All AHL synthase genes, luxI/R homologs, and rsh gene sequences were retrieved from NCBI Genome Database (Microbial Genomes www.ncbi.nih.gov) using the keywords “AHL synth(et)ase,” “quorum sensing,” “luxI” and “autoinducer.” Detailed information for these resulting genes such as sequences and gene locations were manually checked, and recorded for further analysis.

The amino acid sequences of these luxI homologs were aligned by clustalW using MEGA software (version 11), and the phylogenetic tree was constructed by the neighbor-joining method with 1,000 bootstrap.

Bacterial strains and plasmids used in this study are listed in Table 1. Sphingobium japonicum UT26 and Sphingobium sp. SYK-6 were grown at 28°C in 1/3 LB media and LB media, respectively. Their rsh deletion mutants were grown in respective medium plus 100 μg/mL streptomycin. The AHL reporter strain Agrobacterium tumefaciens A136 and all Escherichia coli strains were grown at 37°C in LB medium with proper antibiotics (Lu and Huang, 2018).

In-frame deletions of rsh gene in SYK-6 and UT26 were performed by employing the markerless gene deletion system for Sphingomonads (Kaczmarczyk et al., 2012; Lu and Huang, 2018). Primers used for gene deletion are listed in Supplementary Table S1. Briefly, flanking regions of rsh, approximately 500 bp up- and down-stream each, were amplified by PCR with the primer pairs syk6-5O/-5I and syk6-3O/-3I, then these two fragments were joined using overlap PCR with the primer pair syk6-5O/-3O. The resulting fragment was then cloned into the plasmid pAK405 via the BamHI/HindIII restriction site. Subsequently, the pAK405 derivative was delivered into the wild type strain via conjugal transfer from E. coli S17-1 (λ pir). After the conjugal transfer, bacteria were plated on LB media supplemented with kanamycin (100 μg/mL) and rifampicin (50 μg/mL). Individual colonies were re-streaked once on the same medium and then plated on LB media supplemented with 100 μg/mL streptomycin to select for the second homologous recombination event. The resulting colonies were re-streaked on both LB media supplemented with 100 μg/mL streptomycin and LB media containing 100 μg/mL kanamycin. Kanamycin-sensitive clones were analyzed by colony PCR using primers syk6-FO/-RO to confirm the successful deletion of rsh.

Bacterial strains were routinely grown in respective media at the shaking speed of 200 rpm, and their growth curves were drawn by measuring OD₆₀₀ values at regular time intervals using a Spectrophotometer (Metash, Shanghai, China). All measurements were performed in triplicates.

Congo red is an azo dye that can form complexes with different components in extracellular polymeric substances (EPS) (Semedo et al., 2015). The Congo red binding assay was carried out to evaluate bacterial EPS production (Lu and Huang, 2018).

Cultures were inoculated from a single colony and grown to an OD₆₀₀ of 1 in LB broth; 10 μL of suspensions were spotted on LB plates containing 40 μg/mL Congo red. The plates were grown at 30°C to assess the morphology of colony. For quantitative analysis, the same number of cells from different culture conditions were suspended in 100 μg/mL Congo red in 0.9% saline and then incubated for 90 min with shaking at 30 rpm at room temperature. After incubation, the supernatant was obtained by centrifugation, and then the OD₄₉₀ of the supernatant was measured as A₁. The OD₄₉₀ of Congo red solution without bacterial fluid was recorded as A₀. All measurements were performed in triplicates. The percentage of Congo red stain was calculated to represent the abundance of EPS, by the formula:

The AHL production was analyzed either by cross-feeding assay or paper disk assay, according to our previous publication (Lu and Huang, 2018). Briefly, in the cross-feeding assay, the AHL reporter strain A136 and the tested strains were streaked side by side on the agar plates pre-covered with 50 μL X-gal (20 mg/mL stock solution in dimethylformamide). In the paper disk assay, ethyl acetate extracts of samples were spotted onto paper disks and incubated in LB soft agar plates mixed with A136 plus X-gal. The blue coloration results were recorded accordingly. LC–MS/MS was employed to identify the AHL profiles according to Huang et al. (2019).

qRT-PCR on luxI/R homologs was done as described previously (Lu and Huang, 2018). Briefly, cells of UT26 and SYK-6 at different growth phases and culture conditions were harvested and their total RNA was extracted with RNAprep pure Cell/Bacteria Kit (Tiangen Biotech, Beijing, China) according to the manufacturer’s instructions. cDNA was synthesized using FastKing gDNA Dispelling RT SuperMix (Tiangen Biotech, Beijing, China) and was used as the template for qRT-PCR amplification by using SYBR Premix Ex TaqTM Kit (Tli RNaseH Plus) (TaKaRa, Dalian, China) in Applied Biosystems StepOnePlus TM Real-Time PCR System (Thermo Scientific, Waltham, MA, United States). Primers were designed on the Sangon website¹ and listed in Supplementary Table S2. The relative gene expression level of the mutant to the wild-type strain was analyzed using the 2^−∆∆Ct analysis method (Lu and Huang, 2018).

