Multiple Regulatory Mechanisms Control the Expression of the Geobacillus stearothermophilus Gene for Extracellular Xylanase*

Background: Geobacillus stearothermophilus secretes an extracellular xylanase (Xyn10A, XT-6) for the utilization of xylan. Results: The expression of xynA is regulated by XylR, CodY, and XynX, catabolite repression (CcpA), and quorum sensing. Conclusion: The expression of the extracellular xylanase gene (xynA) is highly regulated by many transcriptional schemes. Significance: The xynA gene is regulated by unique and not yet fully characterized mechanisms. Geobacillus stearothermophilus T-6 produces a single extracellular xylanase (Xyn10A) capable of producing short, decorated xylo-oligosaccharides from the naturally branched polysaccharide, xylan. Gel retardation assays indicated that the master negative regulator, XylR, binds specifically to xylR operators in the promoters of xylose and xylan-utilization genes. This binding is efficiently prevented in vitro by xylose, the most likely molecular inducer. Expression of the extracellular xylanase is repressed in medium containing either glucose or casamino acids, suggesting that carbon catabolite repression plays a role in regulating xynA. The global transcriptional regulator CodY was shown to bind specifically to the xynA promoter region in vitro, suggesting that CodY is a repressor of xynA. The xynA gene is located next to an uncharacterized gene, xynX, that has similarity to the NIF3 (Ngg1p interacting factor 3)-like protein family. XynX binds specifically to a 72-bp fragment in the promoter region of xynA, and the expression of xynA in a xynX null mutant appeared to be higher, indicating that XynX regulates xynA. The specific activity of the extracellular xylanase increases over 50-fold during early exponential growth, suggesting cell density regulation (quorum sensing). Addition of conditioned medium to fresh and low cell density cultures resulted in high expression of xynA, indicating that a diffusible extracellular xynA density factor is present in the medium. The xynA density factor is heat-stable, sensitive to proteases, and was partially purified using reverse phase liquid chromatography. Taken together, these results suggest that xynA is regulated by quorum-sensing at low cell densities.

Plant biomass degradation is a pivotal step in the carbon cycle on earth and is mediated mostly by microorganisms, which are found either free in the environment or part of the digestive track of higher animals (1)(2)(3)(4)(5). Plant-based biomass is considered a readily available renewable energy source that can be turned into liquid biofuel for transportation without contributing much net CO 2 to the atmosphere (6,7). Thus, there is a great incentive to explore the enzymology of biomass degradation (8,9). The main sugar-based components of the plant cell wall are cellulose and hemicellulose. Hemicellulose is a heterogeneous group of branched and linear polysaccharides that can include up to 35% of the total plant cell wall dry mass, and one of its major constitutes is xylan (10). Xylan is a polysaccharide made of a backbone of ␤-1,4-xylose units that are decorated by side chains such as arabinose, methylglucuronic acid, and acetate groups (10). The complete degradation of xylan requires the action of specific glycoside hydrolases, including xylanases, which hydrolyze the main chain backbone, and accessory enzymes that act on the side chains, such as ␣-Larabinofuranosidase, ␣-glucuronidase, ␤-xylosidase, and xylan acetyl esterase (11).
Utilization of high molecular weight polymers imposes several intrinsic challenges for the microorganism. First, the microorganism needs a mechanism to sense the presence of the extracellular polymer and to induce the expression of appropriate genes, which include hydrolytic enzymes and transport systems. Second, because the chemical structure of the polysaccharides is complex, efficient degradation requires the involvement of several hydrolytic enzymes with different specificities. Finally, because the polysaccharides are of high molecular weight, a minimum mass of enzymes is required to initiate growth based on the polymers, resulting in cell density-dependent growth (12).
Nature evolved a variety of enzymes and different physiological schemes for the microbial degradation of the plant cell wall polysaccharides (1,13). Anaerobic bacteria, such as Clostridia, have evolved unique multienzyme complexes, named cellulosomes, that integrate many cellulolytic and hemicellulolytic enzymes (14 -16). Cellvibrio japonicas, a Gram-negative soil bacterium, secretes around 30 enzymes that mediate the initial degradation of the integral plant cell wall polysaccharides. The resulting oligosaccharides are then hydrolyzed by surfacebound proteins (17). In both cases, the hydrolysis of polysaccharides occurs at the cell-substrate interface; as a result, the sugar products remain close to the cells and are maintained at relatively high concentrations. A different approach is to completely hydrolyze the polysaccharides outside of the cell. For example, aerobic filamentous fungi, such as Trichoderma and Aspergillus, secrete a large variety of free cellulases, hemicellulases, and ligninases that work synergistically to completely degrade the polymers into mono-and disaccharides (18). Consequently, these extracellular soluble products can also be utilized by other surrounding microorganisms. In Geobacillus spp. the initial breakdown of the polymer is executed by a limited number of extracellular endo-type enzymes that degrade the polysaccharide backbone into relatively large oligosaccharides (19 -22). These oligosaccharides enter the cell via dedicated sugar transporters, and their final breakdown is carried out by intracellular enzymes (20,22). In this strategy, much of the extracellular saccharide products are not available to most competing microorganisms, because they lack the appropriate sugar transporters. Recent studies demonstrated that microbial utilization of the plant cell wall biomass involves extracellular sensing systems for polysaccharides. In most cases, residual sugar monomers resulting from the hydrolytic activity of the surrounding microorganisms can induce the corresponding utilization systems. Other microorganisms have developed dedicated systems to sense the high molecular weight polysaccharides in the surrounding environments and to activate the corresponding genes. For example, in the rumen anaerobe Prevotella bryantii B14, the xylanase gene is up-regulated in response to xylan in the environment (23). This regulation is mediated by a multidomain protein, XynR, that contains a transmembrane domain, histidine kinase motif, response regulator sequence, and a DNA binding domain in its C terminus. In the human gut bacterium, Bacteroides thetaiotaomicron, part of the glycan degradation system is regulated by extra-cytoplasmic function factors and anti-factors and by a hybrid two-component system that incorporate all domains found in classical two-component environmental sensors into one polypeptide (24,28). Alternative factors are also involved in the regulation of the cellulosomal genes in Clostridium thermocellum via an external carbohydrate-sensing mechanism (25)(26)(27).
The hallmark of the xylanolytic system in G. stearothermophilus T-6 is the extracellular xylanase, Xyn10A (also named XT-6) (29,30,44,45). In this study, we demonstrate that the regulation of the extracellular xylanase (xynA) gene is mediated by several mechanisms that in part are connected to the xylan utilization strategy of this bacterium, thus allowing it to successfully compete in its natural niche.

MATERIALS AND METHODS
Growth Conditions-Growth media for G. stearothermophilus was modified Basic Salt Medium (mBSM) 2 supplemented with 1% glucose or xylose. Liquid mBSM contained the following per liter: KH 2 PO 4 , 0.4 g; (NH 4 ) 2 SO 4 , 2 g; MOPS-sodium salt, adjusted to pH 7.0, 10 g. After autoclaving, mBSM was supplemented with (per liter) the following: nitrilotriacetic acid, 0.2 g, NaOH, 0.125 g, 1 ml of amino acid stock solution (0.05 g/liter of 19 amino acids except tyrosine), and 1 ml of three different trace elements solutions. Solution I contained per liter the following: MgSO 4 ⅐7H 2 O, 0.145 g. Solution II contained per liter the following: CaCl 2 ⅐2H 2 O, 0.132 g. Solution III contained in g per liter the following: FeSO 4 ⅐7H 2 O, 0.998 g; MnSO 4 ⅐4H 2 O, 0.592 g; ZnCl 2 , 0.42 g; and CuSO 4 ⅐5H 2 O, 0.624 g. The pH of the latter solution was adjusted to 2.0 with sulfuric acid. An additional source for amino acids was 0.3% of acid hydrolysate of casein (casamino acids) (Difco).
