Fitness benefits of a synonymous substitution in an ancient EF-Tu gene depend on the genetic background

ABSTRACT Synonymous mutations are changes to DNA sequence, which occur within translated genes but which do not affect the protein sequence. Although often referred to as silent mutations, evidence suggests that synonymous mutations can affect gene expression, mRNA stability, and even translation efficiency. A collection of both experimental and bioinformatic data has shown that synonymous mutations can impact cell phenotype, yet less is known about the molecular mechanisms and potential of beneficial or adaptive effects of such changes within evolved populations. Here, we report a beneficial synonymous mutation acquired via experimental evolution in an essential gene variant encoding the translation elongation factor protein EF-Tu. We demonstrate that this particular synonymous mutation increases EF-Tu mRNA and protein levels as well as global polysome abundance on RNA transcripts. Although presence of the synonymous mutation is clearly causative of such changes, we also demonstrate that fitness benefits are highly contingent on other potentiating mutations present within the genetic background in which the mutation arose. Our results underscore the importance of beneficial synonymous mutations, especially those that affect levels of proteins that are key for cellular processes. IMPORTANCE This study explores the degree to which synonymous mutations in essential genes can influence adaptation in bacteria. An experimental system whereby an Escherichia coli strain harboring an engineered translation protein elongation factor-Tu (EF-Tu) was subjected to laboratory evolution. We find that a synonymous mutation acquired on the gene encoding for EF-Tu is conditionally beneficial for bacterial fitness. Our findings provide insight into the importance of the genetic background when a synonymous substitution is favored by natural selection and how such changes have the potential to impact evolution when critical cellular processes are involved.

translation rate for cellular functions, the use of synonymous codons has the potential to impact fitness in a variety of ways.However, the impact of codon bias on translation speed, accuracy, and efficiency is not yet clear.Studies have correlated highly expressed genes with high levels of codon bias which may provide an advantage for cellular fitness (16,(19)(20)(21)(22)(23).Conversely, some studies suggest that codon bias has minimal or no effect on fitness (24)(25)(26), while others show examples of beneficial synonymous mutations with less-preferred codons (6).Taken together, little is known about the molecular mechanisms and potential of beneficial or adaptive fitness effects of such changes within evolved populations.
In this study, we explored the degree in which a key synonymous mutation influences essential gene adaptation in bacteria building on an experimental system whereby an Escherichia coli (E.coli) strain harboring an engineered essential gene was subjected to laboratory evolution (27).Specifically, an essential protein in translation, elongation factor Tu (EF-Tu), was directly replaced with a phylogenetically inferred ancestor, AnEF, leading to a fitness decrease of the engineered strain (27,28).EF-Tu is an essential GTPase that binds to amino-acylated tRNAs and shuttles them to the A-site of the ribosome (Fig. 1A) (29).EF-Tu is critical in mediating the rate and accuracy of translation elongation (30) and is one of the most abundant proteins in the cell, making up ~6% of the total protein in E. coli (31,32).
We report a single synonymous mutation discovered in the anEF gene coding region in one of the experimentally evolved engineered E. coli populations.Whole genome sequencing reveals that the synonymous mutation in anEF appears after generation 2,000 and sweeps to fixation within the population by the 2,500th generation.Genetic and cellular analyses demonstrate that this synonymous mutation is beneficial during laboratory growth and increases the mRNA and protein levels of AnEF, yet these fitness effects are contingent on the presence of one or more additional mutations in the evolved genetic background.Furthermore, we show that this synonymous mutation increases the abundance of polysomes during translation.Our results demonstrate that synonymous mutations may play a key role when affecting the levels of proteins that limit key cellular processes.

