Regular paper on-line at: www.actabp.pl Purification and characterization of GlcNAc-6-P 2-epimerase from Escherichia coli K92

N-Acetylmannosamine (ManNAc) is the first committed intermediate in sialic acid metabolism. Thus, the mechanisms that control intracellular ManNAc levels are important regulators of sialic acid production. In prokaryotic organisms, UDP-N-acetylglucosamine (GlcNAc) 2-epimerase and GlcNAc-6-P 2-epimerase are two enzymes capable of generating ManNAc from UDP-GlcNAc and GlcNAc-6-P, respectively. We have purified for the first time native GlcNAc-6-P 2-epimerase from bacterial source to apparent homogeneity (1 200 fold) using Butyl-agarose, DEAE-FPLC and Mannose-6-P-agarose chromatography. By SDS/PAGE the pure enzyme showed a molecular mass of 38.4 +/- 0.2 kDa. The maximum activity was achieved at pH 7.8 and 37 degrees C. Under these conditions, the K(m) calculated for GlcNAc-6-P was 1.5 mM. The 2-epimerase activity was activated by Na(+) and inhibited by mannose-6-P but not mannose-1-P. Genetic analysis revealed high homology with bacterial isomerases. GlcNAc-6-P 2-epimerase from E. coli K92 is a ManNAc-inducible protein and is detected from the early logarithmic phase of growth. Our results indicate that, unlike UDP-GlcNAc 2-epimerase, which promotes the biosynthesis of sialic acid, GlcNAc-6-P 2-epimerase plays a catabolic role. When E. coli grows using ManNAc as a carbon source, this enzyme converts the intracellular ManNAc-6-P generated into GlcNAc-6-P, diverting the metabolic flux of ManNAc to GlcNAc.