DNA and RNA Isolation and Manipulation-G. stearothermophilus T-6 genomic DNA was isolated according to the procedure published by Marmur (46) and as outlined by Johnson (47). Plasmid DNA was purified using the DNA Clean-Up System (Promega). DNA was manipulated by standard procedures (48). Total RNA was isolated with the RNeasy kit (Qiagen GmbH, Hilden, Germany) according to the manufacturer's protocol.
Transcriptional Analyses-Transcriptional analyses were performed on total RNA extracted from exponentially growing cells. Northern blot analysis was conducted following the procedure described by Moran et al. (49). For determination of the xynA transcription start site, primer extension reactions were carried out as described previously (50) with avian myeloblastosis virus reverse transcriptase (Promega), 40 g of total RNA, and the relevant primers listed in Table 1. Rapid amplification of cDNA ends (RACE) technique was used to amplify portion of FIGURE 1. Schematic view of the xylanolytic system from G. stearothermophilus T-6. The utilization of xylan by G. stearothermophilus is initiated by residual levels of xylose or xylo-oligosaccharides present in the environment. These sugars interact with the extracellular domain of the class I histidine kinase sensor protein XynD, which triggers phosphorylation of the response regulator XynC (21). The phosphorylated XynC activates the expression of a dedicated ABC xylo-oligosaccharides transporter, XynEFG, which facilitated the entrance of xylosaccharides into the cell. Inside the cell, xylose serves as the molecular inducer and interacts with the master repressor XylR, which negatively regulates five transcriptional units. The expression of the extracellular xylanase provides the cell with ample amounts of decorated xylo-oligosaccharides that enter the cell via two dedicated ABC sugar transporters, XynEFG for xylooligosaccharides and AguEFG for aldotetrauronic acid. The decorated xylooligomers are hydrolyzed to their corresponding monomers by intracellular side chain-cleaving enzymes, including ␣-glucuronidase (Agu67A) (35,39), two ␣-L-arabinofuranosidases (Abf51A and Abf51B) (36), two xylan acetylesterases (CE4) (31,42), an intracellular xylanase (Xyn10A) (43), and three ␤-xylosidases (Xyn39B, Xyn52B2, and Xyn43B3) (33,34,37,38,41). The extracellular xylanase gene (xynA) is subjected to carbon catabolite repression mediated by the global negative repressor CcpA and most likely by CodY. Furthermore, xynA is also regulated by XynX and cell density by uncharacterized mechanisms. the putative transcript at xynX 5Ј-UTR (Clontech). Briefly, reverse transcription was generated using random hexamers using a reverse transcriptase (RT) that exhibits terminal transferase activity. The cDNA was amplified with universal primers as well as specific primer anti-parallel and complementary to a sequence downstream of the xynX gene (Table 1). Continuous and Batch Cultures-Batch fermentations were conducted in a 10-liter fermenter (BIOSTAT ED, Braun) with 8-liter working volume or in 1-liter baffled flasks containing 100 ml of medium and reciprocally shaking at 250 rpm, 60°C in a dry shaker (Lab-Line). The inoculum for all fermentations was originated from logarithmic cultures. Chemostat cultures were run in a 2-liter vessel fermenter (BIOSTAT MD, Braun) with a working volume of 1.5 liter, operating at agitation of 700 -1200 rpm, air flow of 1.5 liters/min maintaining oxygen saturation over 80%. A pH of 7.0 was maintained by automatic addition of 5 M NaOH, and the growth temperature was 60°C. The dilution rate was 0.2 h Ϫ1 , and the turbidity was controlled by the concentration of the limited nutrient. The concentrations of O 2 and CO 2 in the outlet gas were measured by a gas analyzer URAS10E (Hartman and Braun). Samples were taken at steady state after determining that turbidity remained similar for over three generations.
Expression of XylR and Extraction of His 6 -tagged CodY and XynX-The xylR, codY, and xynX genes were cloned using two PCR primers ( Table 1) that allow in-frame cloning of the gene into the T7 expression vector pET11d (Novagen) (linearized with BamHI and NcoI), resulting in pET11d-xylR, pET11d-codY, and pET11d-xynX. Sequences encoding a His 6 tag peptide were added to the N-terminal primer of xynX and the C-terminal primer of codY. Expression of xylR was carried out by growing cultures of Escherichia coli JM109(DE3)(pLysS) (Promega) carrying pET11d-xylR in 200 ml of Terrific Broth (48) supplemented with chloramphenicol (25 g/ml) and carbenicillin (50 g/ml) in 2-liter shake flasks shaken at 230 rpm at 37°C. To improve the solubility of XylR, growth was carried out at 18°C in the presence of 0.4 mM isopropyl 1-thio-␤-D-galactopyranoside for 48 h. CodY and XynX were produced in E. coli BL21(DE3) in 500 ml of Terrific Broth supplemented with ampicillin (100 g/ml) in 2-liter shake flasks shaken at 230 rpm at 37°C. Induction by 0.4 mM isopropyl 1-thio-␤-D-galactopyranoside was carried out at a cell turbidity of 0.6 units of absorbance at 600 nm for 16 h. Cells expressing the xylR gene were harvested, suspended in 20 ml of solution containing 50 mM Tris-Cl (pH 7.5), 100 mM KCl, 10% glycerol, 1 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, and 1 mM dithiothreitol, and disrupted by a single passage through a French press (Spectronic Instruments, Inc., Rochester, NY). Following centrifugation of the cell extract (14,000 ϫ g for 15 min), the soluble fraction was used for gel retardation assays. The cultures expressing the codY and xynX genes were harvested, resuspended in 30 ml of buffer (20 mM imidazole, 20 mM phosphate buffer, 500 mM NaCl (pH 7.0)), disrupted by two passages through a French press, and centrifuged (14,000 ϫ g for 15 min) to obtain soluble extracts. The His-tagged proteins, CodY and XynX, were isolated using a 1-ml HisTrap column (GE Healthcare) mounted on an AKTA-avant FPLC system (GE Healthcare), according to the manufacturer's instructions. The puri-fied CodY and XynX proteins were dialyzed overnight against 2 liters of buffer containing 50 mM Tris-HCl (pH 7.0) and 100 mM KCl, followed by the addition of EDTA and glycerol to final concentrations of 1 mM and 10%, respectively. The proteins were stored at Ϫ80°C in 100-l aliquots.
Mobility Shift DNA Binding Assays-For binding analysis with XylR, four double-stranded DNA probes (corresponding to the xynA, xynX, xylA, and xylR promoter regions) were each composed of two synthetic complementary oligonucleotides ( Table 1). The double-stranded probes contained two noncomplementary T nucleotides at the 5Ј end for end labeling with Klenow fragment (Fermentas) in the presence of [␣-32 P]dATP. For DNA binding assays with His 6 -CodY, an 82-bp DNA probe (from position Ϫ96 to position Ϫ14 relative to the transcriptional start site of the xynA gene), was generated via PCR using a primer labeled at the 5Ј end with biotin ( Table  1). The PCR product was purified from 1% agarose gel. For His 6 -XynX, five DNA fragments of 212, 110, 123, 70, and 72 bp corresponding, respectively, to positions Ϫ191 to ϩ21, Ϫ191 to Ϫ81, Ϫ102 to ϩ21, Ϫ49 to ϩ21, and Ϫ30 to Ϫ102 with respect to the transcriptional start site of xynA were generated via PCR with biotin 5Ј end-labeled primers listed in Table 2. These PCR products were purified with the Wizard SV Gel PCR Clean-Up system (Promega).
