Oligonucleotides were from Integrated DNA Technologies (Coralville, IA). All other chemicals were obtained from either Sigma Aldrich (St. Louis, MO) or Fisher Scientific (Pittsburg, PA). Ribosomes from early log phase growths of P. aeruginosa strain PA01 were prepared in the laboratory of Walter Hill at the University of Montana (Missoula, MT) as previously described [22]. DNA sequencing was at the Howard Hughes Medical Institute (HHMI) laboratory at The University of Texas – Pan American. The plasmid pQE60-RRF(C-His) containing the gene encoding the E. coli ribosome recycling factor (RRF) was a kind gift from Dr. Nono Tomita-Takeuchi at the University of Tokyo (Kashiwa, Chiba, Japan).

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed using 4 to12% polyacrylamide precast gels (Biorad). Benchmark unstained protein molecular weight markers were from Invitrogen (Madison, WI). Protein concentrations were determined by the method of Bradford [23] using Coomassie Protein Assay Reagents (Thermo Scientific) and bovine serum albumin as the standard.

Genes encoding both forms of EF-G were amplified by PCR (Bio-Rad MJ Mini Thermo Cycler) from P. aeruginosa PAO1 (ATCC) genomic DNA. EF-G1A was amplified using the forward primer (5’-ctgagctagcgctcgcaccactcccat-3’) and the reverse primer (5’-gactaagcttcatcagcggccctgcct-3’). EF-G1B was amplified using the forward primer (5’-ctgagctagcgcccgtactacacccatca-3’) and the reverse primer (5’-gactaagcttatcaaccttgttttttaaccagc-3’). The correct DNA sequence of PCR products was confirmed by DNA sequencing (Howard Hughes Medical Institute (HHMI) laboratory at The University of Texas – Pan American). The PCR products were inserted between the NheI/HindIII restriction sites in a pET-28b(+) plasmid (Novagen) and transformed into E. coli Rosetta™ 2(DE3) Singles™ Competent Cells (Novagen). This process placed the genes downstream of a sequence encoding six histidine residues.

Cultures were grown in F-medium (yeast extract,14 g/L, tryptone, 8 g/L, potassium phosphate-dibasic, 12 g/L, potassium phosphate-monobasic, 1.2 g/L and 1% glucose) containing 25 μg/ml of kanamycin and 75 μg/ml of chloramphenicol at 37 °C. Expression of the target proteins was induced at an optical density (A₆₀₀) of 0.6 by the addition of isopropyl β-D-1-thiogalactopyranoside (IPTG) to 0.25 mM. Growth of the bacterial culture was continued for 3 h post induction and the bacteria were harvested by centrifugation (4000 x g, 60 min, 4 °C). The cells were lysed and Fraction I lysate was prepared as previously described [24]. Both forms of EF-G were precipitated between 45 and 60% saturation of ammonium sulfate and the precipitated protein was collected by centrifugation (23,000 x g, 60 min, 4 °C). Both forms of EF-G were further purified to more than 98% homogeneity using nickel-nitrilotriacetic acid (NTA) affinity chromatography (Perfect Pro, 5 Prime) followed by dialysis (two times) against a buffer containing: 20 mM Hepes-KOH (pH 7.0), 40 mM KCl, 1 mM MgCl₂, 0.1 mM EDTA, 10 % glycerol. Purified proteins were fast frozen in liquid nitrogen and stored at -80 °C.

Assays to determine the ribosome-dependent GTPase activity of EF-G were carried out in 50 μl reactions containing: 50 mM Tris-HCl, (pH 7.5), 10 mM MgCl₂, 70 mM NH₄Cl, 1 mM dithiothreitol (DTT) and 1.8 mM GTP. EF-G1A and EF-G1B concentrations were held constant at 1.0 and 0.3 μM in assays in which P. aeruginosa ribosomes were titrated and ribosomes were held constant at 0.4 μM in assays in which EF-G was titrated. Velocity assays contained indicated amounts of GTP and assays were stopped each min between 1 and 6 min. Assays were stopped by the addition of 150 μl of 50 mM ethylenediaminetetraacetic acid (EDTA). The amount of GTPase activity was determined by measurement of the amount of P_i liberated using a colorimetric GTPase assay kit (Novus Biologicals) per manufacturer’s directions. Fusidic acid (FA) effects were determined using the same assays but containing from 4 to 250 μM FA.

