Journal of Applied Microbiology 2001, 90, 771±778
Induction of complement sensitivity in Escherichia coli by citric acid and low pH C. OcanÄa-Morgner and J.R. Dankert Department of Biology, University of Louisiana at Lafayette, LA, USA 550/9/00: received 27 September 2000, revised 17 January 2001 and accepted 24 January 2001
Ä A - M O R G N E R A N D J . R . D A N K E R T . 2001. C. OCAN
Aims: The lytic functions of the complement system play an important role in the control of Gram-negative infections. Complement-resistant Escherichia coli LP1395 (O18) grown under normal conditions can survive the bactericidal action of complement present in human serum. Towards elucidating the mechanisms of complement resistance, the resistance of E. coli LP1395 grown under conditions of low pH and in the presence of citric acid was tested. Methods and Results: E. coli LP1395 becomes sensitive to complement after growth in the presence of citric acid at pH 5. Complement resistance could be restored when the cells were transferred to pH 7 media. However, this recovery was greatly impaired when the cells were transferred to pH 7 media with chloramphenicol. This implies that protein synthesis may be involved in complement resistance. The cells exposed to citric acid at pH 5 showed no indication of a generalized outer membrane (OM) permeability when compared with those grown under normal conditions in terms of sensitivity to lysozyme, uptake of lipophilic dye, or sensitivity to a number of antibiotics. Conclusions: Complement-resistant LP1395 may acquire a sensitivity to complement due not to a generalized disruption of the OM barrier, but possibly to the alteration of the activity of one or more normal complement resistance factors. Signi®cance and Impact of the Study: The elucidation of the mechanisms of complement resistance of Gram-negative pathogens would bring important information about bacterial infections. Complement resistance factors could also be potential targets in antimicrobial therapies.
INTRODUCTION The complement system provides humans with an innate defense mechanism against a number of pathogenic organisms. This effect is accomplished by three functions: (i) opsonization and phagocytosis; (ii) direct killing by the membrane attack complex (MAC); and (iii) activation of in¯ammatory processes. The complement components in the serum can be activated by Gram-negative bacteria via the classical pathway in the presence of speci®c antibodies, or directly by the alternative pathway. In complementsensitive bacteria, this leads to the deposition of the terminal proteins of the complement cascade (C5b±C9), or membrane
Correspondence to: Dr J.R. Dankert, Department of Biology, University of Louisiana at Lafayette, PO Box 42451, Lafayette, LA 70504±2451, USA (e-mail:
[email protected]). ã 2001 The Society for Applied Microbiology
attack complex (MAC), which results in the death of the cells (Taylor 1983). C9 is the major membrane-perturbing element of the MAC (Esser et al. 1979; Dankert and Esser 1987; Dankert 1991; Taylor 1995). One key aspect of successful infection by pathogenic bacteria is their avoidance of killing by the opsonic and lytic functions of the complement system. Although opsonization plays a major role in combating infectious diseases, MAC-mediated killing is important in the control of Gram-negative infections (Taylor 1995). The mechanisms by which pathogenic microbes resist the direct killing of complement can involve several processes: failure to activate either early or later complement components, production of intrinsic complement regulatory proteins, and blockage of antibodies (Mof®t and Frank 1994; Taylor 1995). Studies on mechanisms of serum resistance of Gram-negative bacteria have focused mainly on components
772 C . O C A NÄ A - M O R G N E R A N D J . R . D A N K E R T
and properties of the outer membrane (OM). The length of the lipopolysaccharide (LPS) O side chain may be an important factor in complement resistance (Taylor 1995). Other mechanisms of resistance have also been studied, which involve the possible acquisition of resistance through plasmid- (Pramoonjago et al. 1992) and non-plasmid1 encoded OM proteins (Weiser and Gotschlich 1991). These mechanisms imply that modi®cations made to intrinsic elements and properties of the OM can play a role in complement resistance (Feingold 1969; Taylor 1983; Taylor 1995). Previous studies on complement-resistant Escherichia coli LP1395 indicate that this strain goes through a transient inhibition of inner membrane activity after exposure to complement, in contrast to complement-sensitive strains which suffer permanent damage (Dankert 1989). This indicates that a possible component of cellular resistance may be due to a reversal of the inhibitory