RESULTS
In previous work, we engineered and experimentally evolved an E. coli strain widely used for laboratory evolution studies, REL606, to replace a single copy of ancestral EF-Tu by substituting the tufB gene with an inferred ancestral allele in a strain where the homologous gene copy tufA was deleted (∆tufA, tufB::anEF) (27).We refer to ∆tufA tufB::anEF as the ancestor strain hereafter (Table 1).E. coli's EF-Tu and AnEF share ~95% protein sequence identity (Fig. S1).Additionally, the predicted structure of AnEF exhibits a root mean square deviation (RMSD) value of 1.63 Å when overlaid with E. coli EF-Tu (33) (Fig. 1B).
We assessed the effect of AnEF allele replacement into E. coli in glucose minimal medium via co-culture competition and compared its fitness with a wild-type strain containing a single copy of the wild-type EF-Tu (E. coli REL606 ∆tufA).The relative fitness of the modified ancestor strain was ~0.87 [n = 5, P < 0.01, analysis of variance (ANOVA)] (Fig. 1C).Eight lineages propagated from a single colony of REL606 ∆tufA, tufB::anEF was then subjected to 3,000 generations of laboratory evolution through serial propagation of bacterial populations (Fig. 2A) (34).An E. coli REL606 ∆tufA strain was evolved in parallel as a control.Using co-culture competition assays, we quantified the fitness of the evolved population relative to the ancestor strain every 1,000 generations of the evolution experiment (Fig 2B).Genomes of evolved populations were sequenced at six different time points: generations 500, 1,000, 1,500, 2,000, 2,500, and 3,000.
Population sequencing of one particular lineage reveals a "fixed" (reached a fre quency of 100% in the population) synonymous mutation in the ancient EF-Tu gene coding region in generation 2,500.In this lineage, generation 500 had three fixed nonsynonymous mutations in genes ydjN and fadA and a genomic deletion in the rbs operon (rbsD-[rbsR]) (Fig. 2C and D; Table S3).Generations 1,000 to 2,000 acquired three additional fixed nonsynonymous mutations in genes mrdA, ydfI, and yfaS (Fig. 2C; Table S3).Finally, in generations 2,500 and 3,000, we identified two more fixed mutations, a non-synonymous mutation in gene pykF and a synonymous mutation in the replaced gene anEF (Fig. 2C; Table S3).Substitution dynamics reconstruction associated with each generation time point (Fig. 2D) shows that nonsynonymous mutations arose independ ently in four other genes ECB_01992 (arose in generation 1,000), hupA (generation 1,500), caiD (generation 1,500), and hslU (generation 2,000) (Fig. 2D).However, these mutations did not reach fixation and, instead, were eventually lost from the population.The patterns we observed in genotype dynamics is consistent with each detected fixed mutation (Fig. 2C).

The rise and effect of the AnEF synonymous mutation
Intriguingly, at generation 2,500, a synonymous mutation appeared in the anEF gene coding region, in one out of eight parallel evolved lineages.Notably, the synonymous mutation reached 100% fixation rapidly, appearing between generations 2,000 and 2,500 (Fig. 2C and D; Table S3) and, ultimately, swept the entire population by generation 2,500 (Fig. 2D).Specifically, we identified a C > T mutation at nucleotide 45, GTC (Val)→GTT (Val) mutation, in the N-terminal coding region of the anEF gene (Fig. 3A; Table 1).The presence of the mutation at each generation was successfully verified with whole genome sequencing (Fig. 2C and D; Table S3).We hypothesized that if the synonymous mutation is beneficial, reverting the evolved mutation in the anEF gene back to the ancestral nucleotide would negatively affect fitness.To test this hypothesis, we engineered the codon containing the synonymous mutation back to the ancestral codon sequence while keeping the rest of the genetic background (including all other fixed mutations) constant.This new strain is referred to as evolved + AnEF T45C (Table 1).Furthermore, to test if the fitness impact of the synony mous mutation was background dependent, we introduced the synonymous mutation (C45T) into the anEF unevolved allele in the ancestral background (referred to as ancestor + AnEF C45T ) (Table 1; Materials and Methods).All isogenic constructs were confirmed via local DNA sequencing and whole genome sequencing.

