Characterization of the GDP-D-Mannose Biosynthesis Pathway in Coxiella burnetii: The Initial Steps for GDP-β-D-Virenose Biosynthesis

Coxiella burnetii, the etiologic agent of human Q fever, is a Gram-negative and naturally obligate intracellular bacterium. The O-specific polysaccharide chain (O-PS) of the lipopolysaccharide (LPS) of C. burnetii is considered a heteropolymer of the two unusual sugars β-D-virenose and dihydrohydroxystreptose and mannose. We hypothesize that GDP-D-mannose is a metabolic intermediate to GDP-β-D-virenose. GDP-D-mannose is synthesized from fructose-6-phosphate in 3 successive reactions; Isomerization to mannose-6-phosphate catalyzed by a phosphomannose isomerase (PMI), followed by conversion to mannose-1-phosphate mediated by a phosphomannomutase (PMM) and addition of GDP by a GDP-mannose pyrophosphorylase (GMP). GDP-D-mannose is then likely converted to GDP-6-deoxy-D-lyxo-hex-4-ulopyranose (GDP-Sug), a virenose intermediate, by a GDP-mannose-4,6-dehydratase (GMD). To test the validity of this pathway in C. burnetii, three open reading frames (CBU0671, CBU0294 and CBU0689) annotated as bifunctional type II PMI, as PMM or GMD were functionally characterized by complementation of corresponding E. coli mutant strains and in enzymatic assays. CBU0671, failed to complement an Escherichia coli manA (PMM) mutant strain. However, complementation of an E. coli manC (GMP) mutant strain restored capsular polysaccharide biosynthesis. CBU0294 complemented a Pseudomonas aeruginosa algC (GMP) mutant strain and showed phosphoglucomutase activity (PGM) in a pgm E. coli mutant strain. Despite the inability to complement a manA mutant, recombinant C. burnetii PMI protein showed PMM enzymatic activity in biochemical assays. CBU0689 showed dehydratase activity and determined kinetic parameters were consistent with previously reported data from other organisms. These results show the biological function of three C. burnetii LPS biosynthesis enzymes required for the formation of GDP-D-mannose and GDP-Sug. A fundamental understanding of C. burnetii genes that encode PMI, PMM and GMP is critical to fully understand the biosynthesic pathway of GDP-β-D-virenose and LPS structure in C. burnetii.


Introduction
Lipopolysaccharide (LPS) is a complex molecule and represents the major component of the outer leaflet of the outer membrane of Gram-negative bacteria. The LPS molecule consists of three structural domains: (1) lipid A, which represent the hydrophobic anchor of the LPS molecule and is responsible for the endotoxic properties, (2) a short non-repeating inner and outer core oligosaccharide, which is attached to lipid A and extends outwardly and (3) the O-specific polysaccharide chain (O-PS), which is composed of repeating sugar units and determines the serological heterogeneity among bacterial isolates. The primary function of LPS is to serve as a permeability barrier against external agents such as hydrophobic antibiotics and to maintain the structural integrity of the Gram-negative cell wall [1].
C. burnetii, a Gram-negative small pleomorphic coccobacillus, is the causative agent of the zoonosis Q fever. Q fever manifests in humans generally as an acute, debilitating flu-like illness or less common as chronic Q fever, which develops mainly as endocarditis or hepatitis. C. burnetii is a naturally obligate intracellular bacterium and so far no method for generation of specific mutants has been established. C. burnetii is considered a potential biological weapon because it consistently causes disability, can be manufactured on a large scale, remains stable under various conditions and can be efficiently disseminated [2]. The U.S. Centers for Disease recently designated C. burnetii as a category B bioterrorism agent. There is no licensed vaccine for C. burnetii infection in the U.S. because of adverse reactions to killed whole cell vaccination. Therefore, the understanding of C. burnetii physiology and vaccine development remains an important public health and U.S. national security objective [3].