GlcNAc-6-P 2-epimerase assays.GlcNac-6-P 2-epimerase activity was evaluated using an assay coupled to Neuac synthesis, as previously described (Rodríguez-aparicio et al., 1999).Briefly, the incubation mixture contained the following in a final volume of 50 μl: 0.125 M Tris/HCl pH 7.5, 5 mM GlcNac-6-phosphate; 10 mM MgsO 4 , 10 mM NaCl and protein extract.After incubation at 37 o C for 30 min, the reaction was stopped by cooling on ice.The enzymatically generated ManNac-6-phosphate was dephosphorylated to ManNac and coupled to Neuac synthesis by simultaneous addition of 1 U of alkaline phosphatase, 8.75 mM of pyruvate and 0.05 U of Neuac lyase from C. perfringens (sigma).The final volume of this reaction was 60 μl and after incubation at 37 o C for 30 min (or the required time) the NeuAc synthesized was quantified colorimetrically according to the 2-thiobarbituric acid methodology (Warren, 1959) described elsewhere (Rodríguez-aparicio et al., 1999).
GlcNac-6-phosphate 2-epimerase activity was also evaluated by a radiometric assay using commercial radioactive pyruvate (1 μCi).In this case, the Neuac generated was chromatographed and evaluated by scintillation counting (Rodríguez-aparicio et al., 1987).In all cases, reaction mixtures without GlcNac-6-P or with boiled protein extracts (95 o C for 10 min) were used as controls.
One unit of each enzyme was defined as the amount of 2-epimerase that synthesizes 1 nmol of ManNac-6-phosphate/min at 37 o C under each assay condition.Specific activity (SA) was expressed as units/milligram of protein.
For enzymatic quantification, in all cases control 2-epimerase reaction assays containing heat-denatured protein extracts were used as blanks.In the colorimetric reactions, we also used reaction mixtures without substrate (GlcNac-6-P or pyruvate) as controls.
Proteins were measured by the method of bradford (1976), using bsa as standard.
Purification of GlcNAc-6-P 2-epimerase.E. coli was grown in the ManNAc-Asn-defined chemical medium up to OD 540 = 2.5.bacteria were collected by centrifugation (10 000 × g, 10 min at 2 o C), washed twice with saline solution, and resuspended in 50 mM Tris/HCl buffer, pH 7.6.Cells (20 g wet mass/100 ml of buffer) were subjected to four rounds of sonication (30 s of sonication followed by 30 s on ice in a model B12 Branson:Sonifier) and the homogenate was centrifuged at 17 000 × g, 10 min at 2 o C. The pellet was discarded and the supernatant fluid was treated with 0.75% streptomycin sulphate to remove the nucleic acids.After centrifugation (17 000 × g, 10 min at 2 o C), the supernatant was treated with ammonium sulphate.The protein fraction precipitating at 45% (containing 96% of GlcNAc-6-P 2-epimerase activity) was collected by centrifugation (17 000 × g, 15 min at 2 o C).The pellet was then dissloved in 50 mM Tris/HCl (pH 7.5), and passed through a sephadex G-25 (PD-10) column to eliminate the excess of ammonium sulphate.ammonium sulphate was added to the desalted extract to 35% saturation and applied to a column (10 × 2 cm) of butyl-agarose (from sigma) equilibrated with the same buffer containing ammonium sulphate (25% satn.).The proteins were eluted with 30 ml of a linear gradient of ammonium sulphate (25-0% satn.) in 50 mM Tris/HCl buffer, pH 7.5.Fractions containing maximal enzyme activity were pooled, desalted by passing through a PD-10 column, and subjected to DEaE-FPLC chromatography.Two protein pool fractions from the butyl-agarose chromatography were introduced into an FPLC system (Waters-Millipore) equipped with an anion exchange column (Waters Protein Pack DEAE 8HR 1000 A 8 μm: 10 x 100 mm) equilibrated with 50 mM Tris/HCl buffer, pH 7.5.After a wash with the same buffer, the proteins were eluted with a 0 to 0.18 M NaCl linear gradient and fractions (1 ml) were collected and assayed for GlcNac-6-P 2-epimerase activity.Fractions containing maximal enzyme activity were equilibrated against Tris/HCl buffer 10 mM, pH 8.0, desalted (PD-10 column chromatography) and introduced on a mannose 6-phosphate-agarose column (10 × 5 mm) equilibrated with 10 mM Tris/HCl buffer, pH 8.0.After a wash with 5 ml of the same buffer, GlcNAc-6-P 2-epimerase was eluted with 2 ml of 25 mM Tris/HCl, pH 8.0, containing 0.5 M NaCl.Fractions of 0.5 ml that contained the maximum enzyme activity were tested in a 12% SDS/PAGE system.
Agarose gels were between 0.8% and 1.0% according to the procedure of sambrook et al. (1989).The DNa markers used to estimate the size of unknown DNa were 1kb DNa Plus Ladder (Invitrogen) and Lambda DNa/Hind III (Promega).
Amino acid sequencing.The purified enzyme was desalted by HPLC on a vydac reverse phase 5μm C-4 column (type 214TP54, 0.46 × 25 cm) using a 30 min linear gradient (1 ml/min) from 20 to 95% acetonitrile (0.1% trifluoroacetic acid present).The protein was dissolved in 30 μl of 0.15% trifluoroacetic acid and loaded onto the sequencer.amino acid sequencing was carried out as described elsewhere (Hewick et al., 1981).
Tryptic digestion and mass spectrometry analyses of protein.a gel piece from sDs/PaGE containing purified GlcNAc-6-P 2-epimerase was carefully excised and washed twice in water.Protein was reduced in-gel, alkylated and digested with trypsin according to shevchenko et al. (2000).Briefly, the gel piece was washed twice with water, shrunk for 15 min with 100% acetonitrile and dried in a Savant speedvac for 30 min.Then, the sample was reduced with 10 mM dithioerythritol in 25 mM ammonium bicarbonate for 30 min at 56 o C and subsequently alkylated with 55 mM iodoacetamide in 25 mM ammonium bicarbonate for 20 min in the dark.Finally, the samples were digested with 12.5 ng/μl sequencing grade trypsin (from Roche Molecular biochemicals) in 25 mM ammonium bicarbonate (pH 8.5) overnight at 37 o C.After digestion, the supernatant was collected and 1 μl was spotted onto a MaLDI target plate and allowed to air-dry at room temperature.Then, 0.4 μl of a 3 mg/ml α-cyano-4hydroxy-transcinnamic acid matrix (from sigma) in 50% acetonitrile was added to the dried peptide digest and allowed to air-dry again at room temperature.MaLDI-TOF Ms analyses were performed on a 4 700 Proteomics analyzer MaLDI-TOF/TOF mass spectrometer (applied biosystems, Framingham, Ma, Usa) at the Genomics and Proteomics Centre, Complutense University of Madrid (spain), operated in positive reflector mode, with an accelerating voltage of 20 000 v. all mass spectra were calibrated internally using peptides from the autodigestion of trypsin.analysis by MaLDI-TOF/TOF mass spectrometry affords peptide mass fingerprints and the peptides observed can be collated and represented as a list of monoisotopic molecular masses.For protein identification, the non-redundant NCBI database was searched using MasCOT 1.9 (matrixscience.com)through the Global Protein server v3.5 from applied biosystems.The search parameters were: carbamidomethyl cysteine as the fixed modification and oxidized methionine as the variable modification, a peptide mass tolerance 100 p.p.m. and 1 missed trypsin cleavage site.