The synonymous mutation is conditionally beneficial depending on the genetic background
We next measured the impact of the C45T synonymous mutation on organismal fitness via co-culture competition assays by calculating relative fitness (W).In agreement with the previously published studies, the replacement of native E. coli EF-Tu with the ancestral EF-Tu AnEF causes a 13% decrease in fitness (W = 0.87, P < 0.01, ANOVA Tukey's Honestly Significant Difference, HSD) (Fig. 1C) (27, 28), which upon 3,000 generations of laboratory evolution is recovered and increased by 60% relative to the ancestor (evolved AnEF C45T , W = 1.6, P < 0.01, ANOVA Tukey's HSD) (Fig. 3B).Substitution of the synonymous mutation with the native codon significantly decreases the fitness of the evolved microbe by 30% (evolved AnEF T45C , W = 0.7, P < 0.01, ANOVA Tukey's HSD), whereas introducing the synonymous mutation in the ancestral background has no relative fitness benefit (ancestral AnEF C45T , W = 1.00,P = 0.80, ANOVA Tukey's HSD) (Fig. 3B).These results show the strong epistasis between the synonymous mutation and the evolved background, demonstrating that the fitness effect of the synonymous mutation depends on the genetic background.

AnEF C45T synonymous mutation is associated with increased mRNA and protein levels
Previous studies have reported that synonymous mutations can impact mRNA and protein levels (6-10).To assess the change in transcript level, we measured mRNA levels via quantitative polymerase chain reaction (qPCR) and calculated ∆∆Cq values to compare AnEF mRNA with and without the evolved synonymous mutation (n = 3).
Relative to the ancestor, we observed a threefold increase in the evolved strain (P < 0.05, t-test) and a two-fold increase in the ancestor with the synonymous mutation (ancestor + AnEF C45T , P < 0.05, t-test) (Fig. 4A; Fig. S2).The mRNA levels of the evolved strain with the synonymous mutation reverted back to the ancestral nucleotide (evolved + AnEF T45C ) displayed a 30% decrease in AnEF mRNA relative to the ancestor; however, this observation was not significant (P = 0.06, t-test) (Fig. 4A; Fig. S2).Furthermore, we assessed whether the C45T synonymous mutation has changed the protein levels relative to the ancestor strain (n = 3) (Fig. 4B; Fig. S3).AnEF protein levels increase 32.2% in the evolved strain with the C45T synonymous mutation (P < 0.01, t-test; Materials and Methods) (Fig. 4B; Fig. S2).Similarly, the engineered ancestor with the AnEF C45T allele exhibits a 34.2% increase in AnEF protein levels (P < 0.01, t-test) (Fig. 4B; Fig. S2).
Interestingly, reversion of the synonymous mutation back to the ancestral nucleotide in evolved strain (evolved + AnEF T45C ) leads to a 15.3% decrease in AnEF protein levels (P < 0.01, t-test) (Fig. 4B; Fig. S2).Taken together, these data suggest that the evolved synonymous mutation in AnEF is correlated with an overall increase in AnEF mRNA and protein levels in both the ancestor and evolved genetic backgrounds.

The presence of the synonymous mutation in AnEF leads to an increase in ribosomal abundance in translation elongation
Various studies have demonstrated the link between bacterial growth and ribosomal abundance, showing that growth (and death) rate of bacterial populations is affected by the number of ribosomes (35)(36)(37)(38).Elongation factor proteins are highly expressed in the cell (39) and fulfill a crucial role in the ribosome (40).Thus, upon determining the significant impact the synonymous mutation has on AnEF mRNA and protein levels, we asked if the synonymous mutation affects ribosomal abundance.We hypothesized that the evolved strain with the synonymous mutation would exhibit greater number of ribosomes as it displays an increase in population growth fitness, gene, and protein expression.
We inferred the cellular translatome, which refers to all mRNAs associated with ribosomes in protein synthesis, using polysome profiling to assess the impact the AnEF synonymous mutation has on translation.The translation process is made up of four main steps: initiation, elongation, termination, and recycling, where initiation is the assembly and preparation of the ribosome on a transcript and elongation is the decoding process of protein synthesis (41,42).Polysome profiling provides insight into the translation process by first separating ribosomes via RNA sedimentation and then providing comparative ribosomal abundance for each step of protein synthesis for the corresponding strain.We first assessed the polysome profile of the wild-type REL606 strain (n = 3).As shown in Fig. 5, we used a 10%-40% sucrose gradient for sufficient separation of RNA molecules, of which free RNAs (mRNA, tRNA, etc.) are the lightest and, therefore, will sediment toward the top of the gradient and the heaviest RNA molecules being the longer stretch of polysomes actively translating along an mRNA transcript.Therefore, the first peak detected on a polysome profiles represents the free RNAs in a cell, followed by peaks associated with the small and large ribosomal subunits, 30S and 50S, respectively.The highest peak shown here represents the assembled 70S monosome, which is a single ribosome on an mRNA transcript, or a ribosome in the initiation phase of translation.Finally, each peak following the monosome represents the stretch of polysomes in the elongation phase of translation (Fig. 5; Materials and Methods).