Upon serial passage in an immune-incompetent host, virulent C. burnetii undergoes a shortening of its LPS, traditionally referred to as phase variation in Enterobacteriaceae. Phase variation of C. burnetii is characterized by a non-reversible switch from virulent phase I smooth LPS (S-LPS), which has a full length O-polysaccharide (O-PS) chain to an avirulent phase II rough-LPS (R-LPS). The R-LPS variant is missing the O-PS chain and unknown sugar residues located within the outer core oligosaccharide [4]. Previous studies showed that no significant loss of protein content on the surface of C. burnetii occurred during phase variation and the only characterized difference between virulent phase I and avirulent phase II isolates is LPS [5,6,7,8]. Furthermore, vaccine studies showed that BALBc mice vaccinated with formalin killed whole cell phase I bacteria were protected from C. burnetii challenge while mice vaccinated with whole cell phase II bacteria were not protected [8]. These studies highlight the importance of C. burnetii LPS.
Phase variation in C. burnetii is accompanied by the deletion of a large chromosomal fragment which contains glycosyl transferases and sugar processing genes required to complete b-D-virenose biosynthesis, O-PS chain elongation and inner membrane transport [21,22]. This deletion is likely the O-PS operon and is responsible for the loss of O-PS in the C. burnetii Nine Mile strain RSA439 [21]. Based on the structure of b-D-virenose and the genes located within the deleted region of the C. burnetii phase II variant, the in Figure 1 presented GDP-b-D-virenose biosynthesis pathway is proposed. The aim of this study was to demonstrate the biological significance of three C. burnetii enzymes for the biosynthesis of GDP-D-mannose and examine the initial steps of GDP-b-D-virenose biosynthesis. The presented data provide fundamental knowledge necessary to further characterize the formation of GDP-b-D-virenose, a novel saccharide, and may help develop potential vaccine candidates such as in vivo and in vitro generated glycoconjugates.

Results
Bioinformatic analysis was carried out on the C. burnetii enzymes predicted to be responsible for GDP-D-mannose biosynthesis ( Table 1). Amino acid sequence alignments indicated that each enzyme had a high degree of similarity to characterized GDPmannose biosynthesis enzymes. C. burnetii CBU0671 has the bioinformatic signatures of a type II phosphomannose isomerase (PMI), a small but growing class of PMIs identified in Gramnegative bacteria [23]. Type II PMIs are bifunctional enzymes that catalyze the isomerisation of fructose-6-phosphate to mannose-6-phosphate and the transfer of GDP to D-mannose-1phosphate to form GDP-D-mannose [24]. However, C. burnetii CBU0671 appeared to be unrelated to a type I PMI from E. coli EDL933, but contains the conserved PMI active site, which is characteristic of the type II PMIs [23]. C. burnetii CBU0294 is predicted to catalyze the second step in the GDP-mannose biosynthesis pathway, the conversion of D-mannose 6-phosphate to D-mannose 1-phosphate. Amino acid sequence alignment indicated a high degree of identity to P. aeruginosa AlgC, which was shown to be bifunctional and exhibits phosphoglucomutase (PGM) as well as phosphomannomutase (PMM) activity [1,25]. CBU0689 is annotated as GDP-mannose-4,6-dehydratase (GMD) and might provide the virenose biosynthetic intermediate GDP-6-deoxy-Dlyxo-hex-4-ulopyranose (GDP-Sug) by conversion of GDP-Dmannose. Further CBU0671 and CBU0689 are located within a genomic region associated with O-PS biosynthesis [26].