Protein synthesis assays were carried out in 50 μl reactions containing: 50 mM Tris-HCl, (pH 7.5), 10 mM MgCl₂, 25 mM KCl, 4 mM phosphoenolpyruvate (PEP), 0.025 U/ml pyruvate kinase (PK), 1.5 mM ATP, 0.5 mM GTP, 40 μM [³H]phenylalanine (75 cpm/pmol), 0.3 mg/ml poly(U) RNA, 0.03 mM spermine, 1 mM DTT, 0.05 μM P. aeruginosa Elongation Factor-Ts (EF-Ts), 1.0 μM P. aeruginosa Elongation Factor-Tu (EF-Tu), 0.1 μM P. aeruginosa phenylalanyl-tRNA synthetase (PheRS), 0.2 μM P. aeruginosa ribosomes and the indicated amounts of EF-G1A or EF-G1B. Reactions were started by the addition of E. coli tRNA to a final concentration of 0.5 μM tRNA^Phe and continued for 1 h at 37 °C. Reactions were stopped by the addition of 2 ml 10% trichloroacetic acid (TCA) and filtered through glass fiber filters (Whatman) as previously described [25]. Retention of [³H]Phe represents the amount of poly-phenylalanine, poly(Phe), synthesized. The effects of FA on protein synthesis were determined using the same assays but contained from 4 to 250 μM FA. In these assays, the concentration of EF-G1A and EF-G1B were 1 μM and 0.2 μM, respectively. Assays to determine the effect of RRF on the activity of EF-G1B were the same as described above, with RRF titrated into the assay between 0.2 and 6.4 μM. Assays to determine the ability of EF-G1A/RRF to affect the activity of EF-G1B contained 0.2 μM EF-G1B, 1 μM EF-G1A and 2 μM RRF. In these reactions EF-G1A and/or RRF were pre-incubated in the reaction mix at 37 °C for 5 min, EF-G1B was then added and incubation was continued for 1 h.

EF-G is a protein with a molecular mass of approximately 77 kDa. The crystal structure for EF-G from T. thermophilus has been solved bound to GDP and in the nucleotide free form; it appears to be an elongated protein composed of five structural domains [6,7]. More recently, the structure of P. aeruginosa EF-G1 was determined at 2.9 Å [26] and the structure was shown to be similar to that of T. thermophilus EF-G1. The structure of EF-G2 from T. thermophilus has also been determined and is similar to the structure to EF-G1 even though the sequence homology is only 30% [17]. Other organisms studied also indicate that there is only a modest level of overall amino acid sequence conservation between EF-G1 and EF-G2 proteins from the same organism (Table 1). In Table 1, the sequence identity of EF-G1 and EF-G2 from the same organism ranges from 29-56 % and the sequence similarity ranges from 44-68 %. Unlike the homologs shown in Table 1, the two EF-G-like proteins from P. aeruginosa (EF-G1A and EF-G1B) have a much higher level of amino acid sequence conservation; with the amino acid sequences being 84% identical and 90% similar (Figure 1). This is a much higher level of sequence conservation than observed when EF-G1 from different organisms are compared to each other and is similar to that observed when EF-G1 from different strains of the same bacteria are compared (data not shown). In the phylum Proteobacteria only two bacteria that were analyzed (Bordetella bronchiseptica and Burkholderia rhizoxinica) contained two EF-G molecules that contained a high level of amino acid sequence conservation. When P. aeruginosa EF-G1A and EF-G1B were compared to the four homologs from these two bacteria a high level of homology was observed in which the similarity only varied from 76 to 81% (Table 2). The EF-G molecules from Bordetella bronchiseptica and Burkholderia rhizoxinica were slightly more homologous to each other than either was with P. aeruginosa EF-G proteins, with the similarity only varying between 86 to 91%. This closer homology might be expected as both of these bacteria belong to the beta-subdivision of the Proteobacteria phylum while P. aeruginosa is a member of the gamma-subdivision.