effects of the MAC (Dankert 1989). The same strain became sensitive to the bactericidal action of serum when grown in carbon- and magnesium-limited conditions (Taylor 1978; Taylor et al. 1981). This sensitivity seemed to be related to a lower lipopolysaccharide O side chain sugar to core-sugar ratio, assessed after one hour of serum incubation (Taylor et al. 1981). In an investigation of the effect of nutrients on complement sensitivity, it was found that citric acid at pH 5 could alter the complement sensitivity of this normally complement-resistant strain. Citric acid is a hydroxylic and multicarboxylic weak acid and has been used mainly as a food preservative due to its capacity to inhibit bacterial growth (O'Sullivan 1969; Salmond et al. 1984; Cherrington et al. 1991a; Roe et al. 1998). It can also act as a sequestering agent of divalent ions, such as Ca2+ and Mg2+, and have a disrupting effect on the OM of Gram-negative bacteria. When Pseudomonas aeruginosa was exposed to sodium citrate at pH 7á8 and 9á0, it became more permeable to the lytic action of lysozyme (Ayres et al. 1993, 1998). This effect was prevented by the addition of Ca2+ and Mg2+ ions (Ayres et al. 1993, 1998). The authors concluded that the chelating effect of citrate at these pH levels made the OM more permeable to the lytic action of lysozyme. Similar results were found in another work with E. coli O157:H7 (Cutter and Siragusa 1995). When the cells were grown in the presence of 100 mmol l±1 citrate at pH 7 they became more permeable to the bacteriocin, nisin and this effect was also prevented by the addition of Ca2+ and Mg2+ ions. As stated, it was found that complement-resistant E. coli LP1395 grown in media containing citric acid at pH 5 can lose complement resistance. The induction of the sensitivity required both citric acid and a pH of 5, and was not due to the chelating properties of citric acid.
MATERIALS AND METHODS Reagents Microbiological media were obtained from Difco (Detroit, MI, USA). Normal human serum (NHS) and complement protein C9 were prepared from human blood as described previously in Biesecker and MuÈller-Eberhard (1980). Polyclonal antibodies against C8 and C9 were obtained as described previously in Dankert (1989). Secondary peroxidase-conjugated antibody (anti-rabbit IgG), citric acid, lysozyme and other chemical reagents were obtained from Sigma (St. Louis, MO, USA). Bacteria and growth conditions Complement-resistant E. coli LP1395 (O18) was grown in Luria-Bertani broth plus glucose and 20 mmol l±1 citric acid at pH 5 (LBCA5), pH 6 (LBCA6) and pH 7 (LBCA7). Cells were also grown in Luria-Bertani broth plus glucose without citric acid at pH 7 (LB7) and pH 5 (LB5). Before incubation with serum, the cells were centrifuged, washed and suspended in imidazole buffer (IB++): 10 mmol l±1 imidazole, 150 mmol l±1 NaCl, 0á1 mmol l±1 CaCl2, 1 mmol l±1 MgCl2, at pH 7, to an optical density (O.D.) of 0á5±0á6 at 600 nm. Escherichia coli C600, a complementsensitive strain, was also used. Bactericidal activity of serum Washed and suspended cells were incubated with different concentrations (0á5, 1, 5, 10, 25 and 50%) of NHS at 37°C for 30 min. Cells grown with and without 60 lg ml±1 chloramphenicol were incubated with 50% NHS, and samples were taken at different time intervals. Heating serum at 60°C for 30 min (hd-NHS) abrogated complement activity. hd-NHS was used as control at a concentration of 50%. Serial dilutions were plated and incubated at 37°C overnight. Colony-forming units (cfu) were counted for viability (compared with cells treated the same way but incubated with IB++ alone), which was obtained from quadruplicate platings and expressed as mean standard deviation. Cells were also incubated with 5% NHS plus polyclonal antibodies against C9 (approximately 24 lg ll±1) for 30 min at 37°C. Serial dilutions were plated and incubated at 37°C overnight. Colony-forming units were counted for viability (compared with cells treated the same way but incubated with IB++ alone) as described above. Recovery of resistance to serum Escherichia coli LP1395 was grown in LBCA5 to early log phase. The cells were centrifuged and suspended in LB7,
ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 771±778
LOSS OF COMPLEMENT RESISTANCE IN E. COLI
LB5 or LBCA7 to the original volume. Cells were also suspended in LB7 plus 60 lg ml±1 chloramphenicol. Thereafter, the cells were incubated at 37°C and samples were taken at various time intervals. The samples were washed, resuspended in IB++ and incubated with 50% NHS at 37°C for 30 min. Serial dilutions were plated and incubated at 37°C overnight. Colony-forming units were counted for viability (compared with cells treated the same way but incubated with IB++ alone) as described above.