C. burnetii CBU0671 exhibited GDP-mannose pyrophosphorylase (GMP) but not PMI activity
In order to characterize the enzymatic activities of C. burnetii CBU0671, this protein was expressed in its native form and used for complementation of E. coli manA and manC mutant strains, defective for O-PS or CPS synthesis, respectively. To test for PMI activity the CBU0671 containing plasmid pCN606_2 was introduced into the manA mutant strain E. coli CWG634 and O-PS patterns compared to wild type E. coli CWG28 O9a. Inactivation of manA in E. coli CWG364 was shown to abolish synthesis of mannose-6-phosphate, the precursor of GDP-Dmannose and resulted in a R-LPS phenotype [27]. Complementation of E. coli CWG634 with CBU0671 did not result in restoration of an S-LPS phenotype. Analysis of LPS from wild type, mutant and complemented strains using silver stained SDS-PAGE detected only revealed R-LPS chemotypes (data not shown). However, complementation of E. coli CWG634 with E. coli DH5a manA (pCN601a_5), which encoded a type I PMI, resulted in a smooth LPS phenotype (data not shown).
To test if CBU0671 exhibits GMP activity, plasmid pCN606_2 was introduced into the E. coli manC mutant strain CWG152 and the CPS pattern was compared to wild type E. coli CWG44 K30 [28,29]. As a positive control, manC from E. coli DH5a was cloned and expressed in its native form and the resulting plasmid, pCN603_1, was introduced into CWG152. CPS isolated from wild-type strain E. coli CWG44, manC mutant strain E. coli CWG152 and complemented strains E. coli CWG152/pCN606-2 and E. coli CWG152/ pCN603_1 were analyzed using silverstained SDS-PAGE and immunoblot with anti-K30 antiserum. Complementation of E. coli CWG152 with CBU0671 or E. coli DH5a manC resulted in typical high and low molecular mass CPS bands as detected for the wild type strain (Fig. 2). Taken together these data clearly demonstrated that CBU0671 exhibits GMP activity, but could not complement a type I PMI.

C. burnetii CBU0294 exhibits PMM and PGM activity
The enzymatic function of C. burnetii CBU0294 was evaluated by complementation of an algC mutant of P. aeruginosa PAO1 serotype O5. PMM function of AlgC catalyzes the formation of mannose-1-phosphate, which is a metabolic precursor for synthesis of GDP-D-mannose [25]. The latter is converted to GDP-Drhamnose, the sugar residue composing the O5 A-band homopolymer [30]. It has been shown that PGM function of P. aeruginosa AlgC is required for formation of D-glucose-1-phosphate, which is necessary for biosynthesis of UDP-D-glucose, a component of the core heterooligosaccharide [25,31]. Therefore both, PMM and PGM functions of AlgC are required to visualize the P. aeruginosa PAO A-band. C. burnetii CBU0294 was cloned into the P. aeruginosa shuttle vector pUCP20 and the resulting plasmid, pCN620, used for transformation of P. aeruginosa PAO1 algC::tet. LPS samples prepared from transformed PAO1 algC::tet strains were separated by SDS-PAGE and visualized by silver staining (Fig. 3). Both, the parental and complemented mutant strains produced a typical LPS banding pattern, while PA01 algC::tet alone as well as the vector control did not produce A-band LPS.
To evaluate the PGM activity of C. burnetii CBU0294, pCN620 was transformed into E. coli W1485 pgm::tet and selected on MacConkey agar for the ability to metabolize galactose. E. coli W1485 pgm::tet that harbored pCN620 generated deep red colonies identical to E. coli W1485 wild type, whereas E. coli W1485 pgm::tet alone or the empty vector control E. coli W1485 pgm::tet/pUCP20 produced light pink colonies (data not shown).
These data indicate that CBU0294 might encode a bifunctional enzyme, which exhibits PMM as well as PGM activity. Thus, CBU0294 is likely to catalyze the second step in the GDPmannose biosynthesis pathway of C. burnetii.
Determination of PMI, PMM and GMP activities of purified C. burnetii proteins C. burnetii CBU0671 and CBU0294 were expressed as Histagged proteins in E. coli DH5a and isolated to near homogeneity. Kinetic analyses of purified enzymes were carried out by measuring initial enzyme specific activity relative to varying concentrations of substrate. Km and Vmax values were determined by Lineweaver-Burk Plot analysis with a regression coefficient greater than 0.99 (Fig. 4). Obtained Km and Vmax values as well as specific enzymatic activity for purified CBU0671 with mannose-1-phosphate or GDP-D-mannose as substrate indicate PMI and GMP activity ( Table 2). PMI and GMP activity were also detectable in bacterial crude extracts comparable to E. coli ManA and ManC (Table 3). Specific enzymatic activity obtained for purified CBU0294 with mannose-1-phosphate indicates PMM activity, which is also detectable in crude extracts ( Table 2 and 3).
