A novel threecomponent systembased ... - Wiley Online Library

Molecular Microbiology (2015) 95(4), 576–589 ■

doi:10.1111/mmi.12894 First published online 20 December 2014

A novel three-component system-based regulatory model for D-xylose sensing and transport in Clostridium beijerinckii Zhe Sun,1 Yixiong Chen,2 Chen Yang,1 Sheng Yang,1 Yang Gu1,3* and Weihong Jiang1,4* Key Laboratories of 1Synthetic Biology and 2Insect Developmental and Evolutionary Biology, Institute of Plant Physiology and Ecology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai 200032, China. 3 State Key Laboratory of Motor Vehicle Biofuel Technology, Nanyang 473000, China. 4 Shanghai Collaborative Innovation Center for Biomanufacturing Technology, 130 Meilong Road, Shanghai 200237, China.

Summary D-Xylose is the most abundant fermentable pentose in

nature and can serve as a carbon source for many bacterial species. Since D-xylose constitutes the major component of hemicellulose, its metabolism is important for lignocellulosic biomass utilization. Here, we report a six-protein module for D-xylose signaling, uptake and regulation in solvent-producing Clostridium beijerinckii. This module consists of a novel ‘three-component system’ (a putative periplasmic ABC transporter substrate-binding protein XylFII and a two-component system LytS/YesN) and an ABC-type D-xylose transporter XylFGH. Interestingly, we demonstrate that, although XylFII harbors a transmembrane domain, it is not involved in D-xylose transport. Instead, XylFII acts as a signal sensor to assist the response of LytS/YesN to extracellular D-xylose, thus enabling LytS/YesN to directly activate the transcription of the adjacent xylFGH genes and thereby promote the uptake of D-xylose. To our knowledge, XylFII is a novel single transmembrane sensor that assists two-component system to respond to extracellular sugar molecules. Also of significance, this ‘three-component system’ is widely distributed in Firmicutes, indicating that it may play a broad role

Accepted 10 November, 2014. *For correspondence. E-mail [email protected]; Tel. (+86) 21 54924172; Fax (+86) 21 54924015. E-mail [email protected]; Tel. (+86) 21 54924178; Fax (+86) 21 54924015.

© 2014 John Wiley & Sons Ltd

in this bacterial phylum. The results reported here provide new insights into the regulatory mechanism of D-xylose sensing and transport in bacteria.

Introduction D-Xylose

is the most abundant monosaccharide in nature after glucose and can serve as a cheap, fermentable, lignocellulose-derived sugar for many bacterial species (Aristidou and Penttila, 2000; Rubin, 2008). The uptake of D-xylose into bacterial cells mainly depends on ABC-type transporter and proton symporter systems. As an energyconsuming (by coupling ATP hydrolysis) sugar transporter, the ABC-type transporter is composed of three subunits: the substrate-binding protein XylF, the membranespanning protein XylH and the nucleotide-binding protein XylG (Davidson and Chen, 2004; Davidson and Maloney, 2007). The ABC-type transporter has a much higher affinity for D-xylose than the proton symporter and therefore often plays a more important role in D-xylose uptake in bacteria (Ahlem et al., 1982; Davis and Henderson, 1987). However, despite extensive studies, the mechanisms by which bacteria sense the extracellular D-xylose signal and regulate the downstream D-xylose uptake process are not well understood. The predominant signal sensing and transduction pathways in bacteria are two-component systems (TCSs) (Stock et al., 2000; Wuichet et al., 2010; Jung et al., 2012), which enable cells to flexibly respond to environmental changes, including solutes, pH, temperature and nutrients (Mascher et al., 2006). The classical TCSs are composed of a histidine kinase (HK) normally responsible for sensing specific stimuli, and a response regulator (RR) involved in regulating gene expression. So far, examples of ABC-type transporters that are controlled by adjacent TCSs were only characterized in Firmicutes (Joseph et al., 2002; Coumes-Florens et al., 2011; Dintner et al., 2011), such as the ABC transporters BceAB and PsdAB, which are activated by the TCSs BceRS and PsdRD, respectively, and then confer peptide antibiotics resistance to Bacillus subtilis (Ohki et al., 2003; Staron et al., 2011). With regard to sugar uptake, only an ABC transporter system (xynEFG) for xylo-oligosaccharides, regulated by a TCS (XynDC), has been described in Geobacillus stearothermophilus

A novel D-xylose sensing and transport model 577

recently; however, its extracellular signal and the related signal-sensing mechanisms are still unclear (Shulami et al., 2007). Besides independently responding to signals, HKs of some TCSs can employ auxiliary proteins to sense extracellular signals (Tetsch and Jung, 2009; Buelow and Raivio, 2010). For example, in Escherichia coli, C4-dicarboxylate transporter DctA could form a sensor complex with the fumarate responsive sensor kinase DcuS in the cytoplasm to co-sense fumarate under anaerobic conditions (Witan et al., 2012); the periplasmic binding protein ChvE improves sensitivity of the HK VirA in response to the phenolic signals in Agrobacterium tumefaciens, resulting in an increased expression of virulence genes (Zhao and Binns, 2011). However, an integrated signal sensing and TCS-based regulatory model that is essential for bacterial survival is poorly defined and understood so far. Here, we discovered a gene cluster xylFII-lytS-yesNxylFGH in Firmicutes, which contains a diverse group of important industrial bacteria, such as Bacillus and Clostridium. As exemplified by Clostridium beijerinckii, a typical Clostridium strain, this gene cluster consists of six neighboring genes (i.e. a putative D-xylose transporter substrate-binding protein XylFII, a TCS LytS/YesN and an ABC-type D-xylose transporter XylFGH). A detailed study in C. beijerinckii revealed that these genes formed a sixprotein module for accomplishing D-xylose signaling and uptake regulation by employing XylFII as an auxiliary protein to independently sense the extracellular D-xylose signal, which then activates the adjacent TCS LytS/YesN to regulate the expression of the ABC-type D-xylose transporter. In contrast to previously identified sensing proteins, XylFII is not a transporter, although it is predicted to harbor a transmembrane domain. This suggests that XylFII might be a novel type of D-xylose signaling sensor, which, combined with the neighboring TCS LytS/YesN, forms a novel ‘three-component system’ for the regulation of D-xylose transport. Given the widespread distribution of this ‘three-component system’ in Firmicutes, it may exert a conserved regulatory mechanism in this bacterial phylum. Our studies provide new insights into D-xylose sensing, signal transduction and the following uptake regulation in Firmicutes bacteria.

Results The xylFII-lytS-yesN-xylFGH-like gene cluster related to D-xylose uptake is widely distributed in Firmicutes Genes which share similar or related functions often gather together to form gene clusters. In our previous study (Gu et al., 2010), we observed a gene cluster ‘xylFII-lytS-yesNxylFGH’ in the genome of C. beijerinckii, which is made up © 2014 John Wiley & Sons Ltd, Molecular Microbiology, 95, 576–589

