INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

During the growth transition of bacteria from exponential to stationary phase, the expression of growth-related genes is mostly turned off and instead a set of genes required for stationary-phase survival is switched on (19, 23, 24). For this drastic change in gene expression pattern, structural and functional modulations take place on both transcriptional and translational apparatuses (23, 24). Previously, we identified a 100S form of ribosome dimers in stationary-phase Escherichia coli cells and the association of a small basic protein of 55 amino acid residues, ribosome modulation factor (RMF), with those ribosome dimers (47). RMF is one of the stationary-phase-specific gene products (47, 51), and the rmf mutant strain loses cell viability in stationary-phase culture more rapidly than the wild-type strain (51). The addition of RMF to 70S ribosomes promotes their conversion in vitro into 100S dimers and inhibits translation in vitro (46), supporting the idea that the 100S dimers are a storage form for ribosomes in stationary phase (46, 51).

In order to explore this hypothesis, we have examined the relationship between the formation of ribosome dimers and the fate of E. coli cells. Up to the present time, data concerning stationary-phase adaptability and its variation among E. coli strains have been obtained using a small number of E. coli laboratory strains. It is noted, however, that the laboratory strains often carry mutations in the genes which are expressed only in the stationary phase, probably because these strains have never been exposed to lengthy culture. Surprisingly, some laboratory strains contain mutations that inactivate the RNA polymerase ^S subunit (25), which is required for recognition of some stationary-phase-specific gene promoters (19, 23, 24). In this study, we have analyzed some representative strains from the ECOR collection, which comprises a set of 72 E. coli reference strains isolated from a variety of animal hosts and geographical locations (34), as well as some laboratory strains with or without mutations affecting the ability to undergo ribosome dimerization.

Previously, the growth characteristics of some of the ECOR strains as well as some fresh natural isolates were compared with the kinetic properties of their ribosomes in vitro (30, 31, 35). The variability in the kinetic parameters characteristic of ribosome functions indicates that there is no unique wild-type ribosome phenotype nor is there a unique growth phenotype for the bacteria. As an extension of this line of study, we have carried out a systematic analysis of the time-dependent change in ribosome patterns for some of the ECOR strains. The results described here indicate that the overall pattern of ribosome profile change is similar among strains but that the total length of time for this cycle of ribosome change differs from strain to strain. Since the decrease in 100S ribosome dimers is always accompanied by the initiation of cell death, we propose that the cycle of ribosome profile change is closely related to the life cycle of E. coli. This relationship was tested for two different types of E. coli mutant defective in ribosome dimerization, one devoid of RMF and the other carrying mutant ribosomes defective in binding RMF.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and culture. The bacterial strains used were 19 representatives (Ecor2, -11, -16, -20, -28, -32, -36, -42, -43, -46, -48, -49, -52, -57, -60, -62, -63, -64, and -68) from the ECOR collection and some laboratory strains, including W3110 lineage A (25), Q13 (43), and HMY15, lacking the rmf gene (51). Overnight cultures in Luria-Bertani (LB) medium were inoculated into 1.7 liters of fresh LB medium, and the cells were grown at 37°C with shaking at 100 cycles per min with a TAITEC MM-10 water bath shaker (Koshigaya, Japan). This shaking speed gave maximum stationary-phase survival for most E. coli strains. In time course analyses of ribosome profiles during lengthy culture, medium E (44) supplemented with 2% polypeptone was also used.

Preparation of ribosomes. Cells were ground with an approximately equal volume of quartz sand and extracted with Noll's buffer I (10 mM Tris-HCl [pH 7.6], 10 mM magnesium acetate, 100 mM ammonium acetate, and 6 mM 2-mercaptoethanol) (33). Preparation of both crude and high-salt-washed ribosomes from the cell extracts was carried out essentially according to the method of Noll et al. (33) with slight modification as described by Wada (45).

