Hydrolysis of lactose by immobilized microorganisms.

Cells of Lactobacillus bulgaricus, Escherichia coli, and Kluyveromyces (Saccharomyces) lactis immobilized in polyacrylamide gel beads retained 27 to 61% of the beta-galactosidase activity of intact cells. Optimum temperature and pH and thermostability of these microbial beta-galactosidases were negligibly affected by the immobilization. Km values of beta-galactosidase in immobilized cells of L. bulgaricus, E. coli, and K. lactis toward lactose were 4.2, 5.4, and 30 mM, respectively. Neither inhibition nor activation of beta-galactosidase in immobilized L. bulgaricus and E. coli appeared in the presence of galactose, but remarkable inhibition by galactose was detected in the case of the enzyme of immobilized K. lactis. Glucose inhibited noncompetitively the activity of three species of immobilized microbial cells. These kinetic properties were almost the same as those of free beta-galactosidase extracted from individual microorganisms. The activity of immobilized K. lactis was fairly stable during repeated runs, but those of E. coli and L. bulgaricus decreased gradually. These immobilized microbial cells, when introduced into skim milk, demonstrated high activity for converting lactose to monosaccharides. The flavor of skim milk was hardly affected by treatment with these immobilized cells, although the degree of sweetness was raised considerably.     

Hydrolysis of lactose by immobilized microorganisms.

Cells of Lactobacillus bulgaricus, Escherichia coli, and Kluyveromyces (Saccharomyces) lactis immobilized in polyacrylamide gel beads retained 27 to 61% of the beta-galactosidase activity of intact cells. Optimum temperature and pH and thermostability of these microbial beta-galactosidases were negligibly affected by the immobilization. Km values of beta-galactosidase in immobilized cells of L. bulgaricus, E. coli, and K. lactis toward lactose were 4.2, 5.4, and 30 mM, respectively. Neither inhibition nor activation of beta-galactosidase in immobilized L. bulgaricus and E. coli appeared in the presence of galactose, but remarkable inhibition by galactose was detected in the case of the enzyme of immobilized K. lactis. Glucose inhibited noncompetitively the activity of three species of immobilized microbial cells. These kinetic properties were almost the same as those of free beta-galactosidase extracted from individual microorganisms. The activity of immobilized K. lactis was fairly stable during repeated runs, but those of E. coli and L. bulgaricus decreased gradually. These immobilized microbial cells, when introduced into skim milk, demonstrated high activity for converting lactose to monosaccharides. The flavor of skim milk was hardly affected by treatment with these immobilized cells, although the degree of sweetness was raised considerably.     

High-Level Production of β-Galactosidase by Escherichia coli Merodiploids

Two merodiploids of Escherichia coli that contain genes for the lac operon on both chromosome and episome were tested for production of lac enzymes after growth on various carbon sources. The specific activity of beta-galactosidase (and of thiogalactoside transacetylase) was about twice that from haploid cells when grown on glycerol. With succinate as carbon source, the specific activity increased by an additional factor of 3. Up to 25% of the soluble cell protein is beta-galactosidase in these strains, one of which is inducible and the other constitutive. The enzyme is purified easily in high yield by ammonium sulfate fractionation and electrophoresis.     

A single-cell assay of beta-galactosidase in recombinant Escherichia coli using flow cytometry

A flow cytometric method was developed for the assay of beta-galactosidase in single Escherichia coli cells. A new fluorogenic substrate for beta-galactosidase, C(12)FDG, contains a lipophilic group that allows the substrate to penetrate through cell membranes under normal conditions. When the substrate is hydrolyzed by intracellular beta-galactosidase, a green fluorescent product is formed and retained inside the cell. Consequently, the stained beta-galactosidase-positive cells exhibit fluorescence, which is detected by flow cytometry. This new assay was used to analyze the segregational instability caused by a reduction in specific growth rate of the plasmid-bearing cells in the T7 expression system. Induction results in a substantial accumulation of intracellular beta-galactosidase along with a rapid increase in the fraction of plasmid-free cells. Once the cells lose the plasmid, they no longer produce beta-galactosidase, which is reduced by at least half every generation; thus, after staining, the fluorescent, plasmid-bearing cells can be distinguished from the nonfluorescent, plasmid-free cells using flow cytometry. This article describes the feasibility of the flow cytometric assay for single E. coli cells and reports the optimal assay conditions. A direct relationship between beta-galactosidase activity and green fluorescence intensity was found, and the fractions of recombinant cells in batch cultures were analyzed after various levels of induction.     