M.a. Ferrero and others
In the protein identification, the probability scores were greater than the score fixed as significant, with a p-value of less than 0.05.
PCR analyses and DNA sequencing.The GlcNAc-6-P 2-epimerase gene was amplified from chromosomal DNa of E. coli K92.Amplification reactions were carried out in a Perkins Elmer DNa Thermal Cycler 2400 with primers engineered to in-clude desired restriction sites for cloning into pET-42b(+) expression vector (Novagen).The primers used were GTCaGTTGCGaaTTCCaaaGGaGC (forward) and GGaGaGaCTGCaGCGGGTaaaT-CaG (reverse), which contained an EcoRI and a PstI site, respectively (underlined).PCR reactions were performed under standard conditions.
Specific primers were annealed to chromosomal DNa and incubated with DNa polymerase (Biotools B&M Labs.S.A.) or Pfu DNa polymerase from Pyrococcus furiosus (Promega) over 35 cycles (30 s at 24 o C, 63 s 55 o C and 120 s at 72 o C).PCR amplification products were analyzed by horizontal agarose gel electrophoresis.
After amplification, the PCR product was digested with EcoRI and PstI, and the resulting fragment was cloned into the corresponding sites of the pET-42b(+) expression vector from Novagen according to the manufacturer's guidelines.The resulting plasmid, pETepim, was transformed into E. coli BL21(DE3) (Novagen) and DH5α.
The plasmid sequence was verified and the insert deemed to be free of PCR errors by sequencing at the Laboratorio of Técnicas Instrumentales (University of León, León, spain) using DYEnamic™ ET Dye Terminator Kit (amersham biosciences Part of GE Healthcare) according to the sanger method (sanger et al., 1977).

Time course of the appearance of GlcNAc-6-P 2-epimerase during culture
The GlcNac-6-P 2-epimerase from E. coli K92 began to be synthesized from the early logarithmic phase of growth when this bacterium was incubated at 37 o C in both Xyl/asn, an ideal medium for polysialic acid (Pa) production (González-Clemente et al.,

OD
GlcNAc-6-P 2-epimerase from Escherichia coli K92 1990), and ManNac/asn, which is ideal for induction of the specific ManNAc transport system (Revilla-Nuín et al., 1999).The level of the enzyme increased linearly during the early logarithmic phase (Fig. 1a), a maximum being reached after 10-11 h (in Xyl-Asn medium) or 29-31 h (ManNac-asn medium) of incubation.Thereafter it decreased continuously.This kind of kinetic behaviour, which parallels that of bacterial growth, is very similar to that observed for Neuac lyase (Ferrero et al., 1996), CMP-synthetase (González-Clemente et al., 1989) and sialyltransferase (Ortíz et al., 1989), other enzymes directly involved in the metabolism of Pa by E. coli.a comparative analysis of such kinetic behaviour revealed that the level of the enzyme was significatively higher (1.9 fold) when E. coli K92 was grown in ManNac-asn medium.However, as previously described (Revilla-Nuín et al., 1998b), a dramatic decrease in Pa production was recorded (90%) (see Fig. 1b).These results confirm the notion that ManNac is involved in the regulation of capsular polysialic acid biosynthesis (Revilla-Nuín et al., 1998b) and suggest that GlcNac-6-P 2-epimerase from E. coli K92 is a ManNac-inducible enzyme.Moreover, the fact that ManNac could be the molecule that induces the ManNac-phosphotransferase transport system (ManNac-PTs) in this bacterium (Revilla-Nuín et al., 1999), implicates GlcNac-6-P 2epimerase in the metabolism of ManNac and Glc-Nac when these sugars are used as carbon sources (see below).