GMD activity of C. burnetii CBU0689 was tested using the method described by Alberman et al. [32] by expression of the native protein in E. coli. Therefore C. burnetii CBU0689 was cloned into pBAD and the resulting plasmid, pCN608c-1, transformed E. coli DH5a. Enzyme activity was then measured directly in crude extracts by monitoring the increase in GDP-Sug at OD 320 (e 320 nm = 2.20 L?mmol 21 ?cm 21 ) in alkaline conditions. GMD activity for CBU0689 was determined as 14 NKat/mg. Taken together these data clearly show that all three C. burnetii open reading frames (orf's) exhibit the necessary enzymatic activities for formation of GDP-D-mannose and GDP-Sug as intermediates for virenose synthesis.

Discussion
The goal of this work was to characterize the enzymatic steps responsible for formation of GDP-D-mannose in C. burnetii, which were bioinformatically predicted as the initial steps of GDP-b-Dvirenose biosynthesis. Structural evidence of b-D-virenose isolated from the virulent phase I C. burnetii RSA493 O-PS further supports this hypothesis [19]. Although C. burnetii CBU0671, a predicted bifunctional type II PMI, failed to complement an E. coli manA mutant strain (PMI), it did complement an E. coli manC mutant strain (GMP). An exhaustive bioinformatic search of the annotated genome failed to reveal an alternative C. burnetii PMI. Clustal analysis showed that CBU0671 contained the signature sites observed in other type II PMIs, such as WbpW or AlgA of P. aeruginosa; pyrophosphorylase signature, GMP active site, nucleotidyl transferase domain, mannose-6-phosphate isomerase domain, zinc binding motif and PMI active site [23]. Althrough C. burnetii failed to complement a manA mutation in E. coli in vitro assays using natively formed and His-tagged C. burnetii CBU0671 showed specific activities, Km and Vmax values, comparable to previously reported values for PMI and GMP [33]. Differences in regulation or catalytic process might explain the observed distinct activities for CBU0671 in a manA deficient background or in vitro observed enzymatic activities, respectively. Both, type I and type II PMI's possess a highly conserved motif within the active side, but other proteins have lost the specific catalytic function despite the shared motif [24]. Further investigations, such as complementation of other type II PMI's are necessary to clearly identify the catalytic activities of CBU0671. Taken together, we report that CBU0671 is a new member of the small and poorly characterized class of proteins known as type II PMIs, based on complementation assays and biochemical characterization.
C. burnetii CBU0294 successfully complemented a P. aeruginosa O5 algC mutant strain and restored expression of a smooth LPS.  Previous studies showed that algC of P. aeruginosa O-serotype O5 is involved in formation of D-mannose and D-glucose-1-phosphate, both necessary intermediates for synthesis of the O5 O-specific chain and core oligosaccharide. Therefore inactivation of algC leads to a deep rough phenotype in P. aeruginosa O5 [25,30]. Restored expression of a smooth LPS by complementation indicates that C. burnetii CBU0294 simultaneously carried out PGM and PMM activities in this strain. To further demonstrate that C. burnetii CBU0294 also exhibits phosphoglucomutase activity, a pgm E. coli mutant strain, W1485 pgm::tet, was successfully complemented with CBU0294. This finding supports the notation C. burnetii CBU0294 is bifunctional and carries out PGM and PMM activities as described for algC.
