JOURNAL QF BACITN~~~Y, Jan. 1968, D. 64-77 Vol. 97. No. 1 Copytigbt @ 1969 American Society for Miibioloay Primed In U.S.A. Effects of Colicins El and K on Cellular Metabolism KAY L. FIELDS AND S. E. LURIA Department of Biology, Massachusetts Institute of Technology, Cambridge, Massachusetts 0.2139 Received for publication 7 October 1968 Colicins El and K inhibited a whole series of energy-dependent reactions in Escherichiu cofi cells, including motility, biosynthesis of nucleic acids, proteins and polysaccharides, and the conversion of omithine to citrulline. Respiration was only partially affected, and substrates such as glucose continued to be catabolized through the normal pathways, albeit with reduced COP production. The soluble products of aerobic glucose catabolism by colicin-treated cells were analyzed. Pyruvate replaced acetate as the major excreted product, and the following intermediates of glycolysis were excreted in significant amounts: glucose&phosphate, fructose-l ,6-diphos- phate, dihydroxyacetone phosphate, and Sphosphoglycerate. Anaerobically grow- ing cells manifested a somewhat enhanced tolerance to the colicins. This protection by anaerobiosis appeared to depend on the exclusion of oxygen more than on the extent of fermentative catabolism versus catabolism of the respiratory type. These results are interpreted in terms of possible functions of colicin in lowering the adenosine triphosphate (ATP) content of the cells and in terms of the role of lowered ATP levels in inhibiting many of the energy-requiring reactions. In the preceding paper (16), which described studies of the effects of colicins El and K on active transport systems in Ekcherichiu coli, it was concluded that the inhibition of the accumulation of thiomethyl galactoside (TMG) and certain other substrates (27, 31) was due to the effect of these colicins on energy metabolism, reflected by the sharp reduction in adenosine triphosphate (ATP) levels. The fact that accumulation of cu-methyl-o-glucoside (cxMG) was only slightly affected by colicins or by NaN, was attributed to the fact that (uMG accumulation is carried out by a phosphoenolpyruvate-dependent phos- photransferase system (26), which was presum- ably less affected by colicin treatment. While colicin causes ATP levels to fall, oxygen con- sumption continues (21). Levinthal and Levin- thal (unpublished dota, cited in 27) made the sig- nificant discovery that under conditions of strict anaerobiosis colicin El did not inhibit bio- synthetic reactions in cells of E. coli K-12 strain C600. This suggested a selective inhibition of oxidative phosphorylation. The findings with transport systems led us to study further certain aspects of catabolism and of some other energy-requiring cellular processes in colicin-treated E. coli cells. The results, pre- sented in this paper, reveal some novel features of colicin action and suggest possible mecha- nisms of this action. 1 Present address: Univenitd de G&he, hrtitut de Biologic Moleculairc, Ocoeva, Switzerland. MATERIALS AND METHODS Baeteda. The bacterial strains used are listed in Table 1. Medla. Media and several procedures used were described in the preceding paper (16). The minimal phosphate medium of Kornberg et al. (24) was used for growth of cells with acetate as sole carbon source. For experiments at a low pH, medium 63 was ad- justed to pH 6.2 and supplemented with 2 X lO+ M F&O,. For anaerobic growth, cells were grown in 50-ml tubes fitted with bubbling tubes. Nitrogen, nitrogen- 5% CO,, argon, or argon-5Y0 CO, gas was used for vigorous bubbling. IneorporatIon of radIoaetIve substratw In add-ln- soIuhIe form. Uptake of W-labeled leucine, isoleucine, uracil, thymidine, or glucose by cell suspensions was measured by placing samples in 5% trichloroacetic acid at 0 C. After 20 min, the acid-insoluble material was collected on cold