Isolation and Characterization of Respiratory-Deficient Mutants of ...

JOURNAL

OF

BACTERIOLOGY, Jan. 1988, P. 78-83

Vol. 170, No. 1

0021-9193/88/010078-06$02.00/0 Copyright X) 1988, American Society for Microbiology

Isolation and Characterization of Respiratory-Deficient Mutants of Escherichia coli K-12t HERB E. SCHELLHORN AND HOSNI M. HASSAN*

Departments of Microbiology and Food Science, North Carolina State University, Raleigh, North Carolina 27695-7624 Received 18 May 1987/Accepted 30 September 1987

Several mutants of Escherichia coli K-12 defective in aerobic metabolism were isolated. One such mutant was found to be deficient in cytochromes, heme, and catalase. Aerobically grown cells did not consume oxygen and could grow only on fermentable carbon sources. Supplementation of the growth medium with deltaaminolevulonic acid, protoporphyrin IX, or hemin did not restore aerobic metabolism. The lack of heme and catalase in mutant cells grown on glucose was not due to catabolite repression, since the addition of exogenous cyclic AMP did not restore the normal phenotype. When grown aerobically on complex medium containing glucose, the mutant produced lactic acid as the principal fermentation product. This pleotropic mutation was attributed to an inability of the cells to synthesize heme, and preliminary data mapped the mutation to between 8 and 13 min on the E. coli genome. Escherichia coli, a facultative anaerobe, is able to utilize both fermentable and oxidizable carbon sources. Oxidative metabolism of nonfermentable carbon sources requires either as a terminal electron acceptor or, under anaerobic conditions, alternative electron acceptors such as fumarate or nitrate. Aerobic metabolism exposes the cell to the deleterious effects of endogenously generated oxygen-derived free radicals (for a review, see reference 19). In response, the cell has evolved several antioxidant enzymes to counter the effects of these oxyradicals. E. coli possesses superoxide dismutase (SOD) and hydroperoxidase, which are specific for the removal of superoxide anions (02-) and hydrogen peroxide (H202), respectively. Mutations in loci affecting superoxide dismutases (6, 18) and hydroperoxidases (18) have been identified in E. coli which render the cell sensitive to oxygen toxicity. Hydroperoxidases are known to be coregulated with components of the respiratory chain (17) and a product of the oxyR gene (8), whereas Mn SOD is thought to be induced via some product of oxygen metabolism (16) that affects the redox state of the cell (28; H. M. Hassan and C. S. Moody, J. Biol. Chem., in press). Though lesions in both aerobic and anaerobic respiratory pathways have been described (for a review, see reference 13), the effect of such mutations on the regulation of enzymes involved in aerobic survival has received little attention. In this work, we isolated mutants which were unable to grow aerobically on media devoid of a source of fermentable carbohydrate. Changes in the aerobic metabolism and heme and cytochrome content were evaluated in one of the mutants. We also examined aerobic induction of MnSOD, catalase, and peroxidase in this mutant. The results of these studies form the body of this report.

peroxidase, and delta-aminolevulonic acid were obtained from Sigma Chemical Co., St. Louis, Mo. Media and other chemicals were obtained from Fisher Scientific Co., Pittsburgh, Pa. Stock solutions of kanamycin and tetracycline were filter sterilized (0.2-,um pore size, Acrodisc; Gelman Sciences) and added to media at final concentration of 25 ,ug

02

ml-'.

