A route for utilization by Escherichia coli involving the regulon

Hans Kornberg* and Christopher Lourenco

Biology Department, Boston University, 5 Cummington Street, Boston, MA 02215

Contributed by Hans Kornberg, October 25, 2006 (sent for review October 17, 2006) Fructose can be taken up by Escherichia coli via a variety of operon, a number of Hfr strains of E. coli transferring DNA membrane-spanning proteins that recognize with the 3,4,5- from different points of origin and in different directions and D-arabino- configuration. Here, we describe a mutant that is carrying the Tn10 transposon at known sites (3) were crossed devoid of those proteins but takes up fructose via the FucP carrier with strain HK 2578, a derivative of HK 2298 in which FruA normally used for the transport of ␣-L-fucose; this implies that the activity had been deleted by the introduction of fruA::kanR; all fructose is taken up in the ␣-or␤-fructopyranose form. For growth, exconjugants were tested for retention of the kanR marker. the assimilated fructose is sequentially phosphorylated by ATP and These tetracycline-resistant colonies were screened for loss of (i) manno(fructo)kinase, to form fructose 6-phosphate, and (ii) their ability to grow on 2.5 mM fructose. We found that the Hfr phosphofructokinase, to form fructose 1,6-bisphosphate, which is strain BW6169, which carries argA::Tn10 located at minute a member of central routes of glycolysis and gluconeogenesis. The 63.5 on the E. coli linkage map (4) and transfers its DNA mutation that confers on the organism the ability to take up counterclockwise from approximately minute 84.8, gave rise to fructose via the fucose regulon was located as a deletion of the tetracycline-resistant colonies, of which 40% failed to grow on fucA gene with consequent induction of the proton-linked fucose fructose and which also took up 0.5 mM [14C]fructose at only transporter, FucP. a negligible rate. This result placed the presumptive fructose uptake system at a site close to argA. This location (5) was fructose metabolism ͉ fructose uptake more precisely defined by transduction of HK 2578 with phage carrying fucP::Tn10, a deletion of the gene specifying the e have previously (1) described the properties of a proton-linked transport of fucose, which is sited at min 63.2. Wmutant of Escherichia coli designated HK 2148 that is None of the 116 transductants obtained grew upon, or took up, able to grow on fructose as a sole source of carbon despite fructose. inactivation of all known routes involving the phosphoenol- The identity of the fructose uptake system with a component of the fucose regulon (6) was investigated with two argA::Tn10 pyruvate/glycose phosphotranferase (PT) system for the up- ϫ take and utilization of this (2). The entry of fructose transductants from the cross [P1 phage grown on BW6169 HK into this strain is achieved by its facilitated diffusion through 2578], one (HK 2621) having retained the ability to grow upon fructose and the other (HK 2622) being among the 40% that had an isoform of the normally PT-linked transporter, lost it. We observed that the uptake of 0.5 mM [14C]fructose by PtsG, which therefore requires that relatively high concentra- LB-grown HK 2621 was powerfully inhibited by 0.1 mM-␣-L- tions of fructose be provided in the medium (Km for growth ϭ (Ϫ)-fucose but not by D-(ϩ)-fucose (Fig. 1A). HK 2622 took up Ϸ8 mM). As would be expected from this mode of fructose the labeled fructose at Ͻ20% of the rate observed with HK 2621, uptake, growth on this substrate is powerfully inhibited by which was virtually unaffected by the presence of L-(Ϫ)-fucose analogs of glucose, such as methyl-␣-D-glucoside (␣MG). (Fig. 1B). However, cultures of HK 2148 exposed for several days at 37o Furthermore, suspensions of LB-grown HK 2621 took up to 20 mM fructose plus 1 mM ␣MG gave rise to further ␣-L-[3H]fucose to a significant extent, whereas similar suspen- mutants, some of which were not only impervious to the sions of HK 2622 did not (Fig. 2). presence of ␣MG when growing on fructose but grew on this This uptake by HK 2621 (and by its parent, HK 2578) ketose at concentrations much lower (K for growth Ͻ 1mM) m occurred rapidly during only the first few minutes of exposure than those observed with the parent HK 2148. to the isotope and thereafter increased to only a minor extent, In the present article, we report the properties of one such which indicated that the fucose taken up was not further mutant, designated HK 2298, and its derivatives, which provide metabolized. In agreement with this interpretation, we ob- evidence for the operation of a pathway for the uptake and served that neither strain, HK 2578 nor HK 2621, grew on utilization of fructose that has not been previously described. ␣-L-fucose, whereas their fructose-negative parents HK 2148 ␣ Results and HK 2632 did. Furthermore, addition of 1 mM- -L-fucose 14 to cultures of HK 2578 growing on 1–10 mM fructose greatly Uptake of [ C]Fructose by HK 2298. Whereas suspensions of LB- impeded growth, with kinetics indicating competitive inhibi- grown HK 2298 readily took up and incorporated 14C from ␣ 14 tion (Fig. 3); whereas addition of -L-fucose to cultures of HK [ C]fructose at concentrations as low as 0.1 mM, similar sus- 2632 growing on 5–20 mM fructose had no such effect (data pensions of HK 2148 did so to only a negligible extent (data not not shown). shown). Neither PtsG nor the proteins specified by the PT- These observations suggested that the ability to take up and related fructose operon were involved in this uptake: A deriv- grow upon fructose via the fucose regulon was associated with ative of HK 2298 that also carried ptsG::kanR (strain HK 2385) a mutation in some member of that regulon. This was con- grew on fructose and took up this ketose at rates indistinguish- able from HK 2298, as did other derivatives of HK 2298 carrying either fruA::kanR (strain HK 2578) or a deletion covering the Author contributions: H.K. designed research; H.K. and C.L. performed research; H.K. promoter region of the fructose operon as well as fruB and fruK analyzed data; and H.K. wrote the paper. (strain HK 2544). The authors declare no conflict of interest. *To whom correspondence should be addressed at: Boston University, 5 Cummington Location of the Gene Responsible for Fructose Uptake. To avoid Street, Boston, MA 02215. E-mail: [email protected]. possible errors due to introduction of the PT-linked fructose © 2006 by The National Academy of Sciences of the USA