The XylR-binding reaction mixture (30-l total volume) contained 20 l of binding solution (50 mM Tris-Cl (pH 7.5), 100 mM KCl, 10% glycerol, 1 mM EDTA, 2 g of salmon sperm DNA, 0.66 mM dithiothreitol, 33 g of bovine serum albumin, 0.08 ng of labeled probe), and the indicated concentration of crude protein extract. The binding mixture was incubated for 30 min at 45°C, separated on a 6.6% nondenaturing polyacrylamide gel prepared in Tris borate/EDTA buffer (48), and run for 1 h. Gels were dried under vacuum and exposed to a phosphorimager screen before analysis with a Fuji BAS2000 phosphorimager. Binding of XylR to the xylR operator in the presence of various sugars was carried out under the same conditions described above but with 100 ng/l of crude protein extract and 0.12 ng of labeled probe.
The binding reactions for CodY and XynX were carried out under the conditions described above but with 20 fmol of labeled probe. The binding reactions for CodY also included 2 mM GTP and 10 mM each of the three branched-chain amino acids (isoleucine, leucine, and valine). The binding mixture was incubated for 30 min at 45°C and then separated on a 6.6% nondenaturing polyacrylamide gel prepared in Tris borate/ EDTA buffer (48) and run for 1 h (set to 100 V for 8 ϫ 10 ϫ 0.1-cm gel). Binding reactions were transferred to a nylon membrane at 380 mA for 50 min and then was cross-linked with UV light at 0.32 J/cm 2 . For XynX, the detection of biotinlabeled DNA by chemiluminescence was performed using the Phototope Star Detection kit (N7020, New England Biolabs), according to the manufacturer's protocol, and for binding assays with CodY, the DNA migration was detected using Biotin Chromogenic Detection kit (Thermo Scientific).
Protoplast Transformation of G. stearothermophilus T-6 -Protoplast plasmid DNA transformation to G. stearothermophilus T-6 was based on the protocol described by Wu and Welker (51) with the following modifications. Cells were converted to protoplasts by incubating in fresh lysozyme solution at a final concentration of 2.5 g/ml. One microgram of plasmid DNA was gently mixed with 0.2 ml of protoplast suspension, and transformation was induced by adding 0.9 ml of 40% (w/v) PEG 6000 (Merck), filter-sterilized, and freshly prepared. Plating the transformants on regeneration plates was performed in two steps. First, transformants were plated on regeneration plates containing 0.4% agar and 10% lactose, and the colonies were transferred to regeneration plates with 0.4% agar and 4% lactose (instead of 10% lactose). Plates were incubated at 60°C throughout the experiment. Regeneration plates con-tained the following (per liter): tryptone (Difco), 10 g; yeast extract (Difco), 5 g; NaCl, 5 g; and bacto agar (Difco), 4 g. After autoclaving, the medium was supplemented with the following (per liter): lactose, 100 g, nitrilotriacetic acid, 0.2 g, NaOH 0.125 g, and 1 ml of three different of the following trace elements solutions. Solution I contained per liter, MgSO SEPTEMBER 12, 2014 • VOLUME 289 • NUMBER 37

JOURNAL OF BIOLOGICAL CHEMISTRY 25961
Construction of xynX Null Mutant of G. stearothermophilus T-6 -To construct a xynX-null mutant, we designed a deletion using an overlap PCR technique resulting in removing the entire xynX gene (792 bp) and leaving 450 bp of flanking DNA upstream and downstream of the deletion site (52). The PCR product was digested with SacI and KpnI and ligated to pNW35N, which had been similarly digested to obtain pNW36N (Table 1). This plasmid was transformed to G. stearothermophilus T-6 by protoplast transformation and selection for Cm r clones. Cm r colonies were tested by colony PCR for a single crossover event in which pNW36N integrated into the chromosome and the presence of the vector. To allow a second double crossover event and thereby obtain a chloramphenicolsensitive strain, a colony, positive for single integration event, was grown for five generations at 68°C without selection and then plated to obtain single colonies and screened for sensitivity to chloramphenicol. About 8 -10% of the Cm s colonies were xynX null mutants, as confirmed by sizing and sequencing the PCR fragments flanking the xynX region.
Xylanase and Isocitrate Dehydrogenase Activity Assays-Xylanase activity was determined by measuring the hydrolysis of p-nitrophenyl cellobioside (Sigma) or p-nitrophenyl xylobioside (pNPX 2 ) (Megazyme). The extracellular xylanase (Xyn10A) has a high specificity constant toward pNPX 2 (k cat /K m ϭ 8.6 ϫ 10 3 min Ϫ1 mM Ϫ1 ) (53). The reaction mixture contained 160 l of an appropriately diluted enzyme, 160 l of 0.1 M phosphate buffer (pH 7.0), and 80 l of p-nitrophenyl cellobioside (4 mg/ml) in the same buffer. Tubes were incubated at 50°C until a slight yellow color appeared, and then the reaction was terminated by the addition of 160 l of 1 M Na 2 CO 3 . Absorbance at A 420 was determined spectrophotometrically (Ultrospec 2100pro, Amersham Biosciences). One unit of activity was defined as (A 420 /min/ml of culture per A 600 ) ϫ 1000 (Miller units). Isocitrate dehydrogenase activity was assayed as described previously (54). Unit of activity was defined as nanomoles of NADPH per min per cell culture turbidity at A 600 .
RNase Protection Assays-Strain T-6 was grown in mBSM in a 10-liter fermenter (Braun ED). During the logarithmic growth, samples at different cell densities were collected, and total RNA was extracted with the RNeasy kit (Qiagen) following the supplier protocol. RNA from very low density cultures was extracted from at least 200 ml of culture volume. RNase protection assays were performed using the RPAII kit (Ambion Inc., Austin, TX). Radioactive antisense RNA probes were generated according to the Ambion MAXIscript T7/T3 in vitro transcription kit protocol. [␣-32 P]UTP (800 Ci/mmol) was obtained from Amersham Biosciences. The DNA fragment used for the transcription was cloned in pSL301 (Stratagene, La Jolla, CA), and the construct included the region from Ϫ87 bp to ϩ227 relative to the initial ATG site of the xynA gene. The full-length antisense probes were isolated from polyacrylamide gels according to the supplier's instructions. RNase protection assays were performed with a 467-nt 32 P-labeled antisense RNA probe, of which 227 nt were specific to the 5Ј xylanase transcript and 240 nt were specific to chromosomal and vector sequences. The probe (0.5 ng) was hybridized with 4 g of total RNA. The protected RNA was dissolved in 8 l of sample buffer of which 6 l was subjected to electrophoresis in a 5% polyacrylamide gel containing 7 M urea and visualized using a phosphorimager system.
Determination of the Amount of xynA mRNA Molecules per Cell-Total RNA for each of the cultures was determined using the reagent orcinol monohydrate (Sigma) (55). To determine the amount of total RNA per cell, total cell number was estimated microscopically using a Petroff-Hausser counting chamber. The number of transcripts per cell of the xynA mRNA was derived based on the counts/min per g of total RNA, the specific activity of the probe, the molecular weight of the probe, and the average amount of total RNA per cell (1.78 ϫ 10 Ϫ7 g).
To quantify the results and to ensure that the amount of the probe in the hybridization reaction was not a limiting factor, different amounts of isolated RNA were hybridized with the 32 p-labled antisense probe. Following digestion with RNase, the RNase protection assay product was precipitated, and radioactivity was assessed in a scintillation counter. A linear correlation was obtained between the amount of total RNA used in the assay and radioactivity of the protected product.
Preparation of Conditioned Medium-Conditioned medium was prepared by growing strain T-6 in mBSM containing 1% xylose to a density of ϳ0.8 A 600. The cells were removed by centrifugation; the supernatant was passed through a 0.45-m pore size filter (Millipore). The soluble content was concentrated 2-fold using a freeze dryer and then boiled for 10 min (to inactivate residual extracellular xylanase and proteases).