of six tandem genes, including a predicted D-xylose ABC transporter XylFGH, a TCS LytS/YesN and a putative D-xylose ABC transporter periplasmic binding protein XylFII, and possibly involved in D-xylose utilization. Interestingly, we recently determined that this gene cluster appears to be widely distributed in Firmicutes (Fig. 1A and Table S1). A certain number of homologs with the same gene arrangement were identified in class Clostridia and class Bacilli, occupying approximately 10% (56/574) of Firmicutes bacteria (Fig. 1B and Table S1). Especially among Thermoanaerobacteriales, an order of bacteria within the polyphyletic class Clostridia, up to 37.5% (12/32) of the bacteria contain the homologs (Fig. 1B). Besides, this gene cluster occurs in many lignocellulose-degrading bacteria, such as Thermoanaerobacter thermohydrosulfuricus WC1 and Paenibacillus sp. JDR-2 (Table S1). Notably, the cellulolytic and hemicellulolytic bacterium Paenibacillus barengoltzii G22 contains two homologs of the gene cluster (Table S1). Given the widespread occurrence of this gene cluster, it may play a broad role in Firmicutes. The xylFII-lytS-yesN-xylFGH gene cluster is involved in D-xylose utilization of C. beijerinckii Given the short intergenic regions and even overlapped genes in this gene cluster (Fig. 2A), we first identified its transcription units. A β-galactosidase reporter system was employed to assay the three putative promoter regions, PxylFII (379 nt), PlytS (112 nt) and PxylF (293 nt) (i.e. the upstream region of xylFII, lytS and xylF, respectively) in vivo. The results showed that only PxylFII and PxylF conferred significant higher β-galactosidase activity to the Clostridium strains harboring the reporter system compared with the control strain (harboring the plasmid pIMP1-lacZ) (Fig. 2B). Subsequently, RT-PCR analysis confirmed that yesN and xylF in this gene cluster were not co-transcribed (Fig. 2C). These data indicate the presence of two transcriptional units: xylFII-lytS-yesN and xylF-xylG-xylH (Fig. 2A). Since this gene cluster is adjacent to D-xylose degradation pathway genes in the genome of C. beijerinckii (Gu et al., 2010), we next sought to determine if it is involved in D-xylose utilization. Totally, six mutant strains (8052xylFII, 8052lytS, 8052yesN, 8052xylF, 8052xylG and 8052xylH), in which each of the six genes was inactivated (Fig. S1), were examined in using D-xylose. All of the mutant strains showed obviously impaired growth rate and substrate consumption compared with the wild-type strain (Fig. 2D and E), although the 8052xylFII showed a similar phenotype to the TCS mutants and less impaired phenotype compared with the other three mutant strains (8052xylF, 8052xylG and 8052xylH) (Fig. 2D and E). Simultaneously, no significant difference was observed in D-glucose consumption

578 Z. Sun et al. ■

Fig. 1. Arrangement of the ‘three-component system’ and distribution of the system in Firmicutes. A. Arrangement of the ‘three-component system’ in some examples of Firmicutes bacteria. Homologous genes are colored identically to distinguish each part of the system: periplasmic binding protein (grey), two-component system (white) and D-xylose ABC transporter (black). Detailed data are presented in Table S1. B. Distribution of the bacterial species harboring the system in different orders of Firmicutes. The left and right numbers in ratio mean the number of species harboring this system and all the species with complete genome sequencing data respectively.

between the six mutant strains and the wild-type strain (Fig. S2A and B). Moreover, the growth and D-xylose consumption phenotypes of the mutant strains 8052xylFII, 8052lytS, 8052yesN, 8052xylF and 8052xylG could be restored by functional complementation of the individual xylFII, lytS, yesN, xylF and xylG gene, respectively, thus excluding the possible polar effects on the downstream genes (Fig. S3). These results suggest that the xylFII-lytSyesN-xylFGH gene cluster is essential for D-xylose utilization by C. beijerinckii. XylF, but not XylFII, is a substrate-binding protein for D-xylose ABC transporter The common feature of ABC-type transporters is that they consist of two distinct domains, i.e. the transmembrane domain and the nucleotide-binding domain. Some ABC transporters employ additional binding proteins to capture and deliver the substrate to the translocator (Van der Heide and Poolman, 2002). Here, an interesting observation is that two putative periplasmic substrate-binding proteins, XylFII and XylF, occur in this gene cluster. According to genome annotation, xylFII and xylF were both substratebinding units of D-xylose ABC transporter; however, they

show low homology (17.7% amino acid sequence identity) (Fig. 3A). To clarify the function of these two proteins, heterologous host complementation experiments were performed in E. coli. We observed that xylFGH from C. beijerinckii could completely restore the D-xylose utilization ability of E. coli K12 mutant (K12xylEFGH) that is defective in D-xylose uptake (Fig. 3B), whereas the controls (xylGH and plasmid pUC118) could not. The transcript levels of the xylG and xylH genes were also compared between the E. coli K12 mutant harboring C. beijerinckii xylGH and the empty plasmid pUC118. The results showed that xylG and xylH transcripts in the E. coli K12 mutant bearing xylGH were much higher than those in the control strain, indicating an efficient expression of these two genes in the E. coli K12 mutant (Table S2). Furthermore, xylFGH overexpression conferred increased D-xylose utilization ability to C. beijerinckii (Fig. 3C). These data confirmed that XylFGH does act as a D-xylose ABC transporter. However, the function of xylFII in this system is not simple. Introduction of xylFIIGH, in which xylF of the original xylFGH cluster was replaced by xylFII, into the E. coli K12 mutant (K12xylEFGH) could not restore the defected D-xylose utilization ability (Fig. 3B), although an efficient © 2014 John Wiley & Sons Ltd, Molecular Microbiology, 95, 576–589


Fig. 2. Genetic organization and functional analysis of the xylFII-lytS-yesN-xylFGH gene cluster in C. beijerinckii. A. Schematic illustration of the xylFII-lytS-yesN-xylFGH locus in the C. beijerinckii NCIMB 8052 genome. Numbers in square brackets indicate the lengths of intergenic regions in bp. B. Determination of the promoter regions of the locus. Three intergenic regions, PxylFII (379 nt), PlytS (112 nt) and PxylF (293 nt), were separately fused to a lacZ reporter gene. pIMP1-lacZ was used as the negative control. C. Co-transcription analysis of the six genes from xylFII to xylH. RT-PCR was performed using cDNA as the template and RNA as the negative control. The residual D-xylose in the medium (solid lines) and growth curves (dashed lines) were monitored in wild-type 8052, 8052lytS and 8052yesN (D), and 8052xylFII, 8052xylF, 8052xylG and 8052xylH (E) grown in YP2 medium with D-xylose as the sole carbon source. The vertical bars indicate the standard deviation of the mean for three independent replicate cultures.

expression of xylFII, xylG and xylH were observed by transcriptional analysis (Table S2). In addition, overexpression of xylFIIGH or xylFII solely did not yield the same result as that obtained upon xylFGH overexpression in C. beijerinckii (Fig. 3C). Therefore, we reasoned that the encoded XylFII product here might play a new role instead of a D-xylose-binding protein. LytS/YesN regulates expression of ABC transporter XylFGH Since the xylFII-lytS-yesN-xylFGH gene cluster is involved in D-xylose utilization, we next sought to elucidate the mechanism by which LytS/YesN regulates its target genes in this cluster as well as the other adjacent putative D-xylose degradation pathway genes (i.e. xylAI, xylB and xylR). Quantitative real-time reverse transcription polymerase chain reaction (qRT-PCR) analysis was first performed to identify differentially expressed genes in a lytS and a © 2014 John Wiley & Sons Ltd, Molecular Microbiology, 95, 576–589

yesN inactivation mutant respectively. Interestingly, compared with up to 2.5-fold decrease in transcriptional level of the other genes (xylFII, xylAI, xylB and xylR), the transcription of xylFGH was much more strongly downregulated (approximately 175-fold) after inactivation of the TCS LytS/ YesN (Fig. 4A). Using bioinformatics analysis, we further revealed that YesN exhibits typical features of an RR of TCS (i.e. an N-terminal receiver domain and a C-terminal helix-turn-helix DNA binding domain) (Fig. S4). It is therefore possible that LytS/YesN may directly regulate the xylFGH genes. To investigate this possibility, the recombinant YesN protein with a maltose-binding protein (MBP) fused to its N-terminus was purified and then applied in electrophoretic mobility shift assays (EMSAs). The xylFGH promoter region was tested. A protein concentrationdependent shift of the xylFGH promoter probe was observed (Fig. 4B), while retardation of DNA migration was not detected when using only the MBP tag (Fig. S5A). This demonstrated that YesN directly regulates xylFGH expres-


