One step in de novo pyridoxine (vitamin B 6 ) and pyridoxal 5-phosphate biosynthesis was predicted to be an oxidation catalyzed by an unidentified D-erythrose-4-phosphate dehydrogenase (E4PDH). To help identify this E4PDH, we purified the Escherichia coli K-12 gapA-and gapB-encoded dehydrogenases to homogeneity and tested whether either uses D-erythrose-4-phosphate (E4P) as a substrate. gapA (gap1) encodes the major D-glyceraldehyde-3-phosphate dehydrogenase (GA3PDH). The function of gapB (gap2) is unknown, although it was suggested that gapB encodes a second form of GA3PDH or is a cryptic gene. We found that the gapBencoded enzyme is indeed an E4PDH and not a second GA3PDH, whereas gapA-encoded GA3PDH used E4P poorly, if at all, as a substrate under the in vitro reaction conditions used in this study. The amino terminus of purified E4PDH matched the sequence predicted from the gapB DNA sequence. Purified E4PDH was a heat-stable tetramer with a native molecular mass of 132 kDa. ؊1 in steady-state reactions in which NADH formation was determined. From specific activities in crude extracts, we estimated that there are at least 940 E4PDH tetramer molecules per bacterium growing in minimal salts medium plus glucose at 37؇C. Thin-layer chromatography confirmed that the product of the E4PDH reaction was likely the aldonic acid 4-phosphoerythronate. To establish a possible role of E4PDH in pyridoxal 5-phosphate biosynthesis, we showed that 4-phosphoerythronate is a likely substrate for the 2-hydroxy-acid dehydrogenase encoded by the pdxB gene. Implications of these findings in the evolution of GA3PDHs are also discussed. On the basis of these results, we propose renaming gapB as epd (for D-erythrose-4-phosphate dehydrogenase).

Flagellar biosynthesis in Salmonella typhimurium and Escherichia coli represents the best-understood system in which a large regulatory network of over 50 genes is coupled to the assembly of a large, complex structure, the flagellar organelle. The flagellum is required for motility in response to chemical attractants and repellents in a liquid environment (5, 35, 54). The flagellum is commonly divided into three parts: the basal body, the hook, and the filament. The basal body is contained within the cell envelope, and over 30 genes are required for its synthesis (reviewed in references 2 and 35). The flagellar structure is diagrammed in Fig. 1. The basal body is composed of the following components: (i) a reversible rotary motor driven by a proton pump; (ii) a rod structure that transverses the cell wall, cell membranes, and periplasm; (iii) three ring proteins that stabilize the rod structure in the inner membrane, peptidoglycan layer, and lipopolysaccharide; (iv) a switch complex that determines clockwise or counterclockwise rotation in response to the chemosensory system; and (v) a flagellum-specific type III secretion pathway. Attached to the basal body on the outside of the cell is a flexible hook structure, from which the long filament structure extends. The flagellum is hollow, and a flagellum-specific export pathway sends subunits through the hollow passage to be assembled at the tip of the growing structure (35).The flagellar genes are organized into a transcriptional hierarchy of three classes that are coupled to morphogenesis of the flagellar organelle ( Fig. 1) (25). Each class must be functional for the subsequent class to be expressed.Class 1 genes, flhD and flhC, determine whether flagella are made. Expression of the FlhD and -C proteins is dependent on the general state of the cell. In addition, these proteins are required for normal cell growth. The flhDC operon is regulated in response to environmental stimuli such as cyclic AMP levels and temperature (23, 53), heat shock (50), osmolarity and membrane biosynthesis (22,38,50,51), and the H-NS protein (4). In addition, mutants of the flhDC operon affect cell division (45). The FlhD and FlhC proteins make up a heteromultimeric complex that acts as a transcriptional activator to stimulate transcription from promoters of class 2 flagellar genes by 70 -containing RNA polymerase (32). The FlhDC complex binds a ϳ40-bp region upstream of the Ϫ35 promoter sequence and is thought to interact with the C-terminal region of the ␣ subunit of RNA polymerase to initiate transcription (32,33).Class 2 genes (Fig. 1) encode the proteins needed for the structure and assembly of the basal body-hook (BBH) structure as well as the fliA regulatory gene (25). The fliA gene encodes an alternative sigma transcription factor, 28 , necessary for transcription of most class 3 genes (41). In addition,

In Escherichia coli K-12 strain W3110, the amount of 6-phosphogluconate dehydrogenase relative to that of total protein, i.e., the specific enzyme activity, increases about threefold during growth in minimal media over the range of growth rates with acetate and glucose as sole carbon sources. Previous work with gnd-lac operon and protein fusion strains indicated that two steps in the expression of the gnd gene are subject to growth rate-dependent control, with at least one step being posttranscriptional. With both Northern (RNA) and slot blot analyses, we found that the amount of gnd mRNA relative to that of total RNA was 2.5-fold higher in cells growing in glucose minimal medium than in cells grown on acetate. Therefore, since the total mRNA fraction of total RNA is essentially independent of the growth rate, the amount of gnd mRNA relative to that of total mRNA increases about 2.5-fold with increasing growth rate. This indicates that most of the growth rate-dependent increase in 6-phosphogluconate dehydrogenase can be accounted for by the growth rate-dependent increase in gnd mRNA level. We measured the decay of gnd mRNA mass in the two growth conditions after blocking transcription initiation with rifampin and found that the stability of gnd mRNA does not change with growth rate. We also used a gnd-lacZ protein fusion to measure the functional mRNA half-life and found that it too is growth rate independent. Thus, the growth rate-dependent increase in the level of gnd mRNA is due to an increase in gnd transcription, and this increase is sufficient to account for the growth rate regulation of the 6-phosphogluconate dehydrogenase level. The dilemma posed by interpretations of the properties of gnd-lac fusion strains and by direct measurement of gnd mRNA level is discussed.