Bioinformatic analysis indicats that CBU0689 encodes a GDPmannose 4,6-dehydratase (GMD). When compared to E. coli GMD, CBU0689 was 52% identical, 69% similar on the amino acid level with an E value of 5610 212 [32,34]. Specific activity of native C. burnetii GMD in crude extracts and its gene location within the LPS associated genome region in C. burnetii supports its bioinformatic assignment [26]. GDP-Sug formed by GMD is the metabolic intermediate of GDP-L-fucose, GDP-colitose, GDPperosamine, GDP-D-rhamnose and GDP-6-deoxy-D-talose [35]. The enzymes required to generate the final steps required for GDP-perosamine (perosamine synthase CBU0830) and GDP-L-fucose (fucose synthase CBU0688) have been identified in the C. burnetii genome [26]. However, none of these activated saccharides have been observed in C. burnetii with the exception of a single report in which rhamnose was identified by GC-MS in the C. burnetii LPS outer core [12].
Since the characterization of the avirulent C. burnetii RSA439 genomic deletion [21], the enzymatic mechanism of fucose synthase, located within this region has become more clear [36]. Clustal analysis of the C. burnetii fucose synthase indicated that it bears the characteristic Ser-Tyr-Lys catalytic triad necessary to catalyze three reactions within a single active site; epimerization at both C30 and C50 and NADPH dependent reduction of the ketone at C4 [36]. Based on these data, the formation of GDP-b-Dvirenose may ultimately be formed when GDP-L-fucose is modified by the addition of a methyl group at C30 perhaps by CBU0691 and inversion of stereochemistry at the C20 (Fig. S1).
A fundamental understanding of C. burnetii LPS biosynthesis and its structure are lacking. The intracellular nature of C. burnetii, lack of genetic tools and its status as a select agent has made elucidating these basic physiological mechanism challenging. This study establishes the foundation necessary to fully characterize the GDP-b-D-virenose biosynthesis pathway and ultimately the formation of C. burnetii O-PS, which is the only known virulence factor of C. burnetii.
General DNA methods DNA isolation and manipulations were carried out in according to Sambrook and Russel (2001) [37]. Oligonucleotides used in this study are listed in Table S1. DNA restriction endonucleases, T4 DNA ligase and Accuprime polymerase (Invitrogen) were used as advised by the manufacturer. Electrocompetent E. coli and P. aeruginosa cells were prepared as described elsewhere [38,39] and transformed using a Bio-Rad Gene-Pulser Transfection Apparatus (200V, 25 mF, 12.5 kV/cm, 4.7 ms).
Complementation of E. coli manA and manC mutant strains with C. burnetii CBU0671 CBU071, including the native stop codon, was amplified from chromosomal DNA of C. burnetii RSA 439 with CBU0672FNcoI and CBU0671R and cloned into pBAD for native protein expression. The resulting plasmid, pCN606_2, was used for complementation LPS manA or CPS manC E. coli mutant strains CWG634 and CWG152. As positive controls E. coli DH5a manA (EcmanAF/EcmanAR) and manC (EccpsBNcoIF/EccpsBR) were cloned into pBAD and the resulting plasmids, pCN601a_5 and pCN603_1, used for native protein expression in the corresponding manA and manB mutant strains. All generated plasmids were verified by sequencing. Complemented strains were grown in LB broth supplemented with 0.4% glucose, to avoid uptake of exogenous mannose, and protein expression induced with 0.2% arabinose over night [27]. Expression of full length LPS or CPS was determined by analysis of proteinase K-treated whole cell lysates of complemented strains and compared to corresponding LPS and CPS wild type strains E. coli CWG28 (serotype O9a) and E. coli CWG44 (serotype K30), respectively. Lysates were prepared as described elsewhere [40] and analyzed by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) and stained with silver nitrate or transferred to nitrocellulose membrane (Bio-RAD) [40,41,42,43]. O9a (1:2000) or K30-specific antiserum (1:1000) was used for detection of LPS and CPS expression and traced with horse radish peroxidase-conjugated goat anti-rabbit IgG(c) monoclonal antibody (1:5000) with peroxidase substrate in according to the guidelines of the manufacturer (Amersham Bioscience).
