et al. 2021). Briefly, the entire coding sequence of malT was cloned using malT-FL-F/R by PCR amplification and the amplified PCR fragment was sequenced by the Tsingke Corporation (Beijing, China) to confirm its accuracy. The confirmed MalT coding sequence was inserted into pET-28a after cutting with NdeI and XhoI in frame with the C-terminal 6 × His-tag using a one-step cloning kit (Vazyme Biotech, Nanjing, China) following the protocols provided by the manufacturer. The resulting recombinant plasmid of pET-28a-malT-His was then introduced into E. coli BL21 (DE3) and the transformants were cultured at 37 °C in the LB medium containing 100 μg/ml of kanamycin to an OD 600 of 0.5-0.6. Subsequently, 1 mM (final concentration) isopropyl β-d-1-thiogalactopyranoside (IPTG) was added to induce recombinant plasmid overexpression. The over-expressed cells were subsequently harvested by centrifugation at 10,000×g for 20 min and then lysed by ultrasonication in lysis buffer (pH 8.0) containing 10 mM Tris-HCl, 10% glycerol, and 50 mM NaCl. After centrifugation at 10,000×g for 20 min at 4 °C, the supernatant was loaded onto a Ni-nitrilotriacetic acid (Ni-NTA) column and the 6 × His-tagged MalT was eluted using buffers containing varying concentrations of imidazole (50 mM, 100 mM, 200 mM, 300 mM, and 400 mM).
Moreover, the EMSA assay was conducted as previously described methods with a slight modification (Sun et al. 2014). Briefly, the DNA fragments (10 nM) corresponding to the upstream of ompF were amplified with ompF-P-F/R (Additional file 1) and were then incubated with purified 6 × His-tagged MalT protein (0 to 6 μm) at 25 °C for 30 min in a 20 μl reaction system (pH 8.0) containing 25 mM Tris-HCl, 1 mM CaCl 2 , 10 mM MgCl 2 , 50 mM KCl, 5% (vol/vol) glycerol, and 0.1 mM EDTA. The samples were subsequently loaded onto a 5% native polyacrylamide gel that had been pre-run on ice for 1 h in a 0.25 × Tris-Glycine (TGE) buffer (pH 8.0) containing 25 mmol/L Tris base and 192 mmol/L glycines. After 1.5 h of electrophoresis at 100 V, the gels were stained with 0.1% Coomassie brilliant blue R250 to visualize the bands with a molecular imager (ChemiDoc XRS+; Bio-Rad).

Construction of gfp transcriptional fusions and measurement of fluorescence activity
Fragments of the ompF promoter regions were amplified by PCR using primer pairs of ompF-gfp-F/R (Additional file 1) and WT chromosomal DNA as the template. The PCR products were cloned into the plasmid pUCP254-gfp, which was linearized with XbaI and PstI to produce pUCP254-gfp-PompA, which was then transferred into WT and ΔmalT. The overnight cultured bacterial suspensions were diluted to an optical density (OD 600 ) of 1.0 before being added to fresh LB medium for 2 days of cultivation at 30 °C. The intensity of GFP fluorescence in WT and ΔmalT was detected by the Multiskan (Synergy H1, BioTek, USA) with excitation at 479 nm and emission at 520 nm. Meanwhile, the cells were also collected by centrifugation at 10,000 g for 1 min and examined under a fluorescence microscope (ECLIPSE NI, Nikon, Japan) after being washed twice with PBS.

Statistical analysis
All data obtained in this study were recorded as the mean ± standard deviation (SD) and were also subjected to one-way ANOVA followed by a comparison of multiple treatment levels with the control using Fisher's LSD test.
All statistical calculations were performed using data processing system (DPS) software (Tang and Feng 2007) and a p-value < 0.05 was considered significant.

Data accessibility
The transcriptome sequencing data supporting the results of this study have been deposited in the NCBI Sequence Read Archive (SRA) under the BioProject ID PRJNA790646.