Media and growth conditions. Unless otherwise indicated all strains were grown at 37°C in Trypticase soy-yeast extract medium containing 0.25% glucose (TSYG). For the preparation of phage lysates, Luria-Bertani (LB) medium, containing 10 g of tryptone (Difco Laboratories, Detroit, Mich., 5 g of yeast extract (Difco), and 10 g of NaCl per liter, was used. Cell growth was monitored turbidimetrically by measuring optical density at 600 nm (OD6.). One unit of OD6. is equal to 0.280 mg (dry weight) of cells per ml (14). Anaerobiosis was maintained by incubating cultures in a Coy anaerobic chamber as previously described (28) under an atmosphere of 5% C02, 10% H2, and 85% N2. All media and reagents except phage stocks were preequilibrated in the anaerobic chamber for at least 48 h before use. Phage stocks were preequilibrated for 4 h and diluted into preequiibrated buffer-medium immediately before use in anaerobic transduction experiments. For the preparation of crude extracts, overnight (15- to 17-h) cultures grown on TSYG were used to inoculate prewarmed TSYG medium (flask-to-culture ratio, 5 to 1). Aerobic cultures were shaken at 200 rpm. Cultures were harvested at the indicated optical densities by centrifugation (10,000 x g, 10 min, 4°C) and suspended in 0.05 M potassium phosphate buffer (pH 7.8) containing 0.1 mM EDTA. Cell extracts were prepared by sonication and dialyzed as previously described (28). Mutagenesis. Mutagenesis was performed with XplacMu9 and XpMu507 as described previously (4). After infection of strain SS5074, dilutions were plated under anaerobic conditions onto MacConkey plates containing 25 ,ug of kanamycin ml-' and incubated for 48 h. These master plates were then replica plated onto MacConkey plates and incubated aerobically. Small, dark red colonies which appeared on anaerobically incubated MacConkey plates but failed to grow on aerobic plates were considered presumptive oxygen-sensitive mutants and were examined further.

MATERIALS AND METHODS

Bacterial and bacteriophage strains. The bacterial and bacteriophage strains used are described in Table 1. Chemicals and enzymes. Diaminobenzidine, chloramphenicol, kanamycin, tetracycline, xanthine oxidase, horseradish *

Corresponding author.

t Paper

no. 10675 of the journal series of the North Carolina Agricultural Research Service, Raleigh, NC 27695-7601.

78

TABLE 1. Bacterial and phage strains used Strain

SS5074 HS18A BW6160 BW7261

AplacMu9

XpMu5O7 P1 vir

79

HEME-DEFICIENT MUTANT OF E. COLI K-12

VOL. 170, 1988

reference

Genotype

F- A(argF-lac)U169 pro-22 met-90 trpA trpR his-85 rpsL azi-9 gyrA x- Pis As SS5074 but respiratory

defective Hfr Broda 8 zdh-57::TnJO relAl spoTI metBI A- Ar Hfr Cavalli leu-63::TnlO tonA22 A(argF-lac)U169 ompF627 relAl spoTI T2r Mu cIts62 ner+ A' 'ara' Mu 'S 'lacZ lac yF lacA' kan c1857 Sam7 Mu A' B'

S. Short This work B. Bachmann B. Bachmann 4 4 S. Short

To ensure that mutants harbored a single fusion phage, P1 vir lysates of the mutants were prepared as described by Silhavy et al. (32) under anaerobic conditions and were used to transduce the parental strain (E. coli SS5074) to the mutant phenotype (i.e., small, dark red phenotype on anaerobic MacConkey medium and failure to grow aerobically on the same medium). These secondary transductants were used in further studies. Enzyme assays. SOD was assayed by the cytochrome c method (27), catalase was assayed by monitoring the decomposition of hydrogen peroxide at 240 nm (3), and peroxidase was monitored by the oxidation of ortho-dianisidine (35). Oxygen consumption was determined as described before (14) with a Clark oxygen electrode equipped with a 5.0-ml chamber. Protein in crude extracts was measured by the method of Lowry et al. (24) with bovine setum albumin as the standard. Protein in whole cells was solubilized with deoxycholate (7) before quantification. Determination of glucose and organic acids. Glucose, acetate, formate, and lactate in culture supernatants were separated by high-pressure liquid chromatography (model 512; Waters Associates, Inc., Milford, Mass.) with an Interaction Ion-300 column (300 by 7.8 mm; Phenomenex, Inc., Randro Paloverdes, Calif.). Samples (10 F±l) were eluted at a flow rate of 0.6 ml min-' and a temperature of 32°C with 0.01 N H2SO4 as the eluant. Peak elution was monitored with a Waters R-401 differential refractometer. Concentrations of individual components were calculated with a Spectra Physics SP4100 computing integrator with reference to external standards after correction for media controls. Polyacrylamide gel electrophoresis. Catalase and SOD isozymes were resolved by nondenaturing gel electrophoresis on 10% polyacrylamide tube gels as described by Davis (10). Catalase gels were stained by the peroxidase-diaminobenzidine method (9), and SOD gels were stained by the nitroblue tetrazolium method (2). Determination of pyridine hemochromogen. Whole cells from late-stationary-phase cultures were collected by centrifugation (5,000 x g, 15 min, 4°C) and were washed two times with cold 0.05 M potassium phosphate buffer (pH 7.0) containing 1.0 mM MgCl2. Heme extractions were performed essentially as described by Fuhrop and Smith (11). Washed cell suspensions were dried at 60°C in a vacuum oven (model 5831; National Appliance Co.) and extracted three times with 5.0-ml volumes of acetone containing 5% (vol/vol) 12 N HCI. Acetone extracts were neutralized with 2 to 3 ml of 3.0 M sodium acetate and extracted with