filters (Millipore Corp., Bed- ford, Mass.), washed with cold acid, dried, and counted in a gas-flow counter. Incorporation of IC-glucose into glycogen-like polymers was measured according to the method of Abraham and Hassid (1) by resuspending cells in 30% KOH, boiling for 20 min, and precipitating with ethyl alcohol. The precipitate was collected on Whatman filters, dried, and counted. Total lC-glucose uptake was measured by filtering a chilled sample and wasb- ing it with phosphate buffer containing cold glucose (20 r&4. Incorporation of "C-acetate was measured by add- ing cold 2.5% acetic acid, filtering the preparations, and washing them with 5% acetic acid. Respiration and gas profludon. Oxygen uptake and 64 VOL. 97, 1969 EFFECTS OF CGLICINS ON CELLULAR METABOLISM 65 CGr evolution were measured in a Warburg-type reap&meter with standard manometric techniques UnIess otherwise stated, cells were harvested during exponahl growth, washed, and concentrated to an optical density (OD) at 500 nm of 1.0 to 2.0. The rate oi oxygen upialie ir~expressed as tllieruliters of Or per minute per OD unit (500 nm) of cells. (For E. coff K-12 cells, one OD unit corresponds to 5 X 101 cells, or0.1mgofproteinpaml.) Anaerobic conditions in Warburg vessels wcee established by flushing with a stream of nitrogen or argon for 10 min. Hydrogen production was me&red in Sasks containinn KOH. and combined Co. and H. production was &ulated by assuming that equai amounts of the two gases were made. AnalytIcsI m Glucose was determined with Glucostat reagent (Worthington Biochemical Corp., Freehold, NJ.), b&red when necessary with me- dium 63. Citrulline was determined on the supernatant fractions of centrifuged cultures by the method of Archibald (4), and pyruvate was measured by the method of Friedemann and Haugen (18). Theenxymaticassaysusedwaethosegivenby Bergmeyer (10) with minor modiEcations. The enxymw were the beat grade available and were ob- tained as ammonium sulfate suspensions from C. F. Boehringer and soehne, Mannheim, Germany. The samples were the supanatant fractions of centrifuged cdl suspedons. In all assays, the final volume was 1.0 ml, and oxidation or reduction of nicom ad&e dinucIeotide (NAD) or nicotinamide adenine dinucleotide phosphate (NADP) was measured at room tempe&tu& in a- Zeiss- spectrophotometer. EnxymeeoncentratlonsaregivenininternationaIunits (1 unit - 1 vroleofproductperminat25C). Pyruvate Wes measured by the oxidation of &duced NAD (NADH) mediated by lactic dehydrogenase. Bach assay contained: tris(hydroxymethyl)amino- Lurk ltock cdkction no. LA279a L-104 L-A:52 L-A632 L-A630 L-A631 L-A633 L-620 L-622 L-621 L-89 - L-28 strain E. coil K-12 zzl CXOO hmn u160 AB1302B :7=zL2 YzO(Co1 El) E. coli W E. coli W2244 E. coli PDC- E. co11 B E. co11 B hmn E. toll K235 methane (`Dir), pH 7.5, 30 rmoles; NADH, 0.05 mle; and Iactic dehydrogenase, 0.2 unit. Pyruvate, phosphoenoIpyruvate, 2-phosphogIycer- ate, and 3-phosphoglycunte wae detmnined in a single away which coupled each compound in succcs- sion to the oxidation of NADH by lactic dehydrogen- we. Each away eontaincd, in micromoles: triethanol- amme-ethyknediamindetrsacetate (EDTA) b&r (pH 7.6), m, MgSG,, 10; KCl, 70; NADH, 0.1; and adenosine diphosphate (ADP), 0.5. Pyruvate was de- tmnined Ilrst by adding 2 units of lactic dehydrogen- ase; phospholnolpyruvate was assayed by the further addition of 0.5 unit of pyruvate kinase; 2-phospho- enolpyruvate, by the addition of 0.4 unit of enolase (Sigma Chemical Co., St. Louis, MO.); and 3-phospho- glycerate, by the addition of 0.1 unit of phospho- glycerate mutase (Sigma Chemical Co.) and about 0.2 rmole of 2,3diphosphoglycerate. 1,3-Diphosphoglycerate and 3-phosphoglycerate de@rmi& by the oxidatiin of NADH mediated E-pbospboglycuntc kinaae and glyceraldehyh phosphate dehydrogenase. Each assay contained, in micromoIes: triethanolamineEDTA buffa bH 7.6), 50, MgSGr, 8; glutathione, 2.5; hydraaine3; ATP; 7.5: and NADH. 0.05. 