5.3. Induced cell lysis of E. coli by phage LL-H lysis genes

The phage LL-H lysin is not a secreting protein. It can not access the cell wall murein unless the cytoplasmic membrane is somehow damaged. During phage infection the permeabilization of the cytoplasmic membrane is achieved through the function of phage-encoded holin protein (Young 1992, Young & Bläsi 1995). As previously shown by Auerbach and Rosenberg (1987), phage lysis genes may be cloned into E. coli in order to obtain bacterial strains capable of lysis under certain conditions. The applicability of the phage LL-H lysis genes (the lysin gene mur and the holin gene ORF107) for lysis of E. coli was therefore studied. Coexpression of these genes could be used for externalization of intracellular proteins of E. coli (Paper III). Actively growing and dividing cells seemed to be most readily lysed by phage LL-H holin and lysin genes. Host lysis was, however, inefficient in bacterial cultures having an optical density higher than 0.5 (measured at 600 nm, results not shown). Probably the efficiency of the phage holin function decreases with increasing culture cell densities. Bläsi et al. (1984) realized that expression of phage &phis;X174 holin gene in E. coli results in efficient cell lysis only if the culture cell density was under 5 x 10¹⁰. Bacteria grown in poor carbon source refuse to lyse by phage &phis;X174. The membrane integration step of the membrane-spanning proteins (like colicins and phage holins) is probably dependent on the membrane energy. During late logarithmic or stationary growth of bacteria the energy state, the composition or the fluidity of the cytoplasmic membrane may not favor membrane integration of holins. Best externalization of the lysin protein Mur was achieved by performing coexpression of the genes hol and mur in a heat inducible gene expression system (results not shown). Heat shock is known to induce changes in composition and fluidity of bacterial membranes (Mejía et al. 1995), which may enhance the functionality of phage LL-H holin. Heat induction has, however, contradictory effects on bacterial lysis. Young et al. (1989) showed that in elevated temperatures lysis of E. coli by the phage &phis;X174 occurs earlier, but on the other hand the autolytic activity of the host bacterium is substantially decreased. Judged by the poor intracellular accumulation of Mur in a heat inducible expression system, Mur may be stabilized by membrane association, which is disturbed by elevating the temperature of the culture. For large scale Mur-production or as a general lysis system for E. coli, better (i.e., culture density independent) methods were needed. In order to maximize the production of recombinant proteins, high culture densities are preferred in fermentation processes. Therefore, alternative methods for permeablization of the bacterial cytoplasmic membrane were studied. Small molecular weight substances that either destroy the membrane potential (energy poisons) or dissolve the membrane lipids could be used to release intracellularly accumulated cell wall hydrolases to their murein substrate. Chloroform-treatment disrupted mur-expressing bacteria very efficiently (Papers III and IV). Unfortunately chloroform is poorly soluble in aqueous media and has to be very thoroughly removed from the feedstock before purification of proteins in a chromatography column (Paper III). Rapid change in the membrane electric potential is thought to trigger lysis of the host bacterium during phage infection. Similar changes may be induced by “energy poisons” like cyanide, arsein, and azide compounds (Jolliffe & Doyle 1981, Young 1992). Potassium cyanide at 10 mM concentration did not lyse E. coli cells expressing the gene mur. Sodium azide treatment (75 mM) had to be supplemented with the detergent Triton X-100 in order to obtain lysis, indicating that the outer membrane of E. coli prevents the intake of azide into the cell. Alcohols are known to cause changes in the membrane electric state (Paterson et al. 1972). Thymol was found to functionally replace the effect of phage holin or chloroform-treatment in a cell-density-independent manner (Paper IV). The phage LL-H lysin Mur, although shown to be inefficient on the E. coli cell wall (Paper II), was able to lyse E. coli cells after the addition of chloroform or thymol. An increased amount of cell wall hydrolyzing activity inside the bacterial cells (by overexpression of the lysin gene mur) was necessary for lysis by thymol. Suprisingly, the E. coli strains BL21(DE3)pLysE and BL21(DE3)pLysS (Novagen) expressing a low amount of phage T7 lysin and supposed to be easily lysed (Moffat & Studier 1987, Fidler & Dennis 1992), were not lysed by these treatments. Thymol was capable of penetrating the outer membrane and cell wall of E. coli. Being a bactericidal but nontoxic compound, thymol could be safely used for large scale lysis. Thymol was evenly dissolved in the medium (added as a stock solution in ethanol), and it did not disturb chromatographic purification of proteins (Paper IV).

The thymol-triggered lysis system described above could be well applicable for lysis of bacteria when purifying intracellular proteins produced in E. coli. The ability of thymol to trigger lysis without the presence of a membrane-spanning protein suggests that there is no need to clone the most harmful component of the phage lysis system, the holin gene. Bacterial clones containing phage holin gene may suffer from poor stability (Oki 1996, 1997). Instead of cloning all the phage lysis genes as is the case with the defective lambda lysogen (Auerbach & Rosenberg 1987), cloning of the lysin gene would be adaquate in a thymol-triggered lysis system. A lysin gene could be cloned under control of an inducible promoter and introduced into the bacterial cell by a plasmid or integration vector.