peroxide-free ether. The ether extract was evaporated to near dryness in a vacuum oven (40°C) and then extracted into a solution of 2.1 M pyridine-0.075 N NaOH. Pyridine hemochromogen samples were reduced by adding 25 ,ul of a freshly prepared sodium dithionite solution (200 mg ml') to 2.5 ml of sample in an anaerobic cuvette (Precision Cells, Inc. Hicksville, N.Y.) which had been previously bubbled with N2 for 1 to 2 min to displace dissolved oxygen. Identical samples were oxidized by vigorous shaking and served as reference controls. The difference spectrum (i.e., reduced minus oxidized) for the hemochromogen was determined between 400 and 600 nm by using a dual-beam Kontron 810 spectrophotometer. Protoheme IX concentration was calculated by using the absorption difference between 558 and 542 nm assuming an extinction coefficient of 20.7 mM'1 cm-1 (11).

Cytochrome determination. Cytochromes were determined from the difference spectra of reduced minus oxidized cell suspensions (33). Cultures were grown aerobically in TSYG medium, harvested, and washed as described above for the preparation of pyridine hemochromogen. Samples were diluted to the indicated optical densities and scanned from 400 to 600 nm. Test samples were reduced by the addition of sodium dithionite or lactate (20 mM), whereas reference samples were oxidized by vigorous shaking (33). RESULTS Isolation of presumptive oxygen-sensitive mutants. Eighteen independent kanamycin-resistant insertion mutants were isolated. The mutants formed small, dark red colonies when grown anaerobically on MacConkey-lactose plates but failed to grow aerobically on the same medium. Despite the fact that these mutants could grow and were red in color on MacConkey lactose medium under anaerobic conditions, they did not produce measurable amounts of 3-galactosidase activity. Further studies demonstrated that 3 of the 18 mutants were unable to grow aerobically on nutrient agar plates even when the plates were supplemented with deltaaminolevulonic acid, hemin, or nonfermentable carbon sources (data not shown). One of these three mutants, strain HS18A, was chosen for further study. P1 lysates of strain HS18A, prepared under anaerobic conditions, were used to transduce the parental strain to the mutant phenotype, i.e., inability to grow aerobically on MacConkey lactose plates. All kanamycin-resistant transductants were unable to grow aerobically on MacConkeylactose plates. Strain HS18A was able to grow, albeit poorly relative to strain SS5074, on aerobic plates of nutrient agar or LB medium supplemented with 0.5% glucose. However, strains HS18A and SS5074 grew equally well under anaerobic conditions (Table 2). TABLE 2. Growth kinetics of E. coli SS5074 and HS18A grown under anaerobic and aerobic conditions on TSYG mediuma

Strain

Growth conditions

Generation time (min)

Finalb

HS18A

Anaerobic Aerobic

64 96

1.10 0.82

SS5074

Anaerobic Aerobic

62 36

11.10

a b

Preequilibrated medium was inoculated at Determined after 17 h of growth.

an initial

OD6w

1.45

OD6w of 0.020.