Bacterial cells at the early stationary phase were collected for ppGpp detection using the formic acid method (Lu and Huang, 2018). The presence of ppGpp in the extract was then determined by HPLC (Agilent, Santa Clara, CA, United States) analysis using a ZORBAX SB-C18 column (4.6 × 150 mm, 5 μm, Agilent). The mobile phase consisted of 125 mM KH₂PO₄, 10 mM tetrabutyl ammonium dihydrogen phosphate, 60 mL/L methanol, and 1 g/L KOH (pH 6.0). The column temperature was 40°C and the flow rate was 1.0 mL/min. The ppGpp was monitored at 254 nm and identified by comparison with the retention time of 100 μM ppGpp standards (TriLink Biosciences, San Diego, CA, United States). The ppGpp standard was eluted at around 71 min under the above-mentioned conditions to confirm the presence of ppGpp.

A Student’s t-test with a p-value cut-off of 0.05 was used to test significance of differences using the software SPSS (version 25).

Up to November 2023, 158 strains of sphingomonads had their whole genome sequenced. Among them, 79 strains were found to harbor luxI homologs, including 10 Novosphingobium spp., 18 Sphingopyxis spp., 26 Sphingobium spp., and 25 Sphingomonas spp. (Supplementary Table S3). The number of luxI homologs in each strain ranged from 1 to 4. 36 out of 79 strains contained more than one luxI homolog, and 13 out of 79 strains contained 3 luxI homologs, such as Sphingobium japonicum UT26 and Sphingobium sp. SYK-6 (Figure 1).

The majority of these luxI homologs (110 out of 131) were located on chromosome. More than two thirds, 59 out of 79, of these strains harbored luxI homologs in chromosome only, while 14 out of 79 strains harbored luxI homologs in both chromosome and plasmid (Figure 1). The phylogenetic tree showed that these luxI homologs in sphingomonads were not phylogenetically related (Figure 1). LuxI homologs from different genera mingled with each other and even luxI homologs from the same strain rarely clustered together.

Sequence analysis on genomes of UT26 and SYK-6 indicated that both UT26 and SYK-6 contained two AHL synthase genes on the chromosome and one on a plasmid. These genes were named sjaI 1–3 for UT26 and sphI 1–3 for SYK-6. Except for sjaI3, other luxI homologs had corresponding luxR homologs in their close vicinity. On the other hand, one rsh gene around 2 kb was found on the chromosome of either UT26 or SYK-6. The detailed sequence information and location of rsh and these QS genes were plotted in Figure 2A, with reference strains US6-1 and Rr-2-17.

In UT26, both sjaI/R1 and sjaI/R2 located on the chromosome, and the solo sjaI3 located on the plasmid. The sjaI/R1 located distantly upstream to rsh, around 2.5 million bases apart, while the sjaI/R2 located closely downstream to rsh, around 50 kb apart. Whereas in SYK-6, both sphI/R1 and sphI/R2 located distantly upstream to rsh, around 1.5 and 0.8 million bases apart, respectively. The sphI/R3 located on the plasmid. For the reference strains US6-1 and Rr2-17, the distances between rsh and the luxI/R homologs were 1.5 million bases and 60 kb apart, respectively. The physical distance between the rsh and QS genes on the chromosome were rather random among species.

Successful deletion of the rsh gene in UT26 and SYK-6 was confirmed by PCR amplification of the sequence around the rsh gene. PCR products from both mutants were about 2 kb less than those of their wild type strains, due to the deletion of the rsh gene fragment (Figure 2B). The mutant was deficient in ppGpp production compared to the wild type strain, further proved the successful deletion of rsh (Supplementary Figure S1).