DISCUSSION
Prior studies demonstrate that E. coli can respond to EF-Tu gene perturbations in multiple, contingent ways.For example, a previous study experimentally evolved multiple parallel populations of E. coli harboring an ancient EF-Tu gene for over 3,000 generations and reported that the majority of the populations had gained mutations in the promoter region of the EF-Tu gene, resulting in an increase of protein expression (27).Overexpressing the AnEF partially increases E. coli fitness (28), demonstrating that fitness may be restricted by the availability of key enzymes that mediate critical cellular processes.Subsequent studies have replaced EF-Tu with several of its extant and extinct orthologs, demonstrating evolutionary distance, together with horizontal gene transfer constraints genomic compatibility and interchangeability (28).Successive generations of synthetic laboratory evolution of EF-Tu variant-harboring strains in different growth conditions were tracked for fitness and mutation accumulations in sweep populations (43).These results demonstrated that there is a measurable negative correlation between the rate of beneficial mutation accumulation and fitness advantage in the translation module, and due to this stalling effect, the translation module could not reach an optimal fitness peak.Even under the most controlled circumstances, it is difficult to predict which array of responses a system will take.
We suggest that when protein levels are the limitation for cell growth, synonymous mutations that increase protein levels may lead to beneficial and adaptive outcomes.Indeed, previous laboratory passage studies have reported that synonymous mutations directly influence growth as a result of increased mRNA and protein levels for proteins that sit at bottlenecks in pathways contributing to flux through metabolic systems (6,44,45).While these previous studies demonstrate that the amount of the protein available in the system impacts substrate turnover at key nodes in pathways, our study differs in that the amount of the focal protein directly impacts overall cellular translation rate, thus generally influencing the translation rate of all proteins.Although translation can be sustained in E. coli by the reconstructed ancient allele of EF-Tu (AnEF), this interchange leads to a decrease in cellular protein synthesis and population fitness.(Fig. 4).Intriguingly, a previous study identified that the first 65 codons of tufB are important for tufB regulation in Salmonella and that various synonymous mutations near the start codon (including C45) promote an open conformation of the tufB mRNA, which leads to increased EF-Tu expression (46).Since AnEF was inserted in place of tufB and shares a nucleotide sequence identity of 87%, it is possible that the observed increase in AnEF mRNA and protein levels can thus be explained by the favored open conformation of AnEF mRNA secondary structure.The resulting anEF C45T mRNA sequence between regions −73 and +96 exhibits an unaltered secondary structure when compared to E. coli tufB (Fig. S4; Materials and Methods).Moreover, the wild-type E. coli tufB codon for valine (position 45) is GTC, which is reported at a frequency of 13.5% in codon usage bias frequency index (47).Interestingly, the synonymous mutation in AnEF changed the codon from GTC to GTT, which is reported at a higher frequency of 39.8%, the second most frequent codon used for valine in E. coli for highly expressed genes (47).
To what degree this conversion may be attributable to the codon bias requires further mechanistic investigation.
The synonymous mutation detected in AnEF indicates a trending increase in polysome abundance (Fig. 6; Fig. S3), defined by an increase in the number of elongating ribosomes on global mRNA transcripts.Based on this, we develop three scenarios to interpret our results: (i) The increase in polysomes may reflect the observed increase in AnEF protein levels.The transition of translation initiation to elongation is marked by the first translocation via EF-G (41,42), which is dependent on EF-Tu for bringing a cognate amino-acylated tRNA to form the first dipeptide bond (41,42).