Identification of genes related to xylose utilization in P. thermoglucosidasius DSM 2542

To verify the putative genes related to xylose metabolism, homologs of reported proteins involved in xylose uptake and utilization were searched in the genome of P. thermoglucosidasius DSM 2542 (CP012712.1). Besides previous reported genes xylAB (AOT13_11570 - AOT13_11575, encoding xylose isomerase and xylulokinase, respectively) and rbsACB (AOT13_01805 - AOT13_01795, encoding D-ribose ABC transporter) (Liang et al. 2022a; Wang et al. 2022a), homologic genes encoding an ROK family transcriptional regulator (AOT13_16590) and a set of sugar ABC transporter (AOT13_11435, AOT13_11440, and AOT13_11445) were identified in the genome of P. thermoglucosidasius strains DSM 2542 (Fig. 1a and Table S4).

Fig. 1

Identification of xylose uptake and utilization genes in P. thermoglucosidasius DSM 2542. a Schematic representation of predicted xylose uptake and utilization genes in P. thermoglucosidasius DSM 2542. b Growth profile (blue open circle) and xylose consumption (blue closed circle) of DSM 2542 wild-type strain, and growth profile (red open triangle) and xylose consumption (red closed triangle) of DSM 2542 strain ∆ABCT, using xylose as the sole carbon source

The gene AOT13_16590 encoded an ROK family transcriptional regulator, which was not co-localized with xylAB as observed in Geobacillus kaustophilus HTA426 (Gu et al. 2010). Its protein sequence showed a sequence identity of 68.1% to XylR of G. kaustophilus HTA426, and 37% to 70% to previously identified XylR orthologs from Bacillus subtilis 168, Bacillus cereus ATCC 10987, Bacillus licheniformis ATCC 14580, Bacillus halodurans C-125, Bacillus clausii KSM-K16, and Geobacillus thermodenitrificans NG80-2 (Fig. S1 and S2) (Gu et al. 2010). Phylogenetic analysis further confirmed that AOT13_16590 clustered within the clade of XylRs from Geobacillus (Fig. S2). Furthermore, our previous RNA-seq analysis revealed that the transcription level of AOT13_16590 showed 9.7-fold increase when cultured with 1% xylose compared to those without xylose (Table S4) (Wang et al. 2022a). These findings suggested that AOT13_16590, designated as XylR, was involved in the regulation of xylose metabolism.

A set of sugar ABC transporter encoding genes, including AOT13_11435, AOT13_11440, and AOT13_11445 (encoding ABC transporter substrate-binding protein, ATP-binding protein, and permease, respectively), were found to be directly upstream of the ara operon (AOT13_11450 to AOT13_11485) (Liang et al. 2022a), which showed moderate sequence identities to XylF, XylG, and XylH of G. kaustophilus HTA426 (35.8%, 54.6%, and 43.1%, respectively, Fig. S3 and Table S4) (Gu et al. 2010). Furthermore, these genes showed 118- to 149-fold increase in their transcription levels when cultured with 1% xylose compared to those without xylose, which were about an order of magnitude higher than those of predicted ABC transporters for ribose and arabinose (Table S4) (Liang et al. 2022a; Wang et al. 2022a).

To verify the role of these sugar ABC transporter genes, we generated the mutant strain DSM 2542 ∆ABCT (Fig. S4). We devised an efficient reporter-guided gene knockout method for P. thermoglucosidasius, which used sfGFP and mCherry as selecting markers to enable fast screening of desired crossover strains and then eliminated the need for additional screening methods, such as colony PCR. This gene knockout method used fluorescent proteins sfGFP (at temperatures of 60−70 °C) and mCherry (at temperatures lower than 50 °C) as selecting markers in P. thermoglucosidasius (Fig. S4A) with optimized expression of mCherry using various combinations of promoters and RBSs (Fig. S4B and S4C). In brief, this method included three steps, (1) selecting transformant with the gene knockout plasmid using green fluorescence of sfGFP; (2) selecting first double crossover strains using red fluorescence of mCherry and the loss of green fluorescence of sfGFP, in which the target DNA segment was replaced with the segment containing mCherry; (3) selecting second crossover strains using the loss of red fluorescence of mCherry, in which the inserting segment containing mCherry was removed to generate a markerless gene knockout strain (Fig. S4D and S4E). With this method, we have constructed the xylR gene knockout strain DSM 2542 ∆xylR and the sugar ABC transporter (AOT13_11435, AOT13_11440, and AOT13_11445) gene knockout strain DSM 2542 ∆ABCT in a highly convenient and efficient manner (Fig. S4F and S4G).

