Abh, AbrB3, and Spo0A play distinct regulatory roles during polymyxin synthesis in Paenibacillus polymyxa SC2

ABSTRACT Polymyxins exhibit antibacterial activity against various gram-negative bacteria. However, the regulatory mechanisms of polymyxin synthesis remain unclear. Here, Abh, AbrB3, and Spo0A of Paenibacillus polymyxa SC2 were found to bind to P pmx via DNA pull-down and electromagnetic mobility shift assays. Oxford Cup and liquid chromatography-mass spectrometry assays showed that antibacterial activity and polymyxin production increased in the Δabh strain. Overexpression of abh and abrB3 significantly decreased the antibacterial activity and polymyxin production. The transcription of pmxA determined via quantitative reverse transcription-PCR supported these results. Moreover, the mutation and complementation of spo0A revealed that Spo0A was an activator of polymyxin synthesis. Spo0A could bind to the abh promoter to activate abh expression. Spo0A could bind to the abrB3 promoter to inhibit abrB3 transcription and mitigate the inhibitory effect of AbrB3 on abh transcription. IMPORTANCE Polymyxins are considered the last line of defense against multidrug-resistant bacteria. The regulatory mechanism of polymyxin synthesis is poorly studied in Paenibacillus polymyxa. In this study, we found that Abh and AbrB3 negatively regulated, whereas Spo0A positively regulated polymyxin synthesis in P. polymyxa SC2. In addition, a regulatory relationship between Abh, AbrB3, and Spo0A was revealed, which regulate polymyxin synthesis via multiple regulatory mechanisms in P. polymyxa.

P. polymyxa SC2 was previously screened from the rhizosphere of pepper plants cultivated in Guizhou Province, China, and has been reported to exhibit growth promotion and biological control activities (16,17).Whole-genome sequencing of SC2, which contains the polymyxin synthase gene cluster (pmxA-E) via anti-SMASH analysis, was first completed in 2010 (17,18).P. polymyxa is susceptible to morphological changes and functional degradation during culturing (19,20).We previously reported that SC2 can be differentiated into two morphologically distinct strains, SC2-M1 and SC2-M2, both of which lose the ability to form spores (19).As the wild-type strain SC2 cannot be easily manipulated and is susceptible to developing spontaneous mutations, the strain SC2-M1 was previously used for molecular manipulation to study the regulatory mechanisms of polymyxin synthesis in P. polymyxa.In our previous study, we successfully knocked out msmR1 in SC2-M1 and demonstrated that MsmR1 (a global transcriptional regulator) indirectly affects polymyxin synthesis (21).
In this study, we screened and identified transcription factors related to polymyxin biosynthesis and revealed their regulatory mechanisms.Abh and AbrB3 directly inhibited polymyxin synthesis, whereas Spo0A directly activated polymyxin synthesis.In addition, electromagnetic mobility shift assay (EMSA) and quantitative reverse transcription-PCR (qRT-PCR) showed that Spo0A directly inhibited the expression of abrB3 and activated the expression of abh, whereas AbrB3 and Abh inhibited the expression of each other by directly binding to their promoters.Collectively, these findings provide new insights into the regulation of polymyxin synthesis.

RESULTS
Transcriptional activity of promoter P pmx P. polymyxa SC2 contained five open reading frames consisting of a polymyxin synthe tase gene cluster with a size of 41.2 kb (Fig. 1A).We performed promoter prediction for the upstream region of pmxA using Softberry and identified two −35 (TTTAAG; TTGAGC) and −10 regions (GCGTAGAAC; TCGTCACAT) of P pmx (Fig. S1A).To monitor the transcriptional activity of P pmx , a reporter fusion vector of P pmx and GFP was constructed and introduced into SC2-M1 cells.Fluorescence intensity detected using a fluorescence microscope indicated that this region has transcriptional activity (Fig. 1B).

