Actinosynnema pretiosum X47 and its derivatives were cultured at 30°C on solid ISP2 medium (Ma et al., 2007) for spore production and conjugation. Escherichia coli strains DH5α, BL21, and ET12567 (pUZ8002; Kieser, 2000) were grown in Luria-Bertani (LB) medium or on LB agar at 37°C, for genetic engineering, protein expression, and conjugation, respectively. Antibiotics were used to select genetically modified A. pretiosum and E. coli strains. A. pretiosum is sensitive to apramycin and hygromycin B. Apramycin is used to screen mutant strains of A. pretiosum and hygromycin B is for complementary strain screening.

To construct the phoP mutant strain ΔphoP, flanking sequences of approximately 1,500 bp from the left and right sides of phoP were amplified from the X47 genome using primers PhoP-L-F/L-R and PhoP-R-F/R-R (Supplementary Table S1). An apramycin resistance cassette sequence was amplified from pSET152 (Bierman et al., 1992) using primers Apra-F and Apra-R. Then, the left arm, resistance cassette, and right arm were ligated and inserted into pMD18-T (TaKaRa) using the ClonExpress II One Step Cloning Kit. Subsequently, the fragment constituted by the left arm, resistance cassette, and right arm was released by digestion with XbaI and HindIII and inserted into pJTU1278 (He et al., 2010) to generate plasmid pM-phoP. The resulting plasmid was then transferred into E. coli ET12567 (pUZ8002), and the transformants were used as donors to conjugate with A. pretiosum X47, as described by Kieser (2000). Apramycin-resistant conjugates were selected and the genomic structure of the ΔphoP strains was verified by PCR with the primers PhoP-V-F/R. The phenotypes of the strains were observed when cultured on solid ISP2 medium.

To complement the phoP deletion mutant, a 2,270-bp fragment containing the coding region of phoP and phoR and a 300-bp region upstream of phoR was amplified from the X47 genome using primers PhoP-Com-F/R. The PCR product was inserted into pMD-18T and cloned as a Hind III fragment into pMS82, resulting in plasmid pC-phoP. Plasmid pC-phoP was transformed into E. coli ET12567 (pUZ8002), and hygromycin B was used to select the transformants. Conjugation between E. coli and the ΔphoP strain was carried out, conjugants were selected by hygromycin B, and the genomic structure of the complemented strain C-ΔphoP was confirmed by PCR analysis.

Spores of X47 and its derivatives were cultured on ISP2 medium, and sterile glass coverslips were inserted into the agar (Supplementary Figure S1). The coverslips were removed when the strains had grown for 48 or 72 h, and the coverslips with cultures were soaked in 2% glutaraldehyde for 2 h at room temperature. The fixed coverslips were washed three times with 0.1 M PBS buffer and treated with 1% osmic acid. Then, the coverslips were dehydrated in a critical point dryer (Quorum K850) then coated with gold and imaged with a scanning electron microscope (HITACHI Regulus 8,100).

AP-3 produced by A. pretiosum X47 and ΔphoP strains was extracted with ethyl acetate and detected by HPLC analysis as described (Zhong et al., 2019). HPLC was performed on a Diamonsil C18 Column (250 mm × 4.6 mm) with acetonitrile-water gradient as flow phase and UV detector at 254 nm.

The A. pretiosum X47 and ΔphoP strains were cultivated on solid ISP2 medium for 60 h. To isolate RNA, the mycelium was collected and mixed with Trizol reagent. RNA extraction and DNA removal were conducted by RNA extraction kit (Vazyme). The integrity of the RNA samples was analyzed by agarose gel electrophoresis, and the sample concentrations were detected using an Agilent 2,100 bioanalyzer. RNA sequencing was conducted by Novagene Bioinformatics Technology Co., Ltd. (Beijing, China).

The coding region of the phoP gene of A. pretiosum X47 was amplified using primers PhoP-pET15bF/R, and the PCR products were inserted into pET-15b (Novagen) to generate the phoP expression plasmid pPhoP. Then, pPhoP was introduced into BL21 (DE3), and the transformants were selected using ampicillin. The expression of PhoP protein was induced by 0.5 mM isopropyl β-D-1-thiogalactopyranoside at 25°C for 3–4 h when cell density reached an OD₆₀₀ reading of 0.4–0.6. Cell pellets were collected, and cell lysates were prepared by sonication. PhoP protein was purified on a Ni-NTA column (Sangon) and dialyzed in dialyzing buffer (50 mM NaH₂PO₄ and 50 mM NaCl, pH 8.0). The concentration of PhoP protein was determined with the BCA Protein Assay Kit (Beyotime).

