GnuR Represses the Expression of Glucose and Gluconate Catabolism in Pseudomonas putida KT2440 Authors: Wenbo Chen, Rao Ma, Yong Feng, Yunzhu Xiao, Agnieszka Sekowska, Antoine Danchin, Conghui You Published: 2024 ABSTRACT ABSTRACTIn Pseudomonas putida KT2440, a prime chassis for biotechnology, the clustered distribution of glucose catabolism genes and four related transcription factors (TFs) may facilitate the tight regulation of glucose catabolism. However, the genes under the direct control of these TFs remain unidentified, leaving their regulatory roles elusive. Furthermore, the carbon source gluconate was metabolised similarly to glucose in KT2440, but the responses of these catabolic and TF genes to gluconate were unclear. Here, these mysteries were unravelled through multi‐omics analysis integrated with physiological studies. First, we found that the expression of these catabolic and TF genes were significantly induced by both glucose and gluconate in KT2440. The independent responses of these genes to glucose and gluconate were differentiated in the gcd deletion mutant. We then defined the regulon of GnuR, one of the four related TFs, and discovered that GnuR directly repressed the expression of catabolic genes involved in the Entner–Doudoroff and the peripheral glucose and gluconate metabolism pathways. These results were further confirmed by physiological studies. Finally, a regulatory mode of an incoherent feedforward loop involving GnuR is proposed. Introduction Pseudomonas putida is a Gram‐negative bacterium of the genus Pseudomonas commonly found in polluted soils (Nikel and de Lorenzo 2018). Strain KT2440, one of the best known strains of Pseudomonas, derived from P. putida mt‐2 through the removal of its TOL plasmid (Nakazawa 2002) is a model laboratory species that retains the ability to survive in natural environments (Nikel and de Lorenzo 2018). It is a well‐established host for cloning and gene expression due to its diverse metabolic capabilities, high tolerance to oxidative stress and biosafety and is therefore an ideal chassis in synthetic biology (Nikel et al. 2016; Nikel and de Lorenzo 2018; Turlin et al. 2023; Martínez‐García and de Lorenzo 2024). Furthermore, the characteristic central carbon metabolic pathway of KT2440 when growing on glucose (Kozaeva et al. 2021; Amendola et al. 2024; Pause et al. 2024) provides a theoretical foundation for its use as a bio‐catalytic chassis. The study of its glucose and derivatives metabolism offers an interesting starting point for the design of synthetic metabolism (Nikel and de Lorenzo 2018). Glucose is metabolised in KT2440 through three main catabolic pathways (Regenhardt et al. 2002): the Entner–Doudoroff (ED) pathway, the Embden–Meyerhof–Parnas (EMP) pathway and the pentose–phosphate (PP) pathway (Figure 1) (Nikel, Martínez‐García, and de Lorenzo 2014; Law, Nurwono, and Park 2024). Using the metabolite glyceraldehyde‐3‐phosphate as a node, EMP can be divided into an upper EMP and a lower EMP. In contrast to most organisms, which use the whole EMP as the main pathway for glucose metabolism (Romano and Conway 1996; Bar‐Even et al. 2012), the EMP pathway is incomplete in P. putida since the upper EMP pathway is hardly involved in glucose catabolism in KT2440 (Sudarsan et al. 2014; Kukurugya et al. 2019). Thus, the EMP pathway in P. putida can only run in the direction of gluconeogenesis. In this model organism, 6‐phosphogluconate (6P‐Gluc) is the key metabolic intermediate linking the three pathways of ED, EMP and PP (Fuhrer, Fischer, and Sauer 2005). Catabolic pathways of glucose metabolism in KT2440. The reactions involved in peripheral pathways, upper EMP pathway, PP pathway, ED pathway, lower EMP pathway and TCA cycle are individually grouped in orange, pink, blue, olive green, purple and grey boxes, respectively. The arrows indicate unidirectionality or bidirectionality of the metabolic reactions. The gene annotated to each metabolic protein is placed next to the corresponding reaction arrow. The metabolite abbreviations are as follows: Glucose, Glc; gluconate, Gluc; 2‐keto‐gluconate, 2K‐Gluc; glucose‐6‐phosphate, G6P; 6‐phosphogluconate, 6P‐Gluc; 2‐keto‐6‐phosphogluconate, 2K‐Gluc‐6P; 2‐keto‐3‐deoxy‐6‐phosphogluconate, KDPG; fructose‐6‐phosphate, F6P; fructose‐1,6‐bisphosphate, FBP; glyceraldehyde‐3‐phosphate, GAP; dihydroxyacetone‐phosphate, DHAP; 1,3‐biphosphoglycerate, 1,3‐bisPG; 3‐phosphoglycerate, 3‐PG; 2‐phosphoglycerate, 2‐PG; phosphoenolpyruvate, PEP; ribulose‐5‐phosphate, Ru5P; ribose‐5‐phosphate, R5P; xylulose‐5‐phosphate, Xu5P; sedoheptulose‐7‐phosphate, S7P; erythrose‐4‐phosphate, E4P. Glucose is converted to 6P‐Gluc in KT2440 via three pathways. (I) Glucose enters the periplasmic space via the outer membrane porin OprB (Saravolac et al. 1991), with approximately 10% of glucose being transported into the cytoplasm via the ABC transporter system formed by GtsABC (Cuskey et al. 1985; del Castillo et al. 2007) and subsequently phosphorylated by glucokinase (Glk) to glucose‐6‐phosphate (G6P). G6P is little involved in the upper EMP pathway and is instead converted to 6P‐Gluc by the synergistic action of G6P dehydrogenase (Zwf) and 6‐phosphogluconate lactonase (Pgl) (Daddaoua et al. 2010). (II) Approximately 90% of the glucose entering the periplasmic space is oxidised to gluconate by glucose dehydrogenase (Gcd) in the periplasm, transported to the cytoplasm by the gluconate transporter (GntT) and subsequently phosphorylated to 6P‐Gluc by glucokinase (GnuK) (Nikel et al. 2021). In addition, gluconate assimilated by the cell is mainly metabolised by this pathway via GntT and GnuK. (III) Approximately 11% of gluconate in the periplasm is oxidised to 2‐ketogluconate (2K‐Gluc) by gluconate 2‐dehydrogenase (Gad) (Sudarsan et al. 2014; Wirth et al. 2022), then transported to the cytoplasm by the 2‐ketogluconate transporter (KguT) and converted by 2K‐Gluc kinase (KguK) to 2‐ketoyl‐6‐phosphogluconate (2K‐Gluc‐6P), which is then reduced to 6P‐Gluc by 2K‐Gluc‐6P reductase (PtxD) (Brazma et al. 2001). Therefore, gluconate and 2K‐Gluc accumulate to high levels, when wild‐type P. putida is grown in glucose (Volke et al. 2023). In contrast, gluconate and 2K‐Gluc are not formed when a glucose dehydrogenase deletion strain is grown in glucose. The pathway by which