19496–19499 ͉ PNAS ͉ December 19, 2006 ͉ vol. 103 ͉ no. 51 www.pnas.org͞cgi͞doi͞10.1073͞pnas.0609429103 Downloaded by guest on September 27, 2021 ϩ Fig. 1. Effect of ␣-L-fucose on the uptake of 0.5 mM [14C]fructose. (A) Strain HK 2621 (Fuc-F ) taking up [14C]fructose alone (E) and in the presence of Ϫ D-(ϩ)-fucose (Œ)orofL-(Ϫ)-fucose (F). (B) Strain HK 2622 (Fuc-F ) taking up [14C]fructose alone (E) and in the presence of L-(Ϫ)-fucose (F). Note difference in scale.

firmed by comparison of the DNA sequences of its component Fuc-FϪ strains failed to do so in all strains manifesting the genes fucP, fucK, and fucA, present in the fructose-positive Fuc-Fϩ phenotype. (Fuc-Fϩ) strains HK 2578 and HK 2621 and their fructose- negative counterparts HK 2148 and HK 2622. We found that Metabolism of Fructose Taken Up. If fructose is taken up via a the DNA sequences of fucP and fucK were identical and derepressed fucose carrier, it would enter the cells chemically corresponded exactly to the sequences published for these unmodified and would be expected to undergo phosphorylation genes (7) but that the fucA region of HK 2148 and HK 2622 to fructose 6-phosphate in the presence of ATP and manno- differed markedly from that present in HK 2578 and HK 2621. (fructo)kinase (Mak) (1); the fructose 6-phosphate thus formed The DNA sequence of the fucA in strains HK 2148 and 2622 would then be further phosphorylated by ATP in the presence again corresponded to the published sequence but the PCR of phosphofructokinase (PfkA). Two strategies were used to test product of this region in HK 2578 and HK 2621 was clearly this hypothesis. ϩ ϩ smaller (Fig. 4); moreover, the primers that had yielded In the first, the genes specifying Mak and PfkA in HK excellent PCR products and reproducible DNA sequences in 2298 were selectively replaced by their negative counterparts. Thus, HK 2298 was infected with phage P1 grown on strain HK 2198, which carries mak-o cotransducible with argJ::Tn10 (8); Ͼ70% of the tetracycline-resistant transductants did not grow on fructose. A similar abolition of the ability to grow on

Fig. 3. Hanes plot of the effect of 1 mM-␣-L-fucose on the growth of strain ϩ ϩ Fig. 2. Uptake of 0.5 mM [3H]fucose by strain HK 2621 (Fuc-F )(E) and by HK 2578 (Fuc-F ) at various concentrations of fructose (E). The rates of growth MICROBIOLOGY Ϫ strain HK 2622 (Fuc-F )(■). in the absence of ␣-L-fucose are shown as F.