Fig. 3. Functional analysis of the xylFGH and xylFIIGH genes from C. beijerinckii. A. Amino acid sequence alignment of XylFII and XylF proteins. Red background shows sequence identity, and yellow background shows sequence similarity. The figure was prepared using ESPript 3.0 (Robert and Gouet, 2014). B. Complementation of E. coli mutant K12xylEFGH, which is deficient in xylE (b4031) and xylFGH (b3566, b3567 and b3568) responsible for D-xylose transport, by C. beijerinckii genes. The residual D-xylose in the medium (solid lines) and growth curves (dashed lines) were monitored in strains separately expressing the xylGH, xylFGH or xylFIIGH genes from C. beijerinckii in K12xylEFGH and grown in M9 medium with D-xylose as the sole carbon source. C. Effect of the overexpression of the xylFII, xylFGH or xylFIIGH genes in C. beijerinckii. The residual D-xylose in the medium (solid lines) and growth curves (dashed lines) were monitored in strains grown in YP2 medium with D-xylose as the sole carbon source. Vertical bars indicate the standard deviation of the mean for three independent replicate cultures.

sion. In addition, it should be noted that, with the addition of acetyl phosphate (acetyl-P) in EMSAs, a stronger binding of YesN to the xylFGH promoter probe was observed compared with the conditions in the absence of acetyl-P (Fig. S5B). This indicated a more specific binding of phosphorylated form of YesN to its target DNA fragment. To further understand the role of YesN in regulating xylFGH, high-resolution S1 nuclease mapping was used to determine the transcription start site of the xylFGH genes. We identified two adjacent transcription start sites located −40 and −41 bp upstream of the xylF translation initiation codon (Fig. 4C). The potential −10 region (TATAAT), −35 region (TTGAAA) and ribosome-binding site (GGAGA) were also predicted (Fig. 4E). Subsequently, the precise location of the YesN binding sites in the xylFGH promoter region was determined using DNase I footprinting assays. The analysis revealed a protection area between nucleotides −61 and −102 relative to the transcription start sites (Fig. 4D and E), and mutation of this protected 42 nt sequence resulted in significantly

decreased binding activity of YesN (Fig. S6). Taken together, these data indicate that YesN directly binds to a 42 nt region located upstream of the −35 element and might help recruit RNA polymerase to the promoter (Browning and Busby, 2004; Lee et al., 2012), thereby positively regulating xylFGH transcription and, accordingly, increasing D-xylose transport. LytS/YesN could not directly sense D-xylose signal TCSs often play a major role in bacterial response to extracellular signals. Since we observed that this gene cluster was essential for D-xylose utilization and the TCS LytS/YesN played a transcriptional regulator role for xylFGH, the question has arisen as to whether D-xylose is the signal sensed by LytS/YesN. To explore this possibility, we compared the transcriptional differences of xylFGH, which has been proven to be regulated by LytS/ YesN, in the presence of D-xylose or D-glucose. The results showed that there was an over 22-fold increase in © 2014 John Wiley & Sons Ltd, Molecular Microbiology, 95, 576–589


Fig. 4. The TCS LytS/YesN positively and directly regulates the transcription of the xylFGH genes. A. qRT-PCR of D-xylose utilization-related genes in wild-type 8052, 8052lytS and 8052yesN at 16 h in YP2 medium containing 3 g l−1 D-xylose. Gene expression is represented as fold differences normalized to wild type. The error bars indicate the standard deviation of three independent experiments. B. EMSAs of YesN protein binding to the xylFGH promoter region. The assay was performed using the indicated amounts of purified YesN protein (nM) and Cy5 fluorescence-labeled xylFGH promoter probe (2 nM). EMSAs in the presence of 300-fold unlabeled, specific probe (S) or nonspecific (NS) competitor DNA (sperm DNA) were performed as controls. C. Determination of the transcription start sites of xylFGH using high-resolution S1 mapping. The transcription start sites are indicated by bent arrows with asterisks. D. DNase I footprinting analysis of YesN binding to the xylFGH promoter region. The probe was incubated with increasing amounts of YesN (0, 50, 100, 200, 300, 600 nM) from lanes 1 to 6 respectively. The solid line on the right (labeled −61 to −102) is the segment that was protected from DNase I by YesN. E. The genetic organization of the xylFII-lytS-yesN-xylFGH locus. The −10 and −35 elements of the xylFGH promoter are boxed, and the transcription start sites are indicated by a bent arrow. The −10 region, −35 region, ribosome-binding sites and identified YesN binding sites are underlined.

xylFGH transcription when D-xylose was added into the medium; in the contrary, only slightly down-regulated transcriptional level of xylFGH was observed after the addition of D-glucose (Fig. 5A). Next, β-galactosidase reporter system was used to examine expressional alternation of xylFGH in the presence of three types of monosaccharides (i.e. D-xylose, D-ribose and D-glucose), and again, the result revealed that D-xylose is the only carbon source that strongly induced xylFGH expression (Fig. 5B). According to these findings, in combination with the primary signal sensing and regulatory roles of TCSs in bacteria, we considered the possibility that D-xylose might be the signal to which LytS responds, followed by the initiation of the TCS LytS/YesN-mediated regulation on xylFGH expression. However, bioinformatics analysis showed that LytS here lacks a sensing domain although it exhibits the classic features of a bacterial HK (Fig. S4) using the SMART Web tool (Letunic et al., 2012). In addition, a detailed analysis using the TMHMM 2.0 program and the subcellular localization analysis revealed that LytS only © 2014 John Wiley & Sons Ltd, Molecular Microbiology, 95, 576–589

harbors one transmembrane domain (Figs 6A and S7, amino acid position 135–157) (Krogh et al., 2001) in contrast to all previously identified histidine sensor kinases, which normally have two or more transmembrane domains (Mascher et al., 2006; Cheung and Hendrickson, 2010). Therefore, LytS seems unlikely to be a D-xylose sensor. To clarify this question, we tested interaction between the periplasmic domain of LytS (pLytS) and D-xylose, using isothermal titration calorimetry (ITC). The results showed that no direct interaction was detected between pLytS and D-xylose (Fig. S8), which thus did not support the previous hypothesis that LytS functions as a D-xylose sensor. One possible explanation is that the TCS LytS/YesN here does not directly respond to D-xylose signal and there could be other sensors. XylFII acts as a novel D-xylose sensor Since it is most likely that the HK LytS is not a D-xylose sensor, more attention was paid to xylFII, which is adjacent to and co-transcribed with lytS-yesN (Fig. 2A). Gene


Fig. 5. D-Xylose is the signal for the activation of the TCS LytS/YesN. A. qRT-PCR results for the xylFGH genes in SP2 medium or with addition of D-glucose or D-xylose (7.5 g l−1) at 12 h. The error bars indicate the standard deviation of three independent experiments. B. pIMP1-lacZxylF reporter expression in 8052 wild type in the presence of 15 g l−1 D-ribose, D-glucose or D-xylose. The mean of two or three independent biological replicates and the standard deviation for the specific β-galactosidase activities are shown (***P ≤ 0.001; **P ≤ 0.01; NS, P > 0.05; t-test).

