CRISPR/dCas9-mediated transcriptional improvement of the biosynthetic gene cluster for the epothilone production in Myxococcus xanthus

Background The CRISPR/dCas9 system is a powerful tool to activate the transcription of target genes in eukaryotic or prokaryotic cells, but lacks assays in complex conditions, such as the biosynthesis of secondary metabolites. Results In this study, to improve the transcription of the heterologously expressed biosynthetic genes for the production of epothilones, we established the CRISPR/dCas9-mediated activation technique in Myxococcus xanthus and analyzed some key factors involving in the CRISPR/dCas9 activation. We firstly optimized the cas9 codon to fit the M. xanthus cells, mutated the gene to inactivate the nuclease activity, and constructed the dCas9-activator system in an epothilone producer. We compared the improvement efficiency of different sgRNAs on the production of epothilones and the expression of the biosynthetic genes. We also compared the improvement effects of different activator proteins, the ω and α subunits of RNA polymerase, and the sigma factors σ54 and CarQ. By using a copper-inducible promoter, we determined that higher expressions of dCas9-activator improved the activation effects. Conclusions Our results showed that the CRISPR/dCas-mediated transcription activation is a simple and broadly applicable technique to improve the transcriptional efficiency for the production of secondary metabolites in microorganisms. This is the first time to construct the CRISPR/dCas9 activation system in myxobacteria and the first time to assay the CRISPR/dCas9 activations for the biosynthesis of microbial secondary metabolites. Electronic supplementary material The online version of this article (10.1186/s12934-018-0867-1) contains supplementary material, which is available to authorized users.


Background
Clustered regularly interspaced short palindromic repeats (CRISPR) and the CRISPR associated (Cas) nuclease are widely distributed in bacteria and archaea, playing as an RNA-mediated adaptive immune system to cleave the invasive DNA [1,2]. When the nuclease activity of Cas9 is inactivated, the nuclease-dead protein (dCas9) is able to bind to the target sequence site with no cleavage activity. If guided to a suitable site, the dCas9 fusion with a transcriptional activator domain is able to recruit RNA polymerase (RNAP) to activate and improve the transcription of the target genes [3][4][5][6]. Compared with the traditional expression techniques that are based on the modification of genes and promoters, the CRISPR/dCas-mediated transcription activation method (CRISPRa) is simple, target-specific, programmable and frequently applicable [4,6,7]. Researchers have successfully employed the CRISPRa system to dissect the limiting component of a biochemical process, identify the molecular target of a drug, or activate key rate-limiting steps in a pathway. For example, Bikard et al. reported the

Open Access
Microbial Cell Factories *Correspondence: lilab@sdu.edu.cn 1 State Key Laboratory of Microbial Technology, School of Life Science, Shandong University, Jinan 250100, China Full list of author information is available at the end of the article programmable transcription activation of lacZ and gfp reporter genes in E. coli through a fusion of the ω subunit of RNAP to the C-terminus of dCas9; and a maximum of 23-fold induction was achieved with its Protospacer Adjacent Motif (PAM) positioned at the 59-nt upstream of the − 35 element [3]. However, the CRISPR/dCas9 activation system has not yet applied in complex conditions, such as the biosynthesis of microbial secondary metabolites. The biosynthesis of secondary metabolites in microorganisms often requires a collaboration of multiple enzymes, which are encoded by many genes that are organized in a big gene cluster. Transcription of those genes is complexly regulated [8], and efficient improvement and activation of the transcription of a microbial secondary metabolism is a challenge.
Recently, myxobacteria have been found to be the top microbial producers of secondary metabolites like fungi, actinomycetes and some species of the genus Bacillus [9][10][11]. Due to limited genetic performing techniques, studies of transcriptional regulation in myxobacteria are hampered. There are only a few reports on the transcriptional regulation for the biosynthesis of secondary metabolites in myxobacteria [11,12]. Epothilone is a kind of antitumor polyketides discovered in the broth of Sorangium cellulosum [13,14], biosynthesized by a 56-kb gene cluster [15]. In our previous studies, we integrated the whole gene cluster, including the promoter sequence, from S. cellulosum So0157-2 [16] into the M. xanthus genome by transposition insertion [17]. In this study, we attempted to establish the CRISPR/dCas9-mediated activation technique in M. xanthus, and investigated its improvement efficiency on the transcription of the heterologously expressed biosynthetic gene cluster for the production of epothilones. We compared the improvement efficiencies of different sgRNAs, different activator proteins, and the dCas9-activator expression levels on the transcriptions of different genes in the big gene cluster and the production of epothilones in M. xanthus. Our results suggested that CRISPRa is a simple and broadly applicable technique to improve the transcriptional efficiency of microbial secondary metabolites.