glucose passes through a series of enzymes in the periplasmic space, then converted to gluconate and 2‐ketogluconate and finally metabolised to 6P‐Gluc is also known as the peripheral pathway (del Castillo et al. 2007; del Castillo, Duque, and Ramos 2008). The 6P‐Gluc produced by the above pathways is further metabolised into small molecules with three to five carbon backbones mainly via two metabolic pathways. Approximately 90% of 6P‐Gluc enters the ED pathway, where 6P‐Gluc is converted by 6P‐Gluc dehydratase (Edd) to 2‐keto‐3‐deoxy‐6‐phosphogluconate (KDPG) and then by KDPG aldolase (Eda) to pyruvate and glyceraldehyde‐3‐phosphate (Petruschka et al. 2002; Kukurugya et al. 2019). The remaining ~10% of 6P‐Gluc is converted to 3‐phosphoglycerate and fructose‐6‐phosphate by enzymes in the PP pathway. The 3‐phosphoglycerate metabolite is subsequently metabolised by the lower EMP pathway. In parallel, both 3‐phosphoglycerate and fructose‐6‐phosphate can enter the recently defined EDEMP cycle (a cycle that recruits elements from the ED, EMP and PP pathways) to regenerate G6P (Nikel et al. 2015). The genes encoding catabolic enzymes of glucose metabolism and related transcription factors (TFs) are clustered together to four clusters on the chromosome of KT2440. In parallel, four genes encoding TFs, hexR, gnuR, ptxS and gltR‐II are distributed in these four clusters. The gtrS gene, located immediately downstream of gltR‐II, encodes a sensor kinase of GtrS, which forms a two‐component system with GltR‐II in Pseudomonas sp. (del Castillo et al. 2007). Each of the four TFs has been reported to be involved in the expression regulation of some of the genes in the three glucose catabolic pathways, and the clustered distribution of these catabolic and TF genes could facilitate the co‐expression and tight regulation of catabolic enzymes and transport systems of glucose metabolism (del Castillo, Duque, and Ramos 2008). However, the specific regulons of these TFs remain unidentified, leaving the physiological role of these TFs in glucose metabolism elusive, possibly because previous studies relied primarily on transcriptomic analysis, making it difficult to distinguish whether the observed regulatory effects are direct or indirect. In recent years, the method of integrating high‐throughput ChIP‐seq (Robertson et al. 2007; Peano et al. 2015; Blanco‐Romero et al. 2018) with RNA‐seq (D'Arrigo et al. 2019) has revealed the regulons of many TFs and correspondingly identified their physiological roles in different organisms (Kroner, Wolfe, and Freddolino 2019; Gao et al. 2023; Hou et al. 2023). This prompted us to explore the regulons of these TFs using a similar method in order to identify the specific roles of these TFs in the regulation of glucose metabolism in KT2440. Transcriptomic analysis revealed that the expression of catabolic genes and several TF genes involved in glucose metabolism (del Castillo et al. 2007) were induced by glucose in KT2440. Could gluconate also induce the expression of these genes, given that the glucose and gluconate metabolic pathways overlap (Figure 1)? Furthermore, would glucose or gluconate show specific differences in their induction of these genes? To our knowledge, these interesting questions have not been addressed. Here, we first identified that both glucose and gluconate could induce the expression of catabolic enzymes of glucose metabolism and their four associated TFs in KT2440 by RNA‐seq analysis. The independent responses of these catabolic and TF genes to glucose and gluconate were further differentiated in the gcd deletion mutant, which did not form gluconate when this mutant was grown on glucose. We then focused on GnuR, one of the four TFs induced by both glucose and gluconate in KT2440, and discovered the GnuR regulon, by combining ChIP‐seq and RNA‐seq analysis. We found that the primary role of GnuR was to directly repress the expression of catabolic genes functioning in the ED and the peripheral glucose and gluconate metabolism pathways. This finding was further supported by physiological studies of cell growth. We propose that an incoherent feedforward control with GnuR as a node is one of the features of glucose metabolism regulation in P. putida . Understanding the principles of efficient carbon metabolism is essential in model bacteria for use in engineering applications. Our study sheds new light on the regulation of glucose and gluconate catabolism in P. putida . Results The Expression of Catabolic and TFs Genes Involved in Glucose Metabolism and Regulation Was Induced by Both Glucose and Gluconate in Wild‐Type KT2440 It has been reported that the expression of glucose metabolism genes is induced by glucose in KT2440 (del Castillo et al. 2007). However, to our knowledge, there have been few reports on their response to gluconate. Was the expression of these genes also induced by gluconate, given that the metabolic pathways of glucose and gluconate are highly overlapping (Figure 1)? To answer this question, we performed RNA‐seq to study their expression in KT2440 grown individually in minimal medium with glucose, gluconate or succinate as a sole carbon source. Compared to the succinate growth condition, genes involved in glucose transport and catabolism throughout the peripheral and ED pathways were highly induced by either glucose or gluconate (Figure 2). The expression of several of these genes increased almost 100‐fold. The highly induced expression identified by RNA‐seq was further confirmed by real‐time quantitative polymerase chain reaction (RT‐qPCR) (Figure S1). Glucose catabolism genes belonging to the whole peripheral pathway as well as the ED pathway showed very similar induction changes with both glucose and gluconate. However, we observed that genes gtsABCD, encoding a glucose ABC transporter, and oprB‐I, encoding a glucose porin, showed a much stronger response to glucose than to gluconate. Comparison of the expression of catabolic and TFs genes involved in glucose metabolism in KT2440 grown in glucose or gluconate versus that in succinate. (A) Fold change of gene expression in glucose or gluconate versus that in succinate. Gene expression was determined by RNA‐seq in triplicate, and each value was shown as mean ± SD. The annotated gene function or metabolic pathway of each gene