Fig. 6. Direct interaction between the periplasmic domain of XylFII (pXylFII) and the LytS periplasmic domain (pLytS). A. Subcellular localization analysis of LytS, XylFII and XylF. Flag-LytS, LytS with a Flag-tag fused to the N-terminus; XylFII-HA, XylFII with an HA-tag fused to the C-terminus of XylFII; XylF-HA, XylF with an HA-tag was fused to the C-terminus. Flag-tLytS, truncated Flag-LytS with the predicted transmembrane domain deletion; tXylFII-HA, truncated XylFII-HA with the predicted transmembrane domain deletion. The white boxes in the left represent the whole proteins, in which the grey parts represent the predicted transmembrane domains. T, total cell lysate; S, soluble fraction; M, membrane fraction. B. Escherichia coli two-hybrid assay reveals an interaction between pXylFII and pLytS. The recombinant strains expressing chimeric proteins, as indicated in the table, were separately streaked on nonselective and dual-selective media (3-amino-1,2,4-triazole + streptomycin). The strain expressing chimeric LGF2 and Gal11P was used as a positive control. C. Pull-down experiments probing the interaction between pXylFII-HA and Flag-pLytS. The anti-Flag antibody was cross-linked to protein A+G agarose beads to form an affinity matrix, which is capable of binding to Flag-pLytS. The proteins interacted with Flag-pLytS could be pulled down. The ‘protein control’ meant that the purified proteins were directly used in Western blotting, which was just a preliminary experiment to confirm their positions on the gel. The ‘output’ showed the pull-down experiment results. L, Flag-pLytS; FII, pXylFII-HA; F, XylF-HA. Lane 1, Flag-pLytS; lane 2, pXylFII-HA; lane 3, XylF-HA; lane 4, Flag-pLytS, negative control; lane 5, pXylFII-HA, negative control; lane 6, Flag-pLytS+pXylFII-HA; and lane 7, Flag-pLytS+XylF-HA. © 2014 John Wiley & Sons Ltd, Molecular Microbiology, 95, 576–589


xylFII has been shown not to play an ABC transporter substrate-binding protein role, and moreover, similar growth and D-xylose utilization patterns were observed in the xylFII, lytS and yesN mutants (Fig. 2D and E). Bioinformatics analysis using TMHMM 2.0 provided the first indication that there might be a transmembrane domain (amino acid position 7–24) in XylFII (Fig. S7). The following subcellular localization experiment also confirmed that XylFII did predominantly localize to plasma membrane (Fig. 6A), whereas a truncated XylFII with the transmembrane domain deletion could not (Fig. 6A). This demonstrated that XylFII is a membrane-localized protein. In the contrary, such a transmembrane domain was not found in XylF (Fig. S7). In addition, LytS has also been proven to be a membrane-localized HK here (Fig. 6A). Thus, we considered the possibility that XylFII is a sensor that assists the response of LytS to the D-xylose signal, thereby enabling the activation of LytS/YesN. To explore this possibility, we used a bacterial two-hybrid system to determine if the periplasmic domain of XylFII (pXylFII) interacts with the periplasmic domain of LytS (pLytS). Encouragingly, we observed that only E. coli strain 2 (bearing pXylFII- and pLytS-expressing plasmids) and strain 6 (the positive control) grew on dual-selective medium, whereas all six E. coli strains were capable of growing on nonselective medium (Fig. 6B), indicating an interaction between these two proteins. Subsequently, pull-down experiments were performed to further confirm this finding. Purified Flag-pLytS (pLytS with a Flag tag at its N-terminus), pXylFII-HA (pXylFII with an HA tag at its C-terminus) and XylF-HA (XylF with an HA tag at its C-terminus, negative control) were used for these experiments. As expected, only pXylFII-HA, rather than XylF-HA, was remarkably pulled down by Flag-pLytS (lane 6 and 7) (Fig. 6C), consistent with the bacterial two-hybrid results and indicating that XylFII does interact with LytS directly. To further determine whether XylFII acts as a D-xylose sensor, we examined the interaction between XylFII and D-xylose by ITC analysis. We observed that pXylFII bound strongly to D-xylose molecules (Kd = 3.25 μM) but not to L-arabinose, which is an isomer of D-xylose (Fig. 7A and B). These findings confirm that XylFII responds to the D-xylose signal. We subsequently asked whether XylFII and TCS LytS/ YesN do comprise a ‘three-component system’ that regulates xylFGH adaptation to extracellular D-xylose signal. To explore this possibility, we compared the changes in the transcription and expression of xylFGH after inactivation of xylFII, yesN and xylF. qRT-PCR analysis revealed that transcription of xylFGH was decreased 3.6- and 6.7fold in the xylFII and yesN mutants, respectively, in comparison with the control (wild-type strain 8052) (Fig. 7C). As another control, xylF inactivation caused slightly increased transcription of xylFGH compared with the wild© 2014 John Wiley & Sons Ltd, Molecular Microbiology, 95, 576–589

type strain 8052 (Fig. 7C). Then β-galactosidase gene (lacZ) was used to assay the expressional changes of xylFGH. The promoter of xylF was used to drive the expression of lacZ in the wild-type strain (8052) and three mutant strains (8052xylF, 8052yesN and 8052xylFII), yielding the strains 8052-lacZ, 8052xylF-lacZ, 8052xylFIIlacZ and 8052yesN-lacZ respectively. Both 8052xylFIIlacZ and 8052yesN-lacZ showed significantly decreased β-galactosidase activity against the 8052-lacZ at two time points (i.e. 6 h and 12 h; Fig. 7D), whereas there was no significant β-galactosidase activity difference between 8052-lacZ and 8052xylF-lacZ, which were consistent with that from the qRT-PCR experiment. Moreover, as just described, the 8052xylFII was less impaired in growth and D-xylose consumption compared with the three ABC transporter mutants, and actually, its phenotypes were comparable with the TCS mutants (Fig. 2D and E). Thus, all these results suggested that XylFII and LytS may be involved in the same signaling pathway.

Discussion As an attractive sugar ingredient for industrial-level fermentation, the utilization of D-xylose by bacteria has been studied extensively (Ezeji et al., 2007; Jeffries et al., 2007). However, little is known about signal sensing and downstream regulation of D-xylose uptake in bacteria. In light of the data in this work, we propose a novel ‘three-component system’ that integrates D-xylose sensing and uptake regulation in C. beijerinckii (Fig. 8). Importantly, this ‘threecomponent system’ and the derivative ‘six-protein module’ (XylFII-LytS/YesN-XylFGH) are widely distributed in Firmicutes, thus indicating that they may play a broad role. This module begins with the recognition and direct binding of extracellular D-xylose molecule by XylFII, which was predicted but was proven not to be a substrate-binding subunit of D-xylose ABC transporter. Subsequently, the interaction between the extracellular domain of XylFII and the HK LytS results in LytS activation, possibly through autophosphorylation, and in turn activating the RR YesN. Activated YesN binds to the promoter region of xylFGH to initiate gene expression, thus leading to enhanced D-xylose uptake in C. beijerinckii. TCSs, the predominant signal transduction system in bacteria, typically comprise a transmembrane sensor HK, which can sense specific stimuli, and a cognate RR, which is a DNA-binding protein; together, these proteins regulate gene expression to elicit bacterial responses (Stock et al., 2000). In most cases, the HK directly senses the environmental signals, undergoes autophosphorylation and transfers the phosphoryl group to the RR (Mascher et al., 2006). Moreover, some auxiliary proteins are also employed by TCSs to assist signal sensing, thus enabling TCSs to accomplish its regulatory task (Tetsch and Jung, 2009;


Fig. 7. XylFII assists D-xylose signal sensing by the HK LytS. A and B. ITC studies of D-xylose (A) and L-arabinose (B) binding to pXylFII. The upper panel shows the calorimetric titration of the binding protein with ligand, and the lower panel displays the corresponding integrated heat, which was normalized and corrected for the heat of dilution versus the molar ratio. C. qRT-PCR results for the xylFGH genes in wild-type 8052, 8052xylFII, 8052yesN and 8052xylF at an OD600 of 0.8–0.9 in YP2 medium containing 3 g l−1 D-xylose. Gene expression is represented as fold differences normalized to wild type. The error bars indicate the standard deviation of three independent experiments. D. pIMP1-lacZxylF reporter expression in wild type, 8052xylFII, 8052yesN and 8052xylF. The mean of three independent biological replicates and the standard deviation for the specific β-galactosidase activities are shown (***P ≤ 0.001; **P ≤ 0.01; *P ≤ 0.05; NS, P > 0.05; t-test).