Codon-optimization of cas9 to fit the Myxococcus xanthus cells
The GC-content of the cas9 gene from Streptococcus pyogenes is 35.1%, which is much lower than the 68.9% percentage of M. xanthus DK1622 genome. We optimized the cas9 codon according to the codon usage bias in M. xanthus by using the JCAT software, and named the optimized gene as mxcas9 and the corresponding protein sequence was named as mxCas9. The sequence was deposited in GenBank under the ID number of 2074034.
In addition to the codon optimization, we also removed most commonly used enzymatic sites in the gene to facilitate further genetic manipulation. After the optimization, the GC content of the gene reaches 61.4%, and the codon adaptation index (CAI) value of mxcas9 has been adjusted from 0.05 to 0.97 for M. xanthus.
To generate nuclease-dead Cas9 protein, we made substitutions in mxcas9 (D10A and H840A; referred to Additional file 1: Figure S1) according to the studies by Jinek et al. [18], and named the gene mxdcas9. The gene was inserted into the pSWU30 plasmid under the control of T7A1 promoter [19,20]. The constructed plasmid, named pSW30-mxdCas9, contains an Mx8 site for the insertion in the genome of M. xanthus (Additional file 2: Figure  S2). To check the expression of mxdcas9 in M. xanthus, the gene of fluorescence reporter EGFP was fused to the 3′-end of the mxdcas9 gene to get a new plasmid named pSWdCas9gfp (Additional file 3: Figure S3). This plasmid was transformed by electroporation into M. xanthus DK1622. The RT-qPCR amplification result showed that the mxdcas9-egfp fused gene was able to express (Fig. 1a). We named this strain YL1610. Observed under the 500 ms exposure time, the M. xanthus cells transferred Fig. 1 Expression of mxdcas9 in mutant YL1610 (DK1622::pSWdmxCas9gfp) and the ZE9 control, measured by RT-qPCR amplification of the dcas9 gene (a) or fluorescence under a fluorescence microscope with the 500 ms exposure time (b). In this experiment, the M. xanthus strain ZE9 was used as a control with the pSWdCas9gfp plasmid exhibited green fluorescence, while the cells of M. xanthus without the plasmid did not (Fig. 1b). The result indicated that the mxdcas9 gene was able to express in M. xanthus cells normally.

Effects of different sgRNA-binding sites on the improvement of epothilone production
Judged by their overlapping stop, the presence of the ribosome-binding site of the downstream gene within the coding sequence of the upstream gene and no apparent intergenic transcriptional terminator, the big gene cluster for the biosynthesis of epothilones was suggested to be transcriptionally coupled as an operon under the controlled by the promoter region immediately preceding epoA [15]. We previously determined that, by using the primer extension analysis technique, the 841 nucleotide (nt) promoter sequence of the biosynthetic gene cluster contains two transcriptional start sites (TSSs), one is at 246 bp (TSS1) and the other is at 193 bp (TSS2) in front of the translation start site (TLS) for the transcription [17]. TSS1 has strong promotion activity for transcription while the effect of TSS2 is weak [21]. Figure 2a sketches the gene cluster for the biosynthesis of epothilones and Fig. 2b exhibits the two major components of epothilones (epothilones A and B). In order to activate the whole epothilone gene cluster for the production of epothilones, we designed the spacer target sequences complemented on the promoter sequence (Fig. 2a). Single guide RNA (sgRNA) sequence, a combination of the CRISPR associated RNA (crRNA) and the trans-activation crRNA (tracrRNA), was constructed according to the previous method [18] for to guide mxdCas9 to the target DNA sites. We designed five spacers for the construction of sgRNAs, which contained a 20-nt specific target sequence, a tracrRNA scaffold and a U6 terminator (Fig. 2c). Bikard et al. once demonstrated that approximately 80-110-nt before a TSS provides an optimal distance for a RNA polymerase to bind to the promoter and The molecular structures of epothilones A and B. c Design of sgRNA scalfold, containing a 20-nt specific targeting sequence, the tracrRNA scaffold and a U6 terminator. d The ω subunit protein was designed to fuse to the C-terminal of the mxdCas9 with a linker. This fusion gene is under the control of T7A1 promoter. The two mutation sites (D10A and H840A) are shown to increase gene expression [3]. Because there are two TSSs in front of the TLS of the epothilone gene cluster, the five spacers were designed with a distance ranging from 53 bp (Epoact5) to 103 bp (Epoact4) upstream of TSS1 (Fig. 