Results
ompF contributes to late growth and inhibits the swimming motility of C. werkmanii ∆ompF was constructed using the above-mentioned homologous recombination method, and its growth curves under normal conditions were observed alongside WT and ∆ompFcom. The results showed that there are no significant differences in growth curves among these three strains until 36 h, but ∆ompF obtained lower growth biomass than WT and ∆ompF-com from 36 to 48 h (Fig. 1A), indicating that ompF does not affect C. werkmanii growth at the early and logarithmic stages, but decreases growth at the late stage and accelerates the occurrences of stationary or decline phases. In addition, the swimming diameters of ∆ompF colonies (Fig. 1C) were smaller than those of both WT (Fig. 1B) and ∆ompF-com (Fig. 1D) when grown on semi-solid media WT, ∆ompF, and ∆ompF-com were incubated on a plate reader (Tecan Spark), and their growth rates were measured for successive 48 h with a 30 min interval using a Multiskan GO reader at 600 nm. The swimming abilities of WT, ∆ompF, and ∆ompF-com were determined on semi-solid plates, and the diameters of the colonies were measured with a digital Vernier caliper, and the representative images were presented. The relative expression levels of four genes of flgC, flgH, fliK, and fliF were detected using RNA-Seq and RT-PCR (p < 0.05), indicating that ompF is required for swimming motilities of C. werkmanii. Meanwhile, RNA-Seq results revealed that a large number of flagellar assembly-related genes were repressed in ∆ompF (Additional file 2). Therefore, the four genes flgC, flgH, fliK, and fliF with the greatest fold-change were chosen to determine and confirm their relative expression levels using the qRT-PCR method once more. The results demonstrated that both qRT-PCR and RNA-Seq results showed similar diminishing trends for these four genes (Fig. 1E), confirming the accuracy of RNA-Seq assays and implying that the decreased swimming abilities were caused in part by the down-regulation of flagellar assembly-related genes.

Both cell aggregation and biofilm formation require ompF
When C. werkmanii was cultured in glass tubes in a static condition, the inactivation of ompF from its genome decreased its aggregation ability ( Fig. 2A). Moreover, the biofilm formation of ∆ompF on glass surfaces was significantly reduced when compared to either WT or ∆ompF-com (Fig. 2B). These results suggest that ompF is required for C. werkmanii cell aggregation and biofilm formation.

ompF is involved in the biofilm formation on polystyrene surfaces and osmotic stress responses
Since ompF contributes to swimming abilities, cell aggregation, and biofilm formation on glass surfaces, we hypothesized that this gene might also be involved in other biochemical processes. As expected, we discovered that ∆ompF possessed less planktonic growth and biofilm formation than WT and ∆ompF-com (p < 0.05) when cultured in microtiter plates made of polystyrene (Figs. 3A and 3E).
In the presence of 400 mM sucrose (Fig. 3B) or xylitol (Fig. 3C), there were no differences in planktonic growth among WT, ∆ompF, and ∆ompF-com, but ∆ompF formed fewer biofilms (Figs. 3F and 3G); ∆ompF showed a similar planktonic growth rate to WT and ∆ompF-com (Fig. 3D) but formed more biofilms than these two strains in the presence of 400 mM glucose (Fig. 3H); When exposed to 40 mM (Fig. 3I) or 400 mM (Fig. 3J) Ca 2+ , ∆ompF exhibited a similar planktonic growth rate to WT and ∆ompF-com but formed fewer biofilms than these two strains (Figs. 3M and 3N); In the presence of 160 mM Mg 2+ , both planktonic growth (Fig. 3K) and biofilm formation (Fig. 3O) of ∆ompF are decreased compared to either WT or ∆ompF-com. However, at 320 mM Mg 2+ , these three strains did not vary in terms of planktonic growth (Fig. 3L) or biofilm formation (Fig. 3P).
In particular, the diameters of inhibition zones of WT, ∆ompF, and ∆ompF-com were almost the same in the presence of 20 mM, 40 mM, and 80 mM H 2 O 2 (Additional file 3). Similarly, inhibition zones of polymyxin B at 12 μg/ml, 24 μg/ml, and 48 μg/ml did not differ among WT, ∆ompF, and ∆ompF-com (Additional file 3). These results imply that the susceptibilities of WT, ∆ompF, and ∆ompFcom to H 2 O 2 or polymyxin B were comparable. Fig. 2 ∆ompF showed lower aggregation (A) and fewer biofilms (B) than C. werkmanii WT and ∆ompF-com in glass tubes. WT, ∆ompF, and ∆ompF-com were incubated in glass tubes under static conditions for 4 days, and their aggregation was calculated using the following formula: 100%× (OD 600 post-vortex − OD 600 pre-vortex)/OD 600 postvortex, where OD 600 pre-vortex and OD 600 post-vortex represent the biomass of each strain before and after vortex, respectively. The biofilms that formed in the glass tubes were stained with 0.1% crystal violet, resolved in 95% ethanol, and finally measured at OD 590 . All of the above experiments were repeated at least three times on different days, and the mean results or representative images were presented. Values having different letters are significantly different from each other according to Fisher's LSD test (p < 0.05) In combination, ompF contributes to biofilm formation on both hydrophilic (glass) and hydrophobic (polystyrene) surfaces, but not to resistance to H 2 O 2 and polymyxin B. Specifically, ompF does not contribute to planktonic growth but is involved in biofilm formation in the presence of 400 mM sucrose, xylitol, or glucose. Similarly, when exposed to 40 mM or 400 mM Ca 2+ , ompF is a positive regulator of biofilm formation but not of planktonic growth. In addition, ompF is only important in planktonic growth and biofilm formation at Mg 2+ concentrations of 160 mM but not 320 mM.