80

SCHELLHORN AND HASSAN

Oxygen consumption. Mid-log-phase cultures were examined for oxygen uptake. The endogeneous rate of oxygen consumption by strain HS18A was extremely low (1.0 nmol of 02 min-' mg of dry weight-'). This basal rate did not increase when excess glucose was added to the reaction mixture. In contrast, the parental strain (SS5074) consumed 903 nmol of 02 min-1 mg of cell dry weight-' in the presence of excess glucose. Ninety three percent of this rate was inhibitable by 1.0 mM cyanide. These results indicate that strain HS18A does not possess a functional electron transport chain. Growth kinetics, glucose consumption, and organic acid production. Although strain HS18A was isolated on the basis of a presumptive sensitivity to oxygen, we found that it was able to grow aerobically when glucose was supplied (13.5 mM in TSYG), although at a greatly reduced rate. Strain HS18A formed minute colonies on solid glucose-minimal medium supplemented with the appropriate growth factors. In aerobically incubated rich medium (TSYG), strain HS18A grew to a low final OD6. relative to SS5074 (Table 2 and Fig. 1). In addition, under anaerobic conditions, the generation time of strain HS18A was comparable to that of SS5074, whereas it was about 2.7 times longer under aerobic conditions (Table 2). An analysis of aerobic glucose consumptioh and organic acid production revealed that the parental strain consumed glucose at a slightly faster rate than strain HS18A and initially formed organic acids (lactate, acetate, and formate) that were subsequently metabolized to CO2 and H20. In contrast, strain HS18A produced mainly lactic acid and appeared to be ultimately limited in its growth by the lack of available substrate and by the low pH of the medium (Fig. 1 and 2). Though the high-pressure liquid chromatography system employed was unable to resolve other fermentation products such as ethanol, the organic acids detected accounted for greater than 80% of the glucose consumed during the fermentative phase of growth. Induction of SOD, catalase, and peroxidase. The parental strain, when shifted to aerobic conditions, induced MnSOD,

J. BACTERIOL.

E E

0 C)

w -j

-) 4

0 n C-)

CD

CZ

0

TIME (h) FIG. 2. Glucose consumption and organic acid production by E. coli SS5074 and HS18A during aerobic growth on TSYG medium. Samhples (1.0 ml) of the culture were removed at the times indicated and centrifuged at 12,000 x g for 10 min in a microcentrifuge to remove cells. The supernatants were stored at -20°C until subsequent high-pressure liqulid chromatographic analysis. Symbols: 0, glucose; O, lactate; A, acetate; 0, formate.

catalase, and peroxidase in accordance with previously published data (15). In contrast, strain HS18A, although able to induce MnSOD and peroxidase when shifted to aerobic conditions, appeared to be devoid of catalase under either anaerobic or aerobic conditions as determined by its inability to decompose hydrogen peroxide as measured at 240 nm. In addition, strain HS18A colonies grown aerobically on LB medium containing 0.5% glucose were unable to evolve oxygen when flooded with 30% hydrogen peroxide. The specific activity of SOD was low in strain HS18A in relation to that of the parental strain (Table 3). Exposure of anaerobically grown cultures to air was sufficient to cause SS5074 HS18A induction of the Mn SOD in both strains. The lower absolute enzyme levels in the mutant were probably due to the § v. 30Q reduced growth rate of this strain. Hassan and Fridovich (16) previously observed that the specific activity of SOD increases as a function of specific growth rate. The peroxidase 8.0 activity of strain HS18A was also considerably lower than that of SS5074 when cells were grown under aerobic or X 7X0 anaerobic conditions (Table 3). Polyacrylamide gel electrophoresis of dialyzed cell exL) ~~~~~~~~~~~~~6.0 tracts revealed the presence of a single clear band which corresponded in electrophoretic migration to the hydroperoxi0lase I (Fig. 3), using the designations proposed by Loewen and his co-workers (21). Since the activity was not detectable in the liquid assay (Table 3), it must account for only a small proportion of the total catalase activity normally found 0 in E. coli, as has been suggested by others (22). The absence of detectable catalase activity led us to examine the sensitivity of HS18A to hydrogen peroxide in filter disk inhibition assays. HS18A was found to be much more sensitive than the parental strain, typically exhibiting zones of inhibition of TIME (h) 43 mm compared with 22 mm for the parental when 0.6 mg of FIG. 1. Kinetics of aerobic growth and pH changes of E. coli hydrogen peroxide was applied to 6-mm filter disks (BBL SS5074 and HS18A growing in TSYG under aerobic conditions. Microbiology Systems, Cockeysville, Md.) on nutrient agar Log-phase cells grown on TSYG medium were used to inoculate prewarmed (37'C) TSYG medium (initial OD6w of 0.05). plates containing 0.5% glucose.