1.3Diphosphonlvcerate was detkninai lirst liy adding-O.4 unit of gl@raldehyde- phosphate dehydrogenase; 3-phosphoglycerate was da tamined by the subsequent addition of 0.8 unit of 3-phosphoglycerate kinase. This assay contained no activity for pyruvate or dihydroxyacetone-phosphate. Dihydroxyaceton+phosphate, alyw?wkkb~de~bos- -. phate-and &ctose&hobphate were determined by the oxidation of NADH mediatad bv tdvcerol-I- _ -- phosphate dehydrogenase. Each assay contained, in micromolea: Tris buffer (PH 7.5), 30; and NADH, 0.05. Diiydroxyacetone-phosphate was determined by the addition of 0.1 unit of a-glycerophosphate dehy- drogenase; glycera&hyde-phosphate, by the subse- quent addition of 0.4 unit of triose+hosphate iso- TABLE 1. Bacterial strains Rekvult genotype Prototroph thi thr leu thi thr leu hmn (highly motile) argG argK (derepressed) sIpA aceA thl thr Ieu (Cal El-K30) Prototroph cts ace) hmn Prototroph (Co1 K) J. Monod hi. Behanski L. Fischer-Fantuzzi G. Jacoby T. H. Wilson J. Beckwitb F. Levinthal C. A. Hirsch L. P. Hager M. Beljanski P. FrCdCricq . . . Symbols: ace = acetate requirement (pyruvic dehydrogenase defect); arg = argmine; cts - citrate synthetase; glp = glucose&phosphate permease; hmn = hemin; leu = leucine; thi = thiamine; thr = threonine. 66 FIELDS AND LURIA J. BACI-ERIOL. merase; and fructosediphosphate, by the addition of 0.1 unit of aldolase. Glucosed-phosphate and fructose-6-phosphate were detern+ned by the reduction of NADP mediated by glucose&phosphate dehydrogenase. Each assay contained, in micromoles: Tris (pH 7.5), 30; MgCl,, 5; and NADP, 0.1. Glucosea-phosphate was determined by addition of 0.5 unit of glucose&phosphate de- hydrogenase (Sigma Chemical Co., type V); fructose- Cphosphate by subsequent addition of 0.4 unit of yeast phosphoglucose isomerase (Calbiochem, Los Angeles, Calif.). 6-Phosphogluconate was determined by the re- duction of NADP mediated by phosphogluconate de- hydrogenase. Each assay contained: triethanolamine (pH 7.6), 370 #moles; MgSO,, 5 rmoles; NADP, 0.2 rmole; and 0.5 unit of 6-phosphogluconate dehy- drogenase. The assay was free from activity of glucose- 6-phosphate. Analysis of glucose degradation products. Washed bacterial suspensions (about 5 X I@ glucose-grown cells/ml), either colicin-treated or controls, were placed in Warburg vessels with KOH papers in the center well. 1Qlucose (10-" M, 3.3 @/ml, uniformly labeled, unless otherwise noted) was added from the side compartment, and incubation was continued until OI consumption showed a sharp break. The KOH paper, which trapped the evolved CO*, was counted in 3 ml of ethyl alcohol and 6 ml of toluene scintillation fluid (12). Some samples of the cell suspensions were used to determine incorporation into acid-insoluble products, and other samples were prepared for column chroma- tography as described by Dobrogosz (15). The sus- suspensions were added to a carrier mixture (final con- tent: ethyl alcohol, acetate, pyruvate, formate, lactate, and suchnate, each 0.05 M, with enough HrSOl to give a pH of 1.7). After centrifugation, the supematant fluid, containing all acid-soluble products, was kept at -20 c. Partition chromatography was carried out with silicic acid columns prepared according to Ramsey (33). Silicic acid was separated from fine particles and dried overnight at 1OOC. The exact proportion of solvent to siiicic acid had to be determined for each batch. The column was prepared in a tube (12-mm diameter) with a pierced sintered-glass disc and layers of glass wool, sand, and Celite Super-Gel. Silicic acid (6 g) was mixed with as much 0.5 N H&O, as could be added without loss of its powdery consistency (about 3.2 ml), and the mixture was added to the column tube filled with benzene. Each column