Candidate P pmx -binding proteins
In this study, we aimed to identify key transcription factors involved in the activation or inhibition of polymyxin synthesis by performing DNA pull-down assay.Antibacterial activity was detected when the strain entered the exponential phase at 5 h, followed by a gradual increase in activity (Fig. S2).Therefore, we collected SC2 cells cultured for 5 and 10 h to extract the total protein, which was then used in the DNA pull-down assay.The primer pair pmxABF/R was used to amplify biotin-labeled DNA fragments.Both biotin-labeled DNA fragments and nuclease-free water (negative control) were incubated with beads, as described in the Methods section.Candidate P pmx -binding proteins were analyzed using LC-MS/MS.LC-MS/MS analysis showed that 10 candidate P pmx -binding proteins were specifically detected in the BL 5h group and 25 candidate P pmx -binding proteins were specifically detected in the BL 10h group compared to those in NF group (Table S2).ResD3, YlzA, and FruR1 were present in groups BL 5h and BL 10h , and 32 candidate transcription factors were identified.Both Abh (detected in the BL 10h group) and AbrB3 (detected in the BL 5h group) were found to contain an AbrB domain similar to the transition state regulatory protein AbrB of B. subtilis (22).Previous studies have shown that AbrB can bind to the promoter of pmx cluster in P. polymyxa E681 to negatively regulate polymyxin synthesis (15).In addition, as shown in Fig. S1A, P pmx contains three 0A-like boxes (TGCCGAA; CGTAGAA; TGGCGAA) and one 0A-box: 5′-TTCGACA-3′ (5′-TGTCGAA-3′, complementary sequence) (23) and Spo0A was detected in both BL 10h and NF 10h groups.Therefore, we focused on abh, abrB3, and spo0A in this study.The proteins His6-Abh, His6-AbrB3, and His6-Spo0A were purified (Fig. S3) and analyzed via EMSA, which showed that Abh, AbrB3, and Spo0A were directly bound to P pmx (Fig. 2).

Abh inhibited polymyxin synthesis
To elucidate the mechanism by which Abh regulates polymyxin synthesis, we construc ted a deletion, complementation, and overexpression strain of abh.Our data showed that the mutation in abh led to a minor increase in antimicrobial activity, whereas it significantly increased polymyxin content (Fig. 3A through C). qRT-PCR analyses revealed that the relative transcription level of pmxA was significantly increased in strain Δabh (Fig. 3D).These changes were reversed by introducing pHY-abh into strain Δabh (Fig. 3).However, when abh was overexpressed in strain SC2-M1, the antibacterial activity, polymyxin content, and the relative transcription level of pmxA were significantly decreased (Fig. 3A through D).Therefore, Abh effects polymyxin production by inhibiting transcription of the polymyxin synthetase gene cluster.

AbrB3 inhibited polymyxin synthesis
To determine the relationship between AbrB3 and polymyxin synthesis in vivo, we constructed the overexpression strain M1-pHY-abrB3.The abrB3 overexpression strain exhibited relatively weaker antibacterial activity against E. coli DH5α (Fig. 4A and B).Moreover, polymyxin production and the relative transcription level of pmxA were significantly decreased in the M1-pHY-abrB3 strain (Fig. 4C and D), indicating that AbrB3 inhibited polymyxin synthesis.

Spo0A positively regulated polymyxin synthesis
To determine the mechanism via which Spo0A affects polymyxin synthesis, we construc ted spo0A knockout and complementation strains.Next, we determined the polymyxin content of the spo0A mutant and its antimicrobial activity against E. coli DH5α.Compared with the SC2-M1(Arg) strain (reversion of Spo0A in which the R211H substitution occurred in the HTH region), the antimicrobial activity almost disappeared and the polymyxin content distinctly decreased in the Δspo0A strain (Fig. 5A through C).Further more, the relative transcription level of pmxA in strain Δspo0A was significantly lower than that in the SC2-M1(Arg) strain (Fig. 5D).Both the polymyxin production and the relative transcription level of pmxA were restored by pHY-spo0A into Δspo0A strain (Fig. 5).Therefore, Spo0A positively regulated polymyxin synthesis.

Regulatory relationship of Abh, AbrB3, and Spo0A
Spo0A inhibits the transcription of abrB, while AbrB inhibits the expression of abh in B. subtilis (24).However, the mechanisms underlying the regulation of expression between these three genes in P. polymyxa remain unclear.First, we performed promoter prediction for the upstream region of abh, abrB3, and spo0A using Softberry (Fig. S1B through D).Subsequently, EMSA was performed, which showed that Spo0A was bound to P abrB3 , P abh , and P spo0A , AbrB3 was bound to P abh and P abrB3 , and Abh was bound to P abrB3 (Fig. 6A through C; Fig. S4).In addition, qRT-PCR revealed that the relative transcription level of abh was significantly decreased in the overexpression strain M1-pHY-abrB3 (Fig. 6D), and the relative transcription level of abrB3 was significantly decreased in the overex pression strain M1-pHY-abh (Fig. 6E).In addition, the relative expression level of abrB3 was significantly increased and that of abh was significantly decreased in the Δspo0A strain compared to those in SC2-M1(Arg) (Fig. 6F).In summary, Spo0A positively regula ted abh expression by binding to the abh promoter and indirectly regulated abh expression by directly inhibiting the transcription of abrB3.Spo0A and AbrB3 were found to be self-regulated.