Kornberg and Lourenco PNAS ͉ December 19, 2006 ͉ vol. 103 ͉ no. 51 ͉ 19497 Downloaded by guest on September 27, 2021 Scheme 1.

acquired Makϩ) was infected with phage grown on the Fuc-Fϩ argA::Tn10 strain HK 2621, Ϸ40% of the tetracycline-resistant transductants grew on 2.5 mM fructose.

Discussion We have previously summarized the evidence that fructose can Fig. 4. PCR products of fucA regions of HK 2622 and HK 2621. enter E. coli via a number of membrane-spanning proteins, all of which recognize the 3,4,5-D-arabino-hexose configuration of D- fructose, D-glucose, D-, D-mannitol, and D-glucitol (2). It fructose was observed when HK 2298 (which carries argHBCE, was thus unexpected that in a mutant devoid of all these carriers, located at min 89.5) was infected with P1 phage grown on a ϩ a further mutation in the fucose regulon could efficiently effect the strain that was ArgHBCE and also carried pfkA (located at uptake of fructose. However, although the ␣-L-fucose that has been min 88.5). Transductants were selected on plates containing shown in the present work to compete with the entry of fructose glycerol as carbon source but not containing arginine; all those does not have the 3,4,5-D-arabino-hexose configuration, it is strik- that no longer grew on glucose and were therefore PfkAϪ also ingly similar to the form of ␣-or␤-D-fructose (Scheme 1). failed to grow on fructose. It has also been established (9) that in an aqueous solution, 57% A second strategy to test the postulated route for fructose of fructose is in the ␤-D-pyranose form. It is therefore likely that it utilization consisted of sequential introduction of the genes for ϩ ϩ is this form that enters the cells via the FucA component of the Mak and the fucose-linked fructose transport system Fuc-F fucose regulon and that the relevant mutation of HK 2148 that gave into a recipient strain devoid of both of them. HK 2140 has ϩ these properties (1). In one approach, it was crossed with rise to HK 2298 had led to the derepression of the fucP gene. This phage grown on the Fuc-Fϩ-strain HK 2621 that also carried would be consistent with the view (6) that the genes of the fucose argA::Tn10: None of the tetracycline-resistant transductants system operate as a regulon containing at least three operons, of grew on 2.5 mM fructose. Because it was not feasible to which fucP and fucA form separate parts, and that the substrate on which FucA acts, 1-phosphate, is known to induce the determine which of the many transduced colonies had received ϩ the gene specifying Fuc-Fϩ, a number were pooled and fucose transporter: Deletion of fucA would cause fuculose 1-phos- infected with phage grown on the Makϩ strain HK 2193 that phate to accumulate. This would also be analogous to the obser- also carried argJ::kanR cotransducible with mak. Approxi- vations that growth on fructose via the proteins that specify the mately 40% of the kanamycin-resistant transductants grew on uptake of mannose, glucitol, and mannitol is effected by derepres- fructose. sion of the corresponding genes (2). ϩ ϩ In the reverse sequence, phage grown on the Makϩ kanR Our present demonstration that both Mak and PfkA are strain 2193 were crossed with HK 2140: Again, none of the required for the growth of HK 2298 and its derivatives supports kanamycin-resistant transductants grew on 2.5 mM-fructose, but the conclusion that the fructose taken up by the Fuc-Fϩ system when a colony that grew on 20 mM-fructose (and had therefore subsequently follows the path

Table 1. Bacterial strains used in this study Growth on 2.5 mM Strain Relevant genotype fructose Source

