Molecular cloning of the Escherichia coli B L-fucose-D-arabinose gene cluster.

To metabolize the uncommon pentose D-arabinose, enteric bacteria often recruit the enzymes of the L-fucose pathway by a regulatory mutation. However, Escherichia coli B can grow on D-arabinose without the requirement of a mutation, using some of the L-fucose enzymes and a D-ribulokinase that is distinct from the L-fuculokinase of the L-fucose pathway. To study this naturally occurring D-arabinose pathway, we cloned and partially characterized the E. coli B L-fucose-D-arabinose gene cluster and compared it with the L-fucose gene cluster of E. coli K-12. The order of the fucA, -P, -I, and -K genes was the same in the two E. coli strains. However, the E. coli B gene cluster contained a 5.2-kb segment located between the fucA and fucP genes that was not present in E. coli K-12. This segment carried the darK gene, which encodes the D-ribulokinase needed for growth on D-arabinose by E. coli B. The darK gene was not homologous with any of the L-fucose genes or with chromosomal DNA from other D-arabinose-utilizing bacteria. D-Ribulokinase and L-fuculokinase were purified to apparent homogeneity and partially characterized. The molecular weights, substrate specificities, and kinetic parameters of these two enzymes were very dissimilar, which together with DNA hybridization analysis, suggested that these enzymes are not related. D-Arabinose metabolism by E. coli B appears to be the result of acquisitive evolution, but the source of the darK gene has not been determined.

Isolation and characterization of Escherichia coli mutants able to utilize the novel pentose L-ribose.

Wild-type strains of Escherichia coli were unable to utilize L-ribose for growth. However, L-ribose-positive mutants could be isolated from strains of E. coli K-12 which contained a ribitol operon. L-ribose-positive strains of E. coli, isolated after 15 to 20 days, had a growth rate of 0.22 generation per h on L-ribose. Growth on L-ribose was found to induce the enzymes of the L-arabinose and ribitol pathways, but only ribitol-negative mutants derived from strains originally L-ribose positive lost the ability to grow on L-ribose, showing that a functional ribitol pathway was required. One of the mutations permitting growth on L-ribose enabled the mutants to produce constitutively an NADPH-linked reductase which converted L-ribose to ribitol. L-ribose is not metabolized by an isomerization to L-ribulose, as would be predicted on the basis of other pentose pathways in enteric bacteria. Instead, L-ribose was metabolized by the reduction of L-ribose to ribitol, followed by the conversion to D-ribulose by enzymes of the ribitol pathway.

D-arabinose metabolism in Escherichia coli B: induction and cotransductional mapping of the L-fucose-D-arabinose pathway enzymes.

D-Arabinose is degraded by Escherichia coli B via some of the L-fucose pathway enzymes and a D-ribulokinase which is distinct from the L-fuculokinase of the L-fucose pathway. We found that L-fucose and D-arabinose acted as the apparent inducers of the enzymes needed for their degradation. These enzymes, including D-ribulokinase, appeared to be coordinately regulated, and mutants which constitutively synthesized the L-fucose enzymes also constitutively synthesized D-ribulokinase. In contrast to D-arabinose-positive mutants of E. coli K-12, in which L-fuculose-1-phosphate and D-ribulose-1-phosphate act as inducers of the L-fucose pathway, we found that these intermediates did not act as inducers in E. coli B. To further characterize the E. coli B system, some of the L-fucose-D-arabinose genes were mapped by using bacteriophage P1 transduction. A transposon Tn10 insertion near the E. coli B L-fucose regulon was used in two- and three-factor reciprocal crosses. The gene encoding D-ribulokinase, designated darK, was found to map within the L-fucose regulon, and the partial gene order was found to be Tn10-fucA-darK-fucI-fucK-thyA.

Isolation of a mutation resulting in constitutive synthesis of L-fucose catabolic enzymes.

A ribitol-positive transductant of Escherichia coli K-12, JM2112, was used to facilitate the isolation and identification of mutations affecting the L-fucose catabolic pathway. Analysis of L-fucose-negative mutants of JM2112 enabled us to confirm that L-fucose-1-phosphate is the apparent inducer of the fucose catabolic enzymes. Plating of an L-fuculokinase-negative mutant of JM2112 on D-arabinose yielded an isolate containing a second fucose mutation which resulted in the constitutive synthesis of L-fucose permease, isomerase, and kinase. This constitutive mutation differs from the constitutive mutation described by Chen et al. (J. Bacteriol. 159:725-729, 1984) in that it is tightly linked to the fucose genes and appears to be located in the gene believed to code for the positive activator of the L-fucose genes.