Enzyme purification
a pure and active GlcNac-6-P 2-epimerase preparation was obtained from E. coli K92 extracts growth in ManNAc-Asn medium.The purification steps included butyl-agarose chromatography and anion exchange on DEaE-FPLC (Fig. 2) followed by affinity chromatography on mannose-6-P-agarose.The purification scheme is summarized in Table 1.
In the purification process it was necessary to pool two protein fractions from the butyl-agarose elution procedure to perform the DEaE-FPLC chromatography.Moreover, exhaustive washes in all chromatographic procedures used were crucial for obtaining good purification results.Although after DEAEchromatography we observed a high level of purification, it was necessary to use the affinity column (mannose-6P-agarose) to obtain a single, sharp electrophoretic (sDs/PaGE) protein band (Fig. 3).Under these conditions (Table 1), the final preparation was 1200-fold enriched in enzyme activity, with an overall recovery of 10%.Moreover, the use of other gelfiltration, ionic and hydrophobic chromatographic techniques gave no further increases in the specific activity yield of this enzyme (results not show).

Molecular mass determinations
analysis of pure GlcNac-6-P 2-epimerase from E. coli K92 by SDS/12%-PAGE revealed a single band with a molecular mass of 38.4±0.2 kDa (Fig. 3).This result differs from the expected for a pro-  tein encoded by nanE, the gene whose product has 28 kDa and which in E. coli has been related to Glc-Nac-6-P 2-epimerase activity (Plumbridge & Vimr, 1999;Kalivoda et al., 2003).The electrophoretic mass observed is more similar to the protein encoded by siaA (40-41 kDa), the gene that has been proposed to code for GlcNac-6-P 2-epimerase in Neissera meningitidis (Petersen et al., 2000).Different transcriptional origins or post-translational modifications could account for these molecular mass differences.
The molecular mass of native GlcNac-6-P 2-epimerase from E. coli K92 was determined by gel-filtration chromatography in shephacryl s-200 (using 10 mM Tris/HCl, pH 8.0, as buffer) to be 76 ± 2 kDa.This finding suggests that the active epimerase from E. coli K92 is a dimeric protein with two similar chains.

Physicochemical characteristics and kinetic properties
To establish the best conditions for the enzymatic assays, we studied the effect of pH and temperature on the GlcNac-6-P 2-epimerase reac-  tion.The highest enzymatic activity was observed at pH 7.8 and a temperature of 37 o C (Fig. 4).These results are similar to those previously described for GlcNac-6-P 2-epimerase from E. coli K1 using partially purified protein extracts (Rodríguez-aparicio et al., 1999) and are in the optimum range of Pa production (González-Clemente et al., 1990).a kinetic analysis revealed that under these conditions product generation by GlcNac-6-P 2-epimerase was linear up to 60 min (Fig. 4).Moreover, the enzyme activity displayed a hyperbolic type of behaviour for GlcNac-6-P (Fig. 4); the K m calculated for the substrate was 1.5 mM.This K m value, similar to that described for the E. coli K1 enzyme (Rodríguez-aparicio et al., 1999), is in the range of other proteins involved in PA metabolism (Grosh & Roseman, 1965;Ortíz et al., 1989;Ferrero et al., 1996;bravo et al., 2001).