The upstream regions of ssgB, flgB, flaD, cheA, fliM, swrD, and whiB were amplified by 5’-biotin-labeled primer pairs (Supplementary Table S1) to generate DNA probes. For EMSAs, 100 fmol probes were incubated with PhoP protein, binding buffer, and poly(dI-dC) for 20–30 min at room temperature. The mixtures were loaded on 8% non-denaturing polyacrylamide gels, and then, DNA was transferred to nylon membranes and fixed at 120°C. After blocking and washing, the probe signals were detected by the ECL Western Blotting Analysis System (GE Life).

The conservation of phoP/phoR genes in A. pretiosum X47, S. coelicolor, and M. tuberculosis H37Rv was analyzed by BLAST.¹ PHO box-like motif sequences were predicted by MEME software (Bailey et al., 2015), and searches for PHO box-like motif in the X47 genome were conducted using PREDetector software (Hiard et al., 2007).

Spores of X47 were cultured on ISP2 medium with 0 and 5 mM K₂HPO₄ for 72 h, and then, cultures were collected. Total RNA was obtained by RNA extraction kit (Sparkeasy), and cDNA was synthesized using ReverTra Ace qPCR RT Master Mix (TOYOBO). Real-time PCR was performed by using SYBR Premix Ex Taq kit (TaKaRa) on Roche LightCycler480 thermal cycler. The amounts of cDNA were normalized to the levels of major sigma factor gene hrdB. Results are the means of triplet experiments.

CNX65_RS33265 and CNX65_RS33270 were identified as the phoP and phoR homologs, respectively, in the A. pretiosum X47 genome. Sequence alignments (Figure 1) showed that PhoP and PhoR of A. pretiosum X47 share 82% and 51% amino acid identity with their counterparts in S. coelicolor A3(2; Martin, 2004) and 73% and 35% amino acid identity with the corresponding M. tuberculosis H37Rv proteins (Walters et al., 2006).

To investigate the function of PhoP in A. pretiosum, a ΔphoP mutant strain, lacking part of the phoP coding region from position +147 to +582 downstream of translation start site, was generated, which retains phoR and its promoter (Figures 2A,B). Inactivation of phoP resulted into significant morphological differences compared to strain X47 upon growth on ISP2 solid medium (Figure 2C). The arising of aerial mycelium and production of spores occurred much earlier in the ΔphoP strain than in the wild-type strain. The ΔphoP strain produced white aerial mycelium at 48 h, which turned light yellow during spore formation at 72 h, whereas strain X47 developed mainly vegetative mycelium and had little aerial mycelium at 48 and 72 h. The phoP-complemented strain C-ΔphoP had a phenotype similar to that of X47 (Figure 2C). Altogether these data indicated that PhoP is a crucial regulator controlling the developmental process of A. pretiosum.

Scanning electron microscopy (SEM) was carried out to visualize differences in the morphological differentiation of X47 and ΔphoP grown on ISP2 medium for 48 and 72 h. The mycelium of the X47 strain mycelium grew inside the growth medium (Figure 3A), whereas the ΔphoP strain showed aerial mycelium growing into the air at 48 h (Figure 3B). By 72 h, the X47 strain had produced only a small amount of aerial mycelium, whereas ΔphoP strain exhibited much more aerial mycelium (Figures 3C,D). While as the resolution images of samples cultured by 72 h were enhanced, it was found that the aerial mycelium of ΔphoP has separated into spores, and no spore formation was observed in X47 strain (Figures 3E,F). These results indicated that the formation of aerial mycelium and spores was accelerated in the ΔphoP mutant of A. pretiosum.

The productions of AP-3 in X47 and ΔphoP strains were analyzed by HPLC. The results (Supplementary Figure S2) showed that there is no significant difference in yield of AP-3 between X47 and ΔphoP strains, suggesting that PhoP is not essential for the AP-3 production under tested conditions.