was 15 cm long and was used only once. The acidified, carriercontaining supernatant sample (0.5 ml) was mixed with 1.5 g of silicic acid, and this powder was added to the top of the prepared column. The column was developed with a series of solvents similar to that used by Dobrogosz (15). Acid-washed chloroform (150 ml) was followed by a linear gradient of chloroform (300 ml in the mixing chamber) and 5% t-butyl alcohol in chloroform (300 ml; 4y0 &butyl alcohol in chloroform in the second chamber was used for clean separation of pyruvate and formate). In the fist experiments, lactate was eluted near the end of the gradient, and remaining lactate was eluted with 50 ml of 5% l-butyl alcohol in chloroform. If the gradient was discontinued as soon as formate had been eluted. then all the lactate could immediately be eluted witd 5% r-butyl alcohol in chloroform: 50 to 100 ml of 10% &ityl alcohol in chloroform WA then used to el& succinate. Finally, 50 ml of 30% r-butyl alcohol in chloroform was passed through, followed by distilled water. The chloroform-based-&vents were-used at a rate of 2 ml/min. Water flow was very slow and added pressure was necessary. Fractiops (8 to 10 ml) were collected, and the amount of acid in each fraction was determined by titrating a 4-ml sample under nitrogen with ethanolic KOH (33). This titration established the peaks of carrier acids. The distribution of radioactivity was de- termined by scintillation counting according to Dobrogosz (15) in a Nuclear-Chicago counter, model 724, with corrections applied for chemical quenching. Paper chromatography and electrophoresk. Ma- terials that were eluted from the silicic acid columns with water were partially characterized by paper chromatography. Radioactive compounds were de- tected by using a Nuclear-Chicago strip counter; standard methods were used for detection of reference compounds (11). The sensitivity was insufficient to detect minor radioactive components. To compare the eluted material with known compounds and to test whether the chromatographic properties changed upon incubation with alkaline phosphatase (type 111, E. coli: Sigma Chemical Co.). the following solvent systems-were used: ethyl &thy1 ketone-methylcello- solve-3 M NH, 2:7:3 (29); methanol-NHi-water, 60: lo:30 (8); and butyric acid-O.85% NaOH in water, 69:31, v/v (35). Radioactive samples from the silicic acid columns were subjected to paper electrophoresis at pH 1.8 (4% formic acid) and pH 3.2 (9.2y0 butyric acid in 0.85% NaOH; 35), and in borate buffer, pH 9.2 (17). The migrations of the radioactive samples were com- pared (before and after treatment with alkaline phos- phatase) with those of known phosphate esters. RESULTS Effects of colicins on biosynthetic reactions. It is known that colicins El and K halt biosynthe- sis of protein and ribonucleic acid (RNA) very quickly, as illustrated in Fig. 1. The findings were similar for a variety of colicin-sensitive E. coli strains. Interestingly, colicins K and El caused an arrest of protein or RNA synthesis in Shigella dysenteriae Sh only after a lag of 10 to 15 min. Resistant mutants, which do not adsorb a colicin, do not show any of the effects described. Table 2 illustrates the effects of the colicins on incorporation of "C from glucose or acetate into E. coli cells. Practically complete inhibition occurred, not only of incorporation into acid- insoluble form but also of incorporation into any not readily washable compounds. Chloram- VOL. 97, 1969 EFFBCTS OF COLICINS ON CELLULAR METABOLISM 67 TIME (min) FIG. 1. Effect of cdicin El or K on incorpomtion of isoleucine or uracil. Cells of E. coli 3ooO were grown in meakun 63 with 0.4% glycerol and treated with colicin or buffer for 3 mitt; then W-labeled imleucine (A) or uraciI (B) was aakied. Survival at 4 min was 0.3$& for colicin El and 1% for colicin K. phenicol did