Compared to the wild-type strain DSM 2542, strain DSM 2542 ∆ABCT exhibited a significant impairment when growth with xylose as the solo carbon source (Fig. 1b), while its growth on glucose remained unaffected (Fig. S5). These findings strongly supported that the sugar ABC transporter encoded by AOT13_11435, AOT13_11440, and AOT13_11445 played a crucial role in xylose uptake in P. thermoglucosidasius. However, our results didn’t completely excluded the involvement of other pentose uptake genes in the uptake of xylose, as revealed by previous studies (Liang et al. 2022a, b).

Identification and characterization the core sequence of the promoter of xyl operon

In our previous study, we have identified a set of xylose-inducible promoters in P. thermoglucosidasius DSM 2542 (Wang et al. 2022a). Notably, the promoter of the xylose operon (xyl operon), which includes xylA (AOT13_11570) and xylB (AOT13_11575) genes (Fig. 1a), demonstrated its robust activity when induced by xylose (Wang et al. 2022a). Thus, we have chosen this promoter as the basis for developing a xylose-inducible expression system in P. thermoglucosidasius.

To identify the minimal region exhibiting promoter activity, we systematically truncated the upstream DNA sequence of the xylA gene. PCR amplification resulted in DNA fragments xylA500p, xylA400p, xylA300p, xylA200p, and xylA100p, corresponding to the 500-bp, 400-bp, 300-bp, 200-bp, and 100-bp upstream sequences preceding the xylA gene start codon, respectively (see Fig. S6). These fragments were then inserted into the pUCG18-sfgfp vector and transformed into DSM 2542, generating strains DSM 2542-xylA500p, DSM 2542-xylA400p, DSM 2542-xylA300p, DSM 2542-xylA200p, and DSM 2542-xylA100p (Fig. 2a). Their promoter activities were evaluated by the fluorescence intensities of superfolder green fluorescent protein (sfGFP). As shown in Fig. 2b, while xylA500p exhibited robust promoter activity upon addition of 1% xylose, the truncated fragments (xylA400p, xylA300p, xylA200p, and xylA100p) displayed significantly weaker promoter activity. This suggests that the core promoter sequence is located within the region spanning −500 to −400 bp upstream of xylA, hereafter referred to as xylA1p.

Fig. 2

Identification and characterization of a xylose-inducible promoter from P. thermoglucosidasius. a Schematic representation of the sfGFP reporter system. b Fluorescence intensity levels of DSM 2542 strains carrying plasmids with different promoters at 24 h. c Nucleotide sequences of the core region of the promoter of xyl operon. The transcriptional start site (TSS) is indicated by a red arrow. The putative −10 and −35 regions of the promoter are marked with boxes. Putative XylR binding sequence are labeled in red, together with XylR binding sites from Geobacillus kaustophilus HTA426 and Geobacillus thermodenitrificans NG80-2. d Construction of plasmids containing sfgfp and truncated promoters derived from xylA1p. e Fluorescence intensity levels of DSM 2542 strains harboring plasmids with different promoters at 24 h. The standard deviations were calculated from at least three independent experiments and were represented as error bars. The insulator sequence riboJ was reported to eliminate the interference between promoters and RBSs and improve the modularity of regulatory elements (Lou et al. 2012)

Next, we aimed to identify the core elements of the xylA1p promoter. To pinpoint the transcriptional start site (TSS) of xylA1p, we utilized rapid amplification of 5'-cDNA ends (5'-RACE) analysis (Fig. 2c). With the TSS determined, we identified the putative −10 and −35 regions of xylA1p (Fig. 2c). Furthermore, we identified a conserved motif (CTTTGTTTGTACACTAGACAAACAAAT) within xylA1p, displaying high consistency with the reported XylR recognition sequence of Parageobacillus and Bacillus (Fig. 2c) (Liang et al. 2022a).