Complementation of P. aeruginosa algC mutant strain with C. burnetii CBU0294 CBU0294 was amplified (CBU0294FScaI/CBU0294RXbaI) from chromosomal C. burnetii RSA 439 DNA and obtained DNA fragment was digested with ScaI and XbaI. CBU0294 was subsequently cloned into the ScaI, XbaI treated shuttle vector pUCP20 to generate pCN620 [44]. Correct insertion of CBU0294 in pCN620 was verified by sequencing. For complementation studies P aeruginosa PAO1 algC:tet was transformed with pCN620 and additionally with a P. aeruginosa algC containing shuttle vector, pLPS188 [25]. LPS banding patterns from P aeruginosa PAO1 algC:tet harboring pCN620 or pLPS188 were analyzed as described under 4.3 and compared to wild type LPS from P.aeruginosa PAO1. PAO1 specific antiserum (1:1000) was used for detection of OPS expression and traced with horse radish peroxidase-conjugated goat anti-rabbit IgG(c) monoclonal antibody (1:5000).
Phosphoglucomutase (pgm) complementation of E. coli with C. burnetii CBU0294 CBU0294 was amplified (CBU0294F/CBU0294R) and cloned into pBAD for native protein expression. The resulting plasmid, pCN607a_3, was sequenced and used for transformation of E. coli W1485pgm::tet. Phosphoglucomutase positive wild type E. coli W1485 and complemented mutant strains, were then distinguished from pgm negative strains using MacConkey agar (Difco Laboratories) supplemented with 1% (w/v) galactose and 0.2% (w/v) arabinose as previously described [45].
Cloning and expression of C. burnetii GDP-D-mannose synthesis genes for enzyme activity assays C. burnetii CBU0671 (CBU0671F/CBU0671Rpoly-His) and CBU0294 (CBU0294Fpoly-His/CBU0294Rpoly-His) were amplified and subsequently cloned into pBAD for expression of Histagged proteins. The generated plasmids pCN606c_E1 and pCN607z_A2 were sequenced for correct insertion of target genes and used for transformation of E. coli DH5a. As controls E. coli DH5a manA and cpsB (manC) genes were amplified with EcmanAFpoly-His and EcmanARpoly-His or EccpsBFpoly-His and EccpsBFpoly-His and cloned into pBAD. The resulting plasmids pCN601c_A1 and pCN603a_A4 were sequenced for correct insertion of target genes and used for transformation of E. coli DH5a. Expression was induced with 0.2% arabinose for 4 to 8 h at an OD 600 of 0.6. Bacteria were harvested (10,0006g, 10 min, 4uC), resuspended in 10 mL binding buffer (25 mM NaPO 4 , 0.5 M NaCl, 10 mM imidazole, pH 8.0) with DNase (10 mg/mL), RNase (10 mg/mL) and lysozyme (10 mg/mL) and incubated for 30 minutes on ice. Cells were lysed using French press and cell debris separated by centrifugation (24,4006g, 60 min, 4uC). His-tagged proteins were isolated from supernatants (crudes extracts) using the ProBond purification system as described by the supplier (Invitrogen). Purified proteins were analyzed for purity and size by SDS-PAGE and silver staining or immunoblot analysis with 66His monoclonal antibody (1:5000, Clontech). PMI, PMM or GMD activity was also determined in bacterial crude extracts that contained natively formed C. burnetii proteins.
Phosphomannose isomerase (PMI) in vitro assay PMI enzyme activity was determined by monitoring the reduction of NADP + at 340 nm (e M = 6.22 mM 21 cm 21 ) [46]. One unit of enzyme activity was defined as the detection of 1 mmole of product per minute. Concentration of purified enzyme was determined using the Micro BCA Protein Assay (Invitrogen) as described by the supplier. PMI activity was determined by a modified protocol described by Sa-Correia et al. [33]. The reaction mixture in a 1 mL total volume contained 10 mmol of MgCl 2 , 1.0 mmol of NADP + , 1 unit phosphoglucose isomerase, 1 unit glucose-6-phoshphate dehydrogenase, 1.1 mmol of D-mannose-6phosphate in 50 mM tris HCl buffer pH 7.55. MgCl 2 , NADP + and D-mannose-6-phosphate were dissolved in 50 mM tris HCl buffer pH 7.55 prior to adding them to the reaction mixture. The reaction mixture was equilibrated for 5 min at 25uC and the reaction initiated by adding 50 to 200 mL of crude extract that contained natively formed PMI or 7-15 mg of purified His-tagged PMI (CBU0671).