not antagonize the effect of colicin on incorporation. Effect of colicin 011 motility. Since a low level of aerobic metabolism is sufficient to support movement in E. coli (2, 3), we tested whether colicin-treated cells would retain their motility. E. cofi K-12 U160 is actively motile when grown aerobically in broth. Colicin El was ad&d at room temperature at a multiplicity of 7.5 (determined by survival after complete ad- sorption), and small hanging drops of control and colicin-treated cells were observed in a phasecontrast microscope. Control cells re- mained vigorously motile, whereas colicin- treated cells slowed down, and by 6 min after addition of colicin nearly all cells had stopped moving. Thus, motility is as sensitive to colicin as are other energy-requiring reactions. Convereioa of omithiae to dtrulline. Con- ceivably, the interference of colicins with macro- molecular biosyntheses and with motility may not result from their effect on ATP levels, but may be a product of some other effect, possibly related to membrane association. It seemed desirable to test the effect of colicins on an energy-requiring reaction involving soluble en- zymes and nonpolymeric substrates. The reac- tion chosen for testing was the conversion of omithine to citrulline. This reaction requires carbamyl phosphate, made by carbamyl phos- phate synthetase from bicarbonate, ammonia (or glutamine), and ATP. The strain used was an arginine-requiring mutant AB1302B (22), which lacks arginino- succinate synthetasc, the enzyme that converts citrulline and ATP to the argininosuccinate. This strain is also derepressed for the enzymes of arginine biosynthesis, because of a second mutation in the regulator gene for the arginine enzymes. Intact cells of AB1302B, when pro- vided with exogenous ornithine in the absence of arginine, synthesize carbamyl-phosphate, trans- fer it to ornithine, and excrete the citrulline thus formed into the medium. Thus, the intact cells serve as an in vivo assay system for the synthesis of carbamyl phosphate. As shown in Table 3, the control cells of AB1302B readily converted ornithine into citrulline. If glucose was omitted, there was little or no production of citrulline; thus, the reaction appears to be dependent upon an exoge- nous energy supply. Addition of colicin El re- duced excretion of citrulline below any signifi- cant level. This effect of colicin does not appear to be a by-product of the etfects of colicin on protein synthesis since chloramphenicol did not reduce citrulline excretion. It is also unlikely that colicin prevented the formation of citrulline by inhibiting the accumulation of ornithine since the equi- librium of the transcarbamylase reaction greatly favors citrulline formation; hence, as with o-nitrophenyl-p-n-galactoside (ONPG) hydroly- sis (27), the reaction should not be subject to interference at the level of accumulation. It is likely, therefore, that this reaction is prevented by colicin through its effect on ATP levels. Effects of coBd~ El aod K 011 respirath. Jacob et al. (21) and Nomura (30) reported that colicins El and K allowed continued respiration. Catabolism of glucose by colicin-treated cells, if it occurs through the normal E. coli pathway, requires at least one ATP-linked phosphoryla- tion catalyzed by phosphofructokinase. [The 6rst 68 FIELDS AND LURIA J. BA~EIUOL. TABLE 2. Effect of colicins on incomoration of 1 3ooo' W' a Samples were treats I with colicin or bufl reel ___ . _ . for 5 to 7 min, the %-labeled carbon source was added, and incorporation was measured 30 mm later. Survival after colicin treatment was 0.1 to 1%. Sample volume: experiment 1, 0.05 ml; experiment 2,0.5 ml (total cell material) or 5 ml (glycogen); experiment 3, 0.5 ml. - - Lobding compound -- idC-glucose i4C-glucose ic-acetate Trichloroacetic acid-precipitable Total Glycogen Acetic acid-pre- cipitable