Evolution of penicillin-binding protein 2 concentration and cell shape during a long-term experiment with Escherichia coli

Abstract

Peptidoglycan is the major component of the bacterial cell wall and is involved in osmotic protection and in determining cell shape. Cell shape potentially influences many processes, including nutrient uptake as well as cell survival and growth. Peptidoglycan is a dynamic structure that changes during the growth cycle. Penicillin-binding proteins (PBPs) catalyze the final stages of peptidoglycan synthesis. Although PBPs are biochemically and physiologically well characterized, their broader effects, especially their effects on organismal fitness, are not well understood. In a long-term experiment, 12 populations of Escherichia coli having a common ancestor were allowed to evolve for more than 40,000 generations in a defined environment. We previously identified mutations in the pbpA operon in one-half of these populations; this operon encodes PBP2 and RodA proteins that are involved in cell wall elongation. In this study, we characterized the effects of two of these mutations on competitive fitness and other phenotypes. By constructing and performing competition experiments with strains that are isogenic except for the pbpA alleles, we showed that both mutations that evolved were beneficial in the environment used for the long-term experiment and that these mutations caused parallel phenotypic changes. In particular, they reduced the cellular concentration of PBP2, thereby generating spherical cells with an increased volume. In contrast to their fitness-enhancing effect in the environment where they evolved, both mutations decreased cellular resistance to osmotic stress. Moreover, one mutation reduced fitness during prolonged stationary phase. Therefore, alteration of the PBP2 concentration contributed to physiological trade-offs and ecological specialization during experimental evolution.

FIG. 1.

Mutations identified in the pbpA operon in evolving E. coli populations (adapted from references and 79). (A) Organization of the operon and location of mutations in populations Ara-1, Ara-2, Ara-4, Ara-5, Ara+1, and Ara+6. The operon includes five genes; ybeA and ybeB encode proteins with unknown functions, pbpA encodes PBP2, rodA encodes the RodA protein, and rlpA encodes a rare lipoprotein. Both PBP2 and RodA are involved in cell wall elongation and cell morphogenesis. The size of each coding region is indicated. The locations of evolved mutations are indicated by arrows, and the populations in which they occurred are indicated above the arrows. The two evolved pbpA alleles studied in this work are enclosed in boxes (Ara-5 and Ara+1). The bent arrow upstream of ybeB indicates the operon's putative promoter (2). (B) Characteristics of the evolved mutations, including the populations in which they arose, their positions relative to the first nucleotide of the translation initiation codon of pbpA, their molecular nature, and their consequences for the amino acid sequence. Several populations, indicated by (M), evolved mutator phenotypes and had substantially elevated mutation rates. Population Ara+1 acquired an IS150 insertion (::IS150). Mutations in the regulatory region are indicated by “Non coding.”

FIG. 2.

Effects of pbpA mutations from populations Ara-5 (A) and Ara+1 (B) on the amount of PBP2. Strains were grown in DM250, and cells were collected upon entry into stationary phase. Membrane proteins were extracted, treated or not treated with 10 μM amdinocillin (Mecillinam), and incubated with fluorescein-conjugated ampicillin that bound to PBPs. Immunodetection was performed using 40-μg samples and antifluorescein antibodies. Competition between amdinocillin and ampicillin for binding to PBP2 allowed detection of PBP2 by comparison of samples treated with amdinocillin and samples not treated with amdinocillin (arrowheads). The following strains were used: REL606, the ancestor (pbpA_anc); REL968-8, a 1,000-generation clone from population Ara-5 carrying the evolved allele pbpA_-5; DVS79, same as REL968-8 except with pbpA_anc; LUD5, same as REL606 except with pbpA_-5; REL1158A, a 2,000-generation clone from population Ara+1 with the ancestral allele (pbpA_anc); REL1158C, another 2,000-generation clone from population Ara+1 but carrying the evolved allele pbpA₊₁; and DVS53, same as REL1158C except with pbpA_anc.

FIG. 3.

Effects of evolved mutations in the pbpA operon on cell morphology. Cells were grown in DM250 with or without amdinocillin (Mecillinam) and observed using a Zeiss Axioplan 2 microscope. All photographs were obtained using a magnification of ×1,000, and all images are at the same relative scale. Bars = 10 μm. (A) Effects of inhibition of PBP2 activity with amdinocillin on cell shape. Strains were grown in DM250 for 1 h, and cultures were then split into two halves. One half was treated with amdinocillin (4 μg/ml) and incubated for an additional 5 h, and then cells were collected and observed. (B) Effect of evolved pbpA allele on cell morphology in population Ara-5. (C) Effect of evolved pbpA allele on cell morphology in population Ara+1. The following strains were used: REL606, the ancestor (pbpA_anc); REL1158A, a 2,000-generation clone from population Ara+1 with the ancestral allele (pbpA_anc); REL1158C, a 2,000-generation clone from population Ara+1 with the evolved allele pbpA₊₁; LUD5, same as REL606 except with evolved allele pbpA_-5 from population Ara-5; LUD6, same as LUD5 except with pbpA_anc restored; and DVS53, same as REL1158C except with pbpA_anc.

FIG. 4.

Effect of evolved pbpA_-5 mutation on cell volume. Strains REL606 (pbpA_anc), LUD5 (same as REL606 except with evolved allele pbpA_-5), and LUD6 (LUD5 with the ancestral allele pbpA_anc restored) were grown in DM25 under the conditions used in the long-term experiment and then diluted 1:100 in isotone solution (Beckman), and thousands of cell volumes were measured electronically using a Coulter Multisizer 3 counter (Beckman). The error bars indicate standard errors based on six replicate cultures for each strain.

FIG. 5.

Effects of the evolved pbpA alleles from populations Ara-5 (A) and Ara+1 (B) on sensitivity to osmotic stress. Each strain was grown in liquid LB media containing different NaCl concentrations for 12 h and then plated onto LB agar. The net population growth of each strain, reflecting the effects of NaCl on growth and survival, is expressed relative to the growth of cultures of the same strain in LB medium without NaCl. The error bars indicate standard errors based on four replicates for each combination of strain and NaCl concentration. (A) ⧫, ancestral strain REL606 (pbpA_anc); ▪, LUD5, same as REL606 except with evolved allele pbpA_-5 from population Ara-5; •, LUD6, LUD5 with the ancestral allele pbpA_anc restored. (B) ⧫, REL606; ▪, REL1158A, a 2,000-generation clone from population Ara+1 with the ancestral allele (pbpA_anc); •, REL1158C, a 2,000-generation clone from population Ara+1 with the evolved allele (pbpA₊₁); ×, DVS53, same as REL1158C except with the ancestral allele pbpA_anc.

FIG. 6.

Fitness effects of the evolved pbpA alleles. Competition assays were performed under the conditions used in the long-term evolution experiment, in all cases with the ancestor bearing the opposite Ara marker. The error bars indicate the 95% confidence intervals based on five or six replicate assays. Note the difference in scale between the two panels. (A) Population Ara-5. The following strains were used: REL606, the Ara⁻ ancestor (pbpA_anc); LUD5, same as the ancestor except with the evolved allele pbpA_-5; and LUD6, strain LUD5 with the ancestral allele pbpA_anc restored. (B) Population Ara+1. The following strains were used: REL607, the Ara⁺ ancestor (pbpA_anc); REL1158C, clone from generation 2000 with the evolved allele pbpA₊₁; and DVS53, same as clone REL1158C except with the ancestral allele pbpA_anc introduced.