Sucrose gradient centrifugation of ribosomes. Ribosomes were fractionated by sucrose density gradient centrifugation as described previously (22, 45). Samples were layered on top of a 5 to 20% (wt/vol) sucrose gradient in ribosome buffer (20 mM Tris-HCl [pH 7.6], 15 mM magnesium acetate, and 100 mM ammonium acetate) and centrifuged in a Hitachi RPS50-2 rotor at 40,000 rpm for 80 min at 4°C. After centrifugation, the absorbance of sucrose gradients at 260 nm was measured with a Shimadzu (Kyoto, Japan) UV200 spectrophotometer using a flow cell.

Two-dimensional electrophoresis of ribosomal proteins. Ribosomal proteins were prepared from total cell extracts or ribosomes by the acetic acid method (17). After dialysis against 2% acetic acid, the proteins were lyophilized and stored at 80°C until use. The radical-free and highly reducing (RFHR) method of two-dimensional gel electrophoresis (45) was used for analysis of ribosomal proteins after slight modification as described by Wada et al. (47).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Ribosome pattern differences among E. coli isolates from the ECOR collection. We chose 19 strains from the 72 different isolates of the ECOR collection and used them for an analysis of growth-dependent variation in ribosome profiles. The ECOR strains have been classified into several groups on the basis of pattern differences in multilocus enzyme electrophoresis (34, 40). The 19 strains were selected from these groups, i.e., A, B1, B2, C, and D (for the selected strains, see Materials and Methods). The growth rate was determined for each of the 19 strains cultured by shaking at 37°C in medium E (44) supplemented with polypeptone. The growth rate (GR) varied 2.4-fold, ranging from 0.39 (strain Ecor52) to 0.92 (Ecor29) doublings per h. The maximum saturation density (SD) (A₆₆₀) in medium E ranged from 0.5 (Ecor11) to 3.8 (Ecor16). Previously, Mikkola and Kurland (32) analyzed the growth rate of nearly the same set of E. coli strains in glucose-limited chemostats and found that the maximum growth rates were 0.48 to 1.43 doublings per h, giving about a threefold difference between the lowest and the highest rates of cell growth. The range and order of growth rate differences among the test E. coli strains measured in medium E were essentially the same as those determined in the glucose-limited chemostats. The analysis of evolution in vitro of natural isolates of E. coli in glucose-limited chemostats indicated that, after 280 generations, the maximum growth rates of all the cultures approximated that of standard laboratory wild-type strains, which was close to 1.33 doublings per h (32). We confirmed the increase in the growth rate of ECOR strains after repeated cultures in LB medium with shaking. Thus, it appears that slow growth is characteristic of natural E. coli isolates.

Time course of ribosome pattern change in natural isolates. Previously, we analyzed the change in ribosome profiles during overnight culture of some laboratory E. coli strains (46). Here we examined the ribosome profiles during lengthy culture of the ECOR strains. For comparison, we chose some representative strains with different levels of ribosome dimers at the culture time of 24 h. In the time course experiments, medium E (43) supplemented with 2% polypeptone was used instead of LB medium because the bacterial life span is generally longer in medium E. The test strains were cultured at 37°C for a prolonged time (up to 1 week) with constant shaking. Growth was monitored by measuring the turbidity, while the viable counts were measured on LB plates. Ribosomes were prepared at various growth phases of the culture and analyzed by sucrose density gradient centrifugation. Figure 1 shows the ribosome profiles for four strains, Ecor68 (GR, 0.75 doublings per h; SD [A₆₀₀], 3.4) Ecor48 (GR, 0.50; SD, 2.6), Ecor11 (GR, 0.78; SD, 0.5), and Ecor16 (GR, 0.65; SD, 3.8). The time-dependent change in sedimentation behavior of ribosomes was found to be essentially the same in all of the strains, consisting of four stages: (i) formation of 100S dimers in the early stationary phase, (ii) transient decrease in the dimer level, (iii) return of dimers to the maximum level, and (iv) dissociation of 100S dimers into 70S ribosomes, which are quickly disassembled into subassemblies (Fig. 1; also see Fig. 3 for the ribosome profile of W3110). The duration of this cycle of ribosome profile change, however, varied from strain to strain.