2a). In parallel, the distance range is 106 bp to 156 bp upstream of TSS2. The 20-nt spacers Epoact1, 4 and 5 are on the coding strand, while the Epoact2 and 3 are on the non-coding strand. Each CRISPR target sequence has an "NGG" PAM sequence at the 3′-end of the spacer for proper binding between the protein-RNA complex and the target DNA sequence. We used the pilA promoter from M. xanthus to drive the transcription of sgRNA. The off-target efficiency of spacer was calculated by using the 1.0 version of the software CasOT [22]. Because a spacer with high off-target efficiency does not function well in the transcriptional activation [23][24][25], four spacers (Epoact2, 3, 4 and 5) were designed having low off-target efficiency, and Epoact1, of which the offtarget prediction was high, was employed as a control. Detailed information of these spacers and their distances to TSS1/TSS2 are listed in Table 1.
Bikard et al. demonstrated that the fusion of dCas9 and ω protein could increase the transcription level of the targeted gene and corresponding β-galactosidase activity in either E. coli or S. pneumoniae [3]. We fused the ω protein gene from M. xanthus DK1622 (Table 2) to the 3′-end of mxdcas9 with a 27-nt linker (Fig. 2d), and constructed pSWdCas9-ω by cloning the fused gene into the pSW30 plasmid. The plasmid was transformed by electroporation into M. xanthus ZE9 to obtain the mutant strain YL1611. We inserted different sgRNA sequences, each containing one of the five selected spacers Epoact1-Epoact5, into the autonomously replicative plasmid pZJY41, respectively (the constructed plasmids were named as p41sg1 to p41sg5). These plasmids were then transformed by electroporation into YL1611, respectively, obtaining the YL1612-YL1616 mutant strains (Additional file 4: Figure S4). We found that, except the Epoact1-containing YL1612, the other four mutant strains showed a significant increase in the production of either epothilone A or epothilone B (Fig. 3a). The production of epothilones A and B were 6.08-and 1.70-fold higher in YL1615, or 6.80-fold and 1.77-fold higher in YL1616 than that in ZE9, respectively. The distances between the binding site and TSS1/TSS2 are either 103 bp/156 bp for Epoact4 (in mutant YL1615), or 53 bp/106 bp for Epoact5 (in mutant YL1616). Since an optimal space for RNA polymerase binding to the promoter is approximately 80-110-nt before a TSS [3], we suggested that the production increase in YL1615 was probably majorly due to the increased RNA polymerase at the TSS1 site, while the increase in YL1616 majorly resulted from the increased enzyme at the TSS2 site. Thus, the recruiting efficiency of different sgRNAs was probably similar for either the TSS1 or TSS2 sites. The results from YL1613 (Epoact2) and YL1614 (Epoact3), in which the Epoact2 and Epoact3 spacers were designed on the non-coding strand, showed an agreement with  the suggestion, i.e. sgRNAs having similar recruiting efficiency for either the TSS1 or TSS2 sites. Compared with to the ancestral strain ZE9, the epothilone production did not increase in YL1612. We suggested that the high off-target efficiency of Epoact1 was probably the reason for the inability of Epoact1 in activating the biosynthesis of epothilones. The assembling line for the production of epothilones is a hybrid complex of PKS/NRPS enzymes. The length of the epothilone gene cluster is approximately 56-kb, containing seven unidirectional ORFs (referred to Fig. 2a), encoding five polyketide synthases (PKSs, including EpoA, EpoB, EpoC, EpoD, EpoE), a non-ribosomal peptide synthetase (NRPS, EpoP) and a cytochrome P450 oxidase (EpoF). Compared with the expressions in ZE9, the transcriptions of different genes increased significantly in these mutant strains, except YL1612 (Fig. 3b). Interestingly, the transcriptional levels of the first three genes increased more, while the transcription improvements of the hinder genes were less. For example, compared with that of ZE9, the transcriptional increases of the first three genes, i.e. epoA, epoP and epoB in YL1614 and YL1616 were 2.72-/4.19-/2.90-fold and 5.43-/7.93-/4.46-fold, respectively; whereas the increases of the following epoC, epoD, epoE and epoF genes in these two strains were 2.23-/1.42-/1.83-/1.89-fold and 3.52-/2.97-/3.26-/2.71-fold, respectively. The above results suggested that the fusion of mxdCas9 and the ω protein was able to be guided by the sgRNA sequence bound onto the promoter upstream sequence, and then recruited xanthus ZE9 were each set as 1, and the expressions of the seven genes of mutant strains were shown as the