Buelow and Raivio, 2010). For example, the tripartite tricarboxylate transporter BctCBA was involved in citrate signaling by TCS BctDE in Bordetella pertussis (Antoine et al., 2005); the fumarate/succinate antiporter DcuB and glucose 6-phosphate/inorganic phosphate antiporter UhpC might form a sensor complex with the HKs DcuS and UhpB, respectively, in E. coli (Island and Kadner, 1993; Schwoppe et al., 2002; Kleefeld et al., 2009); both B. subtilis BceRS-BceAB and Staphylococcus aureus BraSR-BraDE systems constituted bacitracin-sensing and detoxification modules (Rietkotter et al., 2008; Hiron et al., 2011). In these examples, the auxiliary proteins, without

exception, play dual roles: substrates sensing and transporting. In addition to above examples, some proteins adhering to the surface of outer membrane or free proteins in periplasmic or intracellular space can also assist HKs for sensing signals in bacteria, such as the outer membrane lipoprotein NlpE and the periplasmic protein TorT in E. coli (Hirano et al., 2007; Moore and Hendrickson, 2012), and the periplasmic protein LuxP in Vibrio harveyi (Neiditch et al., 2006). However, XylFII here is distinct from all these described signal sensors or co-sensors. First, as a signalsensing protein, XylFII is not involved in D-xylose transport; second, XylFII harbors only one transmembrane domain © 2014 John Wiley & Sons Ltd, Molecular Microbiology, 95, 576–589


Fig. 8. Schematic of the ‘three-component system’ for D-xylose sensing and transport in Firmicutes. The left and right sections represent the case in which XylFII is in the soluble fraction and inserted in the membrane respectively. The ‘three-component system’ is encoded by three tandem and closely arranged genes. The predicted xylose transporter substrate-binding protein XylFII is a single transmembrane protein and does not participate in xylose transport but interacts directly with the single transmembrane histidine kinase (HK) to sense D-xylose signal for the TCS LytS/YesN. Receipt of the D-xylose signal stimulates activation of the TCS LytS/YesN, and YesN binds to the xylFGH promoter region, resulting in activation of xylFGH transcription and increased D-xylose transport activity. His_kinase, HATPase_c, REC and HTH_Arac represent the domains of LytS and YesN.

and thus is an intrinsic membrane protein. Therefore, XylFII in this system seems to be a novel type of signal-sensing protein employed by a TCS. Whether such a XylFII-like specific sensor is more efficient than transporter-like co-sensors in responding to environmental stimuli requires further study. Another feature of this model is that LytS has only one transmembrane domain, which is distinct from typical periplasmic-sensing HKs that normally contain two or more transmembrane domains. In addition, typical extracellularsensing HKs often harbor a periplasmic input domain flanked by two transmembrane domains (Mascher et al., 2006; Cheung and Hendrickson, 2010). A recent study in B. subtilis also showed that an HK DctS harboring a periplasmic sensor domain (PASp) could form a complex with both a transporter DctA and a single transmembrane cosensor DctB to sense C4-dicarboxylate (Graf et al., 2014). However, according to our bioinformatics analysis and subcellular localization experiment (Figs S7 and 6A), LytS has only one transmembrane domain and thus lacks such a sensor domain, which may explain its defect in D-xylose sensing (Fig. S8). To date, only a few HKs have been shown to harbor one transmembrane (Winkler and Hoch, 2008; Wang et al., 2013), and moreover, their signal© 2014 John Wiley & Sons Ltd, Molecular Microbiology, 95, 576–589

sensing mechanisms are not well understood. Here, we demonstrate that the single transmembrane LytS requires the auxiliary protein, XylFII, to respond to extracellular D-xylose molecules. These findings offer a novel signalsensing pattern for LytS-like single transmembrane HKs. Of course, we could not rule out the involvement of other factors in regulating all or part of the xylFII-lytS-yesNxylFGH gene cluster. This cluster is closely linked to other D-xylose degradation pathway genes on the chromosome, including enzyme-encoding genes (xylA, xylB, tal and tkt) and a transcriptional regulator gene xylR. A putative candidate regulatory site of XylR was identified in the promoter region of xylFII (Gu et al., 2010), indicating that the xylFIIlytS-yesN may be also regulated by XylR. Besides, we observed that inactivation of the xylFGH genes had a much greater impact on D-xylose utilization than inactivation of the xylFII or lytS-yesN genes (Fig. 2D and E). This indicates that some unknown regulators, besides LytS/YesN, may also participate in controlling xylFGH genes expression. Taken together, these results suggest that the regulatory mechanism related to this ‘six-protein module’ might be more complicated than predicted. In conclusion, we identified a novel ‘three-component system’ that is responsible for D-xylose signaling and


uptake regulation in C. beijerinckii. In this system, XylFII acts as an auxiliary regulator to assist the TCS in D-xylose sensing. Therefore, a XylFII-based model for this signaling pathway and the subsequent regulation of D-xylose uptake was proposed. The adoption of such a signal-sensing mechanism by a D-xylose-utilizing bacterium has not been previously described. Notably, the ‘three-component system’ homologs are present in approximately 10% of Firmicutes bacteria, and thus, this system may also contribute to D-xylose sensing and uptake in other Grampositive bacteria.

Experimental procedures Bacterial strains and plasmid construction The bacterial strains and plasmids used in this study are listed in Table S3. The xylFGH locus of E. coli K12xylE was disrupted by using the λ-Red disruption system (Datsenko and Wanner, 2000). The xylGH and xylFGH genes were directly PCR-amplified from C. beijerinckii NCIMB 8052 genomic DNA. The hybrid xylFIIGH genes were obtained by PCR overlap extension. Then these genes were cloned into pUC118 and used in heterologous host complementation experiments. Plasmid pQ8-yesN was constructed by amplifying the yesN coding region and cloning into pQ8 (Sun et al., 2013), attaching a His6 tag and MBP tag to the N-terminus of YesN. The DNA fragments encoding pXylFII, pLytS, FlagpLytS, pXylFII-HA or pXylF-HA were amplified and cloned into pET32a (Invitrogen) for protein expression in Rosetta (DE3). The group II intron-based Targetron technology was used to inactivate target genes in C. beijerinckii, and the corresponding primers designed were designed using the ClosTron tool (Shao et al., 2007; Heap et al., 2010). Three intergenic regions, PxylFII (379 nt), PlytS (112 nt) and PxylF (293 nt), were obtained by PCR, digested with Pst1 and BamH1 and cloned into the plasmid pIMP1-lacZ to determine whether they are promoters. Plasmids ptb-lytS, ptb-yesN, ptb-xylFII, ptb-xylFGH and ptb-xylFIIGH were generated by amplifying the sequences of the corresponding genes and cloning them into pIMP1. pXY1 was used for the construction of pXY1-Flag-lytS, pXY1-xylFII-HA and pXY1-xylF-HA to perform subcellular localization analysis.