Effect of cations and other molecules
The effect of different cations (5 mM chloride form) on GlcNac-6-P 2-epimerase activity from E. coli K92 was studied.Whereas the addition of monovalent cations such as K + , Li + and Rb + to the reaction mixture had no effect, the presence of Na + caused a significant increase in epimerase activity, which was optimum at 10 mM (145%).This positive effect accounts for the presence of the Na + cation in the reaction mixture (see Material and Methods).
However, the addition of divalent cations to the reaction mixtures, such as Hg 2+ , Cu 2+ , Zn 2+ , Co 2+ , Mg 2+ and Mn 2+ (5 mM) strongly inhibited the enzyme activity (by 95, 70, 85, 72, 53 and 65%).The inhibition caused by the heavy metal ions suggests that, as in other proteins (Martínezblanco et al., 1990;Ferrero et al., 1996;Walters et al., 1999;bravo et al., 2001;solana et al., 2001), the presence of sH groups is essential for catalytic activity.This suggestion is also supported by the strong inhibition (96%) of epimerase activity caused by p-chloromercuribenzoate (1 mM) when added to the epimerase reaction mixture.The significant inhibition caused by Mg 2+ , Mn 2+ and Co 2+ on the GlcNac-6-P 2-epimerase activity may occur through a similar effect and could be related to the specific characteristics of the epimerase reaction.as previously described (Rodríguez-aparicio et al., 1992), the specific effect caused by Mg 2+ on CMP-sialic acid synthetase from rat liver (another enzyme involved in eukaryotic sialic acid metabolism) can be mimicked by Mn 2+ and Co 2+ .Moreover, the inhibitory effect caused by these cations can be related to the strong decrease in Pa production detected (80%) when E. coli is grown in the presence of 10 mM Mn 2+ (Ferrero et al., 1996).Finally, we observed that GlcNac-6-P 2-epimerase from E. coli K92 was inhibited by ammonium sulphate, the salt used during the enzyme purification (see above).A concentration of 1.0 M (25% satn.)caused a 64% inhibition.This effect, although reversible by desalting the enzyme, requires careful elimination of the ammonium sulphate present in the samples to achieve a good quantification of 2epimerase activity.

Substrate specificity and effect of different sugars and sugar derivatives
GlcNac-6-P 2-epimerase from E. coli K92 showed high specificity for GlcNAc-6-P since this phosphoamino sugar could not be replaced by Glc-Nac or UDP-GlcNac (not shown), another two intimately related amino sugars that can be used as substrates for the enzymatic synthesis of ManNac (see Introduction).Furthermore, when we used boiled protein extracts in the reaction mixtures no epimerase activity was observed with any of these sugars, indicating that under our assay conditions no enzymatic hydrolysis or epimerization takes place.When other sugars and sugar derivatives were added to the reaction mixture in the presence of GlcNac-6-P, we observed that monosaccharides (glucose, fructose, mannose, galactose, xylose, arabinose, ribose and sorbose), disaccharides (lactose, melibiose, maltose or sucrose) and sugar derivatives (galactitol, mannitol and sorbitol), or the Nacetyl derivatives N-acetylgalactosamine, GlcNac, N-acetylneuraminic acid and N-acetylglucosaminitol, or the hexosamines glucosamine and galactosamine did not affect GlcNAc-6-P 2-epimerase activity, even at a concentration of 5 mM.However, the addition of mannose-6-P, but not mannose-1-P, to the reaction mixture caused a marked inhibition of enzyme activity.a concentration of mannose-6-P of 5 mM, similar to that of GlcNac-6-P used in the epimerase reaction assays (see Materials and Methods), caused a 50% inhibition.These results suggest that in vivo GlcNAc-6-P 2-epimerase activity may be specifically regulated by variations in the mannose-6-P intracellular pool.

GlcNAc-6-P 2-epimerase and sialic acid metabolism
as shown in Fig. 1, when E. coli K92 used ManNac as a carbon source the amount of GlcNac-6-P 2-epimerase increased up to 1.9 fold.This indicates that our enzyme is directly involved in the intracellular turnover of ManNac, the amino sugar precursor of sialic acid for capsular Pa synthesis.
Different authors have reported that the Man-Nac used by E. coli for sialic and polysialic acid biosynthesis is generated by UDP-GlcNac 2-epimerase (Ringenberg et al., 2003;vann et al., 2004;Tanner, 2005) and that the phosphorylated form of Man-Nac is not involved in this process (Ringenberg et al., 2003).Moreover, the existence in E. coli K92 of a ManNac-inducible GlcNac-6-P 2-epimerase and the fact that when this bacterium grew using Man-Nac as a carbon source a dramatic decrease in Pa production took place (Fig. 1) suggest that, in this bacterium, two different proteins could be involved in sialic acid metabolism: UDP-GlcNac 2-epimerase for the biosynthetic function and the GlcNac-6-P 2epimerase described in this work exerting a catabolic role.Thus, when E. coli K92 uptakes ManNac from the growth medium, the specific and ManNAc-inducible PTs generates intracellular ManNac-6-P (Revilla-Nuín et al., 1998b), which is used by the GlcNac-6-P 2-epimerase to generate GlcNac-6-P.accordingly, the existence of this enzyme, now characterized, may permit this bacterium to use extracellular Man-Nac as a carbon source (Revilla-Nuín et al., 1998b;1999;2002;Ringenberg et al., 2003).Furthermore, it is very likely that the catabolism of ManNac would take place via GlcNac since, as described previously (Revilla-Nuín et al., 1998b;1999;Ezquerro-sáenz et al., 2006), when E. coli K92 uses ManNac, the intracellular pool of ManNac-6-P and GlcNac-6-P is increased.also in favour of this hypothesis is the fact that GlcN-6-P deaminase, an enzyme of the GlcNac metabolism, is also induced by ManNac (biswas et al., 1979).Finally, ManNac induces both Man-Nac-PTs (Revilla-Nuín et al., 1999) and GlcNac-6-P 2-epimerase expression (Fig. 1), suggesting that, as in ManNac catabolism, the genes that encode these proteins form part of a regulon that responds to the same regulation signals.