Influence of amino acids on low-density Escherichia coli responses to nutrient downshifts

A vast bibliography on nutrient effects on high-density cultures exists, while it has been overlooked that low densities of starving cells are often the rule in natural environments. By means of a novel sensitive beta-galactosidase assay, we examined Escherichia coli transitions to minimal media when the cell concentration was 100 to 10,000 cells per ml. As in high-density cultures, the enzyme activity depended on amino acid availability and was subject to catabolite repression and stringent control. In all conditions tested, despite the presence of other nutrient sources, the relationship between beta-galactosidase activity and the l-amino acid pool was hyperbolic. The affinity constant when the amino acid pool was the only nutrient source averaged 14 muM after 90 min and increased up to 222 muM after 4.5 h. While investigating the transition from lag phase to exponential phase, we observed that the cells did not enter into starvation mode in the presence of amino acids, even when the nutrient amount was insufficient to support full survival. Based on these premises, the switch from starvation to hunger was investigated in relation to the amino acid pools. A critical range of concentrations at which E. coli linearly synthesized beta-galactosidase despite, at the same time, suffering a large decrease in cell viability was then recognized. Since both beta-galactosidase production and the dilution rate were reduced by more than half in the absence of leucine, we examined the contribution of leucine to cell recovery capabilities.     

A simple assay for DNA transfection by incubation of the cells in culture dishes with substrates for beta-galactosidase

Transfection efficiency of different cell types as well as promoter strength of cloned genes can be easily determined by direct assay of beta-galactosidase activity encoded from recombinant genes containing the E. coli beta-galactosidase gene. A substrate for beta-galactosidase, o-nitrophenyl-beta-D-galactopyranoside (ONPG), can be added to dishes containing the transfected cells, and the intensity of the colored enzyme product released from either the intact cell or cells lysed in the dishes can be determined. The results obtained by this assay are a reliable measure of transfection efficiency as well as promotor strength of the genes introduced into the cells. In addition, cells expressing the transfected gene can be identified and quantitated under a light microscope after incubation with X-gal. Thus, it is more convenient to use the E. coli beta-galactosidase gene than the chloramphenicol acetyltransferase gene as a reporter gene in the evaluation of DNA transfection.     

Interaction of membrane aminophospholipids of E. coli with fluorodinitrobenzene and trinitrobenzenesulfonate.

E. coli cells were reacted with TNBS in bicarbonate-NaCl buffer, pH 8.5 (buffer A) and in phosphate-NaCl buffer, pH 7.0 (buffer B). In buffer A, DNP-GPE is the major product when FDNB is used. DNP-PE and DNP-LPE are formed in lesser amounts. Phospholipase A activity is high in buffer A. When TNBS is used, the labeling of the lipid components is less than with FDNB and more TNP-PE is formed relative to TNP-GPE. This data suggests that the phospholipases which are located primarily on the outer L-membrane of the cell wall act to a lesser extent on TNP-PE than on DNP-PE. E. coli cells were prelabeled with TNBS and FDNB in buffer A, washed and incubated in buffer A. The endogenous labeled DNP-PE gradually decreased with time with a concomitant increase in DNP-LPE and DNP-GPE due to phospholipase A activity. In contrast, the endogenous labeled TNP-PE also decreased with time as did the endogenous labeled TNP-LPE but a new orange lipid was produced. This lipid is believed to be a derivative of TNP-PE in which one of the nitro groups has been reduced to an amino group by nitroreductase. E. coli cells were prelabeled with TNBS and FDNB in buffer A, washed and incubated in buffer B. Under these conditions with both TNBS and FDNB there is an increase in TNP-PE and DNP-PE with a concomitant decrease in TNP-LPE, TNP-GPE, DNP-LPE and DNP-GPE. These results show that at neutral pH acylation occurs to regenerate TNP-PE and DNP-PE. E. coli cells were incubated with exogenous DNP-GPE or TNP-GPE in buffer A. The DNP-GPE and TNP-GPE were rapidly hydrolyzed by a phosphodiesterase to DNP-ethanolamine and TNP-ethanolamine. An orange derivative was formed which was provisionally identified as a derivative of DNP-ethanolamine or TNP-ethanolamine in which a nitro group has been reduced to an amino group by nitroreductase. The phospholipases and acylating enzymes present in the cell wall of E. coli are active on the dinitrophenyl and trinitrophenyl derivatives of PE and LPE and may act in concert to model and repair the plasma membrane.     

