Azotobacter vinelandii solubilized iron from certain minerals using only dihydroxybenzoic acid, which appeared to be produced constitutively. Solubilization of iron from other minerals required dihydroxybenzoic acid and the siderophore N,N'-bis-(2,3-dihydroxybenzoyl)-L-ly'sine (azotochelin) or these chelators plus the yellow-green fluorescent siderophore azotobactin. In addition to this sequential production of siderophores, cells also demonstrated partial to hyperproduction relative to the iron-limited control. The iron sources which caused partial derepression of the siderophores caused derepression of all the highmolecular-weight iron-repressible outer membrane proteins except a 77,000-molecular-weight protein, which appeared to be coordinated with azotobactin production. Increased siderophore production correlated with increased production of outer membrane proteins with molecular weights of 93,000, 85,000, and 77,000, but an 81,000-molecular-weight iron-repressible protein appeared at a constant level despite the degree of derepression. When iron was readily available, it appeared to complex with a 60,000-molecularweight protein believed to form a surface layer on the A. vinelandii cell.

Ferric reductase activity was examined in Azotobacter vinelandii and was found to be located in the cytoplasm. The specific activities of soluble cell extracts were not affected by the iron concentration of the growth medium; however, activity was inhibited by the presence of Zn2+ during cell growth and also by the addition of Zn2+ to the enzyme assays. Intracellular Fe2+ levels were lower and siderophore production was increased in Zn2+-grown cells. The ferric reductase was active under aerobic conditions, had an optimal pH of approximately 7.5, and required flavin mononucleotide and Mg2+ for maximum activity. The enzyme utilized NADH to reduce iron supplied as a variety of iron chelates, including the ferrisiderophores of A. vinelandii. The enzyme was purified by conventional protein purification techniques, and the final preparation consisted of two major proteins with molecular weights of 44,600 and 69,000. The apparent Km values of the ferric reductase for Fe3+ (supplied as ferric citrate) and NADH were 10 and 15.8 ,uM, respectively, and the data for the enzyme reaction were consistent with Ping Pong Bi Bi kinetics. The approximate K. values resulting from inhibition of the enzyme by Zn2+, which was a hyperbolic (partial) mixed-type inhibitor, were 25 ,uM with respect to iron and 1.7 ,uM with respect to NADH. These results suggested that ferric reductase activity may have a regulatory role in the processes of iron assimilation in A. vinelandii.The presence of iron in many enzymes such as nitrogenase and ribonucleotide reductase, and in many cellular structures such as cytochromes and ferredoxins, make it an essential element for almost all forms of life (26). The extreme usefulness of this particular metal is related to its ability to exist in two stable redox states, Fe3" and Fe2 .

Optimal transformation of Azotobacter vinelandii OP required a 20-min incubation of the competent cells with deoxyribonucleic acid at 30°C in buffer (pH 6.0 to 8.0) containing 8 mM magnesium sulfate. Nitrogen-fixing transformants of nitrogen fixation-deficient recipients could be plated immediately on selective medium, but transformants acquiring rifampin and streptomycin resistance required preincubation in nonselective medium. The three phenotypes achieved an approximately equal and stable frequency after 17 h (six generations) of growth in nonselective medium.

Azotobacter vinelandii forms both catecholate and azotobactin siderophores during iron-limited growth. Azotobactin is repressed by about 3 µM iron, but catecholate siderophore synthesis continues up to a maximum of 10 µM iron. This suggests that catecholate siderophore synthesis is regulated by other factors in addition to the ferric uptake repressor (Fur). In this study the first gene required for catecholate siderophore biosynthesis, which encodes an isochorismate synthase (csbC), was isolated. The region upstream of csbC contained a typical σ 70 promoter, with an iron-box overlapping the N35 sequence and a Sox-box (Box 1) overlapping the N10 sequence. Another Soxbox was found further upstream of the N35 sequence (Box 2). Also upstream, an unidentified gene (orfA) was detected which would be transcribed from a divergent promoter, also controlled by an iron-box. The activity of csbC and a csbC ::luxAB fusion was negatively regulated by iron availability and upregulated by increased aeration and by superoxide stress. The iron-box in the csbC promoter was 74 % identical to the Fur-binding consensus sequence and bound the Fur protein of Escherichia coli with relatively high affinity. Both Box 1 and Box 2 were in good agreement with the consensus sequence for binding the SoxS protein of E. coli and Box 1 was in very good agreement with the Sox-box found in the fpr promoter of A. vinelandii, which is also regulated by superoxide stress. Both Sox-boxes bound a protein found in A. vinelandii cell extracts, with Box 1 exhibiting the higher binding affinity. The Sox protein identified in this assay appeared to be constitutive, rather than inducible by superoxide stress. This indicates that the Sox response in A. vinelandii is different from that in E. coli. These data support the hypothesis that catecholate siderophore biosynthesis is under dual control, repressed by a Fur-iron complex and activated by another DNA-binding protein in response to superoxide stress. The interaction between these regulators is likely to account for the delay in ferric repression of catecholate siderophore production, since these siderophores have an additional role to play in the protection of ironlimited cells against oxidative damage.

Azotobacter vinelandii UWD formed polyhydroxyalkanoate (PHA) copolymers containing 13-hydroxybutyrate and 13-hydroxyvalerate (HV) when grown in a medium containing glucose as the primary C source and valerate (pentanoate) as a precursor. Copolymer was not formed when propionate was added to the glucose medium but was formed when heptanoate, nonanoate, or trans-2-pentenoate was present. Optimal levels of HV were formed when valerate was added at the time of maximum PHA synthesis, although HV incorporation was not dependent on glucose catabolism. IV content in the polymer was increased from 17 to 24 mol% by adding 10 to 40 mM valerate to glucose medium, but HV insertion into the polymer occurred at a fixed rate. Similarly, the addition of valerate to a fed-batch culture of strain UWD in beet molasses in a fermentor produced 19 to 22 g of polymer per liter, containing 8.5 to 23 mol% HV after 38 to 40 h. The synthesis of IV in these cultures also occurred at a fixed rate (2.3 to 2.8 mol% h-'), while the maximum PEIA production rate was 1.1 g liter-' h-1. During synthesis of copolymer in batch or fed-batch culture, the yield from conversion of glucose into PEIA (Yp/s) remained at maximum theoretical efficiency (.0.33 g of PEIA per g of glucose consumed). Up to 45 mol% 1V in the polymer was obtained by growing strain UWD in medium containing c50 mM valerate as the sole C source, but the PHA produced amounted to <1 g/liter. The combination of 30 mM valerate as a sole C source and 0.5 mM 4-pentenoate increased the HV content in the polymer to 52 mol%. The results strongly support a route involving 13-oxidation in the production of 1V in A. vinelandii PEIA. The results show that strain UWD can form PHA copolymers of potential use as bioplastics and that the substrate cost per kilogram of PHA formed in beet molasses medium should be less than half of that per kilogram of PHIA formed in glucose medium.