View larger version (32K): [in this window] [in a new window]   FIG. 1.   Growth phase-coupled change in ribosome profile. Four representative ECOR strains, Ecor68, Ecor48, Ecor11, and Ecor16, were grown in medium E supplemented with 2% polypeptone. At the indicated times (5 h to 6 days [d]) after transfer of the overnight preculture into fresh medium, ribosomes were prepared from aliquots of each culture and analyzed by sucrose density gradient centrifugation. The centrifugal field is directed from left to right. The asterisks indicate the days when the viable count of each strain decreased to less than 1%.

View larger version (27K): [in this window] [in a new window]   FIG. 2.   Growth phase-coupled change in the ribosome profile of strain W3110 and the two-dimensional gel pattern of its ribosomal proteins. E. coli W3110 was grown in medium E supplemented with 2% polypeptone. The sucrose density gradient centrifugation pattern of ribosomes was analyzed at the indicated time of culture (5 h to 6 days [d]), while the ribosomal proteins were analyzed by the RFHR method of two-dimensional gel electrophoresis (45, 47). Parts of the stained gels, including the RMF spots, are shown. The positions of RMF are indicated by arrowheads. Ribosomal protein L16 is split into fragments, and two N-terminal fragments, L16' and L16", are associated even after processing (see days 4 and 5). Coupled with L16 processing, L35 is dissociated from ribosomes. L7/12 are the last proteins associated with disassembled ribosomes (see day 6).

Time course of ribosome profile change in laboratory strains. Since a close correlation exists between the loss of ribosome dimers and the initiation of bacterial cell death, one possible triggering factor for cell death might be the dissociation of ribosome dimers. In order to test this relationship in detail and to identify possible alterations in the protein composition of ribosomes other than the association of RMF, we analyzed the growth phase-dependent changes of the ribosome profile and the two-dimensional gel pattern of ribosomal proteins for some E. coli laboratory strains with defined genetic backgrounds. First, we analyzed strain W3110 lineage A, which we used for analysis of the growth-dependent changes in the RNA polymerase pattern (25) and nucleoid protein composition (42). Under the same culture conditions described above, the W3110 cells retained high levels of the 100S ribosome dimers until day 4 (Fig. 2). At day 5, the amount of ribosome dimers decreased again to 20 to 25% of the maximum level. At day 6, not only 100S dimers but also most 70S monomers disappeared. The viable count at day 6 was about 10% of the maximum level. At day 7, virtually no viable cells were detected (less than 1%). Overall, the patterns of growth phase-dependent changes in the ribosome profile were essentially the same in the natural isolates and the laboratory strain, but the time required for one cycle of ribosome change varied from strain to strain.

View larger version (22K): [in this window] [in a new window]   FIG. 3.   Growth phase-coupled changes in the ribosome profile and the cell viability of strain W3110. (A) Turbidity (squares) was measured at 660 nm, while cell viability (circles) was measured on LB agar plates. (B) Based on the growth-coupled changes in the patterns of ribosome-associated RMF levels and 100S ribosomes, the stationary phase could be divided into four stages (or subphases), I to IV. Stage 0 represents the exponential growth phase. The RMF level represents the total amount of RMF divided by the total amount of 70S ribosome units in 100S dimers.

View larger version (39K): [in this window] [in a new window]   FIG. 4.   Mutants defective in ribosome dimerization. (A) E. coli Q13 was grown in medium E supplemented with 2% polypeptone. At various culture times as indicated, cell lysates were prepared as described in Materials and Methods and immediately subjected to sucrose gradient centrifugation. Viability was determined by measuring the colony formers on LB plates. (B) 70S ribosomes were prepared from E. coli Q13. The ribosomes were incubated for 30 min at 37°C in the absence (control) or presence (+RMF) of 10-fold molar excess of synthetic RMF, and the mixtures were subjected to sucrose gradient centrifugation. (C) Ribosomes were prepared from both exponential- and stationary-phase cultures of E. coli Q13, and ribosomal proteins were analyzed by two-dimensional gel electrophoresis using the RFHR method (45). (D) E. coli HMY13 (51) was grown in medium E supplemented with polypeptone and glucose. At various culture times as indicated, cell lysates were prepared and immediately subjected to sucrose gradient centrifugation. Viability was determined by measuring the colony formers on LB plates.