Characterization of an oxygen-dependent inducible promoter system, the nar promoter, and Escherichia coli with an inactivated nar operon

The nar promoter of Escherichia coli, which is maximally induced under anaerobic conditions in the presence of nitrate, was characterized to see whether the nar promoter cloned onto pBR322 can be used as an inducible promoter. To increase the expression level, the nar promoter was expressed in E. coli where active nitrate reductase cannot be expressed from the nar operon on the chromosome. A plasmid with the lacZ gene expressing beta-galactosidase instead of the structural genes of the nar operon was used to simplify an assay of induction of the nar promoter. The following effects were investigated to find optimal conditions: methods of inducing the nar promoter, optimal nitrate and molybdate concentrations maximally inducing the nar promoter, the amount of expressed beta-galactosidase, and induction ratio (specific beta-galactosidase activity after maximal induction/specific beta-galactosidase activity before induction.)The following results were obtained from the experiments: induction of the nar promoter was optimal when E. coli was grown in the presence of 1% nitrate at the beginning of culture; expression of beta-galactosidase was not affected by molybdate; the induction ratio was maximal, approximately 300, when the overnight culture was grown in the flask for 2.5 h (OD(600) is congruent to 1.3) before being transferred to the fermentor; the amount of beta-galactosidase per cell and per medium volume was maximal when E. coli was grown under aerobic conditions to OD(600) = 1.7; then the nar promoter was induced under microaerobic conditions made by lowering dissolved oxygen level (DO) to 1-2%. After approximately 6 h of induction, OD(600) became 3.2 and specific beta-galactosidase activity became 36,000 Miller units, equivalent to 35% of total cellular proteins, which was confirmed from sodium dodecyl sulfate-polyacrylamide gel electrophoresis. (c) 1996 John Wiley & Sons, Inc.     

The mechanism of the synthesis of enzymes. II. Further observations with particular reference to the linear nature of the time course of enzyme formation

1. The pretreatment induction method of studying the formation of beta-galactosidase in E. coli B has been described. 2. It has been found that E. coli B cells have their maximum capacity to form beta-galactosidase, in response to a constant induction stimulus, when they are in the stationary phase of the growth cycle. 3. The concentration of inductor, the nature of the nitrogen source, the duration of the assimilatory phase, oxygen tension, and temperature are factors which affect, and may limit, the rate of beta-galactosidase formation. 4. When limitations imposed by these factors were removed, the time course of induced beta-galactosidase formation was strictly linear from the onset. 5. The implications of this finding were discussed and a new theory of the mechanism of enzyme formation has been proposed. 6. A very satisfactory method of synthesis of ortho-nitrophenol-alpha-D-galactoside has been described. This substance is a suitable chromogenic substrate for the specific determination of alpha-galactosidase activity. 7. Preliminary experiments using this substrate have confirmed the results of respiration studies and shown that in E. coli B alpha-galactosidase formation may be induced by beta- as well as by alpha-galactosides.     

Phospholipase activity in bacteriophage-infected Escherichia coli. III. Phopholipase A involvement in lysis of T4-infected cells.

Bacteriophage studies with Escherichia coli K-12 (gamma)DR-DS-, a mutant lacking the major known fatty acyl hydrolases (phospholipases), and its wild-type parent showed equivalent phage infection with regard to phage production and time of phage release. Further examination of the DR-DS- mutant, however, revealed that the progeny bacteriophage were released without complete dissolution of the host cell. Prolonged cell integrity of the infected mutant was noted by spectrophotometry and supported by direct microscope examination. The phage release occurred at normal "lysis" time with phage yields comparable to that of the wild-type bacteria. Inner membrane degradation was indicated by the release of beta-galactosidase, a cytoplasmic enzyme, and of trichloracetic acid-precipitable RNA. Thus, outer membrane degradation is required for dissolution of phage-infected cells, and this degradation is at least partly dependent on activation of host phospholipases.     