Genetic study
Different authors have proposed nanE as a putative bacterial gene that codes for GlcNac-6-P 2-epimerase (Plumbridge & Vimr, 1999;Walters et al., 1999;Kalivoda et al., 2003;Ringenberg et al., 2003).To establish whether GlcNac-6-P 2-epimerase from E. coli K92 is indeed codified for this gene, we analyzed the amino-acid sequence of the purified protein.Unfortunately, attempts to sequence the purified GlcNAc-6-P 2-epimerase by N-terminal sequence analysis failed because of amino-terminal blockage.We therefore undertook sequencing by tryptic digestion and mass spectrometry analysis (see Materials and Methods).Table 2 shows the amino-acid sequences of nine peptides obtained from purified GlcNAc-6-P 2-epimerase by tryptic digestion that were compared in a search in a non-redundant NCbI protein sequence database.
a comparative study of these tryptic peptide sequences with a non-redundant Genbank CDs Database (NCbI-bLasT) revealed the existence of a high homology with putative isomerases from different sources (Hayashi et al., 2001;Perna et al., 2001;Jin et al., 2002;Yang et al., 2005) and the pgl gene of E. coli K12 (Thomason et al., 2004) although, surprisingly, not with the epimerases previously related (see above).as shown in Table 3 and Fig. 6, all tryptic peptides obtained showed a complete sequence homology with the recently identified pgl gene from E. coli K12 and isomerases from different GlcNAc-6-P 2-epimerase from Escherichia coli K92 E. coli strains and Shigella species.Moreover, in all cases the molecular mass calculated for the respective expected protein products (36-36.7 kDa) were similar to that of the GlcNac-6-P 2-epimerase purified by us (Table 3 and Fig. 5).However, non-homology was observed when we compared these tryptic peptides sequences with the protein products of nanE and siaA, the two different genes that have been proposed to encode GlcNac-6-P 2-epimerase in bacteria (Plumbridge & Vimr, 1999;Walters et al., 1999;Petersen et al., 2000;Kalivoda et al., 2003;Ringenberg et al., 2003) (not shown).
To establish the identity of the gene that codes for the GlcNac-6-P 2-epimerase from E. coli K92, and using the information obtained from the comparative sequencing studies, we designed different primers that permitted us to amplify, by PCR, a fragment of 1 kbp from E. coli K92 genomic DNa (see Materials and Methods).Cloning and sequencing of this fragment revealed an ORF of 996 nucleotides encoding 331 amino acids (Fig. 5), whose calculated mass (38.4 ± 0.2 kDa) was similar to that obtained by electrophoretic analysis of the GlcNac-6-P 2-epimerase protein purified by us (Fig. 3).A comparison of the amino-acid sequence of this protein with those found in the Genbank CDs Database using blastP (NCbI) program revealed very high homology with the isomerases from E. coli (strain O157: H7) (100%), different Shigella species (98-99%) and the 6-phosphogluconolactonase (pgl gene) from E. coli K12 (Fig. 6).However, no isomerase (evaluated as GlcNac-6-P 2-epimerase) or 6-phosphogluconolactonase activities were detected in protein extracts from E. coli BL21 (DE3) and DH5α transformed with pETepim and induced with IPTG (see Materials and Methods).Moreover, and similar to the comparative tryptic studies, a low degree of amino acid homology was observed with the products of nanE and siaA (12% and 11%, respectively) and even with the related UDP-GlcNac 2-epimerase encoded by the neuC gene (10%), the enzyme that generates ManNAc in polysialic acid biosynthesis pathways (Ringenberg et al., 2003;Murkin et al., 2004;vann et al., 2004).Nevertheless, the fact that by gel-filtration chromatography the native protein (detected by epimerase activity) had a molecular mass of 74-78 kDa -twofold the 38 kDa obtained from sDs/PaGE (denaturing conditions) -precludes the possibility that in the protein amino-acid analysis, a contaminant protein could have been sequenced.If the epimerase purified by us were a product of nanE, the expected molecular mass should be 2 × 28 kDa.However, the molecular mass determined for our native protein was 74-78 kDa.The results obtained from tryptic digestion and peptide sequence analysis revealed the existence of a unique polypeptide chain sequence that showed very high homology with putative isomerases from different bacterial strains.At this juncture, it is also surprising that this protein sequence exhibited high homology with the gene that encodes 6phosphogluconolactonase in E. coli K12 (Table 3 and Fig. 6) i.e., a protein involved in the pentose-P pathway (Thomason et al., 2004).This homology again throws doubt on our results that relate the GlcNac-6-P 2-epimerase with this genetic sequence.However, as reported by Thomason and coworkers in the same work (2004), this genetic sequence product could correspond to a bifunctional protein and, as in Thermoanaerobacter tengcongensis and Vibrio vulnificus CMCP6, could show lactonase and isomerase activity.although nanE has been described in E. coli K12 and, by molecular genetic studies, it has been related to GlcNac-6-P 2-epimerase (Ringenberg et al., 2003), ManNac is metabolized by this bacterium very slowly and a previous mutation is required to derepress the gene (Plumbridge & Vimr, 1999).Nevertheless, E. coli K92 grows very efficiently using ManNAc as a carbon source.as shown in Fig. 1, the use of