HEME-DEFICIENT MUTANT OF E. COLI K-12

VOL. 170, 1988

81

TABLE 3. Aerobic induction of SOD, catalase, and peroxidasea

~~~Activityof (limol min-' mg

50Db (U mg of protein1) SOD'(Umgofprotein-')

Growth conditions

Strain

protein-l)

MnSOD

HySOD

FeSOD

Total

Catalase

(10-s)

HS18A

Anaerobic Shifted to air Aerobic

NDC 0.18 3.21

ND 0.32 2.42

1.60 2.44 8.24

1.60 2.94 13.87

ND ND ND

0.37 0.85 3.86

SS5074

Anaerobic Shifted to air Aerobic

ND 11.54 11.29

ND 0.88 4.22

2.98 7.13 17.02

2.98 19.56 32.48

0.99 6.66

1.49 4.10 12.80

12.35

a Overnight cultures grown anaerobically on TSYG medium were used to inoculate preequilibrated TSYG medium at an initial OD6. of 0.025. After 4 h of anaerobic growth, the culture was shifted to aerobic conditions and was allowed to grow for additional 4 h, a period of time which has previously been shown to be sufficient for the de novo synthesis of induced MnSOD (15). Control cultures were grown in parallel under completely aerobic and anaerobic conditions for 8 h as described in Materials and Methods. b MnSOD, manganese-containing enzyme; FeSOD, iron-containing enzyme; HySOD, hybrid enzyme containing both Mn and Fe. cND, Not detected.

Effect of cyclic AMP on the growth of strain HS18A. Because glucose is known to catabolically repress the synthesis of catalase (17) and cytochromes (5) in E. coli, it was necessary to investigate the effect of exogenously added cyclic AMP on the growth of strain HS18A. The addition of 5.0 mM cyclic AMP did not stimulate growth of strain HS18A in LB medium containing 0.5% glucose, nor did it derepress catalase biosynthesis as determined by direct liquid assay and by gel electrophoresis (data not shown). Absence of cytochroines and protoheme IX in HS18A. An intact cytochrome chain is normally required for aerobic growth with 02 as a terminal electron acceptor. Figure 4A shows the lactate- and dithionite-reduced spectra of whole cell suspensions of strains HS18A and SS5074. The data clearly show that the parental strain possessed the normal complement of b-type cytochromes (a Soret band at 425 nm and a broad alpha-beta band at 558 nm) when grown under aerobic conditions, whereas strain HS18A contained no detectable cytochromes. Hemne is a necessary prosthetic group of cytochromes and hydroperoxidases, and therefore the heme content of the respiratory-defective mutant was examined. The data in Fig. 4fB show that HS18A completely lacked protoheme IX (as indicated by the absence of a Soret band at 407 nm), whereas SS 5074

HS18A

A

B

C

A

B

C

the parental strain (SS5074) possessed 0.179 nmol of protoheme IX per mg of protein. The lack of detectable heme coupled with failure of the heme precursor delta-aminolevulonic acid to restore the wild-type phenotype to strain HS18A suggested that the genetic lesion in the mutant is involved in the latter stages of heme biosynthesis. Such mutants have been observed (26, 31) and reportedly do not respond to exogenous heme supplied in the medium (31), an observation consistent with the growth characteristics of strain HS18A. Genetic mapping. The lack of heme suggested that the lesion was in one of the genes involved in heme biosynthesis. The respiratory mutation was mapped by rapid patch (23) and interrupted conjugation matings (34). In these mating experiments, Hfr strains BW6160 and BW7261 restored the normal respiratory phenotype within the first few minutes of mating and before the introduction of their respective tetracycline markers. This places the mutation within the region of 8 to 13 min of the chromosome, i.e., between the origins of transfer of these Hfr strains, which transfer this region from opposite directioons. Since the hemB and hemH genes map at 8 and 11 min, respectively, P1 crosses were performed to determine which gene may have been mutated. Transductions with the nearby lac and purE markers were unsuccessful. Preliminary data suggest that the mutation may lie very close to the lac deletion in the parental strain, which would be consistent with a mutation in the hemB gene (data not shown).

HP-Il HP-I