DEGs that exist between ∆ompF and WT, as well as their functional enrichments
Since ompF participates in various bioprocesses in C. werkmanii, we hypothesized that the knockout of ompF would result in altered gene expression profiles in WT. To testify this hypothesis, an RNA-Seq technique was employed to find DEGs between ∆ompF and WT. As expected, there are 1,791 DEGs whose expression levels are differentially regulated (Padj < 0.05, | log 2 (fold change) |> 0.6). Moreover, 989 and 802 genes were activated and repressed, respectively, in ∆ompF when compared to WT. Furthermore, Gene Ontology (GO) analyses were also employed to classify the functions of all DEGs (including activated and repressed genes) between ∆ompF and WT. As a result, DEGs were categorized into biological process, cellular component, and molecular function based on their sequence homology (Fig. 4). Moreover, in each group, the terms carbohydrate metabolic process, cell projection, and transcription regulator activity were dominant (Fig. 4). Meanwhile, all of the DEGs (including activated and repressed genes) mentioned above were also mapped to the KEGG pathway terms to better understand their functions and enrich possible metabolic or signal transduction pathways. The top 20 enriched KEGG pathways of DEGs in ∆ompF, including butanoate metabolism (KEGG ID: cfd00650), bacterial secretion system (cfd03070), and protein export (cfd03060), were shown in Fig. 5. Overall, the above results suggest that a great number of genes and related pathways can be affected by ompF deletion. The results were divided into three categories: biological process, molecular function, and cellular component. The x-axis indicates the number of genes in each category. The top ten items from each category were displayed. The number of downand up-regulated genes in each item was represented by red and green columns, respectively

Biosynthesis of maltose
In the genome of C. werkmanii, we discovered two maltose metabolism clusters (Fig. 6A): malM-lamB-malK and malE-malF-malG, whose expression levels were all up-regulated based on RNA-Seq results and confirmed qRT-PCR results (Fig. 6B). Meanwhile, alt, a regulator of maltose metabolism, was also found in the genome of C. werkmanii, though there were no significant differences in the RNA-Seq results (Fig. 6A). All the members of malM-lamB-malK and malE-malF-malG, as well as malt, were all knocked out one by one from WT and ∆ompF, and the maltose metabolic rates of all mutants were determined using high-performance liquid chromatography (HPLC). The results demonstrated that, when compared to the WT, the deletion of ompF, malM, lamB, and malT does not affect maltose metabolic rates on the second or fourth day (Table 1). However, ∆malK, ∆malE, ∆malF, and ∆malG showed lower rates than WT on the 2 and 4 days (except ∆malK, Table 1). While for double mutants, ΔompFΔmalM showed similar maltose metabolic rates to ΔompF and ΔmalM on the 2 and 4 days. ΔompFΔlamB exhibited the highest rates among all the detected strains at all measured time points. ΔompFΔmalK showed a lower rate than WT on the measured days. ΔompFΔmalE has a lower rate than WT on the 2 days, but a larger one on the 4 days. ΔompFΔmalF exhibited no rate differences with WT at all detected times. A lower rate occurred in the 2 days for ΔompFΔmalG, but it increased to the rates of WT in the 4 days. A lower rate was found for ΔompFΔmalT on the assayed days. These results suggested that the malE-malF-malG cluster is primarily responsible for maltose metabolism in C. werkmanii. Meanwhile, ompF participates in maltose transport with lamB, malK, malE, or malG at a specific time point.