Mutational spectrum and recombinogenic effects induced by aminofluorene adducts in bacteriophage M13

Double-stranded replicative form (RFI) DNA of bacteriophage M13 strain M13mp10 which carries partial lacZ gene has been modified in vitro to various extents with N-hydroxy-2-amino-fluorene (N-OH-AF) and then transfected into E. coli cells. High-performance liquid chromatography (HPLC) analysis results demonstrate that the sole adduct (95%) formed in modified DNA is N-(deoxyguanosine-8-yl)-2-aminofluorene (dG-C8-AF). Approximately 20 adducts per RFI molecule constitute 1 lethal event when plaque-forming ability is assayed on E. coli cells which have received no prior SOS induction. The mutagenicity of dG-C8-AF adducts was assayed by measuring loss of beta-galactosidase activity as a function of adducts per molecule. A dose-dependent increase in Lac- mutants was observed, with a 4-fold increase in mutants per survivor at 30 adducts/molecule. The mutations produced, characterized by DNA sequencing, occur predominantly at either G or C positions different from those observed in the spontaneous mutant spectrum. Restriction-mapping results show that in our assay system, dG-C8-AF adducts induce a previously unreported recombinogenic activity.     

Construction and properties of bifunctionally active membrane-bound fusion proteins. Escherichia coli proline carrier linked with beta-galactosidase

For construction of bifunctionally active membrane-bound fusion proteins, we designed plasmids encoding fusion proteins in which the carboxyl terminus of Escherichia coli proline carrier was joined to the amino terminus of E. coli beta-galactosidase directly or with a collagen linker inserted between the two. The expressions of these fusion proteins complemented deficiencies in both proline transport and beta-galactosidase activity in E. coli cells. The fusion proteins were stable and mostly localized in the cytoplasmic membrane. The proline transport activities of the fusion proteins were kinetically similar to that of the wild type proline carrier. The beta-galactosidase moiety of the collagen-linked fusion protein was liberated from membrane vesicles by collagenase treatment. The Km value of released beta-galactosidase for o-nitrophenyl beta-D-galactopyranoside hydrolysis was similar to that of membrane-bound beta-galactosidase in the fusion protein. These results indicated that the fusion proteins are bifunctionally active and exhibit normal proline transport and beta-galactosidase activities. The crypticity of the beta-galactosidase activity associated with the fusion proteins indicated that the carboxyl terminus of the proline carrier was located on the cytoplasmic side of the membrane.     

Inducible synthesis of beta-galactosidase in disrupted spheroplast of Escherichia coli

A membrane preparation obtained from osmotic lysate of spheroplasts of Escherichia coli cells showed an activity of synthesizing beta-galactosidase which was dependent upon oxidative phosphorylation. The synthesis was inhibited by the addition of actinomycin D or of chloramphenicol. The beta-galactosidase synthesized in the membrane preparation was completely released into the medium, while that synthesized in the spheroplasts and intact cells remained within the cells. The minimum concentration of the inducer, methyl-beta-d-thiogalactoside, required for the induction of beta-galactosidase was 5 x 10(-5)m for intact cells, 3 x 10(-4)m for spheroplasts and 1 x 10(-3)m for membrane preparation. Incorporation of labeled glucose into insoluble components in membrane preparation was extremely low compared with that in intact cells or in spheroplasts. Based on these and other observations, the nature of this membrane preparation is discussed in relation to the structure of E. coli cells.     

A bifunctional chimeric protein consisting of MutS and beta-galactosidase

A bifunctional protein consisting of MutS, a mismatch binding protein and a beta-galactosidase reporter domain has been constructed. The fusion of beta-galactosidase to the MutS C-terminus was obtained by cloning the Escherichia coli lacZ gene encoding beta-galactosidase into a plasmid vector carrying the Thermus thermophilus mutS gene. Milligram amounts of this huge chimeric protein (217 kDa monomer) were purified from 1l of overexpressing E. coli cells using metal-chelate affinity chromatography. The mismatch binding properties of the fusion protein were confirmed by DNA mobility shift assay in polyacrylamide gels. Binding to biotinylated mismatched DNA immobilized on streptavidin microplates followed by colorimetric reaction with X-gal (5-bromo-4-chloro-3-indolyl-beta-D-galactopyranoside), demonstrated both mismatch recognition and beta-galactosidase activity of the chimeric protein. The activity of beta-galactosidase domain of the fusion was similar to that of the native enzyme. A colorimetric assay for beta-galactosidase activity using X-Gal supplemented with NBT (nitro blue tetrazolium) allowed detection of 50 and 500 fmol of the chimeric protein with naked eye in 45 microl volumes after 120 and 15 min incubation, respectively.     