(ii) The increase in polysomes may be due to a decline in translation elongation rate in agreement with the previous studies showing that protein synthesis rate is reduced when translation is dependent on the EF-Tu ancestral variant (28).Furthermore, ancient EF-Tu variants were shown to exhibit significantly low K M values for dipeptide formation (48), indicating that ancestral EF-Tus are less efficient at sequestering cognate amino-acylated tRNAs to the ribosome.Altogether, ancient EF variant's reduced efficiency at sequestering cognate amino-acylated tRNAs to the ribosome leads to ribosomal stalling in the decoding process (48).Such stalling increases the presence of elongating ribosomes.(iii) Finally, a combinatorial effect of the scenarios (i) and (ii) collectively increases cellular translation behavior.While the boost in AnEF protein amount would not necessarily boost the efficiency of the AnEF function, it would equate to an increase in EF-Tu availability for ribosomes transitioning from translation initiation to elongation.In sum, the increase in polysome abundance on global mRNA transcripts is very likely due to cumulative effects of the increase in AnEF protein amount coupled and the AnEF's kinetic reduction in translation efficiency relative to its extant counterpart.
The AnEF synonymous mutation (AnEF C45T ) is the cause for the beneficial changes, yet the advantageous impact of the mutation on the population fitness is observed only when the synonymous mutation is present together with the evolved genetic background (Fig. 3).This dependency demonstrates epistatic interactions between the AnEF synonymous mutation and the evolved genetic background and strongly indicates that there are additional changes required to optimize levels of AnEF in the context of the extant E. coli genome.Furthermore, as shown in Fig. 4, the mRNA levels of AnEF C45T are higher in the evolved background compared to when moved to the ancestral background, suggesting that epistasis within the evolved background is reinforcing EF-Tu's role as a limiting factor upon cell growth.
Upon 3,000 generations of bacterial evolution, there were seven additional nonsy nonymous mutations that arose and were fixed prior to the synonymous mutation in anEF (genes: mrdA, ydfI, pykF, ydjN, yfaS, rbsD-[rbsR], and fadA) (Fig. 2C and D; Table S3).The mutations in ydjN, fadA, and rbsD-[rbsR] became fixed in generation 500, where YdjN functions as a L-cysteine transporter (49) and FadA and products of the rbs operon function in metabolism (50,51).The genomic deletion in rbsD- [rbsR] has been observed in previous bacterial evolution experiments, in which multiple experimentally evolved populations acquired a loss in the rbs operon (involved in ribose catabolism) (52).This rbs operon deletion is hypothesized to provide an advantage in glucose minimal medium; thus, we do not suspect it is responsible for the increase in fitness associated with the AnEF synonymous mutation.Mutations fixed in generation 1,000 were in genes mrdA, ydfI, and yfaS.The mrdA gene (also known as pbpA) produces the protein penicillin-bind ing protein 2, which is responsible for maintaining antibiotic sensitivity as well as the rod cell shape in E. coli (53); YdfI functions in metabolism (54), while YfaS remains uncharacterized but predicted to be a part of the alpha-2-macroglobulin family (55).The final mutation to fix was in pykF in generation 2,500 (with anEF), the product of which also functions in cell metabolism (56).Changes in some of these genes, for example, mrdA (57) and pykF (58), were previously shown to increase the fitness of REL606 during laboratory evolution.How these background mutations influence AnEF remains to be explored in detail, but their occurrence in such evolution experiments using native, unaltered E. coli indicates that phenotypic effects of the synonymous mutation in AnEF are not causally linked on prior mutations in mrdA, pykF, or the rbs operon and may have arisen due to chance.
Overall, our data underscore the importance of highly expressed proteins in essential cellular processes as well as the importance of the genetic background when a synonymous substitution is favored by natural selection.The synonymous mutation in AnEF is associated with increased mRNA (Fig. 4A), increased protein (Fig. 4B), and increased polysome abundance (Fig. 6); however, population fitness is only increased when AnEF C45T is coupled with the evolved genetic background (Fig. 3).Taken together, these results demonstrate that synonymous mutations can be beneficial and have the potential to impact evolution when critical cellular processes are involved.