773
360 nm (half bandwith 10 3). Data are expressed as mean standard deviation from six readings. The ®nal methanol concentration did not exceed 0á2% and the total dilution caused by reagent additions was always less than 1%. To avoid recovery of serum resistance in a buffer at neutral pH, the ¯uorescence was measured within 5 min after resuspension in 10 mmol l±1 imidazole and 150 mmol l±1 NaCl at pH 7. Whole-cell ELISA
Lysozyme assay Increased susceptibility to the bacteriolytic action of lysozyme is an indication of permeabilization of the OM (Vaara 1992; Nikaido 1996). After growth in LB7, LBCA7 or LBCA5, the cells were washed and resuspended in 10 mmol l±1 tris, 150 mmol l±1 NaCl at pH 7 (TBS), with 1 mg ml±1 of lysozyme. The suspensions were incubated at 37°C. Samples were taken at different time intervals and serial dilutions were plated and incubated at 37°C overnight. Colony-forming units were counted for viability as described above. 1-N-Phenylnaphthylamine ¯uorescence assay The lipophilic probe 1-N-Phenylonaphthylamine (NPN) was used to assess membrane integrity by measuring the ¯uorescence emitted when the probe accessed hydrophobic regions of bacterial membranes. Disruption of the OM barrier can be detected by an increase in the ¯uorescence emitted by NPN (Tsuchido et al. 1989; Vaara et al. 1990; Nikaido 1996). A Tecan SpectraFluor Plus (Tecan US, 2 Research Triangle Park, North Carolina, USA) with a temperature-regulated chamber was used. After growth in LB7, LBCA5 and Luria-Bertani broth plus 1 mmol l±1 EDTA (LBEDTA), cells were resuspended in 10 mmol l±1 imidazole and 150 mmol l±1 NaCl at pH 7 (O.D. of cells at 600 nm: 0á62±0á77); 199 ll of the suspensions were added to the wells of microtitre plates. NPN dissolved in methanol was added to a ®nal concentration of 20 lmol l±1 and the ¯uorescence was monitored using an emission ®lter of 435 nm (half bandwith 20 5) and an excitation ®lter of
A modi®cation of a whole-cell ELISA (Rautemaa et al. 1998) was used to quantify the amount of C8 deposited on cells exposed to R-9 serum. Washed cells were incubated with a sub-lethal concentration (0á5%) of C9-depleted serum (R9) for 30 min at 37°C. The cells were washed three times with IB++ and resuspended to an O.D. of 0á8 at 600 nm. Aliquots (50 ll) were dispensed into the wells of microtitre plates and allowed to dry overnight at room temperature. The cells were incubated for 1 h at 37°C with primary antibody against C8. After washing, cells were incubated with secondary peroxidase-conjugated antibody (anti-rabbit IgG) for 1 h at 37°C. Absorbance was read using a 405 nm ®lter on an ELISA reader (Tecan SLT Spectra, Tecan US, Research Triangle Park, North Carolina, USA) from quadruplicate readings, expressed as mean standard deviation. Absorbances in wells with hd-NHS-incubated bacteria were subtracted as background. Statistical method For the NPN uptake values and C8 binding values, a two-tailed unpaired Student's t-test was used. Signi®cant difference was considered at a P of < 0á05. RESULTS Loss of resistance to serum Table 1 shows that E. coli LP1395 grown in the presence of 20 mmol l±1 citric acid at pH 5 had a survival of 1% 0á6 after incubation with 50% NHS. When the cells were grown
Table 1 Percentage survival S.D. of Escherichia coli LP1395 grown in different conditions* Luria-Bertani broth with 20 mmol l±1 citric acid at pH:
LBCA5 plus different concentrations of Mg2+
LBCA5 plus 10 mmol l)1
7 (LBCA7)
6 (LBCA6)
5 (LBCA5)
10 mmol l±1
20 mmol l±1
40 mmol l±1
Ca2+
LB5
LBCA5 (hd-NHS)à
124 22
101 22
1 0á6
1 0á3
1 0á3
1 0á3
1 0á3
75 4
84 15
*Cells were prepared and incubated in 50% serum as described in Materials and Methods. Luria-Bertani broth at pH 5. àCells were incubated with heat-denatured serum (hd-NHS). ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 771±778
774 C . O C A NÄ A - M O R G N E R A N D J . R . D A N K E R T
80
60
% Survival