Cell Density Activation Assay in a 96-Well Plate-Xylanase activity was assayed following the hydrolysis of pNPX 2 . A xylose-induced logarithmic culture at 0.1 A 600 turbidity was diluted 200-fold in mBSM supplemented with 0.5% xylose to a final turbidity of ϳ0.001 A 600 . Cells were then transferred to a 96-well plate containing 0.5 mM pNPX 2 and conditioned medium (10 -30 l). The plate (200 l total volume per well) was incubated with shaking at 60°C, and monitored simultaneously for both cell growth (A 600 ) and p-nitrophenol release (A 420 ) in a microplate spectrophotometer (EON, BioTek).
Partial Purification of Xylanase Density Factor (XDF)-Partial purification of the putative XDF was as follows. A T-6 culture was grown on mBSM in a 2.5-liter fermenter (KLF, Bioengineering) to a turbidity of 0.7-0.8 A 600. The culture was centrifuged (7,000 rpm, 5 min); the supernatant was filtered via a 0.4-m pore size filters (Millipore), and the filtrate was loaded on an XAD-2 column (5 ϫ 20 cm) at a flow rate of 2.5 ml min Ϫ1 . After washing the column with 700 ml of 10% ethanol, the remaining content was eluted with 80% ethanol. The eluate was evaporated to 200 ml in a rotary evaporator and then further concentrated in a freeze dryer to ϳ5 ml, resulting in a 500-fold concentration. Aliquots of 0.25 ml of active samples were subjected to reverse phase liquid chromatography using a C18 column (Kromasil, 5-m particle size, pore size 100 Å, 250 mm ϫ 4.6 mm) mounted on an AKTA-Basic FPLC system (Amersham Biosciences). Fractionation was carried out by washing the column with 6% acetonitrile (ACN) in 0.1% aqueous trifluoroacetic acid (TFA), followed by a linear gradient of ACN (6% to 65% in 120 min) in 0.1% aqueous (TFA) at a flow rate of 0.15 ml min Ϫ1 . Eluent absorbance was monitored at 215 nm. Samples (1.5 ml) were lyophilized, re-dissolved in 50 l of 0.1 M MOPS (pH 7.0), and tested for xylanase activation at low density cultures.
Real Time RT-PCR Analysis-Total RNAs from different cell density cultures were extracted with an RNeasy kit (Qiagen) following the supplier's protocol. Extraction of RNA from very low density cultures was done by collecting at least 100 ml of culture volume. The cDNA was obtained with the Verso cDNA kit (Thermo Fisher Scientific) using 1 g of total RNA and random hexamers as primers. To test for DNA contamination, control reactions were carried out in the absence of reverse transcriptase. Real time RT-PCR primers for the xynA, xynA2, xynX, axe2, xynB1, xynB2, and xynB3 and citC (isocitrate dehydrogenase) genes were designed with the aid of the Primers Express 2.0 software (Applied Biosystems) ( Table 1). Relative quantification was performed with the Applied Biosystems 7300 real time PCR system. Each 20-l reaction mixture included template cDNA, 300 nM of each primer, and Power SYBR Green PCR Master Mix (Applied Biosystems). The amplification conditions for all reactions were one cycle at 95°C for 15 min followed by 40 cycles of 95°C for 10 s, 60°C for 20 s, and 72°C for 15 s. Melting curves were analyzed to ensure specificity of primer annealing and lack of primer secondary structure. Data analysis was carried out with the 7300 system software (Applied Biosystems), using citC for normalization.
Nucleotide Sequence Accession Number-The 77,747-bp sequence containing the arabinan, xylan, xylose utilization region, and the XylR repressor from G. stearothermophilus T-6 have been deposited in the GenBank TM under accession numbers DQ868502 and DQ868501.1, respectively.

RESULTS
The efficient utilization of xylan by G. stearothermophilus T-6 is initiated and controlled by the extracellular xylanase Xyn10A (XT-6). Previous studies from our laboratory suggested that the expression of the xylanase gene is regulated at several levels, including induction by xylose and catabolite repression by glucose (56). In this study, we demonstrate that additional regulatory mechanisms control the expression of the xynA gene, and the roles of some of these mechanisms can be explained based on the biomass utilization strategy used by the bacterium.
Transcriptional Analyses of xynA and xynX-axe2-xynB3 Operon-To map the transcriptional start site of xynA, total RNA was extracted from cultures grown in the presence of xylose or glucose and was used for primer extension analysis ( Fig. 2A). The apparent transcriptional start site point of the xynA gene was identified and assigned to a G nucleotide, 90 bases upstream from the ATG initiation codon. The apparent Ϫ35 sequence, TTTACA, resembles the A consensus and is separated by 17 bp from the potential Ϫ10 region (TAAAAT), which differs by one nucleotide from the A consensus sequence, TATAAT (49). An inverted repeat sequence (5Ј-TTAGTATatttaatttacAAACAAA-3Ј) resembling the xylR operator site from Bacillus and other Gram-positive bacteria (57) overlaps the Ϫ10 region (Fig. 2B). In addition, three potential operator sites for CcpA-mediated, catabolite-responsive regulation (catabolite-responsive elements) (58) were identified between position Ϫ130 with respect to the transcriptional start point and the ATG initiation codon (Fig. 2B). The xynA gene lies upstream of and is divergently transcribed with respect to a threegene cluster, xynX-axe2-xynB3, that potentially constitutes a poly- cistronic operon. To determine whether the cluster constitutes a single transcriptional unit, Northern blot analysis was performed using a DNA probe for xynB3. The probe hybridized to RNAs of many different sizes, with a maximum length of about 3 kb, which is close to the calculated size of the xynX-axe2-xynB3 transcript (3.2 kb), suggesting that the three genes are co-transcribed from the xynX promoter (results not shown). Using 5Ј-RACE analysis, the apparent transcriptional start point of the xynX-axe2-xynB3 transcript was identified and assigned to a G nucleotide (Fig.  2B). The apparent Ϫ35 sequence, TTGAAA, differs by one nucleotide from the A consensus sequence and is separated by 17 bp from the potential Ϫ10 region (TATAAT) that matches the A consensus perfectly. It should be noted that there were no detectable transcripts from this region when total RNA was isolated from cultures grown without xylose, indicating that xylose is the molecular inducer of both the xynX-axe2-xynB3 operon and xynA gene (56).
XylR Binds to the Promoters of Xylose and Xylan-utilization Genes in Vitro-The xylR gene from G. stearothermophilus T-6 encodes a 395-amino acid protein and is not part of the xylanolytic gene cluster. The gene was previously cloned and expressed, and we have demonstrated the ability of XylR to bind the promoter region of the two-component, xylo-oligosaccharide-sensing system encoded by xynDC (21). To verify the ability of XylR to bind additional xylan utilization-related promoters, we applied gel retardation analyses (Fig. 3). The mobility of DNA fragments containing the xynA, xynX, xylA, and xylR promoter regions was retarded when the fragments were incubated with a cell extract of E. coli expressing XylR (Fig. 3). No retardation was observed with an unrelated 32-bp fragment containing a 14-bp inverted repeat (the GlcUA operator) (Fig. 3A, lane 6), demonstrating the specificity of XylR. In addition, no DNA-XylR complex was formed when pET11d-containing extract was used (Fig. 3A, lane 7). XylR was also able to bind its own promoter, suggesting that xylR is negatively auto-regulated (Fig. 3D). To test whether xylose is the molecular inducer of xynA, the binding of XylR to the promoter was assayed in the presence of xylose, xylotetraose, and glucose. Binding was completely prevented in the presence of 10 mM xylose; xylotetraose reduced the binding only partially, whereas glucose showed no effect (Fig. 3E). These results are consistent with previous xynA-Northern blot analysis (56) and the primer extension results ( Fig. 2A), indicating that xylose is the molecular inducer of the extracellular xylanase gene. Sequence analysis of the promoter regions of xynD, xynA, xynX, xylA, xylR as well as of the uncharacterized genes xylM and xylE revealed a similar inverted repeat sequence resembling an operator sequence motif (Fig. 3F).