DISCUSSION
Although the polymyxin synthase gene cluster has been identified in several strains, the mechanisms of underlying polymyxin synthesis regulation remain unclear (5,7,8,12).In the present study, we identified multiple candidate transcription factors that could bind to the promoter region of the pmx cluster based on DNA pull-down assays (Table S2).
Using the promoter of yfr2 as a bait, the candidate transcription factor PMM1637 was screened using DNA pull-down assay (25).Our results indicated that this method is effective in identifying transcription factors.Based on our data, the regulatory networks of Abh, AbrB3, and Spo0A were proposed (Fig. 7).Abh and AbrB3 were directly inhibited, whereas Spo0A directly and indirectly activated the expression of the pmx cluster in P. polymyxa SC2.AbrB3 and Abh inhibit each other by directly binding to their respective promoters.Spo0A and AbrB3 were found to be self-regulated.However, the regulatory effects of ResD3, RbsR3, and GlnR1 on polymyxin synthesis require further investigation.
In B. subtilis, AbrB is a transition-state regulator, and the monomer comprises an Nterminal domain (DNA-binding) and a C-terminal domain (multimerization).On the other hand, Abh is a paralogous protein of AbrB showing high similarity with regard to the Nterminal domain (26,27).Abh and AbrB mostly function as repressor that regulate gene expression, but may perform contradictory functions in certain cases (24,28).Abh and AbrB regulate the expression of many antibiotic synthesis genes, including sdpABC, skfABCDEFGH, sboA, and sunA (24,29,30).Although the regulatory functions of Abh and AbrB have been extensively studied in B. subtilis, their functions in P. polymyxa polymyxin synthesis remain unclear.In this study, both Abh and AbrB3 were found to possess an AbrB domain similar to AbrB in the N-terminal domain of B. subtilis, and both could bind to P pmx based on EMSA (Fig. 2A and B) (22).Moreover, the abh knockout increased polymyxin production and pmxA expression compared to those in SC2-M1 cells, which were restored by introducing pHY-abh into Δabh (Fig. 3).In addition, antibacterial activity, polymyxin production, and pmxA expression decreased significantly in abh and abrB3 overexpression strains (Fig. 3 and 4).We could not study the regulatory function of AbrB3 in vivo further owing to the absence of an abrB3 mutant.However, the abrB mutant has been reported to show higher production of polymyxin when heterolo gously expressing the polymyxin synthase gene cluster in B. subtilis (15).In summary, Abh and AbrB3 are negative regulators that directly inhibit polymyxin synthesis in P. polymyxa SC2.
P pmx contained three 0A-like boxes (TGCCGAA; CGTAGAA; TGGCGAA) and one 0Abox: 5′-TTCGACA-3′ (5′-TGTCGAA-3′, complementary sequence) (Fig. S1A).However, it is unclear whether Spo0A directly regulates the expression of this gene cluster.In this study, EMSA revealed that Spo0A was bound to P pmx , indicating that Spo0A directly regulates polymyxin synthesis (Fig. 2C).Previous studies have shown that polymyxin production is Spo0A-dependent because the spo0A mutant strain cannot produce polymyxin in P. polymyxa E681, and Spo0A indirectly and positively regulates the polymyxin synthesis in B. subtilis (15).Here, the antibacterial activity and pmxA expres sion were significantly decreased in the spo0A mutant (Fig. 5A and D).LC/MS results showed that the spo0A mutant did not completely lose its ability to produce polymyxin (Fig. 5C).Therefore, Spo0A positively and directly regulates polymyxin synthesis in P. polymyxa SC2.In this study, compared to the knockout strain Δspo0A, the SC2-M1 strain showed efficient polymyxin production, although polymyxin synthesis in strain SC2-M1 was affected to a certain extent.The relative transcription level of abh was significantly decreased (log2FC = −4.61) in SC2-M1 compared to that in SC2 based on transcriptomic analysis (19).We suggest that Spo0A (R211H substitution occurring in the HTH region of Spo0A) in SC2-M1 positively regulates pmx expression and suppresses the inhibitory effect of Abh on pmx expression.A mutation at position 257 of Spo0A results in loss of sporulation in the strain; Spo0A (A257V) can still inhibit the transcription of abrB (31).In the present study, whether or not Spo0A (R211H) can combine with P pmx, P abh and P abrB3 needs further investigation.In addition, we demonstrated that other candidate binding proteins (ResD3, RbsR3, GlnR1) (data not shown) bind to P pmx except Abh, AbrB3, and Spo0A, indicating that these proteins may also regulate the expression of the pmx cluster.
In B. subtilis, Spo0A directly inhibits abrB expression, whereas AbrB inhibits abh expression (24,32).In P. polymyxa WLY78, Spo0A directly inhibits abrB expression (33).Here, our EMSA and qRT-PCR results indicated that AbrB3 and Abh inhibited each other by directly binding to their respective promoters; Spo0A could bind to the abh promoter to activate abh expression.Spo0A could bind to the abrB3 promoter to inhibit abrB3 transcription and mitigate the inhibitory effect of AbrB3 on abh transcription (Fig. 6; Fig.