A strategy for obtaining active mammalian enzyme from a fusion protein expressed in bacteria using phospholipase A2 as a model

An active preparation of human phospholipase A2 (PLA2) was made after expression as an insoluble fusion protein in Escherichia coli. The new key elements required for PLA2 isolation were the maintenance of the fusion protein in solution after the initial solubilization and the use of a tryptophan cleavage procedure for regeneration of native PLA2 from the fusion protein. The fusion protein was composed of a beta-galactosidase leader peptide incorporating six consecutive threonine residues to aid in insoluble inclusion body formation, followed by a tryptophan adjacent to the N-terminus of PLA2. The fusion protein was purified from cell lysates, and the leader peptide was cleaved on the C-terminal side of the tryptophan residue with N-chlorosuccinimide. The released PLA2 was refolded and renatured to produce an enzyme with activity comparable to that of other phospholipases A2.     

Electrochemical evaluation of cellular physiological status under stress in Escherichia coli with the rpoS-lacZ reporter gene

We developed an electrochemical detection method for evaluating cellular physiological status based on the stringent response as a means to monitor cell viability. A reporter plasmid was constructed by inserting the beta-galactosidase gene (lacZ) under the control of the rpoS promoter, and then used to transform E. coli cells. Electrochemical responses from the products catalyzed by beta-galactosidase expressed by these E. coli cells were detected using the chronoamperometric technique in a nondestructive manner. Comparisons of response currents between the relA-positive strain and relA-negative strain revealed that increases in these currents were caused by the stringent response due to the stressful alcoholic environment, and thus as a model of stressful cultivating conditions. The current was proportional to the beta-galactosidase activity assayed by a conventional method that required the destruction of cells. The cellular physiological status, which depends on the stringent response as a viability marker, therefore, could then be evaluated online with a current using the rpoS-lacZ reporter gene in the relA-positive strain without pretreatment.     

Use of a phoB'-'lacZ fusion gene to determine the N-terminal amino acid sequence of the phoB protein and to prepare antiserum against the protein

phoB is a positive regulatory gene of the phosphate regulon of Escherichia coli. A plasmid carrying a phoB'-'lacZ fusion gene was constructed by in vitro recombination. A PhoB-LacZ hybrid protein was purified from the cells carrying the plasmid by monitoring beta-galactosidase activity. The amino-terminal amino acid sequence of the PhoB protein was determined by utilizing the hybrid protein. Antiserum against the PhoB protein was prepared by immunizing rabbits with the hybrid protein. The serum thus prepared showed specificity for both the PhoB protein and beta-galactosidase.     

The action of human and rabbit serum phospholipase A2 on Escherichia coli phospholipids

We have compared the properties of phospholipase A (E.C. 3.1.1.4) activity in whole human and rabbit serum toward the phospholipids of Escherichia coli. Using as substrate E. coli labeled during growth with either [1-(14)C]-palmitic acid or [1-(14)C]oleic acid, and then autoclaved to inactivate E. coli phospholipases and to render the labeled phospholipids accessible to exogenous phospholipases, we show that the deacylating activity in both human and rabbit serum is almost exclusively of the A(2) type. Rabbit serum is at least 20-fold more active than human serum. Activity in both sera is maximal at physiological Ca(2+) concentrations (2 mM) and is abolished by ethylenediaminetetraacetic acid. To examine hydrolysis of intact (unautoclaved) E. coli treated with 25% serum, use was made of a phospholipase A-deficient E. coli strain (E. coli S17), thereby eliminating the possible contribution of bacterial phospholipases to degradation. Human and rabbit serum are about equally bactericidal toward E. coli and cause comparable structural damage. However, only rabbit serum produces substantial hydrolysis of the phospholipids of intact E. coli S17. Heated (56 degrees C, 30 min) rabbit serum is non-bactericidal and retains phospholipase A(2) activity toward autoclaved, but not intact E. coli. The ability of heated serum to degrade phospholipids of intact E. coli S17 is restored, however, by adding 25% normal human serum, which is bactericidal. In this combination, doses of heated rabbit serum containing as much phospholipase A(2) activity (toward autoclaved E. coli) as is present in 25% unheated rabbit serum, produce roughly the same extent of hydrolysis of intact E. coli as does normal rabbit serum alone. Low doses with a phospholipase A(2) activity comparable to that of normal human serum elicit little or no hydrolysis. These findings indicate that hydrolysis of the phospholipids of intact E. coli S17 by serum occurs when: 1) the serum is bactericidal, and 2) when sufficient phospholipase A(2) is present. The difference in phospholipid hydrolysis that accompanies killing of E. coli by human or rabbit serum appears to reflect, therefore, the different amounts of phospholipase A(2) activity in the two sera. Phospholipid degradation is not required for the bactericidal action of serum. Bacterial phospholipid breakdown may be important, however, in the overall destruction and digestion of invading bacteria by the host.-Kaplan-Harris, L., J. Weiss, C. Mooney, S. Beckerdite-Quagliata, and P. Elsbach. The action of human and rabbit serum phospholipase A(2) on Escherichia coli phospholipids.     