Expression of the Extracellular Xylanase Is Repressed by Glucose and Casamino Acids-The effect of glucose on extracellular xylanase activity was demonstrated on strain T-6 and its constitutive mutant M-7 (30) using continuous cultures (Chemostat System). Continuous culture allows control of the growth rate (via the dilution rate) and maintenance of the limiting nutrient and cell density at steady state conditions. This technique is advantageous (compared with batch cultures)  6 and 7). F, alignment of XylR-binding sites in the xylose and xylan-utilization genes promoter regions.
when examining carbon catabolite repression using different carbon sources, because the growth rate is not affected by the specific carbon source, and the composition of the growth medium remains constant at steady state. Thus, this technique allows the demonstration of catabolite repression independently of the growth rate. G. stearothermophilus T-6 was grown under nitrogen limitation with an excess of xylose at a growth rate of 0.3 h Ϫ1 (doubling time of 140 min). Upon reaching steady state (no change in turbidity), small amounts of either glucose or arabinose were added, and the enzymatic activity of the extracellular xylanase Xyn10A was measured (Fig. 4A). The addition of glucose or arabinose resulted in ϳ70 and 30% decrease in Xyn10A-specific activity, respectively. After several generations, the concentration of the sugars was reduced due to dilution and consumption by the cells, and the specific activity was restored to its initial values. In batch cultures of the wild type strain grown in medium containing xylose and glucose, xylanase activity was repressed by about 50% (Fig. 4B). Interestingly, in the xylanase constitutive mutant M-7 (30), which codes for a defective XylR (59), glucose repression was only about 35%, suggesting that catabolic repression of xylan utilization genes is mediated in part by XylR. The expression of the extracellular xylanase was also repressed by about 3-fold (60%) in defined medium containing casamino acids and highly repressed (87%) in rich medium such as Luria-Bertani (LB) broth (Fig. 4B). In B. subtilis, the expression of many genes is regulated by CodY, a global transcriptional regulator, which is highly active as a repressor in rapidly growing cells in rich medium (60). Indeed, the xynA gene possesses three putative binding sites resembling the CodY-binding consensus motif (AATTTTCWGAAAATT) (61). Two sequences are located between positions Ϫ70 and Ϫ10 with respect to the transcriptional start point (Fig. 2B), and an additional sequence with four mismatches (AACTGTATGAAAAGT) is located inside the xynA gene.
CodY Binds to the xynA Promoter Region-The xynA promoter region contains sequences that resemble the CodY-binding site (Fig. 2B), suggesting that CodY can bind the xynA promoter and regulates its transcription. The codY gene from G. stearothermophilus T-6 codes for a 259-amino acid protein that shares high similarity with CodY proteins from other Grampositive bacteria. The codY gene was fused at its C-terminal sequence to a six-histidine tag, and the gene product was purified from E. coli by nickel-nitrilotriacetic acid chromatography. To test the ability of CodY to bind the putative "codY box" within the xynA promoter, we performed gel mobility shift assays. An 82-bp DNA fragment corresponding to positions Ϫ96 to Ϫ14 with respect to the xynA transcriptional start site was used as a probe for testing CodY-DNA interaction (Fig. 2B). Binding of CodY to the 82-bp xynA promoter fragment, in the presence of 2 mM GTP and 10 mM of mix of the three branchedchain amino acids, was observed at 100 nM of purified His 6 -CodY (Fig. 5A). The binding is specific because no retardation was observed at a concentration of 600 nM CodY with a 79-bp DNA fragment (positions Ϫ172 to Ϫ93), which lacks potential binding sites for CodY (Fig. 5A, lane 6). These results suggest that CodY is a repressor of xynA and may regulate its transcription in response to nutrient excess. Interestingly, at least 21 FIGURE 4. Extracellular xylanase activity is repressed during growth with glucose or in nutrient-rich medium. A, effect of glucose or arabinose on xylanase activity in a Chemostat operating under nitrogen limitation. The dilution rate was 0.3 h Ϫ1 , and the medium contained 0.2 g/liter ammonium sulfate (the limiting nutrient) and 0.4% xylose as sole carbon source. Glucose or arabinose (0.2%) was added directly to the fermenter (indicated by arrows), and the effect on xylanase-specific activity was determined. B, repression of xylanase activity at different nutritional conditions. Cells were grown in minimal medium supplemented with xylose (mBSMX), xylose and glucose (mBSMXG), xylose and casamino acids (mBSMXCA), and on rich medium with xylose (mLBX). When reaching mid-exponential phase (A 600 ranging from 0.7-0.8 to 2 for mLBX media), the specific activity of the extracellular xylanase was measured. Briefly, 1 ml of culture was clarified by centrifugation, and 80 l of the supernatant fluid was mixed with 80 l of 50 mM phosphate buffer and 80 l of chromogenic substrate, p-nitrophenol-cellobioside. The increase in absorbance at 420 nm was measured at 60°C in a plate reader. The specific activity was calculated dividing A 420 min Ϫ1 per cell turbidity (A 600 ). SEPTEMBER 12, 2014 • VOLUME 289 • NUMBER 37

JOURNAL OF BIOLOGICAL CHEMISTRY 25965
hemicellulolytic genes contain putative binding sites for Cody (Fig. 5B).
XynX Binds to the xynA Promoter Region-The xynX gene is located adjacent to xynA in an opposite direction of transcription. Based on sequence homology, the xynX gene product belongs to the Nif3-like protein family (Ngg1p interacting factor 3), which is highly conserved from bacteria to human (62). In yeast, the Nif3 protein is involved in glucose repression of galactose-utilization genes, and in mouse Nif3L1 mediates neural differentiation in the early developmental stages (63). In both cases, the Nif3 proteins cooperate with additional transcriptional factors. In prokaryotes, the biological function of the Nif3 family is enigmatic. Because the xynX gene is induced by xylose and is part of a xylan utilization operon, it is likely that XynX is a transcriptional regulator of xynA. The ability of XynX to bind DNA was tested first by gel mobility shift assays. The tests were conducted using four DNA probes representing different upstream regions of the apparent xynA transcriptional start site (Fig. 6A). XynX binds specifically, albeit weakly, to a 72-bp DNA fragment corresponding to positions Ϫ102 to Ϫ30, suggesting it can function as a transcriptional factor of the xynA gene (Fig. 6B). The 72-bp DNA fragment contained an inverted repeat sequence positioned at Ϫ30 to Ϫ68 with respect to the transcriptional start point (Fig. 6C). Examining the promoter regions of other xylan-utilization genes in G. stearothermophilus T-6 did not reveal sequences resembling this inverted repeat, suggesting that XynX specifically regulates xynA.
Xylanase Activity in xynX Null Mutant-To assess the role of XynX in regulating the xynA gene, we deleted the xynX gene (Fig. 7) and examined the expression of extracellular xylanase in xynX null mutants under different nutritional conditions. At a minimal xylose-containing medium, the level of xynA expression was higher in the mutant compared with the wild type (Fig.  8A), suggesting that XynX either directly or indirectly represses xynA expression. To determine whether the nature of the carbon source influences xynA expression in a xynX null background, we measured the activity of the extracellular xylanase in a xylose-minimal medium supplemented with either glucose or casamino acids. Both wild type and the xynX null mutant cells grown in xylose-containing medium supplemented with glucose exhibited a 2-3-fold decrease in extracellular xylanase activity compared with xylose-grown cells (Fig. 8B). Similar results were obtained when cultures were grown in xylose-containing medium supplemented with casamino acids (data not shown). Therefore, it appears that XynX is not mediating directly the repression of xynA caused by glucose or casamino acids.