GlcNAc-6-P 2-epimerase from Escherichia coli K92
Xyl or ManNac as a source of carbon did not elicit significant delays in growth, suggesting that unlike E. coli K12, in E. coli K92 the genetic derepression through mutation is not necessary for the growth of this bacterium on ManNac.Moreover, as previously described (Plumbridge & Vimr, 1999), for E. coli K12 to grow using ManNac it is also necessary that the multispecific transport system that uptakes this amino sugar in this bacterium (ManXYZ) should be derepressed by an mlc mutation.In E. coli K92 this previous gene derepression is not necessary since in this bacterium the uptake of ManNac is carried out by a different transport system that is very specific, induced by ManNAc-6-P, and much more efficient than the ManXYZ transporter (Revilla-Nuín et al., 1999;2002).All these metabolic differences support the idea that, as proposed, the protein purified by us would be the putative isomerase reported (Table 3) and that it would be responsible for the GlcNac-6-P 2-epimerase activity in E. coli K92.although in E. coli K92 we have observed the presence of nanE (we detected it by PCR analysis, results not shown), it is possible that this gene might be repressed, as in E. coli K12, and that both the specific ManNAc-PTS transport system and the epimerase related in this work could be responsible for the efficient growth of this bacterium when it uses ManNac as a carbon source.Thus, E. coli K92 has evolved by developing a new strategy that permits good cellular growth using ManNac as a carbon source without having to depend on previous mutations.Expression studies by mRNA quantification and by direct reaction of the respective protein products with specific antibodies, analysis of enzymatic activities, and the use of knockout bacteria in these genes should provide key data to confirm these hypotheses and to establish the exact role these genes play in bacterial amino sugar metabolism.Further research in this topic is currently in progress.

Figure 5 .
Figure 5.The nucleotide and translated sequence of the PCR-amplified fragment from E. coli K92 genomic DNA.The upper row shows the nucleotide sequence and the second row shows the amino-acid sequence.The nucleotides underlined show the PCR amplification primers (the reverse primer corresponding to the inverse and complementary chain) and the bold nucleotides corresponding to EcoRI and PstI restriction sites.The underlined amino acids show the tryptic peptide sequences of the protein from E. coli K92.

Table 1 . Purification of GlcNAc-6-P 2-epimerase from E. coli K92
One unit is the amount of enzyme which synthesizes 1 nmol of ManNac-6-P per 30 min at 37 o C. a For DEaE-FPLC chromatography two protein pool fractions from butyl-agarose were used (see Material and Methods section). *