When P. aeruginosa EF-G1A and EF-G1B were compared with EF-G1 and EF-G2 from other organisms, a wide variation in the amino acid sequence similarity was observed (Table 2). A higher degree of similarity was observed when P. aeruginosa EF-G1A and EF-G1B were compared with EF-G from organisms containing only one form of EF-G than with organisms containing multiple forms of EF-G. When compared with other bacteria having distinct EF-G1 and EF-G2 proteins, both P. aeruginosa EF-G1A and EF-G1B were more similar to EF-G1 than with EF-G2 (Table 2).

In a comparison of EF-G1A and EF-G1B, the functional regions (P-loop, switch I and switch II regions) of domain I are strictly conserved with only one residue variation from a methionine to a tyrosine (Figure 1). The five typical motifs (G1-G5) for GTP recognition [2] are also strictly conserved between EF-G1A and EF-G1B. EF-G proteins contain a region within the G-domain termed the G’ insert for which the functional significance is not well understood. It has been speculated that this region may be an internal guanine nucleotide exchange factor (GEF) [27], or possibly a region that specifically functions in ribosome binding [28]. The G’ insert is highly conserved between EF-G1A and EF-G1B, with only modest amino acid variations observed (Figure 1). In domain IV, at position 529 there is a five amino acid insert (KGNIT) observed in EF-G1A that is not present in EF-G1B. Results from alignments with all homologs analyzed indicate that three of these amino acids (N, I andT) are only present in P. aeruginosa EF-G1A (data not shown). These three residues are located in a loop region in domain IV and would therefore probably not affect function. The only region of lower sequence similarity is the region of domain IV at the C-terminus of the protein. This region contains the highest degree of divergence between EF-G1A and EF-G1B; however, the variations seen in these amino acids are moderate. In the structure of EF-G1 (EF-G1A) from P. aeruginosa this region forms an α-helix and is detached from the body of the protein [26].

Fusidic acid inhibits protein synthesis by trapping EF-G in the post-translocation step during elongation [29,30]. Mutations conferring resistance to FA have been mapped in EF-G from T. thermophilus [31,32], S. typhimurium [33] and S. aureus [34]. The positions that these mutations map to in EF-G from P. aeruginosa are shown in Figure 1. In the approximately forty sites in which mutations have been identified, all but two are strictly conserved between EF-G1A and EF-G1B, and these two sites vary from a Ser to Cys at position 115 (EF-G1B numbering) and from an Ala to Val at position 140 (Figure 1). None of the amino acid sequence differences observed would be expected to affect the function of the proteins significantly. Overall, from the amino acid sequence analysis and comparisons with EF-G from other organisms, the differentiation of the roles of EF-G1A and EF-G1B cannot be discerned.

Two forms of EF-G from P. aeruginosa were cloned and expressed. The purification yielded both forms of EF-G in preparations that were greater than 98% homogeneous (Figure 2). GTPase activities of both forms of EF-G were shown to be dependent on the presence of ribosomes (Figure 3A). Only a low level of activity was observed in the absence of ribosomes and as the concentration of ribosomes increased, the level of GTPase activity increased. A ribosomal concentration of 0.4 μM was selected for downstream assays. When compared, EF-G1B exhibited approximately a 2-fold higher GTPase activity than was observed for EF-G1A (Figure 3B). At 0.5 μM (in the linear region of the plot), EF-G1A and EF-G1B were able to catalyze hydrolysis of 25 and 50 μM of GTP in 30 min reactions, respectively. At the inflection point on the titration curves EF-G1B was observed to have the same activity at 0.3 μM as EF-G1A had at 1.0 μM, therefore these concentrations were selected for downstream velocity assays at lower GTP concentrations. Timed assays showed that the GTPase activity of each form of EF-G was linear out to 30 min (Figure 3C).

To determine the kinetic parameters governing the ribosome dependent hydrolysis of GTP by the two forms of P. aeruginosa EF-G, initial velocity assays were carried out at varying concentrations of GTP (from 25 to 600 μM). The kinetic parameters K_M and V_max, for the GTPase activity of P. aeruginosa EF-G1B, were determined from Lineweaver-Burk analysis to be 71 μM and 3.7 μM/min, respectively (Figure 4). From these data the observed turnover number (k_cat^obs) was calculated to be 0.2 s^-1 and the specificity constant (k_cat/K_M) was calculated to be 3.0 x 10³ s^-1 M^-1 (Table 3). The same procedure was used to obtain these parameters for EF-G1A (Figure 5). The K_M and V_max values were 85 μM and 2.4 μM/min, respectively, and the k_cat^obs and k_cat/K_M for the hydrolysis of GTP by EF-G1A were 0.04 s^-1 and 0.5 x 10³ s^-1 M^-1, respectively. The K_M was observed to be very similar for both forms of EF-G; however, from these data the observed ability for the turnover of substrate (k_cat^obs) of EF-G1B is 5-fold greater than that of EF-G1A.