involved is highlighted. (B) Genes in panel A were mapped to the catabolic pathways of glucose metabolism. When a gene shows a fold change greater than 2, the value is shown in parentheses, with the blue (red) numbers indicating the fold change comparing the gene expression level in glucose (gluconate) to that in succinate. We found that the majority of glucose catabolism genes in the PP and EMP pathways did not respond to either glucose or gluconate. However, there were five exceptional genes, namely, glk, zwfA, pgl, gapA and PP_3443, which also consistently responded to glucose and gluconate. Intriguingly, the function of these five genes was also closely linked to the ED pathway. The glk gene in the upper EMP pathway encodes glucokinase, which converts glucose to G6P, and G6P is converted to 6P‐Gluc by glucose 6‐phosphate 1‐dehydrogenase and 6‐phosphogluconolactonase, which are respectively encoded by zwfA and pgl in the PP pathway. The 6P‐Gluc metabolite is the entry substrate of the ED pathway. GAP is the end product of the ED pathway. It is converted to 3‐PG by glyceraldehyde‐3‐phosphate dehydrogenase, encoded by gapA or PP_3443, which encodes an isozyme of GapA in the lower EMP pathway. We also found that all the reported TFs involved in the regulation of glucose metabolism, including gltR‐II‐gtrS, hexR, ptxS and gnuR, were synchronously induced by both glucose and gluconate to similar levels in KT2440. Specific Differences of Glucose and Gluconate Induction on These Catabolic and TFs Genes Was Identified by Deleting the Glucose Dehydrogenase Gene Gcd The studies reported above showed that the majority of the expression of catabolic and TF genes involved in glucose metabolism and regulation was induced similarly by both glucose and gluconate in KT2440. However, since the samples used in these studies were withdrawn at the mid‐exponential phase, when cells grown on glucose had already accumulated higher amounts of gluconate with the help of glucose dehydrogenase (Gcd) enzyme (Figure 3A), it was not surprising that these catabolic and TF genes showed very similar responses. In order to differentiate whether their responses to these two carbon sources were independent, we measured the expression response of these genes to glucose and gluconate in the gcd deletion mutant strain CY1850 (Figure 3B), which did not form gluconate, when grown in glucose. Identification of the independent responses of catabolic and TF genes involved in glucose metabolism to glucose or gluconate by studying their expression in the gcd KO strain of CY1318 grown individually in glucose, gluconate and succinate. (A) This reaction shows that glucose is oxidised to gluconate by glucose dehydrogenase (Gcd). (B) Schematic diagram of the DNA region neighbouring gcd in the gcd wild type and the gcd knockout mutant. (C) Fold change of gene expression in glucose or gluconate versus that in succinate. Gene expression was determined by RT‐qPCR in two to three independent assays, and each value was shown as mean ± SD. The annotated gene function or metabolic pathway of each gene involved is highlighted. Genes were grouped into three groups based on their specific response to these carbon sources. To guide the eyes, the dashed black line points to the cutoff (fold change = 2). (D) Genes in panel C were mapped to the catabolic pathways of glucose metabolism. The blue arrow indicates that the gene is highly induced by glucose. The green arrow indicates that the gene is slightly induced by glucose. The red arrow indicates that the gene is highly induced by gluconate. The purple arrow indicates that the gene is slightly induced by gluconate. Strain CY1850 was grown individually in minimal medium with glucose, gluconate or succinate as a sole carbon source (Figure S2). Total RNA from CY1850 in each condition was extracted, and the expression of these catabolic and TF genes previously identified as induced by glucose and gluconate in the wild‐type strain of KT2440 (Figure 2A) was studied by RT‐qPCR. Intriguingly, these genes showed distinct responses to glucose and gluconate. They were divided into three groups according to their specific response patterns (Figure 3C). Consistent with previous findings in the wild‐type strain KT2440 (Figure 2A), in the gcd mutant strain CY1850, the expression of the catabolic genes eda, edd, zwfA, pgl, glk, gapA and PP_3443, which function in the ED, PP or EMP pathway (Figure 3D), and the TF genes of gltR_II‐gtrS and hexR were similarly induced by both glucose and gluconate and were grouped in Group 1. In contrast, the expressions of the catabolic genes PP_3382, PP_3383, PP_3384, kguT, kguK and ptxD and two TF genes ptxS and gnuR were only induced by gluconate and were grouped in Group 2. In addition, although the expression of gntT and gnuK was slightly induced by glucose, to a 2–3 fold change, their expression was strongly induced by gluconate to a 40–180 fold change. Hence, gntT and gnuK were also grouped in Group 2. The catabolic genes in Group 2 all function in the peripheral pathway (Figure 3D). Therefore, the highly increased expression of these genes in Group 2 observed in KT2440 when it was grown in glucose was actually caused by high levels of gluconate accumulation. Finally, similar to the observations in KT2440, we found that genes encoding the glucose transporter and porin (gtsABCD, oprB‐I) were much more responsive to glucose than to gluconate in CY1850 and these five genes were grouped in Group 3 (Figure 3D). Given the distinct responses of these catabolic genes in these three groups to glucose and gluconate, different TFs could be involved in and mediate their response to these carbon substrates, and therefore the regulatory roles of these TFs need to be further investigated. In the following studies, we focused on the TF of GnuR whose expression was indeed induced by gluconate (Figure 3C) and explored its contribution to gluconate or glucose metabolism in P. putida . GnuR Binding Targets Were Enriched in the Glucose and Gluconate Metabolism Pathways To identify genes that might be controlled by GnuR, we used the technique of ChIP‐seq to identify the DNA binding targets of GnuR in vivo. For this assay, we constructed a derivative strain of KT2440 encoding a C‐terminal 6×His‐tagged GnuR, namely, CY1563. The CY1563 strain