To identify the minimal promoter region capable of driving reporter gene expression in response to xylose, we systematically truncated the 100-bp promoter xylA1p while retaining its promoter core region. This resulted in xylA2p, xylA3p, xylA4p, and xylA5p variants, while xylA6p was used as a negative control in which the −10 region and the TSS were deleted (Fig. 2d). As assessed by the fluorescence intensity of sfGFP, all variants except xylA6p exhibited promoter activities similar to that of xylA1p (Fig. 2e). These findings established xylA5p as the minimal promoter sequence (Fig. 2e), which we subsequently employed in the subsequent investigations.

Characterization of the regulatory mechanism of XylR

Then we try to evaluate the interaction between XylR and xylA5p. To eliminate any naturally occurring regulation mediated via XylR, its encoding gene xylR was knockout in P. thermoglucosidasius DSM 2542, resulting strain DSM 2542 ∆xylR (Fig. S4F). Under the control of the promoter xylA5p, the sfgfp gene was expressed in DSM 2542 ∆xylR, resulting in strain DSM 2542 ∆xylR-xylA5p-sfgfp displaying fluorescence (Fig. 3a and b). Upon the addition of the xylR gene controlled by promoter p43 (a sequence from B. subtilis with promoter activity) (Wang et al. 2019; Wang and Doi 1984) amplified from plasmid puBS4.4-p43 (see Table S2), the resulting strain DSM 2542 ∆xylR-p43-xylR-xylA5p-sfgfp exhibited a significant reduction in fluorescence in the absence of xylose (Fig. 3a and b). Notably, upon the addition of xylose, the fluorescence was restored in a dose-dependent manner (Fig. 3b). These findings demonstrate that the xylA5p-xylR system is capable of sensing and responding to xylose, thereby derepressing the transcription of xylA5p.

Fig. 3

The regulatory mechanism of XylR. a Schematic representation of the sfGFP reporter system. b Fluorescence levels of strain ∆xylR-p43-xylR-xylA5p-sfgfp with varying concentrations of xylose at 24 h. Strain DSM 2542 ∆xylR-xylA5p-sfgfp was used as a positive control. c EMSA depicting the interaction between XylR and a probe containing xylA5p (PxylA5). d Dissociation effect of xylose to the complex of XylR and probe PxylA5 at different concentration of xylose. Each lane contains 6 ng of PxylA5, 1 μM XylR and different amounts of xylose. The standard deviations were calculated from at least three independent experiments and were represented as error bars

The direct interaction between XylR and xylA5p was evaluated by electrophoretic mobility shift assay (EMSA) in vitro. XylR was expressed in E. coli and purified to homogeneous via Ni-NTA chromatography (Fig. S7). EMSA was performed using purified XylR and a probe containing promoter xylA5p (PxylA5), revealing that XylR could bind to PxylA5 in a concentration-dependent manner (Fig. 3c). No retardation was observed for the probe containing promoter p43 (Fig. S8), suggesting the binding between XylR and xylA5p was specific. The responses of XylR/PxylA5 complexes to xylose were also evaluated by EMSA. The XylR/PxylA5 complexes dissociated gradually with the increased xylose concentration (Fig. 3d). These results, together with the in vivo results of sfGFP reporter system, confirmed that XylR could directly bind to promoter xylA5p and repress its promoter activity. This repression effect could be relieved by xylose, thus inducing the transcription of genes downstream to promoter xylA5p.

Development of a xylose-inducible expression system IExyl*

The xylose-concentration-dependent GFP fluorescence observed in DSM 2542 ∆xylR-p43-xylR-xylA5p-sfgfp indicated the potential of XylR and xylA5p as an inducible expression system in P. thermoglucosidasius. Full induction was achieved at a xylose concentration of 10 mM (Fig. 3b). Interestingly, the fully induced fluorescence surpassed that driven by the commonly used promoters pLdh (the promoter of lactate dehydrogenase encoding gene) (Cripps et al. 2009) and pRplsWT (a constitutive promoter selected from a promoter library) (Reeve et al. 2016), while the fluorescence of pLdh is significantly lower than that of pRplsWT, indicating the remarkably strong activity of xylA5p (Fig. 4a).