FIG. 3. Catalase isozymes of E. coli SS5074 and HS18A produced during aerobic and anaerobic growth on TSYG medium. Cell extracts (100 ,ug of protein in a volume of 80 ,ul) were applied to 10% polyacrylamide gels and electrophoresed at 4 mA per gel. Anaerobically grown cultures were divided into two equal volumes. Chloramphenicol (150 ,ug ml-, final concentration) was added to one fraction 15 min before removal from the anaerobic chamber (A). The second portion was transferred to a sterile flask (culture-to-flask volume ratio, 1 to 5) which was subsequently agitated (200 rpm) under aerobic conditions to allow the synthesis of oxygen-induced proteins (B). A control culture was continuously incubated under aerobic conditions (C).

DISCUSSION The use of oxygen as an electron acceptor, although energetically favorable to many cells, may exert deleterious effects since the partially reduced oxygen intermediates are toxic and mutagenic (15). To cope with this oxidative stress, E. coli, like most other aerobic organisms, has evolved defense mechanisms which include the detoxifying enzymes SOD, catalase, and peroxidase. Cytochrome oxidase, the terminal component of the aerobic respiratory chain, can also be thought of as a part of the response of the cell to oxidative stress since it reduces dioxygen to the more safe water molecule. In view of the toxicity of oxygen, it is not surprising that a deficiency in one or more of these components can impair the growth of the cell in the presence of oxygen (6, 18).

82

SCHELLHORN AND HASSAN

WAVELENGTH (nm)

FIG. 4. (A) Difference spectra (reduced minus oxidized) of washed cell suspensions of E. coli SS5074 and HS18A. Cells were reduced by the addition of 20 mM lactate (final concentration), whereas the reference cell was oxidized by vigorous shaking. Cell suspensions were diluted to an OD6T of 20 (corresponding to approximately 4.0 mg of protein ml ) before scanning. Similar results were obtained with sodium dithionite as the reductant. (B) Difference spectra (reduced minus oxidized) of pyridine hemochromogen samples prepared from cell extracts of E. coli SS5074 and HS18A. Reduced samnples were prepared by the addition of 25 ,ul of sodium dithionite (200 mg ml-') to a sample volume of 2.5 ml. Reference samples were oxidized by vigorous shaking.

Several lines of evidence suggest that respiratory metabolism is coupled to the expression of defense enzymes such as SOD and catalase. Using ubiquinone-requiring mutants of E. coli, Hassan and Fridovich (17) have shown that respiring cells have higher levels of catalase than cells growing fermentatively. Poole et al. (29) have presented evidence that a previously described cytochrome species has catalase activity and appears, on the basis of structural and kinetic evidence, to be identical to the HPI catalase. Increased manganese SOD activity may also be correlated to respiration since nitrate, in the absence of oxygen, induces the enzyme (Hassan and Moody, in press). In this study, several mutants were isolated on the basis of their failure to grow aerobically in MacConkey-lactose medium. Morphologically, these mutants appeared as small, dark red colonies when grown anaerobically on the same

J. BACTERIOL.

medium. Physiological and biochemical characterization of one of the mutants, HS18A, revealed that it lacked the ability to synthesize heme, and the mutation was mapped between 8 and 13 min on the E. coli genome. Two genes involved in heme biosynthesis have been identified (1) in this region of the E. coli genome; delta-aminolevulonic acid dehydratase at 8 min (hemB) and ferrochetalase at 11 min (hemH). Phenotypically, the nmutant described in this work appears to be similar to a hemB mutant first described by Sasarman and co-workers (31). Thus strain HS18A produces dwarf colonies on complex medium, is deficient in heme, cytochromes, and catalase, and does not respond to exogenously supplied heme or protoporphyrin IX. In contrast, mutants deficient in ferrochetalase (hemH) are not impaired in their ability to use oxidizable carbon sources and produce red-brown colonies when grown on complex medium due to the intracellular accumulation of protoporphyrin IX (30). Though preliminary mapping and the phenotype of HS18A suggest a mutation in hemB, this does not rule out the possibility that the mutation is in a regulatory locus affecting heme biosynthesis. Mutants in catabolite repression