LYSOSOMAL PHOSPHOLIPASES A1 AND A2 OF NORMAL AND BACILLUS CALMETTE GUERIN-INDUCED ALVEOLAR MACROPHAGES

A single intravenous injection of 0.1 mg of heat-killed Bacillus Calmette Guérin (BCG) in 0.1 ml of Bayol F produced an accumulation of activated alveolar macrophages (BCG induced). Cells were collected 3.5–4.0 wk after injection. Phospholipases A and three lysosomal marker enzymes (acid phosphatase, β-glucuronidase, and lysozyme) were measured in homogenates, and the distribution of the phospholipases A and lysosomal, mitochondrial, and microsomal marker enzymes were examined after sucrose gradient centrifugation of a postnuclear (1,000 g) supernatant. Homogenates of normal and BCG-induced macrophages contained phospholipases A₁ and A₂ which had optimal activity at pH 4.0 in the presence of 2.0 mM ethylenediaminetetraacetate (EDTA). These activities were inhibited 50–70% by 2.0 mM CaCl₂. Homogenates of BCG-induced macrophages had specific activities of β-glucuronidase, acid phosphatase, and lysozyme, which were increased 1.5- to 3.0-fold over the controls, whether expressed as activity per mg protein or activity per 10⁷ cells. The specific activities of the phospholipases A, on the other hand, were consistently lower than those of the control. Distribution of the phospholipases A and the lysosomal marker enzymes after sucrose gradient centrifugation suggested that the phospholipases A active at pH 4.0 in the presence of EDTA are of lysosomal origin since: (a) BCG treatment caused a selective increase in the density of particles which contained both the phospholipases A and three lysosomal marker enzymes; and (b) since the density of mitochondria and microsomes were not affected by BCG treatment. The increase in the density of lysosomes seen here may be related to previously described morphologic changes of BCG-induced alveolar macrophages.     

Cross-adaptive response in Escherichia coli caused by pretreatment with H2O2 against formaldehyde and other aldehyde compounds.

A cross-adaptive response (CAR), defined as a reduction of the effects of an agent by pretreatment with another agent, was demonstrated when E. coli WP2 cells were pretreated with hydrogen peroxide (H2O2) followed by challenging treatment with aldehyde compounds. Pretreatment with a sublethal dose (60 microM) of H2O2 for 30 min made WP2 cells resistant to the killing effects of formaldehyde (FA), and 4 other mutagenic aldehydes: glutaraldehyde, glyoxal, methyl glyoxal and chloroacetaldehyde. CAR was also observed in WP2uvrA (uvrA-) and ZA12 (umuC-) cells, but not in ZA60 (recA-) and CM561 (lexA- (Ind-] cells. A role of recA and lexA in CAR was further suggested by the lack of beta-galactosidase induction in recA- and lexA- cells by H2O2. CAR and beta-galactosidase induction, however, were found to be separate events since CAR was recovered by introducing the recA+ gene into lexA- cells, but no induction of beta-galactosidase by H2O2 was observed in cells with the same gene transfer. These results suggest that H2O2 has the capacity to induce a function which reduces the killing effects of aldehydes, and the function is controlled by the recA gene without involvement of SOS response.