without citric acid at pH 5, the survival was 75% 4. Citric acid did not have the same effect at pH 6 or 7. The addition of calcium and magnesium ions to the media containing citric acid at pH 5 did not inhibit the effect of citric acid (Table 1). Figure 1 shows that at serum concentrations between 5 and 10%, the cells lost resistance to serum. Agglutination of bacteria due to the use of high concentrations of antibacterial antibodies in the serum may cause errors in the counting of surviving colonies. No agglutination of the cells was observed by microscopy after incubation with serum (data not shown). Moreover, no signi®cant decrease in colony counts was observed using heat-treated serum (Table 1). Figure 2 shows that the survival was sensitive to the concentration of citric acid. At a concentration of 5 mmol l±1, survival was less than 5%. The kinetics of the onset of complement sensitivity induced by citric acid can be seen in Fig. 3. The sensitivity induced by citric acid was not immediate, becoming apparent after the cells had been exposed to these conditions for at least 10 min. The addition of 60 lg ml±1 chloramphenicol (an inhibitor of protein synthesis) seemed to delay the onset of sensitivity (Fig. 3). The mechanism of killing was dependent on the presence of C9, as the survival of E. coli LP1395 grown in LBCA5 exposed to 5% NHS supplemented with polyclonal antibodies against C9 was greatly improved to 88 4%.
40
20
0 0
1
5
10
15
20
-1
Citric acid concentration (mmol l )
Fig. 2 Survival of Escherichia coli LP1395 when grown in LuriaBertani broth with different concentrations of citric acid at pH 5, after incubation with NHS. Cells were grown in Luria-Bertani broth + glucose with different concentrations of citric acid at pH 5. The cells were then washed and resuspended in IB++. Cells were incubated with 50% normal human serum and 50% hd-NHS at 37°C for 30 min. Serial dilutions were plated and incubated at 37°C. Viability was obtained as described in materials and Methods. The results are the means of four separate experiments, and vertical bars represent standard deviation
120
100
% Survival
80
60
40
20
0 0
Fig. 1 Survival of Escherichia coli LP1395 when grown in LuriaBertani broth with citric acid at pH 5, after incubation with different concentrations of NHS. Cells were grown in Luria-Bertani broth + glucose and 20 mmol l±1 citric acid at pH 5 (LBCA5). The cells were then washed and resuspended in IB++. Cells were incubated with 0á5, 1, 5, 10 and 25% NHS at 37°C for 30 min. Serial dilutions were plated and incubated at 37°C. Viability was obtained as described in materials and Methods. The results are the means of four separate experiments, and vertical bars represent standard deviation
5
10 20 Time (min)
40
60
Fig. 3 Survival of Escherichia coli LP1395 over time after exposure to NHS when grown in LBCA5. Cells were grown in LB7. Cells were then washed and transferred to LBCA5 (j) and LBCA5 with 60 lg ml±1 chloramphenicol (h). Aliquots of cells were taken, washed and suspended in IB++. Cells were then incubated with 50% NHS at 37°C for 30 min. Serial dilutions were plated and incubated at 37°C. Viability was obtained as described in Materials and Methods. The results are the means of four separate experiments, and vertical bars represent standard deviation
ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 771±778
LOSS OF COMPLEMENT RESISTANCE IN E. COLI
Recovery of resistance to serum Figure 4 shows the kinetics of the recovery of resistance of E. coli LP1395 in LB7 with and without 60 lg ml±1 chloramphenicol. The recovery of resistance to serum was not immediate. There was a signi®cant recovery of resistance after 20 min when the cells were resuspended in LB7. However, the rate of recovery was drastically reduced when chloramphenicol was added to the media, implying that protein synthesis may be involved in the re-acquisition of serum resistance. The use of chloramphenicol, under these conditions and not exposed to serum, did not have a lethal effect on the cells (data not shown). Recovery of resistance was sensitive to pH as resuspension in a medium without citric acid, but with a pH of 5, prevented the cells from recovering resistance (Fig. 5). The presence of citric acid at pH 7 delayed the recovery of resistance (Fig. 5).