Fig. 4

Development and characterization of the inducible expression system ∆xylR-xylRp-xylR-xylA5p-sfgfp. a GFP fluorescence of strains with and without xylose induction. b Plasmid map depicting the components of the inducible expression system. c GFP fluorescence of DSM 2542 strains harboring inducible expression system with different promoters for xylR when cultured without xylose at 24 h. d GFP fluorescence of strain ∆xylR-xylRp-xylR-xylA5p-sfgfp in response to different doses of xylose at 24 h. The concentrations of xylose were 0, 0.001, 0.005, 0.01, 0.05, 0.1, 0.5, 1, 5, 10, 30, and 50 mM. e Quantitative analysis of the relative transcript levels of sfgfp determined by RT-qPCR in strain ∆xylR-xylRp-xylR-xylA5p-sfgfp in response to different doses of xylose at 24 h. The transcript strength of sfgfp in strain ∆xylR-p43-xylR-xylA5p-sfgfp was arbitrarily assigned a value of 1. f GFP fluorescence in strain ∆xylR-xylRp-xylR-xylA5p-sfgfp cultured in medium with glucose as the sole carbon source at 24 h. The standard deviations were calculated from at least three independent experiments and were represented as error bars. The control referred to a DSM 2542 ∆xylR-sfgfp strain containing a plasmid (pUCG18-sfgfp) with a promoterless sfgfp

However, a higher level of leaky expression was observed in the absence of xylose induction compared to the control (Figs. 3b and 4a). To address this issue, we explored the inducible expression system by replacing promoter p43 with promoters xylRp, pLdh, and pRplsWT, respectively, with the aim of achieving more stringent control over the promoter activity of xylA5p by increased expression level of XylR (Fig. 4b). Through a comparison of GFP fluorescence in strains DSM 2542 ∆xylR-xylRp-xylR-xylA5p-sfgfp, DSM 2542 ∆xylR-pLdh-xylR-xylA5p-sfgfp, and DSM 2542 ∆xylR-pRplsWT-xylR-xylA5p-sfgfp in the absence of xylose, we observed a significant reduction in leaky expression, reaching a level similar to that of the control (Fig. 4b and c). Subsequently, we selected DSM 2542 ∆xylR-xylA5p-xylRp-xylR for further investigations.

As shown in Fig. 4d, the fluorescence intensity of the strain ∆xylR-xylRp-xylR-xylA5p-sfgfp exhibited an increase in response to escalating xylose concentrations. To assess the transcriptional activity of xylA5p in strain ∆xylR-xylRp-xylR-xylA5p-sfgfp under different xylose doses, RT-qPCR was performed to measure the transcript levels of the sfGFP gene. The results demonstrated a xylose concentration-dependent enhancement in sfGFP transcript abundance (Fig. 4e), indicating the sensitive and controlled induction of promoter xylA5p in strain ∆xylR-xylRp-xylR-xylA5p-sfgfp at the transcriptional level in P. thermoglucosidasius. Notably, we observed a significant inhibitory effect of glucose, a commonly utilized fermentation substrate, on the transcriptional activity of promoter xylA5p in strain ∆xylR-xylRp-xylR-xylA5p-sfgfp (Fig. 4f). This observation posed a significant challenge to the practical application of the system when employing lignocellulosic biomass or other glucose-containing substrates as carbon sources. The detrimental effect of glucose, known as carbon catabolite repression (CCR), has to be addressed to overcome this limitation.

To develop a xylose-inducible expression system that is not repressed by glucose, we constructed an expression system named IExyl3, in which the transcription of the xylose transporter genes was controlled by a constitutive promoter pLdh (Fig. 5a). The performance of IExyl3 was evaluated in P. thermoglucosidasius. Strain DSM 2542 IExyl3 was cultured in medium with or without 4% glucose, supplemented with xylose concentrations ranging from 1 μM to 10 mM. Unfortunately, only weak fluorescence was observed in the presence of glucose after 10 hours of culture regardless of the xylose concentration added, indicating the promoter activity was repressed by remaining glucose in the fermentation medium (Fig. 5b and S9). In contrast, varying fluorescence intensities were observed in glucose-free medium, indicating that the transcription of these xylose transporter genes was still inhibited by glucose (Fig. 5b).