Expression of the Extracellular Xylanase Gene xynA Is Controlled by Cell Density-In early studies on the expression of xynA in strain T-6, we suggested that the expression of xynA occurs toward and during the stationary phase (56), similar to extracellular enzymes in Bacillus subtilis (64). In these early experiments, growth was carried out in shake flasks, and xylanase activity was determined by a reducing sugar assay using  1-5). Lane 6 contained a 79-bp DNA fragment corresponding to positions Ϫ172 to Ϫ93 with respect to the xynA transcriptional start site, which lacks potential binding sites for CodY. The shifted bands are indicated by arrows: B, CodY-DNA complex; a, unbound DNA. B, motif logo for the putative CodY-binding site identified in hemicellulolytic genes from G. stearothermophilus T-6. Searching for sequences containing the CodY-binding motif (AATTTTCWGAAAATT) was done using the program Gene Runner version 3.05 allowing five mismatches. The logo was generated by the MEME function of the Galaxy site (96). the dinitrosalicylic acid reagent (29). However, when these experiments were modified and growth was carried out in a well aerated 10-liter fermenter that allows maintaining logarithmic growth for more generations, together with improving the sensitivity of the enzymatic assay using the chromogenic substrate pNPX 2 (53), the specific activity of Xyn10A appeared to increase with cell density even during the very early logarithmic phase (Fig. 9A). The same results were obtained with the xylanase constitutive mutant strain M-7, indicating that this effect is not related to induction. These results suggest that xynA could be regulated by cell density (quorum sensing). If cell density mediates the expression of xynA, it is possible to revert a high xynA-expressing culture (at high cell density) to a low xynA-expressing culture by diluting the high cell density culture with fresh medium. A typical result from such experiments is presented in Fig. 8B. Indeed, following a 500-fold dilution of a 0.1 A 600 culture (grown on minimal medium with xylose), the xylanase-specific activity first decreased and then started to increase again with the increase in cell density (Fig. 9B). Thus, cell density-dependent activation of xynA is a reversible process. To rule out that the increase in xynA expression during batch growth is caused by the changes in medium composition or the growth rate, we measured the expression of xynA in continuous cultures operating at a constant growth rate but at different cell densities (Fig. 9C). As a control, the specific activity of an unrelated intracellular enzyme, isocitrate dehydrogenase, was determined. Again, Xyn10A-specific activity increased with cell density, whereas the specific activity of isocitrate dehydrogenase remained constant (Fig. 9C). Similar results were obtained when the Chemostat cultures were operated under magnesium limitation, showing that the cell density activation is not influenced by the limiting nutrient. Taken together, it appears that the extracellular xylanase gene (xynA) is regulated by cell density (quorum sensing) during early logarithmic phase.
Cell Density Regulation of xynA and xynX-axe2-xynB3 Is on the Transcriptional Level-To study the mRNA expression profiles of specific genes as a function of cell density, total RNA was extracted from T-6 cultures at different cell densities, and the expression levels of the genes coding for the extracellular xylanase xynA, the intracellular xylanase xynA2, the intracellular xylosidases xynB1 and xynB2, the xynX-axe2-xynB3 operon, and the isocitrate dehydrogenase gene citC (used for normalization) were evaluated by real time RT-PCR (Fig. 10). The mRNA expression level of xynA increased with cell density, whereas the expression levels of the intracellular xylanase (xynA2) and the two xylosidases (xynB1 and xynB2) remained unchanged (Fig. 10A). Interestingly, the xynX-axe2-xynB3 transcript level was also activated by cell density (Fig. 10B). Cell density-dependent expression of the xylanase gene was also verified by RNase protection assay using 32 P-labeled antisense RNA probes. The protected RNA products were separated on a polyacrylamide gel, visualized by a phosphorimager, and quantified based on direct reading of the radioactive samples (Fig.  11A). The RNase protection data allowed calculating the absolute amount of xynA-mRNA molecules per cell. On average, noninduced cultures, induced cultures at low densities, and high density cultures contained 0.7, 3, and 75 xynA-mRNA transcripts per cell, respectively (Fig. 11B). These results indi-   Table 2). In the wild type strain the PCR fragment is indicates by dashed line and is expected to be 1502 bp (lane 1) and in xynX-null mutant it is 710 bp. SEPTEMBER 12, 2014 • VOLUME 289 • NUMBER 37 cate that cell density regulation of xynA is at the transcriptional level.

Regulation of the Xylanase Gene from G. stearothermophilus
Preliminary Characterization and Partial Purification of XDF-Cell density-dependent regulation is usually mediated by one or more extracellular factors that accumulate in the local environment (65). Because the extracellular factor is present at . Samples were taken from exponentially growing cells at the indicated time points, and xylanase extracellular activity was measured. Briefly, 1 ml of culture was clarified by centrifugation, and 80 l of the supernatant fluid was mixed with 80 l of 50 mM phosphate buffer, and 80 l of substrate, p-nitrophenol-cellobioside. The increase in absorbance at 420 nm was measured at 50°C using a plate reader. Specific activity was calculated dividing A 420 min Ϫ1 per cell turbidity (A 600 ). Error bars present S.D. of the means of three independent experiments. FIGURE 9. Cell density affects extracellular xylanase-specific activity. A, effect of cell turbidity on xylanase T-6 production in batch growth. Growth was carried out in a well aerated 10-liter fermenter in mBSM with 1% xylose. B, xylanase activity in cycled batch culture of G. stearothermophilus T-6. Strain T-6 culture was grown in mBSM with 1% xylose, and the culture was repeatedly diluted 500-fold into fresh medium. C, effect of cell turbidity on xylanasespecific activity in a carbon-limited Chemostat. Growth was carried out on defined medium at a constant dilution rate (D ϭ 0.2 h Ϫ1 ). The specific activity of xylanase (F) and the intracellular enzyme isocitrate dehydrogenase (E) were determined at various culture densities at steady state. Xylanase-specific activity was determined by hydrolytic activity on the synthetic substrate pNPX 2 at 50°C, measured by absorbance at A 420 nm (OD 420 ). One unit of activity was defined as (A 420 /min/ml of culture per A 600 ) ϫ 100 (Miller units).
a high concentration in conditioned medium, cultures grown in this medium should exhibit high specific activity even at low density. To follow the presence of the XDF for the purification process, we developed a simple and sensitive cell density activation assay in a 96-well plate. In this assay, cells from low density culture were placed in the wells in the presence of a xylanase chromogenic substrate, pNPX 2 , and different elution fractions (200-l total volume). The 96-well plate was incubated with shaking at 60°C in a microplate reader, and growth (turbidity at A 600 ) and extracellular xylanase activity (absorbance at A 420 ) were monitored simultaneously. From the resulting curves, xylanase-specific activity at different stages of growth could be readily deduced. To inactivate the extracellular xylanases, the conditioned medium was heated to 100°C for 10 min. As shown in Fig. 11A, adding boiled-conditioned medium to low cell density cultures resulted in high activity of Xyn10A even at low cell turbidities, indicating that a diffusible heatstable extracellular xynA density factor (XDF) is present in the medium. XDF appears to be hydrophobic because passing the conditioned medium via a hydrophobic matrix such as Sep-Pak C18 cartridge (Waters Corp.) or Amberlite XAD-2 (Sigma) resulted in the loss of cell density activity (Fig. 12A). Treating the conditioned medium with a protease (Staphylococcus aureus V8 attached to agarose beads, Sigma) abolished the induction of extracellular xylanase expression at low cell turbidities (Fig. 12B). Based on these results, it is likely that XDF is a small hydrophobic peptide or peptides. Initial attempts toward XDF identification were performed with a relatively small volume of conditioned medium, which was first concentrated 500-fold by absorption on XAD-2 matrix. Further purification was executed using reverse phase liquid chromatography on a C18 column, and fractions containing XDF activity were eluted with 20 -25% acetonitrile (Fig. 12C). Analyzing the active fraction using MALDI-TOF mass spectrometry revealed about 300 peptides of which only 10% were small molecular FIGURE 10. Expression of extracellular xylanase gene xynA and xynX, axe2, and xynB3 genes is increased by cell density. Strain T-6 cultures were grown in mBSM supplemented with 1% xylose and amino acids in a 10-liter fermenter. Samples (up to 100 ml) were collected at various cell densities along the logarithmic phase, and total RNA was extracted. Relative mRNA levels of the indicated genes at different cell densities were measured by quantitative real time RT PCR using the isocitrate dehydrogenase gene for normalization. A, xynA, xynA2, xynB1 and xynB2; B, xynX, axe2, xynB3. A, strain T-6 cultures were grown in a 10-liter fermenter, and samples (up to 200 ml) were drawn at various cell densities along the logarithmic growth phase for total RNA isolation. RNase protection assays were performed with a 467-nt 32 P-labeled antisense RNA probe. The probe was hybridized with 4 g of total RNA. The protected RNAs were loaded on a 5% polyacrylamide gel, which was further visualized using a phosphorimager system. Mr is the labeled DNA fragments of pUC19 digested with Sau3A. weight peptides. The majority of the peptides were part of a single protein that showed high homology to bacterial flagellin protein. We are now applying additional chromatography steps to further purify XDF to homogeneity.