775
membrane permeability to lysozyme when the cells were grown in LBCA7. This effect was inhibited when magnesium ions were added to the media (data not shown), indicating that chelation of cations is involved at pH 7.
Lysozyme assay To determine if the exposure of E. coli LP1395 to citric acid at pH 5 made the OM more permeable to noxious compounds in general, cells were exposed to 1 mg ml±1 lysozyme after being grown in LB7, LBCA7 and LBCA5. Disruption of cell membrane permeability leads to a rapid sensitivity to lytic enzymes such as lysozyme (Vaara 1992; Nikaido 1996). Figure 6 shows that there was an increase in
Fig. 4 Recovery of resistance of Escherichia coli LP1395 in LB7 (j) and LB7 + chloramphenicol (h) after growth in LBCA5. Escherichia coli LP1395 was grown in LBCA5. Cells were centrifuged and suspended in LB7 and LB7 plus 60 lg ml±1 chloramphenicol in the original volume and incubated at 37°C. The samples were then washed, resuspended in IB++ and incubated with 50% NHS. Serial dilutions were plated and incubated at 37°C overnight. Colony-forming units were counted for viability as described in Materials and Methods. The results are the means of four separate experiments, and vertical bars represent standard deviation
Fig. 5 Recovery of resistance of Escherichia coli LP1395 in LB5 (j) and LBCA7 (h) after growth in LBCA5. Escherichia coli LP1395 was grown in LBCA5. Cells were centrifuged and suspended in LB5 and LBCA7 in the original volume and incubated at 37°C. The samples were then washed, resuspended in IB++ and incubated with 50% NHS. Serial dilutions were plated and incubated at 37°C overnight. Colonyforming units were counted for viability as described in Materials and Methods. The results are the means of four separate experiments, and vertical bars represent standard deviation
Fig. 6 Lysozyme assay. Cells were grown in LB7 (j), LBCA7 ( ) and LBCA5( ). Cells were then washed, suspended and incubated at 37°C in 10 mmol l±1 Tris, 150 mmol l±1 NaCl at pH 7á5 (TBS) with 1 mg ml±1 lysozyme. Samples were taken at different intervals and survival was obtained as described in Materials and Methods. The results are the means of four separate experiments, and vertical bars represent standard deviation
ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 771±778
776 C . O C A NÄ A - M O R G N E R A N D J . R . D A N K E R T
Similar results were obtained in other studies of membrane permeability with citric acid at neutral pH (Ayres et al. 1993, 1998 Cutter and Siragusa 1995). However, the conditions required to induce serum sensitivity (citric acid at pH 5) did not lead to an increased permeability or sensitivity to lysozyme. In another study, lactic acid at pH 4 also failed to increase lysozyme sensitivity in E. coli O157:H7 (Alakomi et al. 2000). Cells were also exposed to a variety of hydrophobic and hydrophilic antibiotics: erythromycin, gentamicin, chloramphenicol, kanamycin, novobiocin, colistin and ampicillin. No differences were observed in the sensitivity of E. coli LP1395 to these antibiotics when grown under the various conditions described (data not shown). Alternatively, LPS release was not found from the membrane of cells grown in LBCA5 as determined by use of silver-stained SDS-polyacrylamide gel of proteinase K-treated cell-free supernatant ¯uids (data not shown). 1-N-Phenylnaphthylamine ¯uorescence assay Disruption of the O antigen of the OM can cause the exposure of the lipid A core of the LPS molecules (Goldman et al. 1984; Nikaido 1996). This allows ¯uorescent lipophilic molecules to partition into the membrane, leading to an increase in the emitted ¯uorescence of the molecule (Tsuchido et al. 1989; Vaara et al. 1990; Nikaido 1996). Figure 7 shows the NPN ¯uorescence for the cells grown in LB7 and LBCA5. The ¯uorescence values of the added probe showed no signi®cant differences. However, a signi®cant increase in NPN ¯uorescence was noted for cells exposed to EDTA (LBEDTA). EDTA has been reported to disrupt the association of LPS molecules, causing an increase in the permeability of the OM (Vaara 1992; Nikaido 1996).