relative expressions to that of their corresponding genes in ZE9. The error bars represent the standard deviation of three independent experiments. For statistical analysis between the ancestral strain and mutant strains, the signals of ** and * mean p < 0.01 and p < 0.05, respectively. c RT-qPCR analysis of the expression levels of the seven genes for the biosynthesis of epothilones of M. xanthus ZE9 at 48 h of incubation. The expression of epoA gene was set as 1, and the expressions of the other six genes in ZE9 were the relative expressions to that of epoA more RNAP proteins to improve the transcription of the epothilone gene cluster. The improvement was efficient for the whole gene cluster.
Notably, although the gene cluster was suggested to be co-transcribed under the control of the promoter preceding epoA [15] and the mxdCas9-activator was able to improve the transcription of the whole epothilone gene cluster, the seven genes significantly varied their expressions (Fig. 3c), which is similar to the transcriptomics analysis result of M. xanthus recombinants with the insertions of epothilone biosynthetic genes [17]. We presumed that there were some possible additional internal promoters within the gene cluster, which might involve in the transcriptional regulation of the clustered genes. The biosynthetic gene cluster has to be rechecked and further studies are required to be made on the possible internal promoters.
Another interesting result was that, while the productions of epothilones A and B were both significantly increased in YL1615 and YL1616 (containing Epoact4 and Epoact5, respectively), the Epoact2 and Epoact3 spacers in YL1613 and YL1614 increased the production of epothilone A significantly only (2.57-and 5.21fold higher than that of ZE9, respectively) but not that of epothilone B (Fig. 3a). It is known that epothilones are normally produced as a mixture of epothilones A and B (two major components) in either the original S. cellulosum producers [13,26] or the heterologous M. xanthus hosts [17,27]. The difference between epothilones A and B is the presence of a methyl group at carbon 12 of epothilone B (referred to Fig. 2b). The biosynthesis of mixed epothilones A and B was suggested to be due to the promiscuity in the selection of malonyl-CoA and methylmalonyl-CoA by the second acyltransferase of the EpoC [15,28,29]. Thus, the availability of malonyl-CoA and methylmalonyl-CoA will affect the ratio of epothilones A and B produced [27]. We presumed that the unilateral significant increase of epothilone A in the YL1613 and YL1614 strains was probably due to the side effects of Epoact2 and Epoact3 spacers on the alteration of malonyl-CoA and methylmalonyl-CoA pools, which, however, requires detailed studies.

Effects of different activation factors on the transcriptional improvement and epothilone production
Dove et al. illustrated that the combination of a DNAbound protein (lambda cI repressor) and any RNAP component can improve the RNAP transcription efficiency in E. coli cells [30] In addition to the ω-subunit of RNAP, we further assayed activation effects of different transcriptional factors, including the α-subunit of RNAP, sigma factors σ54 and CarQ. Sigma factors control the specificity of gene transcription by binding to RNAP and the fusion of a sigma factor to dCas9 was probably able to recruit more RNAP to the promoter sequence. As the specific sigma factor for epothilone biosynthesis was unknown yet, we chose the global sigma factor σ54 and the ECF (extracytoplasmic function) sigma factor CarQ for the assay. CarQ is known to be responsible for the transcription initiation of carQRS, the central genes for carotenogenesis in M. xanthus [31]. We presumed that any activator protein that was fused to dCas9 was probably able to recruit more RNA polymerases to activate and improve the transcription of the target genes when guided by a specific sgRNA to a suitable site. The detailed information of these four activator proteins is shown in Table 2. We fused α subunit, σ54 and CarQ respectively to the dCas9-activator by using the similar process for the ω subunit, and constructed plasmids pSWdCas9-α, pSWdCas9-σ54 and pSWdCas9-CarQ using the similar process for the construction of pSWdCas9-ω (Additional file 3: Figure S3). These plasmids were transformed by electroporation into M. xanthus ZE9 to obtain mutant strains YL1617, YL1618 and YL1619, respectively. We transformed the pZJY41 plasmid with the Epoact5containing sgRNA sequence into these three mutants, obtaining the mutant strains YL1620, YL1621 and YL1622, respectively.