Construction of an improved D-arabinose pathway in Escherichia coli K-12.

A ribitol catabolic pathway was transduced into Escherichia coli K-12 in an effort to determine whether the ribitol pathway would confer an advantage to D-arabinose-positive mutants growing on D-arabinose as the sole carbon source. Competition studies in chemostats showed that ribitol-positive strains, with a selection coefficient of 9%/h, have a significant competitive advantage over ribitol-negative strains. Ribitol-positive strains grown in batch culture also exhibited a shorter lag period than did ribitol-negative strains when transferred from glucose to D-arabinose. Repeated transfer of a ribitol-positive strain of E. coli K-12 on D-arabinose yielded a strain with further improved growth on D-arabinose. This "evolved" strain was found to constitutively synthesize L-fucose permease, isomerase, and kinase but had lost the ability to grow on L-fucose, apparently owing to the loss of a functional aldolase. This constitutive mutation is not linked to the fucose gene cluster and may be similar to an unlinked constitutive mutation described by Chen et al. (J. Bacteriol. 159:725-729, 1984).

Inducible xylitol dehydrogenases in enteric bacteria.

Morganella morganii ATCC 25829, Providencia stuartii ATCC 25827, Serratia marcescens ATCC 13880, and Erwinia sp. strain 4D2P were found to induce a xylitol dehydrogenase when grown on a xylitol-containing medium. The xylitol dehydrogenases were partially purified from the four strains, and those from M. morganii ATCC 25829, P. stuartii ATCC 25827, and S. marcescens ATCC 13880 were all found to oxidize xylitol to D-xylulose. These three enzymes had KmS for xylitol of 7.1 to 16.4 mM and molecular weights ranging from 130,000 to 155,000. In contrast, the xylitol dehydrogenase from Erwinia sp. strain 4D2P oxidized xylitol at the C-4 position to produce L-xylulose, had a Km for xylitol of 72 mM, and had a molecular weight of 102,000.

Characterization of xylitol-utilizing mutants of Erwinia uredovora.

Of the four pentitols ribitol, xylitol, D-arabitol, and L-arabitol, Erwinia uredovora was able to utilize only D-arabitol as a carbon and energy source. Although attempts to isolate ribitol- or L-arabitol-utilizing mutants were unsuccessful, mutants able to grow on xylitol were isolated at a frequency of 9 X 10(-8). Xylitol-positive mutants constitutively synthesized both a novel NAD-dependent xylitol-4-dehydrogenase, which oxidized xylitol to L-xylulose, and an L-xylulokinase. The xylitol dehydrogenase had a Km for xylitol of 48 mM and showed best activity with xylitol and D-threitol as substrates. However, D-threitol was not a growth substrate for E. uredovora, and its presence did not induce either dehydrogenase or kinase activity. Attempts to determine the origin of the xylitol catabolic enzymes were unsuccessful; neither enzyme was induced on any growth substrate or in the presence of any polyol tested. Analysis of xylitol-negative mutants isolated after Tn5 mutagenesis suggested that the xylitol dehydrogenase and the L-xylulokinase structural genes were components of two separate operons but were under common regulatory control.

Production of D- and L-xylulose by mutants of Klebsiella pneumoniae and Erwinia uredovora.

D-Xylulose and L-xylulose were produced biologically by the oxidation of a corresponding pentitol. A Klebsiella pneumoniae mutant was constructed for the oxidation of D-arabitol to D-xylulose. This mutant constitutively synthesized the D-arabitol permease system and D-arabitol dehydrogenase but was unable to produce the D-xylulokinase of the D-arabitol pathway or the D-xylose isomerase and D-xylulokinase of the D-xylose pathway. An Erwinia uredovora mutant which constitutively synthesized a novel xylitol-4-dehydrogenase but could not synthesize L-xylulokinase was used for the oxidation of xylitol to L-xylulose. Washed cell suspensions of either mutant incubated with 0.5% pentitol would oxidize 60 to 65% of the pentitol to the corresponding ketopentose in 18 h and excrete the ketopentose into the medium. Ketopentoses were rapidly purified from the remaining pentitol by hydroxyl affinity chromatography.

Directed evolution of a second xylitol catabolic pathway in Klebsiella pneumoniae.