DISCUSSION
The xylan utilization strategy of G. stearothermophilus is based on several elements. The bacterium first senses the presence of xylan in the environment by detecting minute amounts of xylose, which presumably exist as a by-product of xylan hydrolysis by neighboring microorganisms. Xylose is most likely detected by the two-component system, XynDC, comprised of a class I histidine kinase xylose-sensor protein, XynD, and the response regulator XynC (Fig. 1). In the presence of extracellular xylose, XynD phosphorylates XynC, and as a result, the phosphorylated form activates the expression of a dedicated xylo-oligosaccharide ABC transporter XynEFG (21). The facilitated entrance of xylosaccharides into the cell and subsequently their hydrolysis to xylose allows the induction of the system by xylose via inactivation of the XylR repressor.
XylR Is the Master Repressor of the Xylanolytic System-XylR is the master regulator of both xylose and xylan utilization genes and regulates at least six transcriptional units, including the two-component system (xynDC), an uncharacterized xylose epimerase (xylM), an acetyl xylan esterase (axe1), the extracellular xylanase (xynA), the xynX-axe2-xynB3 operon, a xylose kinase, and a xylose isomerase operon (xylAB) as well as its own gene (xylR). The XylR-binding site consensus sequence was found to be TT(A/T)GTTT-(3-11) n (A/T)(A/T)ACAAA (Fig. 3F). The fact that XylR regulates its own transcription (autoregulation) allows the cells to respond more gradually to increasing concentrations of xylose. This type of autoregulation was also observed with the related repressors AraR (20) and UxuR (22), which negatively regulate the arabinan and glucuronic acid utilization systems in G. stearothermophilus, respectively. Interestingly, the xylR gene from G. stearothermophilus M-7, a xylanase constitutive mutant, has a single G to A transition resulting in the substitution of the highly conserved Gly-146 by Asp. Indeed, XylR (G146D) failed to bind a 90-bp DNA fragment containing the xyl operator (66), indicating that Gly-146 is critical for operator binding.
In strain T-6, the molecular inducer of the xylan-utilization genes is xylose, based on its ability to induce the target genes and more specifically its ability to prevent binding of XylR to the corresponding operators as demonstrated by gel retardation assays (Fig. 3E). Interestingly, G. stearothermophilus grows very well on xylose with growth rates approaching those observed on glucose. This is in contrast to B. subtilis that is FIGURE 12. Partial characterization and purification of XDF. A, activation of xylanase using conditioned medium before and after absorption to XAD-2 matrix. 100 ml of conditioned medium was concentrated 2-fold using freeze dryer and then boiled for 10 min. Cells from early logarithmic growth phase were diluted 200-fold (ϳ0.001 A 600 (OD 600 )) in fresh mBSM, conditioned medium (10 -30 l), and pNPX 2 and placed in a 96-well plate. The plate is then incubated with shaking at 60°C and monitored simultaneously for both cell growth (A 600 ) and p-nitrophenol release (A 420 ). Specific activity ϭ A 420 / (A 600 ϫ min). The remaining medium was loaded on XAD-2 column and the flow-through was used to activate xylanase production at low cell-density. B, effect of proteinase K treatment of conditioned medium on xylanase cell density activation. Conditioned medium was incubated with proteinase K-beaded agarose (Sigma) for 1 h at 37°C, subsequent by bead removal by centrifugation. The treated conditioned medium was used to assay xylanase activation at low cell density. C, elution pattern of the active XDF fraction by reverse phase HPLC on C18 column. 2.5 liters of conditioned medium were concentrated 500-fold by absorption on XAD-2 column. Aliquots of 0.25 ml of concentrated active sample were fractionated by reverse-phase chromatography on C18 column with a linear gradient of ACN in 0.1% aqueous TFA at 0.5% min Ϫ1 . Fraction containing XDF activity was eluted with 20 -25% ACN, and it is marked with dashed circle. mAU, milliabsorbance units. capable of growing on xylan but grows very poorly on xylose due to the fact that it lacks a dedicated xylose transporter (57).
Xylan Utilization Genes Are Repressed by Glucose-An additional layer of regulation that coordinates the expression of the xylan utilization genes is the presence of a preferred carbon source, such as glucose. The xynA gene is repressed in media that contain glucose, as was demonstrated in continuous cultures operating at steady state (Fig. 4A). In many Gram-positive bacteria, carbon catabolite repression is mediated by the global regulator CcpA and a seryl-phosphorylated form of Hpr, which binds to the cre (catabolite-responsive elements) sites located in or near the promoter regions of the target operons (67,68). The xynA gene contains several potential cre sites, suggesting that xynA is subject to CcpA-dependent catabolite repression (56). G. stearothermophilus strain T-6 encodes a putative ccpA gene whose product shows 68% identity to B. subtilis CcpA (69). The xylanase constitutive mutant M-7, which produces a defective XylR, has a lower level of glucose repression (Fig. 4B). Indeed, induction of the xyl operon in B. subtilis is prevented by glucose 6-phosphate, which was shown to function as an anti-inducer of XylR (59). Furthermore, analysis of the phylogeny and specificities of the ROK family suggests that XylR together with glucose-responsive regulators evolved from a common ancestral glucose-responsive regulator (70). Taken together, it is likely that, in strain T-6, carbon catabolite repression is partially mediated by XylR.
Extracellular Xylanase Gene Is Regulated by CodY-In addition to glucose, the extracellular xylanase activity is reduced in rich medium or minimal medium supplemented with casamino acids (Fig. 4B). CodY, a global transcriptional regulator, is active as a regulator in cells growing in an excess of casamino acids or cells that have a high intracellular level of GTP, as found in rapidly growing cells (71). However, at the entrance to stationary phase, when cells are nutritionally limited, many genes repressed by CodY become expressed (60,72,73). Our results suggest that expression of xynA and of additional hemicellulolytic systems is subject to CodY control. Thus, in nutrient excess, the hemicellulolytic genes from G. stearothermophilus T-6 are repressed. In B. subtilis the arabinogalactan oligomer permease (ganP) gene contains a putative binding site for CodY (74), suggesting that additional bacterial hemicellulolytic systems are regulated by CodY.
Extracellular Xylanase Gene Is Regulated by XynX, a Novel Transcriptional Factor-The extracellular xylanase gene xynA is adjacent to the xynX-axe2-xynB3 operon that is negatively regulated by XylR and induced by xylose (Fig. 3). Whereas both axe2 and xynB3 code for hydrolases with known functions (31,33,42), the role of xynX is unknown. Based on sequence homology, the xynX gene product belongs to the Nif3-like protein family (Ngg1p-interacting factor 3, Pfam family PF01784), which is highly conserved from bacteria to humans (75).