control (cya and crp) may phenotypically resemble respiratory mutants (5, 13). The possibility that the mutation in HS18A was in one of these loci was ruled out by the finding that exogeneously added cyclic AMP (5 mM) did not restore the wild-type growth characteristics on complex medium. Furthermore, it is unlikely that the lesion affects the levels or availability of intracellular iron, since this mutant produced a functional iron-containing SOD (FeSOD) (Table 3). Heme is an important prosthetic group of many cellular components including catalases, peroxidases, and cytochromes. Reconstitution experiments have demonstrated that the inactive apoforms of these respiratory components are synthesized even in the absence of heme biosynthesis (12, 26). E. coli mutants impaired in heme biosynthesis have been difficult to isolate since this organism is reportedly impermeable to hemin (31). Our mutagenic protocol, which selects for strains defective in oxidative metabolism rather than those possessing a specific requirement for heme, does not depend on the use of mutants which have been previously rendered heme permeable (25, 26). The data in this work indicate that a deficiency in heme synthesis restricts the cell to an essentially homofermentative type of metabolism (Fig. 2), generating lactic acid as the principal product rather than acetic acid, ethanol, and formate. Although the deficiency in heme biosynthesis of strain HS18A did not affect the degree of induction of MnSOD by oxygen, both Mn and FeSODs were produced at lower levels than in the parental strain. Hassan and Fridovich (16) have previously noted that increases in SOD activity in cells growing under conditions of glucose limitation can be correlated to increased rates of cellular respiration. Although catalase activity was not detected in cell extracts of HS18A, a band of catalase activity was found in nondenaturing gels stained for catalase activity (Fig. 3). Although the reasons for this are not clear, we note that nonheme pseudocatalases have been reported in other procaryotes (20). The nature of the oxygen sensitivity of this respiratory mutant requires further investigation. Though restricted to a fermentative type of metabolism under both aerobic and anaerobic conditions, strain HS18A grows more slowly in the presence of oxygen than in its absence, with generation times of 96 and 64 min, respectively (Table 2). The elucidation of the nature of this sensitivity will be the subject of future studies.

HEME-DEFICIENT MUTANT OF E. COLI K-12

VOL. 170, 1988 ACKNOWLEDGMENTS This work was supported by grant DMB-8609239 from the National Science Foundation. We express our gratitude to T. Melton, S. Short, and T. Blevins for their advice and suggestions. We also thank R. F. McFeeters for his generous loan of the high-pressure liquid chromatography equipment and B. J. Bachmann, B. Wanner, and G. Weinstock for supplying phage and bacterial strains.

1. 2. 3.

4.

5. 6. 7.

8.

9.

10. 11. 12. 13.

14.

15. 16.

LITERATURE CITED Bachmann, B. J. 1983. Linkage map of Escherichia coli K-12, edition 7. Microbiol. Rev. 47:180-230. Beauchamp, C., and I. Fridovich. 1971. Superoxide dismutase: improved assays and an assay applicable to acrylamide gels. Anal. Biochem. 44:276-287. Beers, R. F., and I. W. Sizer. 1952. A spectrophotometric method for measuring the breakdown of hydrogen peroxide by catalase. J. Biol. Chem. 195:133-140. Bremer, E., T. J. Silhavy, and G. M. Weinstock. 1985. Transposable placMu bacteriophages for creating lacZ operon fusions and kanamycin resistance insertions in Escherichia coli. J. Bacteriol. 160:1092-1099. Broman, R. L., W. J. Dobrogosz, and D. C. White. 1974. Stimulation of cytochrome synthesis in Escherichia coli by cyclic AMP. Arch. Biochem. Biophys. 162:595-601. Carlioz, A., and D. Touati. 1986. Isolation of superoxide dismutase in Escherichia coli: is superoxide dismutase strictly necessary for life? EMBO J. 5:623-630. Chaykin, S. 1971. Biochemistry laboratory techniques, p. 18-20. John Wiley & Sons, Inc., New York. Christman, M. F., R. W. Morgan, F. S. Jacobson, and B. N. Ames. 1985. Positive control of a regulon for defenses against oxidative stress and some heat-shock proteins in Salmonella typhimurium. Cell 41:753-762. Clare, D. A., M. N. Duong, D. Darr, F. Archibald, and I. Fridovich. 