Fig. 5

Development of the inducible expression system IExyl*. a Plasmid map of the inducible expression system IExyl3. b Evaluation of IExyl3 in medium with or without 4% glucose as the sole carbon source at 10 h. c Genetic organization of the identified xylose transporter encoding genes (AOT13_11435, AOT13_11440, and AOT13_11445). The putative cre motif at +991 is highlighted in red and located within the AOT13_11435 gene. The lower section presents a comparison of the cre sequence in AOT13_11435 (cre +991) and the conserved cre sequence in B. subtilis (cre motif), in which highly conserved nucleotides are underlined. d GFP fluorescence and corresponding cre site sequences for DSM 2542 strains carrying mutated cre site sequences at 24 h. Mutated nucleotides are highlighted in red. e Evaluation of GFP fluorescence in DSM 2542 creM3 in response to increasing concentrations of xylose, in the presence or absence of glucose, at 24 h. f GFP fluorescence (green) and glucose consumption (blue) of DSM 2542 creM3 in the presence of glucose and 1 mM xylose. g Xylose consumption comparison between DSM 2542 IExyl3 (blue) and DSM 2542 creM3 (red). The standard deviations were calculated from at least three independent experiments and were represented as error bars

Previous studies suggested that the CCR-mediated regulation of xylose metabolism might primarily involve the expression control of pentose transporter operons (Li et al. 2018; Liang et al. 2022a). As depicted in Fig. 5c, an anticipated catabolite responsive element (cre) site, TGGAACCGGTTTCA, was identified at position +991 relative to the ATG start codon of AOT13_11435, which aligns with the reported consensus cre sequence (TGNAANCGNNNNCN) observed in B. subtilis and P. thermoglucosidasius NCIMB 11955 (Liang et al. 2022a; Weickert and Chambliss 1990). Consequently, it can be inferred that this element is subject to CcpA-mediated CCR for the regulation of these xylose transporter genes. Next, we mutated this cre site to afford three distinct mutated versions, namely creM1, creM2, and creM3 (Fig. 5d). As illustrated in Fig. 5d, a substantial 5- to 10-fold enhancement in fluorescence intensity was observed for these mutated strains when glucose was employed as the carbon source. These findings underscore the crucial role of conserved nucleotides within the cre consensus sequence in CCR. Among the mutated sequences, creM3 exhibited the highest increase in GFP fluorescence in response to varying concentrations of xylose, in both the presence and absence of glucose (Fig. 5e, 5f, and S10). Concurrently, the overexpression of xylose transporter genes resulted in an accelerated rate of xylose consumption (Fig. 5g), indicating that the cre mutant enables efficient xylose transport even in the presence of glucose. Accordingly, the corresponding expression system was designated as IExyl*.

Application of IExyl* in metabolic engineering for increased riboflavin yield in P. thermoglucosidasius

Previous studies on riboflavin-producing strains of B. subtilis have identified target genes for enhanced riboflavin production. Overexpression of glucose-6-phosphate dehydrogenase (encoded by zwf gene, Fig. 6a) increased the metabolic flow of the pentose phosphate pathway, resulting in 25% increase in riboflavin production (Duan et al. 2010). Another crucial enzyme, GTP cyclohydrolase II/3,4-dihydroxy-2-butanone-4-phosphate synthase (encoded by ribA, Fig. 6a), acted as the rate-limiting step in riboflavin biosynthesis in B. subtilis (You et al. 2022). Overexpression of ribA led to 25% increase in riboflavin production when glucose was used as the carbon source (You et al. 2022). Recent advances in synthetic biology have demonstrated that fine-tuning gene expression to optimize intracellular metabolic flow is a promising strategy for maximizing target product yield (Gupta et al. 2017).