Despite the fact that XynX has no recognizable DNA-binding motif found in bacterial regulators, it was able to shift a 72-bp DNA fragment in gel retardation assays (Fig. 5). This region, corresponding to positions Ϫ30 to Ϫ102, is AT-rich and contains an inverted repeat sequence that is highly conserved in several Geobacillus spp. (Fig. 13). A similar sequence can be found upstream of xynX, thus raising the possibility that XynX regulates its own expression (Fig. 13). To date, the ability of a Nif3-like protein to bind DNA was demonstrated only with the TTHA1606 protein from Thermus thermophilus HB8 (76).
In the xynX null strain, xylanase-specific activity was higher than in the wild type, indicating that XynX represses the expression of xynA. However, the xynX null strain exhibited normal expression of the xynA gene during the logarithmic growth phase and the same level of repression by glucose or casamino acids as in the wild type strain. Taken together, these results indicate that XynX is involved in xynA regulation but does not contribute directly to carbon catabolite repression. In summary, the mechanism by which XynX controls xynA gene expression is elusive as well as the additional factors that may mediate its effect. Characterization of the XynX mode of action might reveal a novel pathway for the Nif3-like proteins in prokaryotes.
The Extracellular Xylanase Gene Is Regulated by Cell Density (Quorum Sensing)-The least expected result presented here is quorum-sensing regulation of the extracellular xylanase (xynA) gene. In batch cultures, the specific activity of the extracellular xylanase increases over 50-fold during exponential growth, suggesting cell density regulation. This phenomenon is not related to induction or cell adaptation as was demonstrated in continuous cultures operating at different cell densities and with a xylanase-constitutive mutant. Quantitative real time RT-PCR and quantitative measurements of xynA-mRNA using RNase protection assays indicated that the regulation is at the transcriptional level.
In Gram-positive bacteria, quorum-sensing regulation is mediated by small modified peptides and controls many physiological processes, including biofilm formation (77), virulence factor regulation (65,78), competence development (79,80), sporulation (81,82), and induction of conjugation (83,84). Our initial attempts to purify the extracellular XDF suggest that it is heat-stable (100°C for 10 min) and pH-stable (2.0), yet it is sensitive to protease treatment. The factor can be absorbed on a hydrophobic matrix and eluted with 20 -25% acetonitrile using a C18 reverse phase HPLC column. All of these results are consistent with Gram-positive quorum-sensing peptide factors (85)(86)(87)(88), suggesting that XDF is a small hydrophobic peptide.
Analysis of the draft genome of G. stearothermophilus strain T-6 reveals sequence homology to well characterized quorumsensing pathways for competence and sporulation processes, FIGURE 13. Alignment of regulatory regions upstream to xylanases genes from Geobacillus spp. The sequences upstream the putative Ϫ10 regions of several xylanases genes from Geobacillus strains: G. stearothermophilus T-6, G. stearothermophilus T-1, Geobacillus sp. MAS1, Geobacillus sp. Y412MC6, Geobacillus sp. C56-T3, Geobacillus sp. GHH01 and G. thermodenitrificans NG80 -2. The first sequence aligned is the sequence upstream of the xynX gene from G. stearothermophilus T-6. Conserved nucleotides are shown in bold, and the inverted repeat sequence, the putative XynX-binding site, is indicated by arrowheads.
including ComQ and ComC, and the modified peptides ComX and CSF. Strain T-6 also possesses the NrpR-NrpX quorumsensing system, and therefore we tested whether the predicted pentapeptide (AKDEH) can activate xynA expression in low cell density cultures. The synthetic peptide failed to activate xynA expression (data not shown), indicating that strain T-6 utilizes a different cell density regulatory system to regulate xynA. The secreted cell-extracellular signaling peptides can bind to a specific sensor component of a two-component signal transduction system (89) or can be import to the cell via specific oligopeptide transport system (90). Bialaphos is a toxic tri-peptide that is known to enter bacteria via oligopeptide permeases (91). We have isolated bialaphos-resistant T-6 mutants, suspected to code for a defective oligopeptide transport system. These bialaphos-resistant mutants exhibited typical activation of xylanase expression during logarithmic growth, suggesting that in strain T-6 cell density activation of xynA is not solely based on an oligopeptide transport system.
What can be the biological benefit of regulating the expression of a vegetative extracellular enzyme by quorum sensing? Quorum-sensing regulation provides a mechanism by which secretion and production of exoproducts ("public goods") are FIGURE 14. Schematic representation of multiple regulatory mechanisms controlling the expression of xynA. Xylose and short xylo-oligomers are sensed by the two-component system XynDC, which in turn activates the ABC xylo-oligosaccharide transporter, XynEFG, facilitating the entrance of xylose and xylosaccharides into the cell. Inside the cell, xylose serves as the molecular inducer and interacts with the master repressor XylR, which negatively regulates the transcription of the xynA gene. XylR can also interact with glucose 6-phosphate, which exerts inducer prevention. Expression of the xynA gene is also repressed by glucose via carbon catabolite repression, most likely mediated by CcpA. In rapidly growing cells in high nutrition availability, xynA is repressed by CodY, which is subjected to regulation by GTP and branched-chain amino acids. Additionally, xynA is regulated by two yet uncharacterized mechanisms, a NIF3-like protein, XynX, and by cell density. The xynA gene is induced via quorum sensing regulation at relatively low cell density. The XDFs are most likely short modified peptides that can either bind to a specific sensor component of a two-component signal-transduction system or can be imported into the cell by a dedicated peptide transporter. The DNA sequence corresponds to position Ϫ131 bp with respect to the transcriptional start point (ϩ1). The proposed cataboliteresponsive element (cre) is presented in red letters. The potential inverted repeat binding site for the xylose repressor, XylR, and the putative XynX-binding site are indicated by arrows. Two of the proposed CodY-binding site sequences are shown in green letters. The table below summarizes the proposed regulatory proteins and their effectors involved in the regulation of the xynA gene. regulated to benefit the cells in the local population (92)(93)(94)(95). In many Gram-positive bacteria the regulation of "public goods" production is associated with late exponential phase and stationery phase and occurs at high cell densities. During this period, and before the cell is committed to sporulation, it expresses proteins that allow it to adapt to poor nutritional conditions and maintain growth, for example, by degrading macromolecules via expression of extracellular enzymes. However, the quorum-sensing regulation of the extracellular xylanase appears to function at very low cell densities (Ͻ0.01 A 600 ) at the early exponential phase of growth and not at the entrance to the stationary phase. To the best of our knowledge, regulation at these low cell densities has never been reported. A possible biological rationale for cell density activation of the extracellular xylanase stems from the fact that the utilization of high molecular weight polymers, such as xylan, requires a threshold concentration of extracellular enzymes. At very low cell densities, the effective concentration of diffusible extracellular enzymes is too low to support growth because the xylo-oligosaccharides products are too large to import into the cells, resulting in "cell density-dependent growth." Only at moderate or "high cell densities" does the culture produce sufficient amounts of extracellular hydrolyzing enzymes that provide suitable degradation products. Thus, it is advantageous for the cell to produce the extracellular enzymes only when cell density can support growth. Cell density growth dependence was demonstrated elegantly with Myxococcus xanthus, a Gram-negative topsoil bacterium that forms swarms and fruiting bodies (12). When grown on casein, the cells exhibited cooperative behavior by pooling their extracellular digestive enzymes. The extracellular proteinase activity was directly proportional to cell number, and the growth rate was correlated with increased concentrations of hydrolyzed casein.
In summary, we have shown that the expression of the extracellular xylanase gene, xynA, from G. stearothermophilus T-6 is intricate and involves multiple regulatory mechanisms (Fig.  14). This regulation is well suited to the physiological scheme the bacterium use for degrading polysaccharides in the environment. The xynA gene is negatively controlled by the xylose repressor, XylR, and is induced by xylose. Expression of the gene is also repressed by glucose via carbon catabolite repression, most likely mediated by CcpA and casamino acids via the transcriptional regulator CodY. In addition, xynA is regulated by two yet uncharacterized mechanisms, a NIF3-like protein, XynX, and by cell density. The quorum-sensing regulation is effective at very low cell densities and most likely represents a novel system.