Medium and growth conditions All E. coli strains were grown at 37°C in Luria–Bertani (LB) medium or M9 minimal medium. C. beijerinckii strains were cultivated at 37°C in Clostridium growth medium (CGM) anaerobically (Thermo Forma Inc., Waltham, MA) (Wiesenborn et al., 1988). YP2 medium, which was modified from P2 medium (Baer et al., 1987), containing 3 g l−1 D-xylose as the carbon source and 0.05 g l−1 yeast extract, was used for routine fermentation. SP2 medium (Xiao et al., 2012; Zhang et al., 2012), which contains 7.5 g l−1 glycerol as the carbon source, was also adopted for culturing C. beijerinckii to perform qRT-PCR. E. coli strains were grown in LB medium supplemented, when necessary, with ampicillin (100 μg ml−1), kanamycin (50 μg ml−1), chloramphenicol (25 μg ml−1), tetra-

cycline (12.5 μg ml−1) and streptomycin (12.5 μg ml−1); C. beijerinckii strains were grown with erythromycin (15 μg ml−1) added when necessary.

Analysis of cell growth and xylose concentration Escherichia coli cells cultured in LB medium overnight were harvested by centrifugation at 5000 g for 10 min. Cell pellets were washed twice and resuspended in M9 minimal medium containing 3 g l−1 D-xylose as the carbon source. The cells were diluted to the same optical density (OD600) and then transferred to M9 medium for fermentation. C. beijerinckii strains grown in CGM medium were transferred into YP2 medium for inoculum preparation and then incubated into YP2 medium for fermentation. The D-xylose concentration was determined by the HPLC system (Model 1200, Agilent) with a Sugar-PakTM I column (Waters Corp., MA, USA) and a refractive index detector.

β-galactosidase assays Clostridium beijerinckii strains for β-galactosidase assays were cultivated in YP2 medium containing 15 g l−1 glycerol as the carbon source. When optical density (OD600) reached 0.1, D-ribose, D-glucose or D-xylose was added into the medium separately (15 g l−1) to induce lacZ expression. Similarly, to determine the promoter regions, 8052pIMP1-lacZ, 8052pIMP1-lacZxylFII, 8052pIMP1-lacZlytS and 8052pIMP1lacZxylF strains were cultivated in YP2 medium using 3 g l−1 D-xylose as the carbon source for 20 h. The cell pellets were harvested by centrifugation, resuspended in Z Buffer and lysed by sonication. Then the crude extracts were heat-treated at 60°C for 30 min and the precipitates were removed by centrifugation at 12 000 g for 30 min. The supernatants were used for β-galactosidase activity assay as previously reported (Tummala et al., 1999).

High-resolution S1 nuclease mapping High-resolution S1 nuclease mapping was performed as described previously (Wang et al., 2013). Briefly, the reverse primer xylFGHs1a (5′-TCCTGTCAATGCTGAAAAT-3′) was labeled with [γ-32P] ATP and T4 polynucleotide kinase (NEB) prior to PCR. Then, the xylFGH specific probe DNA was amplified by PCR from the genomic DNA of C. beijerinckii NCIMB 8052 using 32P-5′end-labeled reverse primer xylFGHs1a and the nonlabeled forward primer xylFs1s (5′CTAACAACATATCACAACGAT-3′). A total of 40 μg of RNA was hybridized to the DNA probe at 45°C for 15 h and then treated with S1 nuclease (Promega) at 37°C for 1 h. The sequencing ladder was generated with the fmolTM DNA cycle sequencing kit (Promega) and the labeled reverse primer. The reaction products were separated on 6% (w/v) polyacrylamide-urea sequencing gel and then visualized by autoradiography.

Subcellular localization analysis Clostridium beijerinckii cells (mid-log phase) were collected by centrifugation at 5000 g and 4°C for 10 min. Cell pellets © 2014 John Wiley & Sons Ltd, Molecular Microbiology, 95, 576–589


were resuspended in phosphate-buffered saline (PBS) and passed through French Press (Constant Systems Limited, UK) to prepare lysed cells. The unbroken cells and large cellular debris were removed by centrifugation at 13 000 g and 4°C for 10 min. Then the supernatant was centrifuged (100 000 g, 4°C, 1 h) again to pellet the membrane fractions. The resulting supernatant was retained as the soluble fraction. The pelleted membrane fraction was washed by PBS and centrifuged at 100 000 g and 4°C for 1 h. The supernatant was discarded, and the pellet was resuspended in PBS as the membrane fraction.

RT-PCR and quantitative real-time RT-PCR RT-PCR for co-transcription analysis and qRT-PCR were performed using primers listed in Table S4. Cells for RT-PCR and qRT-PCR were grown in YP2 medium containing 3 g l−1 D-xylose and SP2 medium with addition of glycerol, D-glucose or D-xylose (7.5 g l−1). Samples were harvested at the same time or optical density (OD600) as indicated by centrifugation at 4°C. The harvested samples were immediately frozen in liquid nitrogen and ground into powder. RNA was isolated by TrizolTM (Invitrogen) extraction according to the manufacturer’s instructions. Contaminant DNA was removed by treatment with DNase I (Takara), and cDNA was synthesized by reverse transcription for RT-PCR and qRT-PCR. DNase-treated RNA samples were used as the negative controls for RT-PCR. qRT-PCR was conducted in MyiQ2 two-color real-time PCR detection system (Bio-Rad) using iQTM SYBR Green Supermix (Bio-Rad). The 16s rRNA gene was employed as the internal control.

Protein overexpression and purification The Rosetta (DE3) strains harboring the overexpression plasmids were grown in LB medium to optical density (OD600) of 0.6–0.8. Then isopropyl β-D-1-thiogalactopyranoside was added into the medium to induce the expression of proteins. The cells were harvested by centrifugation and suspended in binding buffer (50 mM Tris-HCl, pH 8.0, 300 mM NaCl, 10 mM imidazole, 10% glycerol). Then the cells were disrupted by French Press (Constant Systems Limited, UK) and centrifuged at 4°C to remove the intact cells and debris. The resulting supernatants were loaded on a Ni-NTA agarose column (GE healthcare, Sweden). After washing the column by buffer A (50 mM Tris-HCl, pH 8.0, 300 mM NaCl, 60 mM imidazole, 10% glycerol), proteins bound to the beads were eluted with buffer B (50 mM Tris-HCl, pH 8.0, 300 mM NaCl, 500 mM imidazole, 10% glycerol). The purified His-tagged proteins were checked by SDS-PAGE.

EMSAs All unlabeled primers used for EMSA contained a universal sequence (5′-AGCCAGTGGCGATAAG-3′) at the 5′-end. The DNA fragments of the promoter regions were generated by PCR amplification from the genomic DNA of C. beijerinckii NCIMB 8052. Then, the DNA fragments were labeled with Cy5 by PCR using the Cy5-labeled primer 5′-AGCCAGTGGC GATAAG-3′. The labeled probes (2 nM) were incubated with © 2014 John Wiley & Sons Ltd, Molecular Microbiology, 95, 576–589

indicated amount of purified proteins in the binding buffer containing 20 mM Tris (pH 7.9), 1 mM dithiothreitol (DTT), 40 mM KCl, 10 mM MgCl2, 0.04 mg ml−1 calf BSA, 5% glycerol and 200 μg ml−1 sonicated salmon sperm DNA (Sangon). After incubation at 30°C for 30 min, the DNA–protein complexes were loaded onto a 1.5% agarose gel, electrophoresed in 0.5 × TAE buffer at 4°C and detected by a FLA-9000 phosphorimager (Fujiflim).