S4
). AbrB is a transition phase regulator that mainly plays a regulatory role in the exponential growth of strains (28,34).AbrB levels decrease as the strain enters the stationary phase from the growth phase, whereas Spo0A levels increase after the exponential phase (35).In the present study, only AbrB3 was found in the BL 5h group, whereas Abh, AbrB3, and Spo0A were found in the BL 10h group; AbrB3 and Spo0A were also found in the NF 10h group.Based on these findings, we hypothesize that AbrB3 represents the main inhibitor of polymyxin synthesis at the early stage of strain growth, and that Spo0A combined with Abh dynamically regulates pmxA expression at the later growth stage of the strain.This hypothesis requires further verification in future studies.
In the present study, Abh and AbrB3 directly inhibited the expression, whereas Spo0A both directly and indirectly activated the expression of the pmx cluster in P. polymyxa SC2.We revealed a regulatory relationship between Abh, AbrB3, and Spo0A, which regulate polymyxin synthesis via multiple regulatory mechanisms in P. polymyxa.By studying the regulatory mechanism of polymyxin synthesis, we can provide a theoretical foundation for the future metabolic engineering of strains.To achieve this, proteomics and transcriptomic analyses are required to study the regulatory mechanism.Genes that encode negative regulator proteins can be knocked out, and genes that encode positive regulator proteins can be overexpressed to construct high-yield strains.

Promoter activity verification
The pmxAF/AR primer pair was used to amplify P pmx from the genomic DNA (gDNA) of SC2.After digestion of plasmid pHY300PLK-P gap -gfp-cm using XbaI and BamHI, it was ligated with the amplified fragment at 16°C using T4 DNA ligase (Accurate Biotechnol ogy, Hunan, Co., Ltd., Changsha, China) for 12 h and then transformed into Escherichia coli DH5α and SC2-M1 (19,36).The strains carrying a GFP-fusion vector were cultured in LB solid medium containing 20 µg/mL tetracycline and cultured at 37°C for 24 h.Single colonies were selected to prepare bacterial suspensions, and fluorescence activity was observed under a fluorescence microscope at 1,000× magnification (Olympus BX53, Japan).