1984. Effects of molecular oxygen on the detection of superoxide radical with nitroblue tetrazolium and an activity stain for catalase. Anal. Biochem. 140:532-537. Davis, B. J. 1964. Disc electrophoresis II: method and application to human serum proteins. Ann. N.Y. Acad. Sci. 121: 404-427. Fuhrop, J.-H., and K. M. Snith. 1975. Laboratory methods, p. 806-808. In K. M. Smith (ed.), Porphyrins and metalloporphyrins. Elsevier Science Publishing, Inc., New York. Haddock, B. A. 1973. The reconstitution of oxidase activity in membranes derived from a 5-aminolevulonic acid-requiring mutant of Escherichia coli. Biochem. J. 136:877-884. Haddock, B. A. 1977. The isolation of phenotypic and genotypic variants for functional characterization of bacterial oxidative phosphorylation, p. 95-120. In B. A. Haddock and W. A. Hamilton (ed.), Microbial energetics. Cambridge University Press, Cambridge. Hassan, H. M. 1984. Exacerbation of superoxide radical formation by paraquat. Methods Enzymol. 105:523-532. Hassan, H. M., and I. Fridovich. 1977. Enzymatic defenses against the toxicity of oxygen and of streptonigrin in Escherichia coli. J. Bacteriol. 129:1574-1583. Hassan, H. M., and I. Fridovich. 1977. Physiological function of

17. 18.

19.

20.

21. 22.

23. 24.

25. 26. 27. 28.

29.

30. 31.

32. 33.

34. 35.

83

superoxide dismutase in glucose-limited chemostat cultures of Escherichia coli. J. Bacteriol. 130:805-811. Hassan, H. M., and I. Fridovich. 1978. Regulation of the synthesis of catalase and peroxidase in Escherichia coli. J. Biol. Chem. 253:6445-6450. Hassan, H. M., and I. Fridovich. 1979. Superoxide, hydrogen peroxide and oxygen tolerance of oxygen-sensitive mutants of Escherichia coli. Rev. Infect. Dis. 1:357-367. Hassan, H. M., and I. Fridovich. 1980. Superoxide dismutases: detoxication of a free radical, p. 311-332. In W. B. Jakoby (ed.), Enzymatic basis of detoxication, vol. 1. Academic Press, Inc., New York. Kono, Y., and I. Fridovich. 1983. Isolation and characterization of the pseudocatalase of Lactobacillus plantarum. J. Biol. Chem. 258:6015-6019. Loewen, P. C., J. Switala, and B. L. Triggs-Raine. 1985. Catalases HPI and HPII are induced independently in Escherichia coli. Arch. Biochem. Biophys. 243:144-149. Loewen, P. C., B. L. Triggs, C. S. George, and B. E. Hrabarchuk. 1985. Genetic mapping of katG, a locus that affects the synthesis of the bifunctional catalase-peroxidase hydroperoxidase I in Escherichia coli. J. Bacteriol. 162:661-667. Low, B. 1973. Rapid mapping of conditional and autotrophic mutants of Escherichia coli. J. Bacteriol. 113:798-812. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193:265-275. McConville, M. L., and H. P. Charles. 1979. Isolation of haemin-requiring mutants of Escherichia coli K-12. J. Gen. Microbiol. 113:155-164. McConville, M. L., and H. P. Charles. 1979. Mutants of Escherichia coli permeable to haemin. J. Gen. Microbiol. 113: 165-168. McCord, J. M., and I. Fridovich. 1969. Superoxide dismutase: an enzymatic function of erythrocuprein. J. Biol. Chem. 193: 265-275. Moody, C. S., and H. M. Hassan. 1984. Anaerobic biosynthesis of the manganese-containing superoxide dismutase in Escherichia coli. J. Biol. Chem. 259:12821-12825. Poole, R. K., B. S. Baines, and C. A. Appleby. 1986. Haemoprotein b-590-a reducible catalase and peroxidase: evidence for its close relationship to hydroperoxidase I and a 'cytochrome a,b' preparation. J. Gen. Microbiol. 132:1525-1539. Powell, K. A., R. Cox, M. McConville, and H. P. Charles. 1973. Mutations affecting porphyrin biosynthesis in Escherichia coli. Enzyme 16:65-73. Sasarman, A., and T. Horodniceanu. 1967. Locus determining normal colony formation on the chromosome of Escherichia coli K-12. J. Bacteriol. 94:1268-1269. Silhavy, T. J., M. L. Berman, and L. W. Enquist. 1984. Experiments with gene fusions. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. Smith, L. 1954. Bacterial cytochromes: difference spectra. Biochim. Biophys. Acta 50:299-314. Wanner, B. L. 1986. Novel regulatory mutants of Escherichia coli. J. Mol. Biol. 191:39-58. Worthington Biochemical Corp. 1972. Worthington enzyme manual, p. 43-45. Worthington Biochemical Corp., Freehold, N.J.