Growth rate-dependent regulation of the level of Escherichia coli glucose 6-phosphate dehydrogenase, encoded by zwf, and 6-phosphogluconate dehydrogenase, encoded by gnd, is similar during steady-state growth and after nutritional upshifts. To determine whether the mechanism regulating zwf expression is like that of gnd, which involves a site of posttranscriptional control located within the structural gene, we prepared and analyzed a set of zwf-lacZ protein fusions in which the fusion joints are distributed across the glucose 6-phosphate dehydrogenase coding sequence. Expression of j8-galactosidase from the protein fusions was as growth rate dependent as that of glucose 6-phosphate dehydrogenase itself, indicating that regulation does not involve an internal regulatory region. The level of ,-galactosidase in zwf-lac operon fusion strains and the level of zwf mRNA from a wild-type strain increased with increasing growth rate, which suggests that growth rate control is exerted on the mRNA level. The half-life of the zwfmRNA mass was 3.0 min during growth on glucose and 3.4 min during growth on acetate. Thus, zwf transcription appears to be the target for growth rate control of the glucose 6-phosphate dehydrogenase level.A model system for studying growth rate-dependent regulation of nonribosomal genes is the Escherichia coli gnd gene, which encodes 6-phosphogluconate dehydrogenase (6PGD; EC 1.1.1.44), an enzyme of the pentose phosphate pathway. An interesting feature of gnd regulation is that it occurs at the posttranscriptional level and requires a negative control site that lies deep within the structural gene (3, 4). The site, called the internal complementary sequence (ICS), appears to function by forming a long-range mRNA secondary structure that sequesters the ribosome-binding region of gnd mRNA (9). The proposed regulatory role for the secondary structure is to reduce the translation initiation frequency and/or the stability of the mRNA. The effector of the regulation is unknown.Although the formation of long-range mRNA secondary structures that sequester a translation initiation region on mRNA has been shown in other systems to be involved in mediating translational coupling (6,21,25), no other example of a role for such a structure in growth rate-dependent regulation has yet been reported. Since we were interested in determining whether the mechanism is used for growth rate control of other genes, particularly central metabolism genes, we decided to study the zwf gene, which codes for glucose 6-phosphate dehydrogenase (G6PD; EC 1.1.1.49), the enzyme catalyzing the first step in the pentose phosphate pathway. The physiological response of zwf to changes in growth rate is similar to that of gnd. The specific activity of G6PD increases with increasing growth rate during steadystate growth in minimal media, although the extent of the increase is less than that of 6PGD (33). Also, the accumulation of G6PD after a nutritional upshift follows the same kinetics as that of 6PGD (11).In a previous publication, we ...

The arrangement of the Escherichia coli serC (pdxF) and aroA genes into a cotranscribed multifunctional operon allows coregulation of two enzymes required for the biosynthesis of L-serine, pyridoxal 5-phosphate, chorismate, and the aromatic amino acids and vitamins. RNase T 2 protection assays revealed two major transcripts that were initiated from a promoter upstream from serC (pdxF). Between 80 to 90% of serC (pdxF) transcripts were present in single-gene mRNA molecules that likely arose by Rho-independent termination between serC (pdxF) and aroA. serC (pdxF)-aroA cotranscripts terminated at another Rho-independent terminator near the end of aroA. We studied operon regulation by determining differential rates of ␤-galactosidase synthesis in a merodiploid strain carrying a single-copy {⌽(serC [pdxF]-lacZYA)} operon fusion. serC (pdxF) transcription was greatest in bacteria growing in minimal salts-glucose medium (MMGlu) and was reduced in minimal salts-glycerol medium, enriched MMGlu, and LB medium. serC (pdxF) transcription was increased in cya or crp mutants compared to their cya ؉ crp ؉ parent in MMGlu or LB medium. In contrast, serC (pdxF) transcription decreased in an lrp mutant compared to its lrp ؉ parent in MMGlu. Conclusions obtained by using the operon fusion were corroborated by quantitative Western immunoblotting of SerC (PdxF), which was present at around 1,800 dimers per cell in bacteria growing in MMGlu. RNase T 2 protection assays of serC (pdxF)-terminated and serC (pdxF)-aroA cotranscript amounts supported the conclusion that the operon was regulated at the transcription level under the conditions tested. Results with a series of deletions upstream of the P serC (pdxF) promoter revealed that activation by Lrp was likely direct, whereas repression by the cyclic AMP (cAMP) receptor protein-cAMP complex (CRP-cAMP) was likely indirect, possibly via a repressor whose amount or activity was stimulated by CRP-cAMP.