Construction of high-performance thermo-inducible modules
Inducible gene expression systems typically comprise two main components: a transcriptional regulator (TF) and a cognate target promoter harbouring TF-specific DNA-binding sequences. To construct thermo-inducible modules, the coding region of rheA with the noncoding intergenic region between rheA and hsp18 was assembled with the promoter-less neomycin phosphotransferase gene (neo). It is worth noting that the promoters of rheA and hsp18 are located in a divergent orientation, and the RheA DNA-binding sequences (RheA operator, rheO) is embedded within the intergenic region (Fig. 1a and Supplementary Fig. 1). The construct was transferred into S. albidoflavus J1074 to obtain J1074-RS01-neo. For comparison, J1074-neo and J1074-PhrdB-neo were included to serve as negative control and positive control, respectively (Fig. 1a). J1074-neo contains the promoter-less neo cassette, while J1074-PhrdB-neo contains the neo cassette driven by the constitutive hrdB promoter2. J1074-RS01-neo was then cultivated, along with J1074-neo and J1074-PhrdB-neo, on minimal medium (MM) agar plates supplemented with increasing concentrations of kanamycin. As seen here, J1074-neo displayed resistance to kanamycin at 1.0 µg/mL, whereas it became sensitive at 2.5 µg/mL of kanamycin (Fig. 1b). The phenomenon is consistent with our previous findings that S. albidoflavus J1074 exhibited a low level of intrinsic kanamycin resistance2. When cultivated at 28 °C, J1074-RS01-neo exhibited kanamycin resistance up to 25.0 µg/mL, suggesting that the native hsp18 promoter exhibited a relatively high level of leaky expression (Fig. 1b). To mitigate leakiness, TRS01 was generated by replacing the native promoter of rheA with the constitutive hrdB promoter, and the tfd terminator was included to avoid interference from unwanted transcriptional readthrough (Fig. S1). Moreover, an additional rheO was inserted immediately after the transcription start site (TSS) of hsp18 promoter in TRS01 to generate TRS02 (Supplementary Fig. 1). Transfer of the two constructs into S. albidoflavus J1074 generated J1074-TRS01-neo and J1074-TRS02-neo, respectively (Fig. 1a). When cultivated at 28 °C, both J1074-TRS01-neo and J1074-TRS02-neo can grow on MM agar plates containing at 2.5 µg/mL kanamycin. However, both strains are unable to tolerate 5.0 µg/mL or higher kanamycin concentrations (Fig. 1b), suggesting that a strong promoter driven repressor is an effective way to minimize leaky expression. When cultivated at 37 °C, J1074-TRS01-neo and J1074-TRS02-neo exhibited a high level of expression similar to that of the constitutive hrdB promoter, with kanamycin resistance up to 50 µg/mL (Fig. 1b). Next, electrophoretic mobility shift assays (EMSAs) were performed to examine protein-DNA interactions. For this purpose, recombinant RheA protein (∼25.4 kDa, calculated mass) was purified from E. coli BL21 (DE3) harboring pET28a::rheA (Supplementary Fig. 2). Of interest is that one shift band was observed for the original RS01 probe in a RheA concentration dependent manner, while two shifted bands for the RS02 probe with two copies of rheO was noticed at a lower RheA protein concentration. With an increase in protein concentration, the abundance of the upper band was increased obviously accompanied by disappearance of the lower band (Supplementary Fig. 3). Taken together, the results suggest that TRS01 and TRS02 exhibit tight modulation at 28 °C and a high level of induced gene expression at 37 °C in S. albidoflavus J1074.
Tuning the bio-switch with a physiological range of temperatures
To examine the dynamic range of the thermo-sensing modules, TRS02 was selected to drive the expression of the reporter gene gusA, which encodes the β-glucuronidase enzyme (Fig. 2a). For comparison, the promoter-less gusA and gusA driven by the constitutive hrdB promoter were included to serve as negative control and positive control, respectively5. The resulting strain J1074-TRS02-gusA, J1074-PhrdB-gusA and J1074-gusA were cultivated on R2 agar plates for visualization of the characteristic blue color of GusA activity. Six different temperatures, covering a physiological range from 26 to 37 °C, were used for cultivation of these recombinant strains. A rise in cultivation temperature stimulates gusA expression, which in turn results in a gradual increase in blue color intensity. A vivid blue color was observed when J1074-TRS02-gusA was incubated at 34 and 37 °C (Fig. 2b). The color intensity of J1074-TRS02-gusA is similar to that of the positive control J1074-PhrdB-gusA. To quantify GusA activity, the strains were inoculated in liquid R5A and cultivated at six different temperatures. A gradual induction of GusA activity was observed with an increase in the cultivation temperature, and a full induction of gusA expression was observed at 37 °C when cultivated for 24 h or 34 °C when cultivated for 48 h (Fig. 2c). The regulatory system shows sharp thermal transitions with a dynamic range of ∼62-fold induction.