case or acetate +El- +K +El- - +El +CM- +El +El + CM 65,000 335 305 2,~ ljii -40 17,500 15,900 95 120 * Cells grown in 63-glucose, washed, resuspended in medium 63. ' Cells grown in 63-glucose, washed, and resuspended in phosphate buffer + 10-a M MgSOd. d Cells mown in acetate medium, resuspended in buffer, and starved for 60 min. CM = chlorampheni- co1 (100 &/ml). _ phosphorylation of glucoscd-phosphate can probably be carried out by a phosphoenol- pyruvate-linked phosphotransferase (34)]. If the phosphofructokinase reaction continued, the colicin-treated cells would be an example of an ATP-requiring function not inhibited by these colicins. A typical experiment is shown in Fig. 2. Under growth conditions with glucose as carbon source, colicin El had no effect on 01 consumption. When growth was prevented by removal of required amino acids, respiration by control cells was depressed, but colicin actually stimu- lated respiration. In the absence of growth, glucose catabolism may be slowed down by excess ATP, and colicin may restore a faster rate by reducing ATP levels. The effect of colicin El on respiration was studied on several E. coli strains with a variety of substrates and with suspensions of cells grown on different substrates. The significant results of these experiments can be summarized as follows. Respiration rate is least affected by colicin when glucose is the substrate; the total amount of O2 taken up per mole of glucose, however, is always reduced by colicin treatment. The rate of CO* evolution is more strongly affected than the rate of 0, consumption. Respiration with galactose, arabinose, glycerol, or or-glycerol- phosphate is partially inhibited; respiration with succinate is more strongly affected. TABLE 3. Inhibition of conversion of ornithine to citrulline by colicin EP Additions 0.4% glucose.. . . . . . . . . . . . . . 0.4% glucose + chloramphenicol . . . . 0.4yc glucose + colicin El. . . . . . . 0.4% glucose + colicin El + chloram- 20 20 4 phenicol............................ None. . . . . . . . . . . . . . . . . . . 1 a Cells of strain ABl302B growing in medium 63 (with glucose, threonine, leucine, and arginine) were harvested, resuspended in medium without arginine for 15 min, and then again collected and resuspended in medium 63 with threonine, leucine, and 20 rg of L-ornithine per ml. Glucose (0.4%), colicin El (survival 0.2oJ,), and chloramphenicol (50 #g/ml) were added in various combinations. After 60 min, citrulline was assayed in the super- natant fluid of centrifuged samples. The only substrate for which colicin El com- pletely abolished respiration was acetate; the rate of 0~ consumption by suspensions of grow- ing bacteria with acetate as substrate was re- duced to that of colicin-treated cells without substrate. (In some experiments, in which very concentrated suspensions of starved cells were treated with cdicin El, 0, consumption in the VOL. 97, 198 EFFECTS OF COLICINS ON CELLULAR METABOLISM 69 c C 600 300 TIME (mid FIG. 2. Eflect of colicin El on oxygen uptake by E. coli C&M with glucose. Cells of E. coli C&XI grown in medium 63 with glucose, threonine, and leucine were collected, resuspended in medium without amino acids, and shaken for 30 min at 37 C. Then samples of the suspension were resuspended in fully supplemented medium (complete), or without glucose, or without amino acids (-AA), and placed in Warburg flasks for 0, uptake measurements. Colicin El or buffer was added from the side arm at 12 min. The survival of colicin-treated cefls was about 0.1% (measured at 50 min) . presence of acetate continued for about 20 min before inhibition set in. Incorporation of uracil or of counts from labeled acetate into acid- insoluble materials stopped as promptly as with lighter cell suspensions. The reason for the delayed onset of inhibition of respiration re- mains