showed no growth difference as compared to the KT2440 wild‐type strain (Figure S3), indicating that the 6×His‐tagged GnuR was as functional as the native GnuR. The highly induced expression of GnuR by both glucose and gluconate in the KT2440 wild‐type strain (Figure 2A) was further validated by western blot in CY1563 (Figure S4). In addition, studies have shown that the GnuR homologue of GntR in Pseudomonas aeruginosa uses an effector‐mediated de‐repression mechanism (Daddaoua et al. 2017). The protein is promoter‐bound in the absence of ligand, and the binding of gluconate and glucose to the DNA/protein complex causes the release of GntR. Gluconate has a stronger inhibitory effect on the GntR protein than glucose. The activity of GnuR in P. putida could therefore be controlled in a similar way. We had observed in KT2440 that the expression level of GnuR was as high in glucose as in gluconate (Figure 2), but gluconate would show a stronger inhibitory effect on GntR than glucose. Therefore, in order to obtain a high amount of GnuR protein and also to maintain a relatively high binding affinity of GnuR to DNA, we investigated the DNA binding targets of GnuR in the CY1563 strain grown in the medium with glucose as the sole carbon source by using the ChIP‐seq method. The ChIP‐seq assay identified more than 600 significant binding targets of GnuR with the identified fold enrichment of the ChIP signal higher than 1.3, which were located in the putative promoter regions of genes (meaning that each binding peak covered the upstream of the CDS of its adjacent gene, Figure 4A, Table S1). We performed a motif analysis of the GnuR binding targets using the MEME software, and a consensus sequence was identified in each of three analyses using promoter sequences carrying binding peaks of GnuR with the fold enrichment of the ChIP signal reaching three different cutoffs, which were 2.0, 1.6 and 1.3 (Figures 4B and S5). The consensus sequence of the motif obtained in each analysis was quite similar, but the E‐value of the motif varied. The E‐value of the motif obtained by using promoter sequences carrying binding peaks of GnuR with fold enrichment of ChIP signal higher than 1.3 was the lowest. Therefore, the consensus sequence of the motif with the lowest E‐value was used for further analysis (Figure 4B). The consensus sequence of the GnuR binding motif was GC‐rich with a reverse complementary sequence in its center. Identification of the GnuR binding targets by ChIP‐seq analysis. (A) Distribution of the GnuR binding peaks located in the putative promoter regions of genes identified in the KT2440 genome. The peak fold enrichment parameter indicates the strength of the ChIP‐seq signal. (B) Logo of the consensus sequence of the GnuR binding targets in the promoter regions. The logo was analysed by the MEME Suite (https://meme‐suite.org/) with E‐value = 2.0e−159 using ~600 promoter sequences showing binding peaks of GnuR with fold enrichment of ChIP signal higher than 1.3. Similar logos were also obtained by applying the MEME Suite using promoter sequences showing binding peaks of GnuR with fold enrichment of ChIP signal higher than 2.0 or 1.6 (Figure S5), with E‐value = 2.3e−002 or E‐value = 9.9e−072. (C) Distribution of the GnuR binding targets detected in the four gene clusters involved in glucose catabolism. The red bars in front of the gene symbols indicate that GnuR binding targets are detected in the promoter region of these genes. Interestingly, we found that the identified GnuR binding targets were highly enriched in the promoter regions of metabolism genes involved in glucose catabolism and TF genes involved in the expression regulation of these metabolic genes (Figure 4C). In almost all metabolic pathways related to glucose metabolism, certain genes contained GnuR‐specific binding targets in their promoter regions (Figure S6). In particular, GnuR binding peaks were identified in the promoter region of each gene in all peripheral pathways, except PP_3384, a putative gluconate 2‐dehydrogenase auxiliary subunit. The ChIP‐seq assay suggested the putative importance of GnuR in the regulation of glucose and gluconate metabolism in P. putida. We then used RNA‐seq to compare the expression levels of these genes in the GnuR wild‐type and knockout strains in order to identify which gene with a detected GnuR binding peak would respond to GnuR. GnuR Directly Represses the Expression of Genes Involved in the ED and the Peripheral Glucose and Gluconate Metabolism Pathways We constructed a gnuR knockout (KO) mutant strain derived from KT2440, namely, CY1318. The CY1318 and KT2440 strains were each grown in two different media. One medium was the M9 minimal medium with glucose as the sole carbon source, and the other medium was the LB‐rich medium. The reason for using glucose as the sole carbon source was the same as that when we performed the ChIP‐seq assay. Similarly, although we observed that the expression level of gnuR in KT2440 when it was grown in the LB‐rich medium was approximately 5‐fold lower than that in KT2440 when it was grown in glucose (Figure S7), the DNA binding activity of each molecule of GnuR could still be high in the LB medium without the high amount of inhibitory ligand of glucose (Daddaoua et al. 2017). Thus GnuR may still have a regulatory role when P. putida is grown in the LB‐rich medium, and it would be interesting to explore the regulation of GnuR in this medium. The total RNA of CY1318 or KT2440 grown in each condition was extracted at log phase, and their transcriptomes were analysed by RNA‐seq. Differentially expressed genes (DEGs) of CY1318 versus KT2440 in both conditions were screened using a criterion of fold change greater than 2 and a q‐value less than 0.01. In the glucose minimal medium growth condition, 11 DEGs were identified (Figure 5A, Table S2), and in the LB‐rich medium growth condition, 14 DEGs were identified (Figure 5B, Table S3). The majority of these DEGs were further validated by RT‐qPCR (Figure S8). These DEGs included the catabolism genes in the two operons of ptxS‐kguE‐kguK‐kguT‐ptxD and PP_3384‐PP_3383‐PP_3382 (coding for gluconate 2‐dehydrogenase) together with the two genes of gntT and gnuK operating in the peripheral pathway and the gene eda operating in the ED pathway (Figure 1). The gtsA DEG