To determine the inhibitory effect of the antibiotic fusidic acid on the GTPase activity of both forms of EF-G, assays were performed containing FA at concentrations between 5 and 250 μM (Figure 6). The GTPase activity of EF-G1B was inhibited at the lowest concentration of FA tested and the activity was completely inhibited at FA concentrations above 30 μM. Alternatively, the GTPase activity of EF-G1A was not affected at any concentration of FA up to 250 μM.

An aminoacylation/translation (A/T) system composed of P. aeruginosa protein synthesis components has been developed in our laboratory [35] (Figure 7). To determine the ability of both forms of EF-G to function in protein synthesis, each was tested for the ability to function in the synthesis of poly(Phe). Each form of EF-G was titrated into the assay between 0.05 and 0.5 μM (Figure 7). EF-G1B displayed robust activity in polypeptide synthesis. In contrast, EF-G1A was only observed to have a low level of activity. To ascertain the ability of FA to inhibit elongation, FA was added to protein synthesis assays at concentrations between 5 and 500 μM (Figure 8). As shown in Figure 8A, even though the activity of EF-G1A only yields 0.6 μM poly(Phe) in the elongation phase of protein synthesis, the addition of FA does not appear to affect the activity even at the highest concentration of FA. The initial drop in activity shown in Figure 8A is due to the inhibition of background activity, likely due to small amounts of EF-G1B co-purified with ribosomes. These results are similar to the lack of effect that FA had on the GTPase activity of EF-G1A. However, FA has a profound effect on the ability of EF-G1B to function in protein synthesis (Figure 8B). At the lowest concentration of FA (4 μM) the activity was reduced by 25% and at the highest concentration of FA the activity is completely inhibited. These results suggest that EF-G1B is involved in the elongation phase of protein synthesis and that EF-G1A is not.

In mammalian mitochondria the two forms of EF-G were shown to have different functions. EF-G1_mt functions in the elongation stage of protein synthesis while EF-G2_mt in the presence of RRF functions in ribosomal recycling [14,15]. This was also shown to be the case in B. burgdorferi where EF-G1 was found to function in translocation while EF-G2 was shown to function exclusively in recycling [19]. To determine whether EF-G1A or EF-G1B functions along with RRF in catalyzing the separation of the two ribosomal subunits, each form of EF-G was tested in protein synthesis assays in the presence of increasing amounts of RRF (0.2 to 6.4 μM). Unlike protein synthesis using a natural messenger RNA, in poly(U) directed protein synthesis initiation can begin in the presence of tight-coupled (TC) 70S ribosomes [36]. If ribosomes are dissociated into the individual subunits the synthesis of poly(Phe) would be reduced. Likewise, the assay only detects poly(Phe) bound to ribosomes [37], therefore poly(Phe) synthesized and then released during the recycling event would not be detected and the overall activity would be reduced. When EF-G1B was assayed in this system (Figure 9A), no decrease in the synthesis of poly-Phe was observed at any concentration of RRF, indicating that EF-G1B was not able to act with RRF to recycle or separate the ribosomal subunits. In identical assays, the low level of activity of EF-G1A in protein synthesis was not observed to be affected by RRF at any concentration (data not shown). Figure 9B shows that RRF has no effect on the activity of either EF-G1A or EF-G1B in protein synthesis. When EF-G1A is added to the A/T protein synthesis system along with EF-G1B there is also no decrease in activity detected. However, when EF-G1A and RRF were pre-incubated with ribosomes prior to the addition of EF-G1B, protein synthesis was decreased by 25-30% (Figure 9B). These results indicate that EF-G1A and RRF act in concert to reduce protein synthesis. The recycling of the ribosomal subunits decreases the number of ribosomes available to function in protein synthesis and is the likely mechanism by which this is occurring.