unclarified.) Taken as a whole, the results of the respira- tion studies indicate that catabolism of many carbon sources continues after colicin El treat- ment. Colicin K, whenever tested, gave com- parable results. The lower degree of inhibition observed for glucose than for other carbon sources may be due to the fact that glucose acti- vation can be mediated by the phosphoenolpyru- vate-dependent phosphotransferase, whereas the other substrates require ATP for entry or for activation. The very strong inhibition of respira- tion with acetate by acetate-grown cells may re- fleet some effect of colicin on the functioning of the Krebs cycle or on the uptake of acetate from the medium. Produets of aerobic glucose catabolism. Even though substantial glucose oxidation by colicin- treated cells does occur, its catabolic fate is clearly altered. The linal levels of OS consump- tion are lower, and the production of COz is also reduced. Data from a typical experiment are shown in Table 4. In this and similar experi- ments, light suspensions of washed bacteria from growing cultures were allowed to oxidize a limited amount of *C-glucose in Warburg flasks. OS consumption was monitored to deter- mine when the substrate had been used up, and then the suspensions were analyzed for `Gcon- taining compounds as described in Materials and Methods. Figure 3 illustrates a silicic acid column fractionation of products from labeled glucose obtained with control and El-treated cells in an experiment similar to that of Table 4. The overall effects of colicin El can be sum- marized as follows. The uptake of O2 was re- duced by 40 to 50% and CO* production was reduced by 70 to 80%. As expected, trichloro- acetic acid-precipitable material went from about 40y0 of the added carbon in the controls to less than 1 Y0 in the colicin-treated cells. Pro- duction of acetate, the major nonvolatile product of normal cells, was strongly diminished, whereas pyruvate, absent from the supernatant fluid of the control suspensions, became a major product. In normal cells, a small variable amount of the iC from glucose was found as materials eluted from the silicic acid column by water; this frac- tion was greatly increased by colicin treatment. Thus, the alterations of glucose catabolism caused by the colicins are, first, the replacement of COP and acetate by pyruvate as main products and, second;a substantial increase in substances that are not fractionated by silicic acid column chromatography. To test whether the pyruvate produced by colicin-treated cells was formed from glucose via the glycolytic pathway, as is the case for normal E. cob cells (36), a comparison was made be- tween uniformly labeled and C-l labeled glucose. The distribution of carbon 1 into pyruvate, acetate, and CO* was consistent with glycolytic degradation and not with production of pyruvate by either the Entner-Doudoroff pathway or the hexose-phosphate shunt. Analysis of the water-eluted fraction. Among the soluble substances produced from glucose by colicin-treated cells, the fraction that was eluted with water from the silicic acid column accounted for 45 to 75% of the carbon in different experi- ments. Paper chromatography showed that this fraction did not consist of simple carbohydrates; neither did it form derivatives characteristic of ketones and aldehydes. In chromatography with alkaline solvents, a distinct peak was observed, whose position was altered by pretreatment with alkaline phosphatase. Likewise, the electro- TABLE 4. Effect of colicins El and K on glucose &gradation by E. coli 3ooO Percale of `C input cohmnoe.pontiim(pnmtyaoftotll'~ Time of incubation colicio ritb ;C$rcox Vile da/d Rcco\[ayof gz - ah . -.@ Acidid~bk `"a=' 6) "$gd "`"~gf= Acetate Pyrwate Lactate "t3z ttitl%z& rater -- ----~ Control 3.5 x 101 14 33 47 95 4 1 22 ND ND 17 9 ii I ClO' 3.5 0.1 :1 95 1.6 6 12.4