Ancestral protein sequence and structure reconstruction
Ancestral protein AnEF sequence was inferred through ancestral sequence reconstruc tion of EF-Tu proteins as previously described (59).AnEF protein and nucleotide sequences were aligned to E. coli EF-Tu protein and E. coli tufB nucleotide sequences using MAFFT v7.490 (60).AnEF structures was predicted by LocalColabFold (61) which is a script to use AlphaFold 2.3.1 on local machines (last used April 2023).The AnEF structure was predicted using templates from PDB (--template), default number of prediction recycles (--num-recycle 3), and amber structure refinement method (--amber).

Strain construction
Ancestral and evolved lineages were derived from E. coli B strain REL606 as detailed in reference (27).The genetic marker TP22-amilCP_opt-kan-sacB-T0 was inserted in intergenic region between rpoC and yjaZ via dsDNA recombineering (28,64,65) to link it to the different EF-Tu alleles in strains CH6556 and CH6585 (Table S1).The genetic marker and linked EF-Tu allele were moved between strains by P1 virA phage-mediated transduction (28,66).Constructs were confirmed via PCR, local Sanger DNA sequencing, and whole genome sequencing.A complete list of primers and strain genotypes is listed on Tables S1 and S2.

Experimental evolution
Experimental evolution was carried out in serial dilutions in DM25 liquid medium for 3,000 generations (~6.6 generations per day) as described previously (34) and reported for the AnEF strains (27).

RNA extractions, cDNA synthesis, and quantitative PCR
Strains were grown in LB medium and collected at 0.3-0.6OD 600 for RNA extraction.Pellets were lysed and prepared following the standard protocol in Rneasy kit (Qia gen, 74104).Following RNA extraction, samples were removed of any contaminating genomic DNA using the DNase I standard protocol (Invitrogen, 18068015).Following DNase treatment, RNA extracts were synthesized into cDNA using reverse transcriptase SuperScript IV First-Strand Synthesis System (Invitrogen, 18091050).Finally, all cDNA samples were run on PCR to confirm no genomic DNA contamination before quantitative PCR.All cDNA samples (three biological replicates) were run in SsoAdvanced Universal SYBR Green Supermix (Bio-Rad, 1725271) in three technical replicates for 40 cycles.All primers are listed on Table S2.∆∆Cq values were calculated using housekeeping gene rpoB as a control.

Immunoblotting experiments
Strains were grown in LB-rich medium and collected at 0.3-0.6OD 600 for cell lysis (three biological replicates).Cell pellets were lysed in 300 µL of lysis buffer [10-mL BugBuster (Millipore 70584-4), ¼ tablet cOmplete Protease Inhibitor Cocktail EDTA-Free (Roche)] for 20 min at room temperature and centrifuged at 13,000 rpm at 4℃ for 30 min to clear all cell debris.Protein concentrations were measured using BCA assay (Thermo Fisher 23227) and stored at −20°C.Whole cell lysates were linearized at 95°C for 5 min and were run on a 12% resolving Tris-glycine SDS-PAGE gel at a constant voltage of 125 V. Protein bands were then transferred to a nitrocellulose membrane at a constant amperage of 400 mA for 45 min.All membranes were blocked with 5% nonfat milk for 1 hour at room temperature.To visualize EF-Tu bands, membranes were probed with 1:1,000 EF-Tu monoclonal antibody (Hycult Biotech mAb 900; catalog no.HM6010) for at 4°C overnight, followed by incubation with 1:10,000 IRDye 680RD anti-mouse secondary antibody (LI-COR, 929-70050) for 1 hour at room temperature.All membranes were imaged using the LI-COR Odyssey XF.To normalize for band quantitation, membranes were stained with GelCode Blue Stain (Thermo Fisher, 24590) for 5 min and destained for 10 min using a destaining solution composed of 40% dH 2 0, 10% acetic acid, and 50% methanol.Bands were quantified using ImageJ, and statistics were calculated using Rstudio.