Phosphomannomutase (PMM) in vitro assay
PMM activity was determined by monitoring the reduction of NADP + at 340 nm (4.6.1.) as described by Sa-Correia et al. [33]. The reaction mixture in a 1 mL total volume contained 10 mmol of MgCl 2 , 1.0 mmol of NADP + , 1 unit phosphoglucose isomerase, 1 unit glucose-6-phoshphate dehydrogenase, 5 units of purchased PMI (Sigma), 0.25 mmol of D-glucose-1,6-diphosphate (ADGD) and 5.5 mmol of D-mannose-1-phosphate in 50 mM tris HCl buffer pH 7.55. C. burnetii CBU0294 is annotated as a bifunctional phosphomannomutase (PMM) and phosphoglucomutase (PGM). Therefore, the addition of ADGD moved the kinetics of the reaction towards the formation of D-gluconate-6-phosphate. The reaction mixture was equilibrated for 5 minutes at 25uC and the reaction initiated by adding 50-200 ml of crude extract that contained natively formed PMM or of 7-15 mg of purified Histagged PMM (CBU0294).

GDP-D-mannose pyrophosphorylase (GMP) in vitro assay
GMD activity was determined using a modified protocol described by Munch-Peterson et al. [47], monitoring the reduction of NADP + (4.6.1.). The reaction mixture in a 1 mL total volume contained 10 mmol of MgCl 2 , 1.0 mmol of NADP + , 0.1 mmol of ADP, 2 mmol of PPi, 5.0 mmol of NaF, 1 unit of hexokinase, 1 unit of nucleoside kinase, 1 unit of glucose-6-phoshphate dehydrogenase, 0.8 mmol of glucose, 5.5 mmol of GDP-D-mannose in 50 mM tris HCl buffer pH 7.55. The reaction mixture was equilibrated for 5 minutes at 25uC, and initiated by adding 50-200 ml of crude extract. Endogenous activity in crude extracts of E. coli DH5a carrying the empty pBAD vector were subtracted from the test samples.
GDP-D-mannose 4,6-dehydratase (GMD) in vitro assay CBU0689 was amplified (CBU0689FNcoI/CBU0689R) and cloned into pBAD, retaining the native stop codon. The resulting plasmid, pCN608c_1, was used for transformation of E. coli DH5a. Crude extract GMD activity was determined by a modified protocol described by Albermann et al. [32]. The reaction mixture in a total volume of 300 mL contained 10 mmol of MgCl2, 1.0 mmol of NADP + and 5.5 mmol of GDP-D-mannose in 50 mM tris HCl buffer pH 7.55. After equilibrating for 5 minutes at 37uC, the reaction was initiated by adding 60 mL of prewarmed crude extract. Aliquots of 50 mL were taken every 10 min and added to 950 mL of 37uC 100 mM NaOH. The reaction was incubated for an additional 20 minutes. The formation of GDP-4-keto-6-deoxy-D-mannose was measured directly at OD 320 (e M = 2.2 mM 21 cm 21 ) [34]. Figure S1 Clustal analysis of C. burnetii fucose synthase CBU0688 (GFS). The C. burnetii GFS has the characteristic ''Catalytic Triad,'' Ser (S) 107-Tyr (Y) 136-Lys (K) 140 boxed in black, observed in SDR family enzymes. Additionally, boxed in red are active sites implicated as the acid/bases involved in promoting the epimerization reactions. (TIF)