We have been able to identify in K. aerogenes strain W70 two enzymes, RDH and DRK, which are induced in response to D-ribulose, an
apparent intermediate of a specific ribitol catabolic pathway. A similar pathway is found in A. aerogenes strain PRL-R3 (22) composed of two enzymes similar in terms of induction and activity. Our decision to shift studies of A. aerogenes PRL-R3 to K. aerogenes W70 was influenced by the discovery of a transduction system mediated by phage PW52, which was reported by MacPhee et al. (13) and the discovery that K. aerogenes W70 was similar to A.
aerogenes strain PRL-R3 with respect to its degradation of naturally occurring C, compounds, and in its adaptation to growth on the
uncommon C, sugars and alcohols. The lone dissimilarity in this area was the inability of K. aerogenes to degrade those C5 compounds which
in A. aerogenes appear to be catabolized by a pathway involving L-xylulokinase. L-Xylulokinase activity is demonstrable in A. aerogenes, but we have been unable to demonstrate it in K. aerogenes. The procedures worked out for utilizing A. aerogenes for selection and isolation of mutants were, therefore, modified to work in K. aerogenes W70. The excellent background work in A. aerogenes on ribitol catabolism allowed us to develop similar information for K. aerogenes W70 on induction of the enzymes involved, the apparent inducer, and mode of adaptation to growth on the rare sugars D-arabinose and xylitol. The enzymes RDH and DRK are coordinately controlled, since mutants selected for constitutive synthesis of RDH by their ability to grow with xylitol as a carbon and energy source, or for the ability of starved cells to reduce tetrazolium dyes in the presence of ribitol, are constitutive for DRK, even though we can demonstrate the effectiveness of the selections in mutants lacking a functional DRK. In addition, L-fucose isomerase-constitutive cells, when grown on D-arabinose, produce high levels of RDH and DRK even though RDH is not involved in growth on that substrate. We have found that
L-fucose isomerase in K. aerogenes, as in A. aerogenes, converts D-arabinose to D-ribulose. This leads to induction of the enzymes of the
ribitol pathway. Mutants capable of producing D-ribulose from D-arabinose, but unable to synthesize RDH, still produce DRK in response to
the D-ribulose, indicating that reduction to the pentitol is not required for induction of DRK. Mutants lacking DRK activity, although unable
to grow on D-arabinose alone, produce, when grown on casein hydrolysate supplemented with D-arabinose, high levels of RDH, indicating that
the DRK is not required for production of the inducer from the D-ribulose formed. The role of n-ribulose as an inducer of these enzymes is
consistent with our inability to find mutants which produce DRK but not RDH when grown on casein hydrolysate supplemented with ribitol,
even though mutants which appear to induce DRK but not RDH are found if arabinose rather than ribitol is offered as substrate.
In addition to helping us identify the inducer of the ribitol pathway, growth on D-arabinose shows that K. aerogenes, like A. aerogenes,
mutates preferentially to constitutive synthesis of the L-fucose catabolic enzymes (in particular L-fucose isomerase) rather than to a modification
that allows induction of these enzymes in cells grown on D-arabinose, as found in Escherichia coli (10). The latter alternative, however,
can also be found in this organism.
In a similarmanner, growth on the unusual pentitol, xylitol, is the result of a mutation to constitutive synthesis of the ribitol catabolic enzymes. In
this case RDH converts xylitol to D-xylulose. The D-xylulose produced can be phosphorylated by specific D-xylulokinases. This will be considered
in greater detail in a subsequent paper. Our understanding of growth on D-arabinose and xylitol allowed us to map the sites for phenotypes which we had surmised to represent lesions in structural and regulatory sites of the ribitol pathway enzymes. Prior to setting up the crosses reported, we were obliged to construct a mutant which would allow us to score the various phenotypes on semisolid media. We, therefore, created a mutant with the following
lesions: (i) uracil requiring, to allow us to identify contaminants as such if they arose; (ii) L-ribulokinase negative, to eliminate conversion by this enzyme of D-ribulose to D-ribulose-5-phosphate as reported in A. aerogenes (10); (iii) L-fucose isomerase constitutive, to allow isomerization of D-arabinose to D-ribulose and subsequent growth potential on D-arabinose through the DRK as illustrated in Fig. 1; (iv) L-fucose negative, to prevent shunting of Dribulose through the L-fuculokinase and fucose aldolase, as reported in E. coli (9), in those mutants lacking a functional DRK. In this basic mutant we were able to set up two- and three-point crosses with mutants with the phenotypes listed in Tables 1 and 2, i.e., mutants which constitutively synthesized RDH and DRK or failed to produce one or both of these enzymes under conditions where these enzymes should have been induced. The lesions
were all closely linked genetically, as shown in the two-point crosses, and three-point crosses yielded the map order indicated in Fig. 3.

The map order is consistent with our concept of an operon, and the induction of the enzymes is consistent with the functioning of an operon. We have not examined the direction of transcription or the possibility of positive or negative control. Thus, although we have no data for this system which are inconsistent with the concept of an operon, we do not have sufficient evidence to report it as an operon. We have recently found that there appear to be at least two genetically distinguishable control sites, as defined by mutations to constitutive synthesis of RDH and DRK, but we cannot yet define
them in terms of operator and regulator functions. This aspect of the work will be considered in a subsequent paper.