Klebsiella pneumoniae PRL-R3 has inducible catabolic pathways for the degradation of ribitol and D-arabitol but cannot utilize xylitol as a growth substrate. A mutation in the rbtB regulatory gene of the ribitol operon permits the constitutive synthesis of the ribitol catabolic enzymes and allows growth on xylitol. The evolved xylitol catabolic pathway consists of an induced D-arabitol permease system that also transports xylitol, a constitutively synthesized ribitol dehydrogenase that oxidizes xylitol at the C-2 position to produce D-xylulose, and an induced D-xylulokinase from either the D-arabitol or D-xylose catabolic pathway. To investigate the potential of K. pneumoniae to evolve a different xylitol catabolic pathway, strains were constructed which were unable to synthesize ribitol dehydrogenase or either type of D-xylulokinase but constitutively synthesized the D-arabitol permease system. These strains had an inducible L-xylulokinase; therefore, the evolution of an enzyme which oxidized xylitol at the C-4 position to L-xylulose would establish a new xylitol catabolic pathway. Four independent xylitol-utilizing mutants were isolated, each of which had evolved a xylitol-4-dehydrogenase activity. The four dehydrogenases appeared to be identical because they comigrated during nondenaturing polyacrylamide gel electrophoresis. This novel xylitol dehydrogenase was constitutively synthesized, whereas L-xylulokinase remained inducible. Transductional analysis showed that the evolved dehydrogenase was not an altered ribitol or D-arabitol dehydrogenase and that the evolved dehydrogenase structural gene was not linked to the pentitol gene cluster. This evolved dehydrogenase had the highest activity with xylitol as a substrate, a Km for xylitol of 1.4 M, and a molecular weight of 43,000.

Biosynthesis and catabolism of 6-deoxy L-talitol by Klebsiella aerogenes mutants.

A mutant strain of Klebsiella aerogenes was constructed and, when incubated anaerobically with L-fucose and glycerol, synthesized and excreted a novel methyl pentitol, 6-deoxy L-talitol. The mutant was constitutive for the synthesis of L-fucose isomerase but unable to synthesize L-fuculokinase activity. Thus, it could convert the L-fucose to L-fuculose but was incapable of phosphorylating L-fuculose to L-fuculose 1-phosphate. The mutant was also constitutive for the synthesis of ribitol dehydrogenase, and in the presence of sufficient reducing power this latter enzyme catalyzed the reduction of the L-fuculose to 6-deoxy L-talitol. The reducing equivalents required for this reaction were generated by the oxidation of glycerol to dihydroxyacetone with an anaerobic glycerol dehydrogenase. The parent strain of K. aerogenes was unable to utilize the purified 6-deoxy L-talitol as a sole source of carbon and energy for growth; however, mutant could be isolated which had gained this ability. Such mutants were found to be constitutive for the synthesis of ribitol dehydrogenase and were thus capable of oxidizing 6-deoxy L-talitol to L-fuculose. Further metabolism of L-fuculose was shown by mutant analysis to be mediated by the enzymes of the L-fucose catabolic pathway.

Utilization of D-xylose by wild-type strains of Salmonella typhimurium.

Enzyme studies of strains of Salmonella typhimurium representing biotypes that utilized D-xylose rapidly (xylose strong) or slowly (xylose weak) showed that they were different in the utilization of D-xylose because the xylose-weak strains were deficient in the transport of D-xylose. This observation is consistent with the idea that strains of the different xylose-weak biotypes, e.g. biotypes 17 to 32, were descended from strains of xylose-strong types, particularly from biotype 1.

A comparison of alternate metabolic strategies for the utilization of D-arabinose.

Mutants of Klebsiella aerogenes W70 that metabolize the uncommon pentose D-arabinose were isolated. These mutants were found to be either constitutive or indicible by D-arabinose for the synthesis of enzymes in the L-fucose pathway. Such mutants could then utilize L-fucose isomerase to convert the structurally similar D-arabinose molecule to D-ribulose. D-Ribulose is an intermediate and the inducer of an existing ribitol pathway and could thus be metabolized. In those D-arabinose-positive mutants where the ribitol pathway was blocked by mutation, D-ribulose could alternatively be metabolized by using the remaining L-fucose pathway enzymes. When the two D-arabinose catabolic routes were compared, catabolism of D-arabinose via the ribitol pathway was found to be more efficient. Catabolism of D-arabinose using the L-fucose pathway permitted D-ribulose to escape into the media and produced an unmetabolizable end product, L-glycolic acid. A comparison of growth using constitutive versus inducible control of the borrowed L-fucose isomerase did not reveal an advantage for one control type over the other. Several differences were observed, however, when we determined the degree to which these control mutations perturbed the normal functioning of the L-fucose and associated pathways. Growth of the constitutive mutant was impaired with L-fucose as substrate. The inducible-control mutant had altered growth characteristics on ribitol and L-rhamnose.