To investigate the performance of TRS02 in other Streptomyces species, the construct was also transferred into the model organisms S. coelicolor M1146 and S. lividans TK24 to obtain M1146-TRS02-gusA and TK24-TRS02-gusA, respectively. An obvious induction of GusA activity was observed with M1146-TRS02-gusA (Supplementary Fig. 4a). It should be noted that the strain exhibited a low level of leaky expression as judged by the light blue color of colonies at lower incubation temperatures. TK24-TRS02-gusA exhibited a high level of induced GusA activity with no leaky expression (Supplementary Fig. 4b). Thus, we established a high-performance thermo-inducible system exhibiting low leakiness and high maximal expression. The results suggested that TRS02 provides a switch-like control of gene expression with temperature as stimulation input, and the temperature-sensing module is functional in three model Streptomyces species.
The thermal bio-switch enables bidirectional control of gene expression
As mentioned earlier, existing induction systems have certain limitations. The main disadvantage is a lack of reversibility associated with chemical inducible systems. In terms of reversibility, temperature has superiority over chemical inducers, as it can be easily adjusted. Thus, TRS02 was used to drive the expression of enhanced green fluorescent protein (eGFP). The resulting J1074-TRS02-eGFP was cultivated in liquid R5A with incubation temperatures switched back and forth between alternating temperatures. No eGFP expression was detected when the culture was initially kept at 28 °C for 12 h. An obvious eGFP expression was observed after the culture was shifted to 34 °C (Fig. 3). Reversibility can be achieved when the culture was switched back to 28 °C, followed by fluoresce activation at 37 °C and then fluoresce quenching at 28 °C (Fig. 3). Of note is that residual fluorescent signals were observed with J1074-TRS02-eGFP in the second and third round cultivations at 28 °C. Considering J1074-eGFP, a derivative of S. albidoflavus J1074 containing a promoter-less eGFP, also displayed low intensity of fluorescent signals, we speculate that the fluorescent signals in J1074-TRS02-eGFP may arise from a combination of background signals and the residual eGFP left from the preceding activation event. Our results demonstrated that the thermal bio-switch, designated as StrepT-switch thereafter, enables bidirectional control of gene expression.
The bio-switch facilitates highly efficient CRISPR/Cas9-mediated genome engineering
It is well documented that CRISPR/Cas9 shows high toxicity to the host, thus limiting its application in many Streptomyces strains6. To demonstrate the utility of our bio-switch, StrepT-switch was used for flexible control of the CRISPR/Cas9-mediated genome editing tools. First, StrepT-switch was used for CRISPR/Cas9-mediated target knock-in to activate the expression of indC. The indC gene was identified within a small gene cluster encoding the blue pigment indigoidine7. Under routine laboratory conditions, the gene cluster remained silent in S. albidoflavus J1074 as no pigment formation was observed. A previous study found that substitution of the native indC promoter with constitutive promoter led to the activation of the gene cluster7. For this purpose, the tipA promoter of Cas9 was replaced with StrepT-switch for knock-in of the constitutive kasO* promoter proceeding the coding region of indC (Fig. 4a). When introduced into S. albidoflavus J1074, there is only a few colonies observed on MS agar plates with the construct containing tipA promoter-driven Cas9 (Fig. 4b). This observation is in agreement with the fact that the tipA promoter exhibits a high level of leaky expression even in the absence of the thiostrepton inducer. In contrast, more transformants were obtained with the construct containing StrepT-switch-driven Cas9, indicating that StrepT-switch ensures almost no expression of Cas9 and avoid its toxicity when the transformants were maintained at 28 °C. Once we obtained the transformants, they can be passed on MS agar and then incubate at 34 °C to activation Cas9 expression for CRISPR/Cas9-mediated target knock-in (Fig. 4c). A similar strategy was used for CRISPR/Cas9-mediated target knock-out. An obvious high efficiency was observed with knock-out of approximate 48 kb of the daptomycin gene cluster from the chromosome of Streptomyces roseosporus (Supplementary Fig. 5). StrepT-switch-driven Cas9 has also been used successfully for knock-out of target genes in the industrial Streptomyces strains (Data not shown), while no transformants were obtained with tipA-driven Cas9, suggesting that StrepT-switch is particularly useful for strains with low DNA transformation efficiency. The results showed that StrepT-switch improved the DNA transformation efficiency by reducing the toxicity of Cas9 at lower temperature, and thus greatly increased the chance to obtain CRISPR/Cas9-mediated knock-in or knock-out mutants.