DNase I footprinting assays DNase I footprinting assays were carried out as described before (Wang et al., 2013). Briefly, the forward primer xylFGHfoots (5′-TAACAACATATCACAACGAT-3′) was first labeled with [γ-32P] ATP and T4 polynucleotide kinase (NEB) at the 5′-end. The 32P-labeled forward primer and the unlabeled reverse primer xylFGHfoota (5′-TCTTATTTCCATTG CTAGA-3′) were used to amplify the DPxylFGH probe by PCR from genomic DNA. The labeled DPxylFGH probe (90 000 cpm) was then incubated with indicated amount of purified YesN protein in 50 μl of binding buffer containing 20 mM Tris-HCl (pH 7.9), 1 mM DTT, 10 mM MgCl2, 0.5 mg ml−1 calf BSA and 5% (v/v) glycerol. After binding at 30°C for 30 min, 5.5 μl of RNase-free DNase I buffer and 0.33 U DNase I (Promega) were added to digest the DNA probe for 75 s at 28°C. The reaction was terminated by adding 50 μl of stop solution (20 mM EGTA, pH 8.0), followed by phenol extraction and ethanol precipitation. At last, the samples were run on a 6% polyacrylamide urea gel and analyzed by autoradiography.

ITC ITC experiments were carried out by using a Microcal iTC200 microcalorimeter (GE Healthcare). Briefly, the protein solutions were dialyzed into a buffer containing 50 mM Tris-HCl, pH 8.0, 50 mM KCl, 0.5 mM EDTA and concentrated to 120 μM (pXylFII) and 75 μM (pLytS). The ligand D-xylose was dissolved in the same buffer to the concentration of 360 μM. pXylFII or pLytS was added into the sample cell and then titrated with D-xylose in the syringe. L-Arabinose was used to titrate pXylFII as a negative control. The resulting titration curves were fitted with MicroCal ORIGIN software. Calorimetric data analysis was carried out with ORIGIN 7.0 software (MicroCal). Binding parameters were determined by fitting the experimental binding isotherms.

Bacterial two-hybrid assays The bacterial two-hybrid assays were accomplished using the BacterioMatch II Two-Hybrid System (Stratagene). The DNA sequence coding for the periplasmic domain of LytS was cloned into the bait vector pBT in order to create a fusion protein with λ repressor protein (λcI). The DNA fragments corresponding to pXylFII or XylF were separately inserted into the vector pTRG in frame with the α-subunit of RNA polymerase. The pBT-derived plasmid and pTRG-derived plasmids were then cotransformed into the validation reporter strain. Then, the resulting strains, grown at 30°C in LB medium overnight, were harvested by centrifugation (5000 g, 2 min) and washed twice with M9+ His-dropout broth. The cells were


diluted to different gradients, spotted on the nonselective screening medium and selective screening medium, and finally incubated at 30°C for 2–4 days. The colonies grown on the selective screening medium were selected and streaked on the dual screening medium for further verification (Kruh et al., 2007). The cotransformant containing plasmids both pBT-LGF2 and pTRG-Gal11P was used as the positive control.

Pull-down experiments The protein A+G agarose beads were equilibrated in washing buffer (PBS, 1% Triton X-100), then mixed with anti-Flag antibody (1:100) and rotated at 4°C for 1 h. The corresponding proteins of the same amount were added to the complex and then incubated with constant rotation at 4°C overnight. After that, the beads containing immunoprecipitates were harvested by centrifugation (2500 g, 3 min, 4°C) and washed for eight times with the same buffer. Finally, the proteins binding to the beads were eluted by SDS protein loading buffer and analyzed by Western blotting.

Acknowledgements We would like to thank Prof. Yinhua Lu at Shanghai Institutes for Biological Sciences (SIBS) for helpful suggestions and comments on the manuscript. This work was supported by the National Natural Science Foundation of China (31100062, 31121001), National Basic Research Program of China (2011CBA00800), National High-tech Research and Development Program of China (2011AA02A208) and the State Key Laboratory of Motor Vehicle Biofuel Technology (No. 2013005).

References Ahlem, C., Huisman, W., Neslund, G., and Dahms, A.S. (1982) Purification and properties of a periplasmic D-xylose-binding protein from Escherichia coli K-12. J Biol Chem 257: 2926–2931. Antoine, R., Huvent, I., Chemlal, K., Deray, I., Raze, D., Locht, C., and Jacob-Dubuisson, F. (2005) The periplasmic binding protein of a tripartite tricarboxylate transporter is involved in signal transduction. J Mol Biol 351: 799–809. Aristidou, A., and Penttila, M. (2000) Metabolic engineering applications to renewable resource utilization. Curr Opin Biotechnol 11: 187–198. Baer, S.H., Blaschek, H.P., and Smith, T.L. (1987) Effect of butanol challenge and temperature on lipid composition and membrane fluidity of butanol-tolerant Clostridium acetobutylicum. Appl Environ Microbiol 53: 2854–2861. Browning, D.F., and Busby, S.J. (2004) The regulation of bacterial transcription initiation. Nat Rev Microbiol 2: 57–65. Buelow, D.R., and Raivio, T.L. (2010) Three (and more) component regulatory systems – auxiliary regulators of bacterial histidine kinases. Mol Microbiol 75: 547–566. Cheung, J., and Hendrickson, W.A. (2010) Sensor domains of two-component regulatory systems. Curr Opin Microbiol 13: 116–123. Coumes-Florens, S., Brochier-Armanet, C., Guiseppi, A., Denizot, F., and Foglino, M. (2011) A new highly conserved

antibiotic sensing/resistance pathway in Firmicutes involves an ABC transporter interplaying with a signal transduction system. PLoS ONE 6: e15951. Datsenko, K.A., and Wanner, B.L. (2000) One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc Natl Acad Sci USA 97: 6640–6645. Davidson, A.L., and Chen, J. (2004) ATP-binding cassette transporters in bacteria. Annu Rev Biochem 73: 241–268. Davidson, A.L., and Maloney, P.C. (2007) ABC transporters: how small machines do a big job. Trends Microbiol 15: 448–455. Davis, E.O., and Henderson, P.J. (1987) The cloning and DNA sequence of the gene xylE for xylose-proton symport in Escherichia coli K12. J Biol Chem 262: 13928–13932. Dintner, S., Staron, A., Berchtold, E., Petri, T., Mascher, T., and Gebhard, S. (2011) Coevolution of ABC transporters and two-component regulatory systems as resistance modules against antimicrobial peptides in Firmicutes bacteria. J Bacteriol 193: 3851–3862. Ezeji, T., Qureshi, N., and Blaschek, H.P. (2007) Butanol production from agricultural residues: impact of degradation products on Clostridium beijerinckii growth and butanol fermentation. Biotechnol Bioeng 97: 1460–1469. Graf, S., Schmieden, D., Tschauner, K., Hunke, S., and Unden, G. (2014) The sensor kinase DctS forms a tripartite sensor unit with DctB and DctA for sensing C4dicarboxylates in Bacillus subtilis. J Bacteriol 196: 1084– 1093. Gu, Y., Ding, Y., Ren, C., Sun, Z., Rodionov, D.A., Zhang, W., et al. (2010) Reconstruction of xylose utilization pathway and regulons in Firmicutes. BMC Genomics 11: 255. Heap, J.T., Kuehne, S.A., Ehsaan, M., Cartman, S.T., Cooksley, C.M., Scott, J.C., and Minton, N.P. (2010) The ClosTron: mutagenesis in Clostridium refined and streamlined. J Microbiol Methods 80: 49–55. Hirano, Y., Hossain, M.M., Takeda, K., Tokuda, H., and Miki, K. (2007) Structural studies of the Cpx pathway activator NlpE on the outer membrane of Escherichia coli. Structure 15: 963–976. Hiron, A., Falord, M., Valle, J., Debarbouille, M., and Msadek, T. (2011) Bacitracin and nisin resistance in Staphylococcus aureus: a novel pathway involving the BraS/BraR twocomponent system (SA2417/SA2418) and both the BraD/ BraE and VraD/VraE ABC transporters. Mol Microbiol 81: 602–622. Island, M.D., and Kadner, R.J. (1993) Interplay between the membrane-associated UhpB and UhpC regulatory proteins. J Bacteriol 175: 5028–5034. Jeffries, T.W., Grigoriev, I.V., Grimwood, J., Laplaza, J.M., Aerts, A., Salamov, A., et al. (2007) Genome sequence of the lignocellulose-bioconverting and xylose-fermenting yeast Pichia stipitis. Nat Biotechnol 25: 319–326. Joseph, P., Fichant, G., Quentin, Y., and Denizot, F. (2002) Regulatory relationship of two-component and ABC transport systems and clustering of their genes in the Bacillus/ Clostridium group, suggest a functional link between them. J Mol Microbiol Biotechnol 4: 503–513. Jung, K., Fried, L., Behr, S., and Heermann, R. (2012) Histidine kinases and response regulators in networks. Curr Opin Microbiol 15: 118–124. Kleefeld, A., Ackermann, B., Bauer, J., Kramer, J., and © 2014 John Wiley & Sons Ltd, Molecular Microbiology, 95, 576–589