Fig. 6

Application of IExyl* for the improvement of riboflavin production in P. thermoglucosidasius. a Schematic representation of the riboflavin biosynthetic pathway in P. thermoglucosidasius. b Heat map illustrating the riboflavin yields of DSM 2542 Rib-Gtg-IExyl*-ribA under different xylose concentrations and induction time points. c Heat map illustrating the riboflavin yields of DSM 2542 Rib-Gtg-IExyl*-zwf under different xylose concentrations and induction time points. d Growth profiles and riboflavin yields of DSM 2542 Rib-Gtg, DSM 2542 Rib-Gtg-IExyl*-zwf, and DSM 2542 Rib-Gtg-IExyl*-ribA in shake-flask fermentation at 24 h. e Relative transcription levels of zwf and ribA genes determined by RT-qPCR in DSM 2542 Rib-Gtg, DSM 2542 Rib-Gtg-IExyl*-zwf, and DSM 2542 Rib-Gtg-IExyl*-ribA at 24 h, respectively. The standard deviations were calculated from at least three independent experiments and were represented as error bars

In a previous study, the riboflavin biosynthetic gene cluster of P. thermoglucosidasius DSM 2542 was overexpressed using pLdh promoter in this strain, resulting a riboflavin-producing strain P. thermoglucosidasius DSM 2542 Rib-Gtg (Yang et al. 2021). Therefore, we utilized the inducible expression system IExyl* to finely regulate the expression of zwf and ribA genes in DSM 2542 Rib-Gtg, resulting the engineered strains DSM 2542 Rib-Gtg-IExyl*-zwf and DSM 2542 Rib-Gtg-IExyl*-ribA. The expression levels of zwf and ribA were optimized by adding xylose at different doses and time points. Riboflavin production was subsequently evaluated under various induction conditions. Our results revealed that the optimal conditions for riboflavin production in both DSM 2542 Rib-Gtg-IExyl*-zwf and DSM 2542 Rib-Gtg-IExyl*-ribA were achieved by adding 1 mM xylose at 6 h (Fig. 6b and c).

Growth and riboflavin production were assessed through shake-flask fermentation of the fine-tuning strains DSM 2542 Rib-Gtg-IExyl*-zwf and DSM 2542 Rib-Gtg-IExyl*-ribA, and compared to the starting strain DSM 2542 Rib-Gtg. As shown in Fig. 6d, the biomass of the fine-tuning strains at 24 h was slightly lower than that of the starting strain DSM 2542 Rib-Gtg, while the riboflavin yields of the fine-tuning strains exhibited a 2.3- to 2.8-fold increase compared to the starting strain (Fig. 6d). The specific rates of glucose consumption for the engineered strains were similar to that of the parent strain DSM 2542 Rib-Gtg (Table S5). These findings showed that the inducible expression system IExyl* can serve as an effective tool for rational engineering in Parageobacillus, offering a promising approach for enhancing product yields.

The impact of fine-tuning the zwf and ribA genes using IExyl* was further investigated in the fine-tuning strains. As depicted in Fig. 6e, the transcription levels of both genes were significantly enhanced in the fine-tuning strains, as assessed by RT-qPCR. Additionally, the intracellular concentrations of intermediate metabolites of the pentose phosphate pathway were also measured (Table 1). The pentose phosphate pathway plays a crucial role in providing ribulose-5-phosphate, a direct precursor for riboflavin biosynthesis, as well as ribose-5-phosphate, which is converted to phosphoribosylpyrophosphate (PRPP) and subsequently to GTP, another biosynthetic precursor of riboflavin. In the fine-tuning strains, slight increase in the intracellular concentration of ribulose-5-phosphate was observed, while the concentration of ribose-5-phosphate exhibited more obvious increase (Table 1). Moreover, compared to the parent strain, the intracellular PRPP concentration in the fine-tuning strains was significantly elevated, reaching a maximum concentration of 3.99 nmol/mg (dry weight) (Table 1). These findings indicate that the inducible expression system successfully reconfigures the intracellular metabolite concentrations, thereby promoting riboflavin production. In conclusion, we have designed a plug-and-play plasmid utilizing the newly developed inducible expression system IExyl*, which has been employed for fine-tuning gene expression to coordinate intracellular metabolism and maximize the yield of target products.

Table 1 Intracellular Concentration of Metabolites in P. thermoglucosidasius strains