Mutants defective in the formation of ribosome dimers. We next analyzed mutants defective in ribosome dimerization, in which RMF is involved (46, 47). This involvement was strongly supported by the finding that ribosome dimers could not be detected for RMF null mutants (51). We followed the cell growth and ribosome profile during lengthy culture of mutant HMY15, with rmf deleted (51). As shown in Fig. 4D, no ribosome dimers were found throughout the culture from exponential to late stationary phase. At day 4, viability had already decreased to 15% as measured by colony formation, and at day 5, virtually no colony formers were detected.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Most of what we know about genotypic and phenotypic variation in E. coli has been derived from studies of laboratory strains. For instance, the response of bacteria to environmental stresses has been analyzed for the most part using a small number of laboratory strains (for reviews, see references 23 and 24). The molecular mechanisms observed in such model systems may be different from those expressed by natural isolates. For example, some E. coli W3110 stocks lack ^S and/or ^F subunits of RNA polymerase (25). Even though these  factors are required for transcription of some stress response genes, the relevant stress conditions may normally be absent from laboratory culture conditions, and thus the defective bacterial stocks can be maintained in captivity.

The ECOR collection, a set of 72 reference strains of E. coli isolated from a variety of hosts and geographical locations, was established by Ochman and Selander (34) and has been widely used in studies of variation in natural isolates. Such studies have focused on diverse characteristics of the bacterial genome (1, 4-9, 11, 12, 14, 15, 17, 18, 21, 36, 37, 39, 41, 48), on the distributions of branched DNA-RNA molecules (20), and on variations in the complexities of protein families (3, 7, 13, 16, 29, 38, 49, 50). Previously, Mikkola and Kurland (31) analyzed ribosomes from translation mutants as well as some natural isolates of E. coli. The results indicated that bacteria harboring kinetically impaired ribosomes increase the number of ribosome particles accumulated under poor growth conditions in order to compensate for their defective function. Some of the kinetic characteristics of ribosomes are different among original natural isolates. After growth in glucose-limited chemostats, however, the ribosomes of all of the cultures become kinetically indistinguishable from those of laboratory wild-type bacteria. Thus, bacteria grown under normal laboratory conditions have been selected for maximum growth rates, and this in turn demands maximum translation efficiency (32). In contrast, maximum growth rates do not seem to be strongly selected for in the ECOR collection of natural isolates (30).

In rapidly growing E. coli strains, ribosomes are involved in dynamic cycles of translation. Upon entry into the stationary phase, the step time of the functional ribosome cycle is delayed, and as a result, the level of translation frameshift by ribosomal slippage is markedly increased (2, 10). Excess unused ribosomes are converted into translationally inactive 100S dimers (46, 47). Although a number of changes take place in parallel both in the composition and pH of media and in intracellular compositions in stationary-phase E. coli culture, the ribosome dimerization might be primarily due to the association of RMF, because (i) ribosome dimers can be reconstituted even from chemically synthesized RMF and exponential phase 70S ribosome dimers (46) and (ii) ribosome dimers are not formed in vivo in the absence of RMF (51). The structural and functional modulations of ribosomes appear to be required for prolonged survival in the stationary phase. The RMF null mutants showed reduced viability in stationary phase (51). Reduced viability was also observed for the E. coli Q13 strain (Fig. 4A), which does not form ribosome dimers, presumably due to a ribosomal defect in binding RMF (Fig. 4A and B).

As in the case of translation machinery, RNA polymerase appears to be stored in inactive form in the stationary phase. The core enzyme forms complexes with polyphosphate (28), and the ⁷⁰ subunit for transcription of growth-related genes forms a complex with the anti-sigma factor RSD (regulator of sigma D) (26, 27; also reviewed in references 23 and 24). Thus, in the stationary phase, both translation and transcription apparatuses are converted into inactive stored forms, which can be reused upon reentry into the growth cycle. Since the disassembly of 100S ribosomes is immediately accompanied by degradation of 70S ribosomes, ultimately leading to cell death, it appears that the initiation of 100S ribosome disassembly triggers switching to cell death. Alternatively, an as-yet-unidentified common factor may simultaneously trigger both ribosome disassembly and cell death.