Fig. 7 Average of ¯uorescence values after addition of NPN. A total of 199 ll of cells suspended in 10 mmol l±1 imidazole and 150 mmol l±1 NaCl were added to wells of microtitre plates (O.D. of cells at 600 nm: 0á62±0á77). NPN dissolved in methanol was added to a ®nal concentration of 20 lmol l±1 and the ¯uorescence was monitored at 465 nm (excitation 360 nm). A Tecan SpectraFluor with temperature-regulated chamber was used as described in Materials and Methods. The results are the means of six separate experiments, and horizontal bars represent standard deviation. LBEDTA: LB7 plus 1 mmol l±1 EDTA. *Signi®cant difference P < 0á05; **non-signi®cant difference P > 0á05
Whole-cell ELISA In order to determine whether complement components could be formed on the cells after growth under various conditions, and to quantify the relative amount of MACs assembled on the surface of the cells, antibodies to the C8 component of the MAC were used in an ELISA (Rautemaa et al. 1998). The assay showed deposition of MAC components on the cells grown in LB7, LBCA7 and LBCA5 (Fig. 8). The relative amount of C8 deposited on LP1395 grown at pH 7 was similar to that deposited on the normally complement-sensitive strain (C600). This ®nding agrees with a study by Dankert (1989) who found that the amount of radio-labelled C8 bound on the surface of E. coli LP1395 and C600 is similar. Interestingly, signi®cantly less C8 was deposited on the cells grown under conditions that rendered the cells sensitive to complement (Fig. 8).
Fig. 8 C8 binding ELISA. Cells were incubated with R9 serum, washed three times with IB++ and resuspended to an O.D. of 0á8 at 600 nm. Aliquots were dispensed into the wells of microtitre plates and allowed to dry overnight at room temperature. The cells were incubated with primary antibody against C8 followed by secondary peroxidase-conjugated antibody (anti-rabbit IgG). The absorbances were read using a 405 nm ®lter on an ELISA reader (Tecan SLT Spectra) as described in Materials and Methods. Absorbances in wells with hd-NHS-incubated bacteria were subtracted as background. The results are the means of four separate experiments, and horizontal bars represent standard deviation. C600: E. coli C600 grown in LB7. *Signi®cant difference P < 0á05; **non-signi®cant difference P > 0á05
ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 771±778
LOSS OF COMPLEMENT RESISTANCE IN E. COLI
DISCUSSION Citric acid at concentrations above 5 mmol l±1 conferred sensitivity to the bactericidal action of NHS to complement resistant E. coli LP1395 when grown at pH 5. Several lines of evidence indicate that this acquired sensitivity is not due to a generalized or global change in the permeability of the OM as a result of the chelating action of citrate for Ca2+ and Mg2+ ions. The apparent ion af®nity of citrate is less at pH 5 than at pH 7 (Chaberek and Martell 1959; O'Sullivan 1969; Bell 1977). If loss of resistance was due to a chelating action of citrate at pH 5, it should also be present at pH 7. The suspension of the cells in IB plus Ca2+ and Mg2+ ions (IB++) prior to incubation with serum would also argue that sensitivity is not due to chelation of cations. The OM of Ps. aeruginosa became more permeable to lysozyme when exposed to 25±35 mmol l±1 citrate at pH 7á8 and pH 9, an effect that was reversed with the addition of Mg2+ ions (Ayres et al. 1993, 1998). In the present study, the addition of magnesium and calcium ions to the media did not inhibit the effect of citric acid at pH 5. It was also found that E. coli LP1395, when exposed to lysozyme after growth in citric acid at pH 5, was not as sensitive to the enzyme as the cells grown in LB7. The chelating action of citrate at pH 7 was indeed seen, as cells grown at pH 7 with citrate were more sensitive to lysozyme. The ¯uorescence values of the lipophilic probe NPN when added to cells did not show signi®cant differences between E. coli LP1395 when grown in LBCA5 and LB7. However, when grown in LB7 plus the chelator EDTA (LBEDTA), the NPN ¯uorescence was signi®cantly enhanced (see Fig. 7). Helander and MattilaSandholm (2000) showed an increase in NPN intake in E. coli O157:H7 cells exposed to citric acid at low pH. However, the same strain had a similar increase when exposed to buffer acidi®ed with HCl (Alakomi et al. 2000). The use of a different strain in this study