MalT has the potential to regulate OmpF expression
It has been reported that all the maltose metabolic genes can be regulated by MalT. We intend to investigate whether the expression of OmpF can also be regulated by MalT. EMSA results revealed that the upward promoter of the ompF could be combined with the purified proteins of MalT in a dosedependent manner (Fig. 6C), indicating that the expression of the ompF could be regulated by MalT. To further Fig. 5 KEGG pathway-based functional analysis of DEGs between ∆ompF and C. werkmanii WT. The top 20 pathways were displayed, which included butanoate metabolism, bacterial secretion system, protein export, glycolysis/gluconeogenesis, and flagellar assembly, among others. The KEGG pathway and the richness factor are shown on the y-and x-axis, respectively, which denotes the ratio of the number of DEGs to the number of annotated genes enriched in this pathway confirm that the OmpF promoter is regulated by MalT, we constructed transcriptional fusions between the promoter and a gfp gene, yielding PompF-gfp that could be expressed in both WT and ΔmalT (Fig. 7A). The gfp fluorescence activity of ΔmalT was nearly 3.75 times greater than that of WT (Fig. 7B), confirming the findings that MalT negatively regulates OmpF expression by combining with its promoter. In addition, the transcriptional expression of ompF in ΔmalT was also enhanced by 3.62 times (Fig. 7C), which further confirmed that MalT has the ability to negatively regulate OmpF expression.