LC–MS/MS analysis indicated that both Δrsh_UT26 and Δrsh_SYK-6 produced AHL profiles similar to their wild type strains in rich LB medium, while in the minimal medium 1% LB, Δrsh_SYK-6 almost lost AHLs production (Table 2). Strain UT26 mainly produced C12-HSL, with a small amount of C14-HSL and 3-OH-C8-HSL. Whereas SYK-6 produced C14-HSL, 3-OH-C8-HSL, 3-OH-C6-HSL, and a trace amount of C8-HSL when grown in LB media (Supplementary Figure S2). The AHL signals from WT_SYK-6 and Δrsh_SYK-6 were confirmed by the reporter strain A136, and both triggered a blue coloration in A136, indicating a positive AHL production in both strains (Figure 2C). These results indicated that deletion of rsh did not abolish the AHL production, but did affect their AHL profiles in both strains.

Growth curves of Δrsh_UT26 and WT_UT26 indicated that deletion of rsh rendered a delay in cell proliferation, but their overall biomasses in individual growth phase were similar (Figure 3A). The AHL production analysis indicated that in the middle log phase, AHL productions of Δrsh_UT26 and WT_UT26 were similar, while in early stationary phase, Δrsh_UT26 remained high production of C12-HSL, which was significantly higher than WT_UT26 (Figure 3B).

Relative gene expressions of sjaI1-3 and sjaR1-2 in Δrsh_UT26 compared to WT_UT26 were measured (Figure 3C), results indicated that in the middle log phase, all these five genes were upregulated in Δrsh_UT26, especially sjaI2-3 and sjaR2. Whereas in the early stationary phase, sjaI1-3 was slightly upregulated, but sjaR1-2 was significantly downregulated. The consistent upregulated expression of these sjaIs in Δrsh_UT26 explained the higher production of AHLs in the mutant. The Δrsh_UT26 and WT_UT26 showed similar EPS production in the middle log phase. However, in the early stationary phase, the mutant showed much higher EPS production than the wild type (Figure 3D).

Overall, these results indicated that for strain UT26, rsh negatively regulated the expression of most QS genes, which in turn negatively regulated AHLs and EPS production. These regulatory effects were growth phase dependent.

Since rsh gene was believed to function when nutrition was limited, therefore, 1% LB medium was used as a poor medium to mimic a nutrient-limited culture condition. Regulatory effects of rsh on AHL/EPS production and gene expression were tested in both rich media (LB) and poor medium (1% LB).

In LB medium, growth curves indicated that deletion of rsh significantly promoted the biomass of Δrsh_SYK-6 at stationary phase (Figure 4A). AHL production at 36 and 48 h were visualized. Δrsh_SYK-6 showed robust AHL production, the signals were stronger than WT_SYK-6 at 48 h (Figure 4B). However, relative expressions of most QS genes were significantly downregulated in Δrsh_SYK-6, especially for sphI3 and sphR1. Notably, expression of sphI 2 was not significantly affected (Figure 4C). EPS production of Δrsh_SYK-6 was similar to that of the wild type when grown in the liquid medium, while was significantly higher than WT_SYK-6 when grown in plate biofilm (Figure 4D). These results indicated that deletion of rsh downregulated most QS genes, the higher AHL and EPS production in the mutant might be attributed to the consistent expression of sphI2, and the higher biomass.

In 1% LB broth medium, deletion of rsh did not affect the growth of SYK-6 significantly (Figure 5A). Meanwhile, the mutant did not show AHL signals in broth extract (Figure 5B). LC–MS/MS analysis indicated that only 3-OH-C8-HSL, C14-HSL were produced by the WT_SYK-6 in 1%LB, while only trace amount of 3-OH-C8-HSL was detected in Δrsh_SYK-6 in 1% LB (Table 2). This was contrary to the cross-feeding assay that on 1% LB agar plates, Δrsh_SYK-6 triggered a strong blue coloration of the reporter strain A136 just like the WT_SYK-6 (Figure 2C). Gene expression analysis indicated that most QS genes were downregulated in the mutant, especially sphI1 and sphR3. Notably, sphI2 was the least affected gene (Figure 5C). The downregulated pattern of rsh on QS genes was different from what was found in the LB medium (Figure 4C). On the other hand, the EPS production of the mutant was always higher than that of the wild type in both broth and plate cultivation (Figure 5D).

Indeed, the regulatory effect of rsh on the QS system in SYK-6 was affected by nutrient availability and culture conditions. Specifically, planktonic cultivation in broth or attached growth on agar plate (biofilm) affected the regulatory effect of rsh on QS system.