The cultures of these mutant strains showed different colors on CYE plate, and the strains with the same introduced activators normally had the same colony colors (Fig. 4a). For example, the colonies of the YL1619 and YL1622 strains with pSWdCas9-CarQ were red. This is probably also a side effect of mxdCas9-CarQ on carotenogenesis, suggesting the CRISPR/dCas9-mediated activation is often not strand-specific [32]. However, the fermentation result showed that the three activatorproteins were also able to increase the production of epothilones significantly (Fig. 4b). The improvement efficiencies caused by different activators were similar in the production of epothilone A, but varied in epothilone B. The activation effects of the α-subunit and CarQ were similar, but both were weaker than the ω-subunit did. Among them, the global sigma factor σ54 showed the weakest effect for the production of epothilones. The results suggested that, besides the activation effects on the target epothilone biosynthesis, different activator proteins might have additional activation effects on other genes, and these unrequired additional activation effects were dependent upon the characteristics of the activators, as well as the off-target efficiency of spacers. The RT-qPCR analysis revealed that, consistent with the increase in production of epothilones, the expression levels of the seven ORFs in the epothilone biosynthetic gene cluster were mostly increased (Fig. 4c). Similarly, the transcriptional levels of the former genes were higher than that the hinder genes.

Effects of the expression levels of dCas9-activator-fusion on the gene transcription and epothilone production
To investigate effects of the expression levels of mxd-Cas9-activator-fusion, we replaced the T7A1 constitutive promoter with an inducible promoter. In M. xanthus, four inducible expression systems have been developed, i.e. the light-inducible CarQ promoter [33][34][35], the IPTG-based PpilA-lacO-RBS promoter [36], the MerR like repressor CarH-based promoter [37,38], and the copper-inducible promoter [39]. Among them, the copper-inducible promoter is suitable for the general usage for its simple and highly controllable manner that the cuoA promoter is inducible by the addition of copper and the transcription level by the promoter is linearly dependent upon the copper concentration within a suitable extent [39,40]. There is almost no influence on the growth of M. xanthus cells with the addition of a copper concentration up to 500 μM in medium [39].
We chose the copper-inducible promoter cuoA to drive the expression of mxdCas9-ω subunit with the guidance of the Epoact5 sgRNA. First, we constructed plasmid pSWcuodCas9-ω and electroporated it into ZE9 to (ZE9::pSWmxdCas9-ω-p41sg5), YL1617 (ZE9::pSWmxdCas9-α), YL1618 (ZE9::pSWmxdCas9-σ54), YL1619 (ZE9::pSWmxdCas9-CarQ), YL1620 (ZE9::pSWmxdCas9-α-p41sg5), YL1621 (ZE9::pSWmxdCas9-σ54-p41sg5), YL1622 (ZE9::pSWmxdCas9-CarQ-p41sg5). b The yield and the summation of epothilones A and B in mutants (YL1616, YL1620, YL1621, YL1622) and M. xanthus ZE9. c RT-qPCR analysis of the seven genes for the biosynthesis of epothilones in M. xanthus strains at 48 h of incubation. The expressions of the seven genes in ZE9 were each set as 1, and the expressions of seven genes in mutant strains were the relative expressions to them. The error bars represent the standard deviation of three independent experiments. For statistical analysis between the ancestral type strain and mutant strains, the signals of ** and * mean p < 0.01 and p < 0.05, respectively obtain mutant strain YL1623. Then, the p41sg5 plasmid was electroporated into YL1623, producing the mutant YL1624. In our previous studies on the CRISPR/Cas9induced deletion, we demonstrated that the M. xanthus cells possessing the cuoA promoter-controlled cas9 gene were unable to tolerate a high concentration of copper; the strain grew normally at 15 μM copper concentration and the 25 μM concentration was lethal for the strains [40]. However, when we cultivated the M. xanthus strain with the cuoA-induced mxdcas9 gene in media containing 500 μM concentration of CuSO 4 , the cells showed almost the same growth ability as that of the strain with T7A1-controlled mxdcas9 under the conditions without the addition of copper, which was probably due to the inactivation of the mxdCas9 nuclease activity. We compared the epothilone production abilities of the mutant strains YL1616 (the mxdCas9-ω fusion was driven by the constitutive promoter T7A1) and YL1624. The YL1616 strain was grown in normal cultures without the addition of copper, while the YL1624 strain was grown with the addition of 25, 120 and 200 μM concentration of CuSO 4 . The result showed that, compared with that in the original ZE9 strain, the productions of epothilones in YL1616 and YL1624 were all significantly increased (p < 0.01). However, compared with that in YL1616, the production of epothilones was not significantly changed in YL1624 grown with 25 μM copper concentration (Fig. 5a). With the copper concentration increased, the production of epothilone B in YL1624 was significantly higher than that in YL1616 (p < 0.05 at 120 μM and p < 0.01 at 200 μM). Thus, there was a positive correlation between the epothilone production in YL1624 and the copper concentration in cultures. The RT-qPCR analysis also showed that, compared with that of YL1616, the expressions of most of the seven genes were not significantly changed in the YL1624 mutant grown with 25 μM copper concentration (Fig. 5b). With the copper concentration increased, the expressions of most of the seven genes were significantly increased in YL1624, when compared with that in Yl1616. This result indicated that the expression increase of the mxdCas9-activator fusion is able to improve the target transcriptional efficiency, and consequently increase the production of epothilones.

Conclusions
Efficient expressions of exotic genes in heterologous hosts are not only important for the yield increase of desired products, but also useful in the discovery of novel compounds [41,42]. Promoters play a fundamental role for the regulation of gene expressions, and many efforts have been made to optimize promoters [43,44]. CRISPR/ dCas9 system can regulate gene expression without the editing performance on the genome or genes, like the eukaryotic RNAi technique [45,46]. Here, we described a method using sgRNA to guide mxdCas9-activator to improve the expressions of the exotic epothilone biosynthetic genes in Myxococcus xanthus. This study is the first report exploring the CRISPRa technique in M. xanthus and the CRISPR/mxdCas9-based activation for the gene cluster expression of secondary metabolites. Our results suggested that efficient activation of the CRISPRa technique depends on the capacity of recruiting RNAP, as well as the DNA binding specification of the activator protein. Strong promoter to improve the expression of mxdCas9-activator is beneficial to increase the biosynthesis further. This method provides the inspiration and experience for the exploration of microbial metabolisms. The CRISPR/mxdCas9 technique might be also used for the programmable repression of gene transcription in Myxococcus.

Construction of plasmids and mutants
The plasmids and primers used in this study are provided in Additional file 6: Tables S2 and Additional file 7: S3. Plasmid pSW30-mxCas9 was generated by amplifying codon-optimized mxcas9 with the primers T7cas9Eco-F/cas9Hin-R, which was inserted into the EcoRI/HindIII site of pSWU30. The mutations of D10A (GAC→GGC) and H840A (CAC→GCC) were performed in mxcas9 of the pSW30-mxCas9 plasmid. The mutated gene was amplified from the plasmid with primers M840In-F/ M10In-R and M10Out-F/M840Out-R. Plasmid pSW30-mxdCas9 was constructed with linear-linear homologous recombination mediated by full length RecET in GB05-dir [47]. The gfp reporter gene was inserted into the pSW30-mxdCas9 plasmid by using the RecET linear-linear homologous recombination to generate pSWmxdCas9gfp.
Fusion of the activation factors (ω subunit, α subunit, σ54, CarQ) with mxdCas9 was achieved by amplification of pSW30-mxdCas9 with primers exeffector-F/exeffector-R, and amplification of genes encoding activation proteins with primers ωRed-F/ωRed-R, αRed-F/αRed-R, σ54Red-F/σ54Red-R, CarQRed-F/CarQRed-R to create the C-terminal fusions, respectively, followed by Red/ ET linear-linear homologous recombination. The genes were all PCR amplified from the M. xanthus DK1622 genome template. Plasmid p41sg was used to express sgRNA. The plasmid was constructed by amplifying synthetic pilA promoter and tracrRNA sequence with primers gRNABamHI-F/sgRNAKpnI-R and inserting at the BamHI/KpnI site of pZJY41. Plasmids p41sg1 to p41sg5 were constructed by using p41sg as a template to amplify DNA fragment with primers sgRNA1-F to sgRNA5-F/ sgRNA-R and linked with themselves.