encodes a glucose transporter (Liu et al. 2020), and the ptxS DEG is one of the TF genes that functions in the regulation of the expression of glucose catabolism genes (Daddaoua et al. 2010). The DEG yihS gene encodes an aldose‐ketose isomerase, which catalyses the reversible epimerisation between N‐acetyl‐D‐glucosamine and N‐acetyl‐D‐ mannosamine (Itoh et al. 2008). The DEG yeaD encodes a G6P 1‐epimerase, which is suggested to be a member of the galactose mutarotase‐like superfamily. Galactose mutarotase produces α‐D‐galactose from β‐D‐galactose by stereochemical inversion at carbon 1 of galactose. Then α‐D‐galactose is transformed into glucose 1‐phosphate (G1P) by the Leloir pathway (You et al. 2010). G1P is further converted to G6P, an intermediate compound in glycolysis. Thus the two DEGs yihS (Itoh et al. 2008) and yeaD (You et al. 2010) are involved in the catabolism of glucose derivatives. The PP_3418 DEG encodes a hypothetical protein. All these DEGs involved in the metabolism of glucose and its derivatives were directly or indirectly negatively responsive to GnuR (Figure 5A,B). Identification of genes directly controlled by GnuR. (A, B) Volcano maps comparing the transcriptomes of the GnuR KO strain CY1318 and the wild‐type strain KT2440. Strains were grown in M9 minimal medium supplemented with glucose as a sole carbon source (A) or in LB‐rich medium (B). Horizontal lines indicate q‐value = 0.01, and straight lines indicate log2FC = 1 or −1. Red (A) or purple (B) squares indicate up‐expressed DEGs (Up). Blue (A) or green (B) triangles indicate down‐expressed DEGs (Down), and black dots indicate genes not differentially expressed (noDEG). (C) Regulon genes of GnuR invoved in glucose metabolism. By integrating the ChIP‐seq results in Figure 4 with the transcriptomic results in panels A and B, genes directly controlled by GnuR (GnuR regulon) were identified and mapped to the metabolic pathways, with symbols shown in red colour, and the fold change of gene expression identified by comparing the transcriptome of CY1318 to that of KT2440 was shown in the parentheses, with red (purple) colour indicating value obtained in panel A (B). Integrating our previous ChIP‐seq results (Figure 4) with the RNA‐seq data (Figure 5A,B), we identified genes directly controlled by GnuR. These were the DEGs identified by the RNA‐seq assay, which had GnuR‐specific binding peak in their promoter region identified by the ChIP‐seq assay (Table 1, Figure S9). The expression of ptxS‐kguE‐kguK‐kguT‐ptxD, PP_3383‐PP_3382, gntT, gnuK, yihS, yeaD and PP_3418 was directly controlled by GnuR. The putative DNA binding site of GnuR on each promoter was also detected (Table S4) based on the consensus binding motif of GnuR identified in this work (Figure 4B). Accordingly, we mapped these genes to their metabolic pathways and found that they were enriched in the peripheral and the ED pathways (Figure 5C). Thus, the primary role of GnuR was to repress directly the expression of genes that function in the ED and the peripheral glucose and gluconate metabolism pathways. TABLE 1GnuR regulon genes identified in this study.Fold enrichment of the GnuR binding peak a First gene of the transcription unitAll known gene(s) in the transcription unit1.38 yihS yihS 27.46 yeaD yeaD 1.74 eda eda 1.62 ptxS ptxS‐kguE‐kguK‐kguT‐ptxD 4.82PP_3383PP_3383‐PP_338223.23 gnuK gnuK 23.23 gntT gntT 1.4PP_3418PP_3418 a The fold enrichment of the GnuR binding peak was analysed by MACS2 software. Each peak covered within −500 to −1 bp relative to the start codon of the adjacent gene. GnuR regulon genes identified in this study. The fold enrichment of the GnuR binding peak was analysed by MACS2 software. Each peak covered within −500 to −1 bp relative to the start codon of the adjacent gene. GnuR Affected the Cellular Utilisation of Both Glucose and Gluconate in P. putida Since we had identified the repressive regulation of GnuR on the expression of catabolic enzymes of glucose and gluconate metabolism (Figure 5), will GnuR affect the cellular use of the two carbon sources in P. putida ? To answer this question, we monitored the growth of the wild‐type KT2440 and the GnuR KO strain in glucose or gluconate minimal medium to investigate the putative regulation of GnuR on the cellular use of glucose or gluconate. We grew each strain in LB‐rich medium to stationary phase and inoculated the same number of cells into glucose (Figure 6A,C) or gluconate (Figure 6B,D) minimal medium, ensuring that the starting OD600 of each strain in glucose or gluconate minimal medium was set to the same value. We observed that there were no significant changes in the growth rate of their logarithmic growth phase in either glucose or gluconate, but the length of their lag time before reaching fast exponential growth varied. The GnuR KO strain had a shorter lag time before reaching fast exponential growth as compared to the wild‐type strain. Thus, the regulatory function of GnuR was more important in the early phase when the cells switched from rich medium to glucose or gluconate starvation medium, and it would be more difficult for the wild‐type GnuR strain than the GnuR KO strain to start metabolising glucose or gluconate and grow. This finding further supported the inhibitory effect of GnuR on the expression of the catabolism genes involved in glucose and gluconate metabolism (Figure 5). Growth study of KT2440 and CY1318 in glucose and gluconate. (A, B) The wild‐type strain KT2440 and the GnuR KO strain CY1318 were grown in LB‐rich medium to stationary phase and diluted to an equal OD600 in M9 minimal medium supplied with 22 mM glucose (A) or 4 mM gluconate (B), respectively. Their growth curves in minimal medium were monitored by a microplate reader. Each growth was subjected to triplicate runs, and the growth curve of OD600 was plotted against time using the averaged value ± SEM. at each time point. To visualise the lag time before reaching fast exponential growth of each growth better, the growth curves of the early stage of the two strains grown in glucose and gluconate shown in panel A and panel B were, respectively, re‐plotted and shown in panel (C) (glucose) and (D) (gluconate). Discussion Here, by using global assays together with physiological studies, we discovered several interesting features of the