Fitness assays
All competition assays were followed per published Ara-/+ competition protocol (34).Mixed populations were revived with 100 µL from a glycerol stock into 10 mL of LB medium and grown overnight shaking at 250 rpm at 37°C.Clonal populations were revived with 2 µL from a glycerol stock.The next day, each mixed population or clonal strain was preconditioned by diluting 1:10,000 in 10 mL of DM25 medium (five replicates each) and grown overnight shaking at 250 rpm at 37°C.To begin the competition, two selected strains were combined in equal amounts (50 µL) into 10 mL of DM25, plated 100× diluted on TA agar (d 0 ), and grown for 24 hours shaking at 250 rpm at 37°C.After 24 hours, the competition assay was plated 100× diluted on TA agar a second time (d 1 ) and grown at 37°C overnight.All plates were imaged after 24 hours.To calculate relative fitness (W), we calculated the ratio of each strain's Malthusian parameters (M A and M B ) as follows: M A = ln 100*Ad 0 / Ad 1 M B = ln 100*Bd 0 /Bd 1 W = M A /M B

Polysome profiling
Samples were prepped in accordance with Qin and Frederick (74).Each strain was grown to mid-log (OD 600 0.3-0.6) at 37°C and shaking at 200 rpm for a total of six biological replicates and two technical replicates.At mid-log, 35 mL of culture was collected, pelleted at 4,000 rpm at 4℃ for 5 minutes, resuspended in 500 µL of chilled lysis buffer (10 mM Tris-HCl pH 8.0, 10 mM MgCl 2 , 1 mg/mL lysozyme), and flash frozen in liquid nitrogen.The lysates were then thawed in ice water and immediately refrozen in liquid nitrogen and stored at −80℃.To separate polysomes, lysates were thawed on ice and resuspended with 15 µL of 10% sodium deoxycholate.The lysates were then cleared of cell debris at 10,000 rpm at 4°C for 10 min.A normalized volume of 500,000 ng of RNA was carefully loaded on top a sucrose gradient (10%-40% sucrose, 20 mM Tris-HCl pH 8.0, 10 mM MgCl 2 , 100 mM NH 4 Cl, 2 mM DTT, assembled using Biocomp Gradient Master) and ultracentrifuged (Beckman Coulter Optima XE-90, Rotor SW41Ti) for 3 hours at 35,000 rpm at 4°C.To collect polysome profiles, samples were collected, and RNA was measured at 260 nm using the Biocomp scanner; the overall assay was repeated twice.Ribosome abundance was measured by calculating the area under the curve (AUC) for peaks corresponding to the 30S subunit, 50S subunit, 70S monosome, and polysomes comprising two to four ribosomes.Subsequently, peak areas were normalized by the total area under the curve (30S + 50S + 70S + polysomes).Areas under the curve were generated by averaging across both trials.

FIG 2 Full
FIG 2 Evolutionary dynamics of an E. coli strain harboring an ancestral EF-Tu gene.(A) Engineering and evolution experiment schematics.(B) Change in fitness of evolved population relative to the ancestor.(C) Gene mutation frequencies in evolved population over 3,000 generations.anEF gene synonymous mutation highlighted in red.(D) Muller plot demonstrating the genotype dynamics across laboratory evolution with 3,000 generations.Plotted are the mutations in genes that reached a minimum frequency of 25%, and highlighted in red is the selection frequency of the anEF synonymous mutation.

FIG 6
FIG6 Polysome profiling of the evolved strain (red) relative to the ancestor (black).(B) Quantification of polysome abundance from shaded box in the profiles in generated panel A (n = 6, t-test).

TABLE 1
List of E. coli samples used in the study a Bacterial strain or population name and characteristics are indicated.A complete list of constructs used for genome engineering can be found under Supplementary Information. a