To investigate the mechanism of morphological differentiation caused by PhoP, the transcription patterns of the ∆phoP and X47 strains grown on ISP2 medium for 60 h were compared by RNA-seq analysis. Compare to the X47 strain, 222 genes were differentially expressed in ∆phoP. Among these genes, 122 genes were downregulated, and 100 genes were upregulated. The expression of gene clusters responsible for the structure, assembly, and motility of flagella was remarkably higher in ∆phoP, including flgBCDGKLN, flaD, fliD-R, motA, and swrD (Table 1). In addition, whiB and ssgB, which are involved in cell division and sporulation of Streptomyces, and the chemotaxis-related genes cheA, cheW, cheX, and cheY were significantly upregulated in the ∆phoP strain. Overall, the results of RNA-seq indicated that PhoP negatively regulates the transcription of genes required for chemotaxis and for the formation of flagella and spores.

The above analysis demonstrated that PhoP controls multiple genes related to morphological development. Based on the transcriptional changes (Table 1) and gene arrangements (Supplementary Figure S3), it was deduced that flagellum biosynthesis and motility genes were contained within several operons (cheAW, swrD-motA-ompA-fliL, fliMOPQR-flhB, flaD-fliD-fliS, and flgBC-fliEFGHI-tolA-nlpC-fliK-flgD-flgG). To determine whether PhoP directly regulates the expression of these genes, electrophoretic mobility shift assays (EMSAs) were performed. His₆-PhoP protein was purified, and 200-bp sequences upstream of the flgB, flaD, ssgB, cheA, fliM, swrD, and whiB genes were amplified and labeled with biotin to create the probes. EMSA results (Figures 4A–C) showed that the flgB, flaD, and ssgB probes were shifted when incubated with 1.0 μg His₆-PhoP. As a control, for each target gene, excess unlabeled specific DNA fragments were added to the reactions and resulted into the apparition of free un-shifted probe demonstrating that the binding was specific, while excess unlabeled unspecific DNA fragments in the reaction did not influence the bindings. The above results suggested that PhoP specifically binds to the promoters of flgB, flaD, and ssgB. However, no shift was observed when PhoP was incubated with the promoters of cheA, fliM, whiB, or swrD (Figure 4D), suggesting that PhoP controls the transcription of these genes indirectly.

PhoP controls the transcription of target genes by binding to the PHO box, a conserved motif found in their promoters (Santos-Beneit, 2015). The results of sequence analysis indicated that the putative PHO box-like motifs identified in Streptomyces also existed in the promoter regions of flgB, flaD, and ssgB genes in A. pretiosum. In the flgB promoter, the putative PHO box-like motif (GTTCACCC) was located from position −141 to −134 bp upstream from transcription start position (TSP; Figure 5A), and for the ssgB promoter, the putative PHO box-like motif (GTTCAGGT) was located from position 272 to 279 bp downstream from TSP. Three PHO box-like motifs (GTTCACGC, GTTCACGC, and GTTCAGAC) were identified in the promoter of flaD, with the last one ending −84 bp upstream from TSP (Figure 5A). While eight consensus binding sequences with five conserved nucleotides for PhoP were revealed (Figure 5B). To confirm the role of PHO box-like motifs identified in this study in PhoP binding, EMSAs were performed by using probes containing mutation in the promoters of flgB, flaD, and ssgB genes. Mutation of four conserved nucleotides in the putative PHO box-like motifs severely reduced PhoP binding (Figures 5C,D), suggesting that the motifs are essential for PhoP binding to the promoters. These PHO box-like motifs have the 8-bp consensus motif of GTTCACNC, and hundreds of sites with this motif were identified in intergenic regions in the genome of strain X47 using PREDetector software (Hiard et al., 2007). Supplementary Table S2 shows a partial list of PHO box-like motifs in A. pretiosum, and the range of target genes further indicates that PhoP is a global regulator.

The concentration of Pi in ISP2 liquid medium was determined, which is 0.44 ± 0.05 μM, suggesting that Pi in ISP2 medium is scarce. Real-time PCR assays (Figure 6) showed that the expression of phop/phoR in X47 strain was significantly decreased when 5 mM K₂HPO₄ was added in the medium, which indicated that the transcription of phop/phoR was upregulated under the condition of Pi limitation.