Unden, G. (2009) The fumarate/succinate antiporter DcuB of Escherichia coli is a bifunctional protein with sites for regulation of DcuS-dependent gene expression. J Biol Chem 284: 265–275. Krogh, A., Larsson, B., Von Heijne, G., and Sonnhammer, E.L. (2001) Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes. J Mol Biol 305: 567–580. Kruh, N.A., Rawat, R., Ruzsicska, B.P., and Tonge, P.J. (2007) Probing mechanisms of resistance to the tuberculosis drug isoniazid: conformational changes caused by inhibition of InhA, the enoyl reductase from Mycobacterium tuberculosis. Protein Sci 16: 1617–1627. Lee, D.J., Minchin, S.D., and Busby, S.J. (2012) Activating transcription in bacteria. Annu Rev Microbiol 66: 125–152. Letunic, I., Doerks, T., and Bork, P. (2012) SMART 7: recent updates to the protein domain annotation resource. Nucleic Acids Res 40: D302–D305. Mascher, T., Helmann, J.D., and Unden, G. (2006) Stimulus perception in bacterial signal-transducing histidine kinases. Microbiol Mol Biol Rev 70: 910–938. Moore, J.O., and Hendrickson, W.A. (2012) An asymmetryto-symmetry switch in signal transmission by the histidine kinase receptor for TMAO. Structure 20: 729–741. Neiditch, M.B., Federle, M.J., Pompeani, A.J., Kelly, R.C., Swem, D.L., Jeffrey, P.D., et al. (2006) Ligand-induced asymmetry in histidine sensor kinase complex regulates quorum sensing. Cell 126: 1095–1108. Ohki, R., Giyanto, Tateno, K., Masuyama, W., Moriya, S., Kobayashi, K., and Ogasawara, N. (2003) The BceRS two-component regulatory system induces expression of the bacitracin transporter, BceAB, in Bacillus subtilis. Mol Microbiol 49: 1135–1144. Rietkotter, E., Hoyer, D., and Mascher, T. (2008) Bacitracin sensing in Bacillus subtilis. Mol Microbiol 68: 768–785. Robert, X., and Gouet, P. (2014) Deciphering key features in protein structures with the new ENDscript server. Nucleic Acids Res 42: W320–W324. Rubin, E.M. (2008) Genomics of cellulosic biofuels. Nature 454: 841–845. Schwoppe, C., Winkler, H.H., and Neuhaus, H.E. (2002) Properties of the glucose-6-phosphate transporter from Chlamydia pneumoniae (HPTcp) and the glucose-6phosphate sensor from Escherichia coli (UhpC). J Bacteriol 184: 2108–2115. Shao, L., Hu, S., Yang, Y., Gu, Y., Chen, J., Jiang, W., and Yang, S. (2007) Targeted gene disruption by use of a group II intron (targetron) vector in Clostridium acetobutylicum. Cell Res 17: 963–965. Shulami, S., Zaide, G., Zolotnitsky, G., Langut, Y., Feld, G., Sonenshein, A.L., and Shoham, Y. (2007) A two-component system regulates the expression of an ABC transporter for xylo-oligosaccharides in Geobacillus stearothermophilus. Appl Environ Microbiol 73: 874–884. Staron, A., Finkeisen, D.E., and Mascher, T. (2011) Peptide

© 2014 John Wiley & Sons Ltd, Molecular Microbiology, 95, 576–589

antibiotic sensing and detoxification modules of Bacillus subtilis. Antimicrob Agents Chemother 55: 515–525. Stock, A.M., Robinson, V.L., and Goudreau, P.N. (2000) Twocomponent signal transduction. Annu Rev Biochem 69: 183–215. Sun, P., Zhao, Q., Yu, F., Zhang, H., Wu, Z., Wang, Y., et al. (2013) Spiroketal formation and modification in avermectin biosynthesis involves a dual activity of AveC. J Am Chem Soc 135: 1540–1548. Tetsch, L., and Jung, K. (2009) The regulatory interplay between membrane-integrated sensors and transport proteins in bacteria. Mol Microbiol 73: 982–991. Tummala, S.B., Welker, N.E., and Papoutsakis, E.T. (1999) Development and characterization of a gene expression reporter system for Clostridium acetobutylicum ATCC 824. Appl Environ Microbiol 65: 3793–3799. Van der Heide, T., and Poolman, B. (2002) ABC transporters: one, two or four extracytoplasmic substrate-binding sites? EMBO Rep 3: 938–943. Wang, R., Mast, Y., Wang, J., Zhang, W., Zhao, G., Wohlleben, W., et al. (2013) Identification of two-component system AfsQ1/Q2 regulon and its cross-regulation with GlnR in Streptomyces coelicolor. Mol Microbiol 87: 30–48. Wiesenborn, D.P., Rudolph, F.B., and Papoutsakis, E.T. (1988) Thiolase from Clostridium acetobutylicum ATCC 824 and its role in the synthesis of acids and solvents. Appl Environ Microbiol 54: 2717–2722. Winkler, M.E., and Hoch, J.A. (2008) Essentiality, bypass, and targeting of the YycFG (VicRK) two-component regulatory system in gram-positive bacteria. J Bacteriol 190: 2645–2648. Witan, J., Bauer, J., Wittig, I., Steinmetz, P.A., Erker, W., and Unden, G. (2012) Interaction of the Escherichia coli transporter DctA with the sensor kinase DcuS: presence of functional DctA/DcuS sensor units. Mol Microbiol 85: 846– 861. Wuichet, K., Cantwell, B.J., and Zhulin, I.B. (2010) Evolution and phyletic distribution of two-component signal transduction systems. Curr Opin Microbiol 13: 219–225. Xiao, H., Li, Z., Jiang, Y., Yang, Y., Jiang, W., Gu, Y., and Yang, S. (2012) Metabolic engineering of D-xylose pathway in Clostridium beijerinckii to optimize solvent production from xylose mother liquid. Metab Eng 14: 569–578. Zhang, L., Leyn, S.A., Gu, Y., Jiang, W., Rodionov, D.A., and Yang, C. (2012) Ribulokinase and transcriptional regulation of arabinose metabolism in Clostridium acetobutylicum. J Bacteriol 194: 1055–1064. Zhao, J., and Binns, A.N. (2011) Characterization of the mmsAB-araD1 (gguABC) genes of Agrobacterium tumefaciens. J Bacteriol 193: 6586–6596.

Supporting information Additional supporting information may be found in the online version of this article at the publisher’s web-site.