HK 1796 srlA⌬ fruB fruK⌬ pfkA rpsL (argHBCEϩ) Ϫ Stock HK 2140 srlA⌬ fruB fruA manXYZ::CmR rpsL Ϫ Stock HK 2148 srlA⌬ fruB manXYZ::CmR argHBCE rpsL Ϫ Ref. 1 HK 2193 As 2148 but also makϩ::kanR Ϫ Stock HK 2198 srlA⌬ fruB fruK⌬ manXYZ::CmR mak-o::Tn10 Ϫ Stock HK 2385 As 2298 but also ptsG::kanR Ϫ This work HK 2298 As 2148 but also resistant to ␣MG on 20 mM-fructose ϩ This work HK 2544 As 2298 but also (promoter, fruB, fruK)⌬ϩThis work HK 2578 As 2298 but also fruA::kanR ϩ This work HK 2621 [P1(BW 5659) ϫ 2578]argA::Tn10 (Fruϩ) ϩ This work HK 2622 [P1(BW 5659) ϫ 2578]argA::Tn10 (FruϪ) Ϫ This work HK 2632 As 2148 but also fruA::kanR Ϫ This work BW 6169 Hfr (PO min 84.8, counterclockwise) argA::Tn10 ϩ Ref. 3

Ϫ indicates no growth, and ϩ indicates good growth at 37°C overnight on plates containing 2.5 mM fructose as the sole C source.

19498 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0609429103 Kornberg and Lourenco Downloaded by guest on September 27, 2021 Uptake of Labeled Substrates. The uptake of 0.5 mM [14C]fructose or 0.5 mM [3H]fucose was measured as the quantities of radio- active material retained by cells at 0.2 mg of dry mass per milliter, shaken in shallow vessels in a water bath at 30°C. Samples (0.1 ml) were taken at known times, filtered rapidly through Milli- pore (Billerica, MA) filters (pore size, 0.45 nm) and washed with 2 ϫ 5 ml of 50 mM phosphate buffer, pH 7. The filters were plunged immediately into plastic cuvettes containing 5 ml of Materials and Methods Ecoscint H (National Diagnostics, Atlanta, GA), and their radioactivity was assayed with an LS 6500 MultiPurpose scintil- Strains and Growth Conditions. The bacterial strains used are lation counter (Beckman Coulter, Fullerton, CA). [14C]Fructose listed in Table 1. Bacteria were grown at 37°C either in a liquid and [3H]fucose were obtained from American Radiolabeled culture consisting of LB broth (25 g of LB base per liter of Chemicals, (St. Louis, MO). water) or in defined media (10), or on such media solidified with 2% agar. Cell density was measured in 1-ml cuvettes (1-cm Analysis of Genomic DNA. Procedures for amplification of genomic light path) with absorbance at 600 nm (A600 of 1.0 being DNA and manipulation and sequencing of DNA fragments were equivalent to 0.26 mg of dry mass per milliliter). Procedures those previously described (1). used for high-frequency recombination-mediated conjuga- tions and for phage P1-mediated transductions were those We thank Dr. Mary Berlyn (Yale University, New Haven, CT) for her described by Miller (11). generous and frequent gifts of E. coli mutants.

1. Kornberg HL, Lambourne LTM, Sproul AA (2000) Proc Natl Acad Sci USA Vides J, Glasner JD, Rode CK, Mayhew GF, et al. (1997) Science 277:1453– 97:1808–1812. 1474. 2. Kornberg HL (2002) Adv Enzyme Regul 42:349–360. 8. Sproul AA, Lambourne LTM, Jims Jean-Jacques D, Kornberg HL (2001) Proc 3. Wanner BW (1986) J Mol Biol 191:39–58. Natl Acad Sci USA 98:15257–15259. 4. Berlyn MK (1988) Microbiol Mol Biol Rev 62:814–984. 9. Mathews CK, van Holde KE (1990) in (Benjamin/Cummings, 5. Rudd KE (1998) Microbiol Mol Biol Rev 62:985–1019. Redwood City, CA), p 267. 6. Lin ECC (1966) in Escherichia coli and Salmonella, ed Neidhardt FC, (Am Soc 10. Ashworth JM, Kornberg HL (1966) Proc R Soc London Ser B 65:179–188. Microbiol, Washington, DC), 2nd Ed, pp 307–342. 11. Miller JH (1972) Experiments in Molecular Genetics (Cold Spring Harbor Lab 7. Blattner FR, Plunkett G, Bloch CA, Perna NT, Burland V, Riley M, Collado- Press, Plainville, NY). MICROBIOLOGY

Kornberg and Lourenco PNAS ͉ December 19, 2006 ͉ vol. 103 ͉ no. 51 ͉ 19499 Downloaded by guest on September 27, 2021