regulation of glucose and gluconate catabolism in P. putida . First, we found that both glucose and gluconate could strongly induce the expression of genes of glucose catabolism, especially genes involved in the whole peripheral and the ED pathways (Figure 2) in the wild‐type P. putida strain KT2440. This finding further emphasised the overlapping metabolic pathways of the two carbon sources (Figure 1) (Nikel et al. 2015). We found that genes in the upper EMP pathway could not be induced by either carbon source (Figure 2), suggesting that the upper EMP is not the main metabolic pathway of the two carbon sources, since the pathway induced by a particular carbon source is usually the main metabolic pathway of that carbon source in bacteria, as exemplified by the famous lac operon in Escherichia coli . Therefore, our results support the reported literature showing that glucose is mainly metabolised via the ED pathway but not the upper EMP pathway (Nikel et al. 2015; Kukurugya et al. 2019). Meanwhile, similar to glucose, our study suggests that gluconate is also mainly metabolised via the ED pathway. However, because gluconate is highly accumulated when P. putida is grown on glucose with the help of glucose dehydrogenase Gcd (Volke et al. 2023), the observed high response of these catabolic and TF genes to glucose in KT2440 (Figure 2) could actually be due to the high amounts of gluconate produced. Our subsequent studies of the expression response of these catabolic and TF genes to glucose and gluconate in the gcd deletion mutant strain which does not form gluconate when grown on glucose clearly differentiated their independent responses to these two carbon sources (Figure 3). Interestingly, all the catabolic genes related to gluconate metabolism in the peripheral pathway were highly induced by gluconate and genes involved in glucose transport showed a much stronger response to glucose than to gluconate. In contrast, the catabolic genes involved in pathways downstream of the peripheral pathway (ED, PP and EMP) could be induced to a similar extent by both glucose and gluconate. Second, by applying the integrated analysis of ChIP‐seq and RNA‐seq, we revealed the regulon of GnuR, which was one of the four TF genes involved in glucose metabolism and was initially observed to be induced by both glucose and gluconate in KT2440 (Figure 2A), whereas our study in the gcd mutant strain proved that GnuR was indeed induced by gluconate (Figure 3C). The ChIP‐seq studies revealed that more than 600 genes carried binding peaks of GnuR in their promoter regions in vivo (Figure 4A, Table S1), but only a small fraction of these genes were shown to respond to GnuR (Figure 5C, Table 1). Thus, the majority of the GnuR bindings in vivo had no regulatory effect on the neighbouring genes under the conditions we investigated. What might be the reasons for this? We found that the consensus sequence of the identified GnuR binding motif was GC‐rich (Figure 4B). In parallel, the genome of P. putida is also GC‐rich (Belda et al. 2016), and the probability of having this GC‐rich motif on the genome could be relatively high. Thus, GnuR could bind to many sites with a sequence similar to this GC‐rich motif. Alternatively, genes with identified GnuR binding peaks in their promoter regions could respond to GnuR under other physiological conditions, such as growth conditions with different glucose derivatives, which were beyond the scope of this study. In a previous microarray analysis, GnuR has been reported to repress the expression of two of its neighbouring genes, gntT and gnuK (del Castillo, Duque, and Ramos 2008), but it is difficult to determine whether the regulation is direct or indirect. By integrating the results of ChIP‐seq with RNA‐seq, we have delineated the contours of the GnuR regulon for the first time (Table 1): GnuR is a transcriptional repressor, and genes directly controlled by GnuR are enriched in the whole peripheral and the ED pathways (Figure 5C). The repressive effect of GnuR on the expression regulation of these catabolic genes was further demonstrated by the physiological studies, which showed that the regulatory function of GnuR was significant in the early stage when cells just started to metabolise glucose or gluconate as a sole carbon source (Figure 6). Given that the GnuR KO strain showed better adaptation to these two carbon substrates than the wild‐type strain KT2440, the GnuR KO mutant strain of CY1318 constructed in this study could be used as a starting chassis for further bioengineering studies when using glucose or gluconate as a carbon source. Finally, our study suggests that there is a feedback control in the regulation of glucose or gluconate catabolism in P. putida (Figure 7). In KT2440, the glucose and gluconate catabolism genes together with the TF gene gnuR were induced by both glucose and gluconate, and, in parallel, the induced GnuR further repressed the expression of these catabolic genes. Thus, glucose or gluconate, GnuR and the catabolic genes of glucose and gluconate formed an incoherent feedforward loop (Alon 2007; Reeves 2019), in which the input signal (glucose or gluconate, X) activated both the intermediate (GnuR, Y) and the final output (catabolic genes of glucose or gluconate, Z), while Y repressed Z. The advantage of this feedback control of glucose or gluconate catabolism in P. putida needs further exploration. Schematic diagram of the incoherent feedforward loop proposed in this work. Glucose (Glc) or gluconate (Gluc), GnuR and the catabolic genes of Glc/Gluc form an incoherent feedforward loop. The input signal (Glc or Gluc) induces the expression of the intermediate (GnuR) and the final output (catabolic genes of Glc/Gluc), while GnuR further represses the expression of these catabolic genes. By studying the expression response of glucose catabolism genes to different carbon sources and by demonstrating the existence and function of the GnuR regulon, we have contributed to the growing body of knowledge concerning the intricate control of glucose and gluconate catabolism in P. putida . The results presented here provide a theoretical basis for the design of innovative synthetic metabolism of carbon sources when using KT2440 as a bio‐catalytic chassis. A similar approach can be applied to investigate the