Discussion
In this study, the ompF gene was deleted from the genome of C. werkmanii using a homologous recombination method. The altered phenotypes and transcriptome profiles of ∆ompF were systematically analyzed using biochemical and molecular methods using the WT or ∆ompF-com as controls. The results demonstrated that ompF is involved in biofilm formation on both polystyrene and glass surfaces, cell aggregation, swimming motility, osmotic responses, and the alternation of diverse gene expression levels in C. werkmanii, particularly the flagellar assembly pathway genes and maltose metabolic genes. Moreover, the malEFGK2 cluster contributes to maltose transportation and malT can regulate the expression of ompF by combining its upward promoter.
Bacteria harbor a sophisticated survival strategy for sensing and adapting to environmental changes, particularly in unfavorable conditions (Bible et al. 2012). Swimming motility is regarded as one of the survival strategies for bacteria to acquire nutritious elements and optimal chemical concentrations, as well as overcome adverse environments (Singh and Olson 2008). Moreover, two OMPs, FlgO and FlgP, have been identified to play a crucial role in Vibrio cholerae motility (Martinez et al. 2009). In this study, ∆ompF exhibited smaller swimming colonies than WT and ∆ompF-com (Figs. 1B, 1C, and 1D) with a down-regulation of swimming-related genes (Fig. 1E), which suggests that these flagellar genes, including flgC, flgH, fliK, and fliF, are necessary for motilities in C. werkmanii. As we all know, the bacterial flagellum is made of three parts: the filament (helical propeller), the hook (universal joint), and the basal structure (rotary motor) (Nakamura and Minamino 2019;Terashima et al. 2008). Flagellar formation and function   (Akiba et al. 1991), flgH (Schoenhals and Macnab 1996), fliK (Hirano et al. 1994;Minamino et al. 2004), fliF (Ueno et al. 1992), and more than 45 others (Macnab 2003). The absence of OMPs may have an adverse effect on the physical structure of V. cholerae flagellum and flagellar rotation may not operate functionally in the OMPs mutant (Bari et al. 2012). Taken together, ompF contributes to the swimming abilities of C. werkmanii, and ompF deletion results in the down-regulation of several flagellar assembly pathways related genes (Additional file 2), which are important components for flagellar constructions and eventually lead to a possible alteration of flagellar structures and finally demonstrated impaired swimming motility. Cell aggregation and biofilm formation are reduced when ompF is deleted (Figs. 2 and 3E). Cell aggregation has been proposed as a pre-adoptive strategy for ensuring the survival and growth of bacterial populations in the face of various environmental stresses, such as toxic chemicals or bactericides (Jagmann et al. 2015;Klebensberger et al. 2007). SDS-induced cell aggregations of Pseudomonas aeruginosa involved c-di-GMP signaling with the psl operon as a possible target (Klebensberger et al. 2007). It was also discovered that type I pili were critical in the biofilm formation and aggregation of Xylella fastidiosa (Li et al. 2007). In addition, aggregation formation is an indication of stronger cell-cell interactions and increased biofilm formation (Shanks et al. 2008). In this study, we discovered that the deletion of ompF reduces C. werkmanii cell aggregation ( Fig. 2A), implying that ompF is a positive regulatory factor for aggregations, and fewer aggregations induced a decrease in biofilm formation. In addition, it has been also reported that the deletion of yceP (bssS) increases biofilm formation in continuous-flow chambers with minimal glucose medium since the deletion of yceP increased extracellular autoinducer 2 concentrations (Domka et al. 2006). Accordingly, the expression levels of bssS in ∆ompF increased by appropriately 2.67 folds (padj = 9.24E−05), which may also contribute to the inhibition of biofilm formation.
Moreover, OmpF in C. werkmanii contributes to the permeability of certain carbon and mental sources, particularly in biofilms (Fig. 3). In the presence of 400 mM sucrose or xylitol, ∆ompF formed fewer biofilms (Figs. 3F and 3G); ∆ompF formed more biofilms in the presence of 400 mM glucose (Fig. 3H); When exposed to 40 mM (Fig. 3I) or 400 mM (Fig. 3J) Ca 2+ or 160 mM Mg 2+ (Fig. 3O), ∆ompF formed fewer biofilms (Figs. 3M and 3N). These results indicated that C. werkmanii responded to different carbon sources or osmotic pressure in a dose-dependent mode via OmpF. The above phenomenon is simple to comprehend because numerous studies have yielded similar results. The addition of high concentrations of NaCl, KCl, or sucrose to the growth medium caused a drastic change in the ratio of the two peptidoglycan-associated major outer membrane proteins of E. coli K-12, with the amounts of proteins b and c present in cell envelope preparations decreasing and increasing, respectively (Alphen and Lugtenberg 1977). OmpW and OmpU were found to be highly expressed at higher salinity, whereas Omp26La, the OmpV homolog, functions as a salt-responsive protein at lower salinity (Kao et al. 2009). The balance of major OMPs, known as OmpF and OmpC, is achieved through complex regulation that is often reciprocal in response to many environmental factors, such as osmotic pressure (Liu and Ferenci 2001). At high osmolarity, OmpC becomes the predominant porin, whereas, at low osmolarity, OmpF predominates (Bystritskaya et al. 2016). Calcium ions, in particular, participate in a wide Fig. 7 The activity of the OmpF promoter can be inhibited by MalT. To further confirm that the promoter of OmpF is regulated by MalT, we constructed transcriptional fusions between the promoter and a gfp gene, resulting in PompF-gfp, which could be expressed successfully