may explain the difference in NPN uptake with E. coli O157:H7. Further, using transmission electron microscopy, other authors have found no evidence of OM disruption in E. coli and Salmonella spp. when grown in a yeast extract minimal medium at pH 5 with approximately 50 mmol l±1 citric acid (Cherrington et al. 1991b; Thompson and Hilton 1996). Lactic acid at pH 4 disrupts the OM of E. coli O157:H7 (Alakomi et al. 2000). Since a similar effect was observed with HCl, the authors concluded that this effect could not be due to the chelation of cations from the OM, but rather to the protonation of anionic LPS components such as carboxyl and phosphate (Alakomi et al. 2000). Although the recovery of resistance to serum was delayed for more than 1 h in LB5 (see Fig. 5), a protonation effect would not explain the time course of the loss and recovery of resistance. Further, growth of LP1395 in media with low pH adjusted with HCl (LB5) did not render the cells sensitive to serum (Table 1). This
777
indicates that a mechanism different from protonation, or in addition to it, could be involved in the loss of serum resistance. The killing of the cells was C9-dependent, as the addition of antibodies against C9 to cells incubated with NHS led to an enhancement in cell survival. An explanation for the loss of serum resistance was further complicated by the fact that there was signi®cantly less binding of C8 in cells grown in LBCA5 and exposed to NHS (see Fig. 8). The growing conditions may alter the way MAC components bind to the membrane of the cells, masking some C8 epitopes that are normally detected by antibodies against C8. However, this is unlikely as the antibodies used in this study were polyclonal. Escherichia coli LP1395 has been reported to become more sensitive to the bactericidal action of serum when grown at near maximal rates under carbon- and magnesium-limited conditions (Taylor 1978, 1983; Taylor et al. 1981). The authors suggested that expression of novel proteins might be involved in the lack of resistance (Taylor 1978, 1983; Taylor et al. 1981). For the present work, the cells were not nutrient-limited, but the ability to delay both the loss and the recovery of complement resistance by chloramphenicol would argue that some protein synthesis or protein component is involved in resistance. Identi®cation of factor(s) that are expressed or lost as a result of these sensitivity-inducing conditions kinetically relevant to the induction and loss of sensitivity is being attempted. These factors could be potential targets in antimicrobial therapies. REFERENCES Alakomi, H.-L., SkyttaÈ, E., Saarela, M., Mattila-Sandholm, T., LatvaKala, K. and Helander, I.M. (2000) Lactic acid permeabilizes Gramnegative bacteria disrupting the outer membrane. Applied and Environmental Microbiology 66, 2001±2005. Ayres, H., Furr, J.R. and Russell, A.D. (1993) A rapid method of evaluating permeabilizing activity against Pseudomonas aeruginosa. Letters in Applied Microbiology 17, 149±151. Ayres, H., Furr, J.R. and Russell, A.D. (1998) Effect of divalent cations on permeabilizer-induced lysozyme lysis of Pseudomonas aeruginosa. Letters in Applied Microbiology 27, 372±374. Bell, C.F. (1977) Industrial applications. In Principles and Applications of Metal Chelation ed. Atkins, P.W., Holker, J.S.E. and Holliday, A.K. pp. 132±138. Oxford: Clarendon Press. Biesecker, G. and MuÈller-Eberhard, H.J. (1980) The ninth component of human complement: puri®cation and physicochemical characterization. Journal of Immunology 124, 1291±1296. Chaberek, S. and Martell, A.E. (1959) Commercial applications of 3 chelating agents. In Organic Sequestering Agents. pp. 297±415. New York: John Wiley & Sons. Cherrington, C.A., Hinton, M., Mead, G.C. and Chopra, I. (1991a) Organic acids: chemistry, antibacterial activity and practical applications. Advances in Microbial Physiology 32, 87±108. Cherrington, C.A., Hinton, M., Pearson, G.R. and Chopra, I. (1991b) Short-chain organic acids at pH 5 kill Escherichia coli and Salmonella
ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 771±778
778 C . O C A NÄ A - M O R G N E R A N D J . R . D A N K E R T
4