Programmable control of antibiotic production and morphology differentiation via thermo-inducible CRISPRi
To demonstrate the utility of our thermal bio-switch, StrepT-switch was used to repress gene expression via CRISPR interference (CRISPRi) in a programmable manner. In previous studies, the dCas9 gene, a catalytically dead variant of Streptococcus pyogenes Cas9, was controlled either by the thiostrepton-inducible tipA promoter or the constitutive ermE* promoter2, 8, 9. For programmable control of gene repression, a tight regulatory expression of dCas9 is highly preferred. To this end, pSET152::dCas9-actI1 was selected as it has been used for efficient repression of actinorhodin (ACT) production by targeting actI-ORF18. Thus, the expression of dCas9 was subjected to the control of StrepT-switch rather than the constitutive ermE* promoter (Fig. 5a). The resulting construct was then introduced into S. coelicolor M145 to obtain M145- StrepT-switch-dCas9-actI1. For comparison, M145-dCas9-actI1 was included to serve as a positive control. Unlike S. coelicolor M145, which could produce the blue pigment ACT and the red pigment undecylprodigiosins (RED), M145-dCas9-actI1 lost the ability to produce ACT on R5MS agar plates regardless of the incubation temperature (Fig. 5b). M145-StrepT-switch-dCas9-actI1 retained the ability to produce ACT when cultivated at 28 °C. However, it lost the ability to produce ACT and retained the ability to produce RED when cultivated at 37 °C (Fig. 5b). The same strategy was also used for flexible control of gene expression involved in morphology differentiation. Thermal inducible repression of ftsZ resulted in aerial hyphae that lack both cross-walls and sporulation septa in S. venezuelae (Fig. 5c). These results suggest that StrepT-switch can be used for programmable control of gene expression via CRISPRi based upon the thermo-inducible expression of dCas9.
Induction of ZouA-dependent amplification of act gene cluster multiplies antibiotic yield
A previous study identified a ZouA-dependent DNA amplification system in S. kanamyceticus10. The ZouA amplification system consists of three genetic elements, a site-specific relaxase ZouA, and two oriT-like recombination sites RsA and RsB. It has been used successfully for amplification of the act gene cluster in S. coelicolor MT111010, bleomycin in Streptomyces verticillus11, and spinosad in S. coelicolor M114612. To expand the utility of StrepT-switch, the regulatory system was used to drive the expression of zouA. For this purpose, the act gene cluster was engineered with flanking sequences, including RsA, RsB, and a neo cassette (Fig. 6a). The engineered gene cluster was transferred into S. coelicolor M1146 to obtain M1146-actAB. Incorporation of StrepT-switch-driven zouA generated M1146-actAB-StrepT-switch-zouA. Twelve conjugates were randomly streaked on R5MS agar plates containing 50 µg/mL kanamycin (generations one, G1), and then passaged an additional seven times (generations two–eight; G2–G8) on R5MS agar plates supplemented with increasing concentrations of kanamycin (100–1200 µg/mL) (Fig. 6b). Three strains with more intense blue color were inoculated into R5MS liquid media for ACT quantification. For comparison, the empty vector pSET152 was introduced into M1146-actAB, and the resulting strain M1146-actAB-pSET152 was included to serve as a negative control. A significant increase in ACT production was observed with M1146-actAB-StrepT-switch-zouA, suggesting that ZouA-dependent DNA amplification multiplies ACT yield.
The bio-switch is functional in E. coli
To broaden the utility of the thermal bio-switch, the intact rheA was positioned under the control of T7 promoter, and StrepT-switch was used to drive the expression of neo reporter gene in E. coli. Ideally, at lower temperature, IPTG-induced RheA will repress the expression of the neo cassette, and the cells are unable to grow in the presence of kanamycin. When cultivated at higher temperature, IPTG-induced RehA will relieve its repression on the neo cassette, and the cells grow well in the presence of kanamycin. Unfortunately, E. coli cells grow normally when cultivated either at 28 °C or 37 °C (Supplementary Fig. 6). We speculate that this observation may arise from codon bias correlated with the expression of rheA in E. coli. To address this issue, a codon-optimized version of rheA was synthesized and the corresponding bio-switch was designated as EcoT-switch. When cultivated at 28 °C, the cells are unable to grow in the presence of kanamycin. However, when cultivated at 37 °C, they grow normally in the presence of kanamycin (Supplementary Fig. 6). The results suggested that the codon-optimized version of rheA retain its regulatory function and maintain its ability of thermal stimuli response. Thus, the thermal bio-switch can be adapted for further use in E. coli.