in WT and ΔmalT (A). The gfp fluorescence activity of ΔmalT was nearly 3.75 times greater than that of WT (B). The relative expression levels of ompF in ΔmalT (C) range of biological pathways, serving as an intracellular second messenger and modulating bacterial biofilm formation and architectural integrity (Bootman and Berridge 1995;Patrauchan et al. 2005). In P. aeruginosa FRD1, Ca 2+ addition increased the production of alginate, a component of biofilms, by enhanced ionic crosslinking and gelling of the uronic acid residues . Our previous study has proved that different Ca 2+ concentrations could stimulate the biofilm formation of C. werkmanii in different ways and that the OmpF was induced to higher expression in the biofilm formed at 400 mM Ca 2+ (Zhou et al. 2016b). In this study, we discovered that the biofilm of ∆ompF was consistent under normal conditions and in the presence of 40 mM Ca 2+ (Fig. 3M), indicating that OmpF is an active site for Ca 2+ enhancement biofilms at this concentration. However, when exposed to 400 mM Ca 2+ (Fig. 2D), all three strains of WT, ∆ompF, and ∆ompF-com formed more biofilms than under normal conditions, implying that OmpF was lost the active site for Ca 2+ inducing biofilm enhancement and that 400 mM Ca 2+ also induced some other potential proteins. Aside from Ca 2+ , Mg 2+ was found to have biofilm enhancement ability at concentrations of 160 mM and 320 mM, implying that these two divalent metal ions may share the same biofilms enhancement mechanisms for C. werkmanii. In addition, the OmpR protein is required for ompF transcription, and the EnvZ protein is essential for normal ompF expression regulation, which is affected by the medium osmolarity (Mizuno and Mizushima 1987;Pratt et al. 1996). In this study, RNA-Seq results revealed that ompD (log2FoldChange = 1.32, padj = 0.03), ompR (log2FoldChange = 0.91, padj = 0.12), ompW (log2Fold-Change = 1.21, padj = 0.03), ompX (log2FoldChange = 2.13, padj = 0.00024), mzrA (EnvZ/OmpR regulon moderator; log-2FoldChange = 1.40, padj = 0.01) were also shown different expression levels between WT and ∆ompF. More research is needed to determine whether these outer membrane proteins or moderators regulated or compensated for the expressions and functions of ompF deletion.
Transporters are critical players in bacterial growth and survival since they facilitate nutrient uptake (Piepenbreier et al. 2017). Meanwhile, bacteria prefer primary sources, such as glucose and glucose 6-phosphate to secondary sources such as maltose or maltodextrins (Schaechter 2015). The Mal-system required for maltose-uptake in E. coli, where the import of maltose and maltodextrins is dependent on the ABC transporter MalEFGK2, which is genetically organized in two operons, malEFG and malK-lamB-malM (Boos and Shuman 1998). We discovered a complete maltose transport system in C. werkmanii's genome (Fig. 6A), which is similar to that in E. coli. Moreover, malT is the activator of all mal genes, and malT mutations result in constitutive expression of these mal genes (Dardonville and Raibaud 1990;Richet and Raibaud 1987). Besides malT, the level of mal gene expression in Yersinia enterocolitica may be regulated by the Mlc, a malT repressor (Raczkowska et al. 2008). In this study, the RNA-Seq and RT-PCR results revealed that a total of 6 genes in the maltose system, including lamB, malE, malK, malG, malM, and malF, were upregulated (Fig. 7B). However, there are no differences in maltose consumption rates between WT and ∆ompF on no 2 days or 4 days (Table 1). Moreover, in the presence of a primary nutrient deficiency, the uptake of maltose/maltodextrins in E. coli is facilitated by the synergistic action of the LamB and MalEFGK2 systems (Saier Jr and Crasnier 1996). Subsequently, all 6 up-regulated genes and malT were deleted from the genome of C. werkmanii WT and ∆ompF, respectively. The results demonstrated that ∆malK, ∆malE, ∆malF, and ∆malG showed lower rates than WT on the 2 and 4 days (except ∆malK, Table 1), implying that C. werkmanii BF-6 primarily used MalEFGK2 without LamB to transport maltose. While for double mutants, there are no identical trends for maltose metabolism (Table 1). These results strongly suggest that single and double gene mutants differ in maltose metabolism.
malT has been shown to regulate all genes involved in maltose metabolism (Dardonville and Raibaud 1990;Richet and Raibaud 1987). In this study, we discovered that the upward promoter of ompF could be combined with the purified proteins of MalT in a dose-dependent manner (Fig. 6C), implying MalT may also regulate OmpF expression. Two additional MalT boxes in a direct repeat upstream of the 238 regions were discovered to be essential for MalT-dependent mal gene expression. The analysis of these MalT boxes also led to an extension in the consensus sequence, which is now defined as 5′-GGGGA(T/G)GAGG-3′ (Vidal-Ingigliardi et al. 1991). However, only 5′-GAT GAG G-3′ (− 10) was discovered in the upward of ompF, and whether this is the binding site of MalT requires further investigation. Whereas, GFP activity analyses revealed that the transcription levels of OmpF's promoter could be negatively regulated by MalT (Figs. 6C and 7), implying that MalT regulates OmpF expression by direct combining with its promoter.

Conclusions
In summary, our results demonstrated that ompF in C. werkmanii plays a vital role in a variety of biological processes, including swimming ability, biofilm formation, and mental ions responses. Moreover, C. werkmanii mainly depends on the malEFGK2 cluster for maltose transportation, and the promoter of OmpF could be combined by MalT, resulting in an altered OmpF expression. However, detailed molecular regulation mechanisms of some of the above phenomena should be investigated further.