contribution of incompletely characterised TFs to the catabolism of the two carbon sources, thereby elucidating further principles of efficient carbon metabolism in this model bacterial chassis, which is widely used in bioengineering. Materials and Methods Construction of Bacterial Strains The strains used in this study were all derived from the wild‐type P. putida KT2440, as detailed in Table S5. We constructed derivatives including the gnuR knockout mutant CY1318, the gcd knockout mutant CY1850 and CY1563 carrying the C‐terminal 6×His‐tagged GnuR by the method of bacterial conjugation, a genetic exchange mechanism that requires direct contact between donor and recipient cells (Smillie et al. 2010). To construct these derivatives, we used the spliced overlap extension (SOE) PCR technique to generate the desired DNA fragment flanked by restriction enzymes cutting sites. The DNA fragments used to construct the gnuR knockout mutant CY1318 included an 800 bp sequence located upstream of the gnuR start codon and an 800 bp sequence located downstream of the gnuR stop codon. The DNA fragments used to construct the gcd knockout mutant CY1850 included an 800 bp sequence located upstream of the gcd start codon and an 800 bp sequence located downstream of the gcd stop codon. The fragments used to construct the CY1563 strain carrying the C‐terminal 6×His tagged GnuR included an 800 bp sequence upstream of the gnuR stop codon, followed by a 6×His‐tagged sequence (CATCATCATCATCATCAT), the original gnuR stop codon and an additional 800 bp sequence downstream of the gnuR stop codon. The PCR fragments used to construct CY1318 and CY1850 had EcoRI and HindIII restriction enzyme cutting sites at both ends, and the PCR fragment used to construct CY1563 had BamHI and HindIII restriction enzyme cutting sites at both ends. These three PCR products were then each subcloned into the plasmid pK18mobsacB carrying a kanamycin cassette (Schäfer et al. 1994), and the correct constructions were verified by DNA sequencing. The resulting three plasmids containing the desired DNA fragment were each introduced into the E. coli strain of S17‐1 by electroporation. Each plasmid was then transferred by conjugation from S17‐1 (donor) to P. putida KT2440 (recipient). At least three conjugants were randomly selected and streaked on LB plates supplemented with 10% sucrose to obtain single colonies in each strain construction. Then, in each construction, four single colonies were randomly isolated and plated on Pseudomonas Isolation Agar (PIA) plates plus 50 μg/mL kanamycin and LB plates, respectively. The correct constructions showing kanamycin sensitivity were verified by PCR and DNA sequencing, resulting in strains CY1318, CY1850 and CY1563, respectively. Growth of Cell Culture in a Water Bath Shaker Batch cultures were grown either in M9 minimal medium [Na2PO4 (6 g), KH2PO4 (3 g), NaCl (5 g), NH4Cl (1 g), MgSO4 (0.24 g), BH3O3 (0.75 mg), ZnCl2 (0.125 mg), MnCl2·4H2O (0.075 mg), CoCl2 (0.5 mg), CuCl2·2H2O (0.025 mg), NiCl2·6H2O (0.05 mg), NaMnO4·H2O (0.075 mg) and FeCl3 (0.015 mg) in 1 L] supplemented with one of the three carbon substrates [22 mM glucose (Glc), 20 mM succinate (Suc), 20 mM gluconate (Gluc)] as a sole carbon source or in LB‐rich medium. All growths were carried out in three steps, as previously described (You et al. 2013), but in a 30°C water bath shaker shaking at 150 rpm. The three steps include seed culture in LB broth, pre‐culture and experimental culture in M9 minimal medium or LB‐rich medium. One colony of P. putida from a fresh LB agar plate was inoculated into liquid LB and grown at 30°C with shaking (seed culture). Cells were then diluted 102 to 104 times in M9 minimal medium or LB‐rich medium and grown overnight in a 30°C water bath shaker (pre‐culture). Finally, cells from the overnight pre‐culture were diluted 102 to 104 times in pre‐warmed M9 minimal medium or LB‐rich medium and grown in a 30°C water bath shaker (experimental culture). RNA‐Seq Assay and Analysis of Transcriptome Data To study the transcriptomic response of KT2440 to different carbon substrates, KT2440 were grown individually in M9 minimal medium supplemented with one of the three carbon substrates [22 mM glucose (Glc), 20 mM succinate (Suc) and 20 mM gluconate (Gluc)] as the sole carbon source. To study putative GnuR‐responsive genes, the wild‐type strain KT2440 and the GnuR KO strain CY1318 were respectively grown in M9 minimal medium with 22 mM glucose as the carbon source or were grown in LB‐rich medium, respectively. All growths were carried out in three steps in a 30°C water bath shaker, as previously described. Bacterial samples for total RNA extraction were harvested during the mid‐exponential phase at an OD600 of ~0.4. Total RNA extraction and RNA‐seq assays were performed according to previously established methods (Li et al. 2019). Raw reads obtained from strand‐specific RNA sequencing using Illumina HiSeq 2000, PE125 were filtered, and adaptors were trimmed. Clean reads were mapped to the KT2440 genome (NCBI accession number for chromosome: NC_002947.4) using Bowtie2. The expression of each gene was calculated using the RPKM (reads per kilo bases per million mapped reads) method. Three independent total RNA extractions and transcriptome analyses by RNA‐seq were performed for each condition. DEGs between compared conditions were characterised by DEseq2 with adjusted q‐value < 0.01. ChIP‐Seq Assay and Analysis of ChIP‐Seq Data The ChIP‐seq assay was performed according to the established procedures (Hou et al. 2023), and two independent ChIP‐seq assays were performed in this study. Briefly, CY1563 cells (with a C‐terminal gnuR‐6×his fusion) were grown in M9 minimal medium with 22 mM glucose as the carbon source. Then 40 mL culture at mid‐exponential growth phase was cross‐linked with 1.1 mL 37% formaldehyde. After sonicating the chromatin, the sequencing library was constructed, following the protocol provided by the INEXTFLEX ChIP‐Seq Library Prep Kit for Illumina Sequencing (NOVA‐514120, Bioo Scientific), and was sequenced by Illumina Xten with PE 150. To analyse the sequencing data, Trimmomatic (version 0.38) was used to filter out low‐quality reads. Clean reads were mapped to KT2440 genome (NCBI accession number for chromosome: NC_002947.4) by Bwa (version 