Growth of Klebsiella aerogenes on xylitol: implications for bacterial enzyme evolution.

When Klebsiella aerogenes was grown in continuous culture with xylitol. an unnatural pentitol, as the growth limiting substrate, the structural gene which codes for ribitol dehydrogenase, an enzyme which gratuitously catalyzes the oxidation of xylitol to D-xylulose, was duplicated. It appears that the duplication mechansim only duplicates the gene which is subjected to selective pressure and not any of the other closely linked genes. The degree to which the ribitol dehydrogenase gene is duplicated does not appear to be strictly correlated with the ability to grow faster on xylitol. Duplication mutants do, in fact, grow faster than their parent strain, but when challenged to grow at even higher growth rates there is a catabolic repression of enzyme activity. Thus a situation is created in which a structural gene is duplicated in response to selective pressure; these mutants can grow faster on the new substrate, but faster growth results in a "silencing" of a portion of the genes by catabolite repression.

Natural and altered induction of the L-fucose catabolic enzymes in Klebsiella aerogenes.

Mutants of Klebsiella aerogenes W70 were isolated that had gained the ability to utilize the uncommon pentose D-arabinose as their sole source of carbon and energy. In contrast to the D-arabinose-negative, parent strain, these mutants were found to be either constitutive for certain enzymes of the L-fucose catabolic pathway or inducible for such enzymes when incubated in the presence of D-arabinose. The mutants used L-fucose isomerase to convert D-arabinose to D-ribulose, which is an intermediate and inducer of the ribitol catabolic pathway. The D-ribulokinase of the ribitol pathway was then induced. This enzyme catalyzed the phosphorylation of D-ribulose at the 5-carbon position. Mutants that were negative for D-ribulokinase could still dissimilate D-arabinose slowly by using all three enzymes, the isomerase, kinase, and aldolase, of the L-fucose pathway. Using condition negative mutants, we were able to demonstrate that the natural induction of the L-fucose pathway enzymes by L-fucose required the activity of a functional L-fucose isomerase and a functional L-fuculokinase but not an L-fuculose-1-phosphate aldolase. A metabolic intermediate, L-fuculose-1-phosphate, was thereby shown to be a probable inducer of at least the isomerase and kinase of the L-fucose catabolic pathway. Similar experiments, with D-arabinose-positive mutants, which were induced for the L-fucose pathway enzymes upon incubation with D-arabinose, revealed that the activities of the L-fucose isomerase and the L-fuculokinase were also required for the induction of the L-fucose enzymes. These D-arabinose-positive mutants apparently produced an altered regulatory protein that accepted both L-fuculose-1-phosphate and D-ribulose-1-phosphate as inducers. Examination of constitutive mutants revealed that L-fucose isomerase and L-fuculokinase were both synthesized constitutively, with the aldolase apparently under separate control.

Ribitol catabolic pathway in Klebsiella aerogenes.

In Klebsiella aerogenes W70, there is an inducible pathway for the catabolism of ribitol consisting of at least two enzymes, ribitol dehydrogenase (RDH) and d-ribulokinase (DRK). These two enzymes are coordinately controlled and induced in response to d-ribulose, an intermediate of the pathway. Whereas wild-type K. aerogenes W70 are unable to utilize xylitol as a carbon and energy source, mutants constitutive for the ribitol pathway are able to utilize RDH to oxidize the unusual pentitol, xylitol, to d-xylulose. These mutants are able to grow on xylitol, presumably by utilization of the d-xylulose produced. Mutants constitutive for l-fucose isomerase can utilize the isomerase to convert d-arabinose to d-ribulose. In the presence of d-ribulose, RDH and DRK are induced, and such mutants are thus able to phosphorylate the d-ribulose by using the DRK of the ribitol pathway. Derivatives of an l-fucose isomerase-constitutive mutant were plated on d-arabinose, ribitol, and xylitol to select and identify mutations in the ribitol pathway. Using the transducing phage PW52, we were able to demonstrate genetic linkage of the loci involved. Three-point crosses, using constitutive mutants as donors and RDH(-), DRK(-) double mutants as recipients and selecting for DRK(+) transductants on d-arabinose, resulted in DRK(+)RDH(+)-constitutive, DRK(+)RDH(+)-inducible, and DRK(+)RDH(-)-inducible transductants but no detectable DRK(+)RDH(-) constitutive transductants, data consistent with the order rbtC-rbtD-rbtK, where rbtC is a control site and rbtD and rbtK correspond to the sites for the sites for the enzymes RDH and DRK, respectively.