The plasmid with copper-inducible promoter, pSWcuodCas9-ω, was generated by introducing the Pcuo promoter to the pSWdCas9-ω plasmid through amplification of this plasmid with primers PcuoRed-F/PcuoRed-R and expro-F/expro-R, followed by Red/ ET linear-linear homologous recombination of the two products.
All the plasmids above were electroporated into M. xanthus DK1622 or ZE9. The related genes following its own promoters were inserted at the attB site in the genome, forming the related mutant strains. The mutants were then cultivated on CYE plates containing tetracycline (10 μg/mL) and apramycin (25 μg/mL). Plasmids p41sg1 to p41sg5 were electroporated into the constructed mutants, which were cultivated on CYE plates containing tetracycline (10 μg/mL), apramycin (25 μg/ mL), and kanamycin (40 μg/mL). The mutants were identified and verified by PCR amplification and sequencing. Restriction enzymes, DNA ligase, and other DNA enzymes were used according to the manufacturers' recommendations. All fragments were validated by Sanger 3730 sequencing.

Fluorescence assay
Green fluorescence protein (GFP) assays were performed in M. xanthus DK1622 mutant (YL1610). Fluorescence was measured under a fluorescence microscope. In this experiment, background fluorescence, or auto-fluorescence, was measured using a control strain lacking the GFP reporter (M. xanthus DK1622).

Fermentation and product detection
M. xanthus strains containing mxdCas9-activator, plasmids with sgRNA and the complete epothilone gene cluster were grown for 24 h in 50 mL CYE medium supplemented with Apra Tet and Km antibiotics. The cultures were used as seeds to inoculate at a ratio of 2:100 into the CYE medium containing 2% of the XAD-16 resin for the adsorption of epothilone products. After 7 days rotation at the 250 rpm and 30 °C, the mixtures of cells and resin were harvested by centrifugation and extracted with two volume of methanol by shaking at room temperature overnight [26]. After centrifuged for 20 min, an aliquot of 20 μL of the sample was injected into a HPLC system interfaced with a Finnigan MSQ classic quadrupole mass spectrometer (Thermo Finnigan, USA). The analysis was carried out on a Shim-pack MRC-ODS RP column (4.6 mm × 250 mm, 4.60 μm; Shimadzu, Japan) at a temperature of 28 °C with a mobile phase of 60% methanol (HPLC grade) and ddH 2 O at a flow rate of 1.0 mL/ min. The production levels of epothilones were averaged from three independent cultivations and extractions.

RT-qPCR
After harvested, the M. xanthus mutant cells were adjusted to the concentration of approximately 1.75 × 10 10 cells/mL. The cell suspension was washed three times using TPM buffer. RNA was extracted immediately with a BIOZOL total RNA extraction reagent (BioFlux) following the manufacturer's instructions. Genomic DNA contamination was removed by using DNA eraser supplied in the PrimeScript RT Reagent kit with gDNA Eraser (TaKaRa). The purified RNA extracts were transcribed reversely to cDNA and stored in aliquots at − 80 °C. Quantitative real-time PCR was performed in a total reaction volume of 25 μL, containing 250 nM primers, 12.5 μL of SYBR Premix Ex Taq GC mix (TaKaRa), 10.5 μL of RNase-free water (TaKaRa), and 1 μL of a 10-fold-diluted cDNA template. PCR was conducted in a Roche Light Cycler 480 sequence detection system, following the program: 3 min at 95 °C, followed by 40 cycles of 30 s at 95 °C, 30 s at 55 °C, and 15 s at 72 °C. The gapA gene, encoding for glyceraldehyde-3-phosphate dehydrogenase, was used as the normalization signal. Calibration curves of gapA and mxdCas9 were generated from a 10-fold dilution of M. xanthus ZE9 genomic DNA. The primer pairs used for each gene are listed in Additional file 7: Table S3.
To examine the significance of differences, the Student's t test was conducted using the SPSS software. The significance and the high significance of differences were set as p < 0.05 and < 0.01, respectively. The biological replicates from three independent experiments were employed to calculate the significance values.