DNA affinity chromatography-pulldown (DNA pull-down) assay
The DNA pull-down assay was performed as previously described (37).Briefly, the collected SC2 cells (cultured for 5 and 10 h) were washed twice with phosphate-buf fered saline (PBS) (0.01 M PO 4 3-, 0.8% NaCl, 0.02% KCl, pH 7.2-7.4)and resuspended in BS/THES buffer for sonication (37), and the total protein content in the supernatant was determined after centrifugation.The fragment P pmx was amplified via PCR using a 5′-biotinlabed primer (pmxABF/pmxABR).Biotin-labeled DNA fragments (200 ng/µL) were combined with pretreated Dynabeads M-280 Streptavidin (Invitrogen, Carlsbad, CA, USA) at room temperature for 30 min in a rotor for an experimental group named BL.The aforementioned step was repeated.Biotin-labeled DNA fragments (200 ng/µL) were replaced with nuclease-free water, and similar incubation with beads was performed in the negative control group named NF.Biotin-labeled DNA fragment bead complex incubation with the aforementioned total protein (cultured for 5 and 10 h named BL 5h and BL 10h , respectively) and beads (incubation with nuclease-free water) incubation with above total protein (cultured for 5 and 10 h named NF 5h and NF 10h , respectively) were performed separately.The bait bound to the target protein was eluted using different concentrations of 100, 200, 300, 500, 750, 1,000 mM of NaCl solution.The protein eluent was digested with trypsin and analyzed via liquid chromatography-mass spectrometry (LC-MS/MS) on a Q Exactive (Thermo Fisher Scientific, San Jose, CA, USA) coupled with Easy-nLC 1000 (Thermo Fisher Scientific, San Jose, CA, USA).LC-MS/MS was performed by Shanghai Applied Protein Technology Co., Ltd.(Shanghai, China).LC-MS/MS data were analyzed using mascot 2.2 to obtain qualitative identification information of the target protein polypeptides.

Construction of mutant strains
abh and spo0A in the SC2-M1 strain were deleted via homologous recombination.Briefly, the primer pairs abh-qF1/R1 and abh-qF3/R3 were used to amplify the upstream fragment 1 and downstream fragment 3 of abh, respectively.The primer pairs spo0A-qF1/R1 and spo0A-qF3/R3 were used to amplify the upstream fragment 1 and down stream fragment 3 of spo0A, respectively.The primer pairs abh-qF2/R2 and spo0A-qF2/R2 were used to amplify fragment 2 of the cat cassettes from plasmid pDG1661, respectively (40).Combined fragments 1-3 are generated using fusion PCR (41).Fragments 1-3 and the pRN5101 plasmid (Institute of Plant Protection, Chinese Academy of Agriculture Sciences, China) were digested (NheI/SalI or NheI/BamHI) and ligated overnight at 16°C and transformed into E. coli DH5α to obtain the deletion plasmids pRN5101-abh and pRN5101-spo0A.These plasmids were then transformed into Trans 110 cells to remove methylation and then electroporated into SC2-M1 cells.Double-crossover recombinants were used to screen knockout strains, and PCR was used for identification (19).Since abrB3 in the SC2-M1 strain could not be deleted, the knockout and complementation strains for abrB3 were not obtained.

Quantitative reverse transcription-PCR
Strains were cultured at an initial OD 600 of 0.7 in LB broth, transferred to the fermenta tion medium at a 1:33 dilution, and then cultured at 37°C for 24 h before harvesting.Total RNA was extracted using the Total RNA Kit (Omega Bio-Tek, Norcross, GA, USA).Total RNA (800 ng) was reverse-transcribed into cDNA using the Evo M-MLV RT Kit (Accurate Biotechnology, Hunan, Co., Ltd., Changsha, China) according to the manufac turer's instructions.qRT-PCR was performed using the SYBR Green Premix Pro Taq HS qPCR Kit (Accurate Biotechnology, Hunan, Co., Ltd., Changsha, China) according to the manufacturer's instructions.The primers used for qRT-PCR in this study are listed in Table S1, and the concentration of primers used was 0.2 µM.Relative expression levels were determined using the 2 -ΔΔCt method with three technical replicates and three biological replicates (43).

Data analysis
Promoter prediction was performed using Softberry (http://www.softberry.com/berry.phtml)(44).Sequence alignment was performed using DNAMAN.GraphPad Prism 9.0 software (La Jolla, CA, USA) was for generating plots.Primer Premier 5 was used for primer design.Statistical differences for different treatments were analyzed using the t-test of Microsoft Excel (Microsoft, Redmond, WA, USA).A value of P < 0.05 was considered significant.

FIG 7
FIG 7 Proposed the regulatory model of Spo0A, AbrB3, and Abh in polymyxin synthesis.Abh and AbrB3 directly inhibited polymyxin production by binding to the pmx cluster promoter.Spo0A directly or indirectly promoted polymyxin production by binding to the promoter of pmx cluster and abrB3.AbrB3 and Abh inhibited each other by directly binding to their respective promoters.The regulatory effects of ResD3, RbsR3, and GlnR1 on polymyxin synthesis require further investigation.Arrows represent activation; bar-ended lines represent inhibition; curved arrows represent self-regulated; dotted arrows represent uncertain regulation.