0.7.15), and samtools (version 1.3.1) was used to remove potential PCR duplicates. The MACS2 software (version 2.1.1.20160309) was used to call peaks with default parameters. Input DNA served as a control to represent background noise. The fold enrichment of the peak indicated the strength of the ChIP‐seq signal. Called peaks with fold enrichment higher than 1.3 and p‐value smaller than 0.05 were counted as significant binding peaks in this study. The averaged value of the fold enrichment of the peak was documented when a binding peak was identified by both of the two independent ChIP‐seq assays. In this work, peaks that covered the upstream of the CDS of a nearest gene were examined. The GnuR binding motif was analysed using MEME (https://meme‐suite.org/meme/tools/meme), and the logo with the lowest E‐value was documented. RT‐qPCR Batch cultures of the wild‐type strain KT2440 and the gnuR KO strain CY1318 were grown either in the M9 minimal medium supplemented with 22 mM glucose as the carbon source or in the LB‐rich medium to mid‐exponential phase, and total RNA was extracted, as described previously (Li et al. 2019). Batch cultures of the gcd KO strain CY1850 were grown in the M9 minimal medium individually supplemented with 22 mM glucose, 20 mM gluconate and 20 mM succinate as the carbon source to mid‐exponential phase, and total RNA was extracted (Li et al. 2019). Then, cDNA templates were synthesised according to the instructions of EasyScript All‐in‐One First‐Strand cDNA Synthesis SuperMix for qPCR (One‐Step gDNA Removal) (TransGen Biotech, Beijing, China). RT‐qPCR assays were conducted using the qTOWER real‐time PCR thermocycler (Analytik Jena). The primers used for the assays are reported in Table S6. The expression level of each gene was normalised to the expression level of hslO, which served as the reference. All reactions were run in triplicate. The relative expression was analysed using the ΔΔCt quantification method using qPCRsoft v3.2, where CT refers to the cycle threshold. Two to four independent experiments were conducted in each condition. Western Blot Assay The strain CY1563 carrying the C‐terminal 6×His‐tagged GnuR was grown in M9 minimal medium supplemented with one of the three carbon substrates [22 mM glucose (Glc), 20 mM succinate (Suc) and 20 mM gluconate (Gluc)] and was grown to reach the log phase with an OD600 of approximately 0.4. The cells were centrifuged at 6000 rpm for 3 min and resuspended in 1 × phosphate buffered saline (PBS). Then, 0.1 OD600 of cells were taken from each sample and boiled in SDS‐PAGE sample buffer (GenStar, Beijing, China) for 10 min. All the processed samples were loaded for SDS‐PAGE electrophoresis. To detect target proteins, membranes were incubated with appropriate primary and secondary antibodies and then treated with the SuperSignal West Pico PLUS Chemiluminescent Substrate (Thermo Fisher, Waltham, MA, United States). The treated membranes were imaged using a Tanon‐5200 imaging system (Tanon, Shanghai, China). The primary antibodies used in this study were anti‐6×His tag (Abcam, Cambridge, UK). The secondary antibody used in this study was Goat Anti‐Rabbit IgG‐HRP (Abmart, Shanghai, China). Cell Growth in M9 Minimal Medium With Glucose or Gluconate as a Sole Carbon Substrate To study the putative regulation of GnuR on the cellular utilisation of glucose or gluconate, the wild‐type strain of KT2440 and the GnuR KO strain CY1318 were grown in LB‐rich medium to stationary phase, with OD600 reaching ~4.0, and subsequently, each culture was diluted 104 to 105 times to obtain the same value of starting OD600 in 100 μL M9 minimal medium supplemented with 22 mM glucose or 4 mM gluconate as a carbon source and grown in a 24‐well plate (SORFA, Zhejiang, China) using a microplate reader (BioTek, Vermont, United States) with the temperature set at 30°C and shaking rate at 483 rpm. The OD600 of each well was monitored every 20 min. Each growth was run in triplicate, and the growth curve was plotted against time using the average value of OD600 at each time point. Author Contributions Wenbo Chen: data curation, formal analysis, investigation, validation, visualization, writing – original draft. Rao Ma: data curation, formal analysis, investigation, writing – original draft. Yong Feng: data curation, formal analysis, investigation. Yunzhu Xiao: formal analysis, investigation, data curation. Agnieszka Sekowska: resources, formal analysis, writing – review and editing. Antoine Danchin: resources, formal analysis, writing – review and editing. Conghui You: conceptualization, data curation, formal analysis, funding acquisition, supervision, writing – original draft, writing – review and editing. Conflicts of Interest The authors declare no conflicts of interest. Supporting information Figure S1. Expression of several genes involved in glucose metabolism in P. putida KT2440 grown in glucose versus that in succinate using RT‐qPCR. Figure S2. Doubling time of CY1850 grown individually in glucose, gluconate and succinate. Figure S3. Doubling time of KT2440 and CY1563 grown in glucose. Figure S4. Detection of expression levels of GnuR protein in CY1563 using western blot. Figure S5. Logo of the consensus sequence of the GnuR binding targets in the promoter regions. Figure S6. Distribution of the GnuR binding targets in catabolic pathways of glucose metabolism. Figure S7. Detection of the expression of gnuR in KT2440 grown in LB‐rich medium or in glucose minimal medium by RT‐qPCR. Figure S8. Detection of the expression of glucose metabolic genes in KT2440 and CY1318 by RT‐qPCR. Figure S9. Binding peaks of GnuR identified in the promoter regions of operons. Table S1. Binding targets of GnuR identified in the putative promoter regions of genes in P. putida KT2440 by ChIP‐seq. Table S2. List of differentially expressed genes in the GnuR KO strain CY1318 compared to the wild‐type strain of KT2440 grown in M9 minimal medium with glucose as a sole carbon source by RNA‐seq. Table S3. List of differentially expressed genes in GnuR KO strain CY1318 compared to the wild‐type strain of KT2440 grown in LB‐rich medium using RNA‐seq. Table S4. Predicted binding sequences of GnuR in the promoter region of the GnuR regulon genes. Table S5. Strains used in this work. Table S6. Primers used for RT‐qPCR in this work.