spp. without causing membrane perturbation. Journal of Applied Bacteriology 70, 161±165. Cutter, C.N. and Siragusa, G.R. (1995) Population reductions of Gram-negative pathogens following treatments with nisin and chelators under various conditions. Journal of Food Protection 58, 977±983. Dankert, J.R. (1989) Complement-mediated inhibition of function in complement-resistant Escherichia coli. Journal of Immunology 142, 1591±1595. Dankert, J.R. (1991) Resistance of Escherichia coli to osmotically introduced complement component C9. Infection and Immunity 59, 109±113. Dankert, J.R. and Esser, A.F. (1987) Bacterial killing by complement: C9-mediated killing in the absence of C5b-8. Biochemical Journal 244, 393±399. Esser, A.F., Kolb, W.A., Podack, E.R. and MuÈller-Eberhard, H.J. (1979) Molecular reorganization of lipid bilayers by complement: a possible mechanism for membranolysis. Proceedings of the National Academy of Science USA 76, 1410±1414. Feingold, D.S. (1969) The serum bactericidal reaction IV. Phenotypic conversion of Escherichia coli from serum-resistance to serumsensitivity by diphenylamine. Journal of Infectious Diseases 120, 437±444. Goldman, R.C., Joiner, K. and Leive, L. (1984) Serum-resistant mutants of Escherichia coli O111 contain increased lipopolysaccharide, lack an O antigen-containing capsule, and cover more of their lipid A core with O antigen. Journal of Bacteriology 159, 877±882. Helander, J.M. and Mattila-Sandholm, T. (2000) Fluorometric assessment of Gram-negative bacterial permeabilization. Journal of Applied Microbiology 88, 213±219. Mof®tt, M.C. and Frank, M.M. (1994) Complement resistance in microbes. Springer Seminars in Immunopathology 15, 327±344. Nikaido, H. (1996) Outer membrane. In Escherichia coli and Salmonella: Cellular and Molecular Biology Vol. 1, ed. Neidhardt, F.C. pp. 29±47. Washington D.C.: American Society for Microbiology Press. O'Sullivan, W.J. (1969) Stability constants of metal complexes. In Data for Biochemical Research ed. Dawson, R.M.C., Elliot, D.C., Elliot, W.H. and Jones, K.M. pp. 423±435. Oxford: Oxford University Press. Pramoonjago, P., Kaneko, M., Kinoshita, T. et al. (1992) Role of TratT protein, an anticomplementary protein produced in
Escherichia coli by R100 factor, in serum resistance. Journal of Immunology 148, 827±836. Rautemaa, R., Jarvis, G.A., Marnila, P. and Meri, S. (1998) Acquired resistance of Escherichia coli to complement lysis by binding of glycophosphoinositol-anchored protectin (CD59). Infection and Immunity 66, 1928±1933. Roe, A.J., McLaggan, D., Davidson, I., O'Byrne, C. and Booth, I.R. (1998) Perturbation of anion balance during inhibition of growth of Escherichia coli by weak acids. Journal of Bacteriology 180, 767±772. Salmond, C.V., Kroll, R.G. and Booth, I. (1984) The effect of food preservatives on pH homeostasis in Escherichia coli. Journal of General Microbiology 130, 2845±2850. Taylor, P.W. (1978) The effect of the growth environment on the serum sensitivity of some urinary Escherichia coli strains. FEMS Microbiological Letters 3, 119±122. Taylor, P.W. (1983) Bactericidal and bacteriolytic activity of serum against Gram-negative bacteria. Microbiological Reviews 47, 46±83. Taylor, P.W. (1995) Resistance of bacteria to complement. In Virulence Mechanisms of Bacterial Pathogens ed. Roth, J.A., Bolin, C.A., Brogden, K.A., Minion, F.C. and Wannemuehler, M.J. pp. 49±64. Washington D.C.: American Society for Microbiology Press. Taylor, P.W., Messner, P. and Parton, R. (1981) Effect of the growth environment on cell-envelope components of Escherichia coli in relation to sensitivity to human serum. Journal of Medical Microbiology 14, 9±19. Thompson, J.L. and Hilton, M. (1996) Effect of short-chain fatty acids on the size of enteric bacteria. Letters in Applied Microbiology 22, 408±412. Tsuchido, T., Aoki, I. and Takano, M. (1989) Interaction of the ¯uorescent dye 1-N-Phenylonaphthylamine with Escherichia coli cells during heat stress and recovery from heat stress. Journal of General Microbiology 135, 1941±1947. Vaara, M. (1992) Agents that increase the permeability of the outer membrane. Microbiological Reviews 56, 395±411. Vaara, M., Plachy, W.Z. and Nikaido, H. (1990) Partitioning of hydrophobic probes into lipopolysaccharides bilayers. Biochimica et Biophysica Acta 1024, 152±158. Weiser, J.N. and Gotschlich, E.C. (1991) Outer membrane protein A (OmpA) contributes to serum resistance and pathogenicity of Escherichia coli K-1. Infection and Immunology 59, 2252±2258.
ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 771±778
