1 Introduction

In general, the translation process is produced by ribosomes—complex molecular machines that provide peptide bond synthesis in an accordance with the mRNA nucleotide sequence. The mechanisms of translation in eukaryotes and prokaryotes may be different in many aspects, but fundamentally basic processes are uniform. To ensure an accuracy and an efficiency of the synthesis, in addition to a ribosome, special protein factors are required that are involved at various stages of translation. In prokaryotes one of such factors is three domain protein elongation factor P (EF-P) [1], eukaryotes have a two domain analogue of this factor—eIF5A [2]. There is still no integral comprehensive theory that includes the full mechanism of all the functions of EF-P and its regulation.

EF-P have been discovered in 1975 as a stimulating factor during a methionine-puromycin test—its positive effect on the synthesis of the first peptide bond was found [3]. Several studies showed the structure of EF-P in different organisms: Pseudomonas aeruginosa [4], Thermus thermophilus [5]. An important milestone in the development of our knowledge about EF-P function was the solution of the EF-P complex structure with the T.thermophilus ribosome [5]. Since then we can say that EF-P is located between P and E sites of the ribosome and is directly adjacent to the P-tRNA. EF-P is a three-domain protein resembling tRNA in shape [4, 6]. Each of the domains possesses its own role. Most of all is known about the domain I: it is located in vicinity of the CCA-end of P-tRNA and the peptidyl transferase center (PTC). The domain II is in contact with a L1 protein and the contact is most likely needed to remove the EF-P from the ribosome. The domain III is located near the anticodon loop of P-tRNA, S7 protein and the E-site codon of mRNA. It was proposed that the domain III could play role in frameshift prevention [7, 8]. Overwhelming number of studies showed that that EF-P provides specialized translation of proteins with polyproline type amino acid stalling motifs [9, 10], therefore shift interest from the initiation function of EF-P [5] to the translation function. To maximize the efficiency of the antistalling function, the EF-P needs the special post-translational modification in the conservative region of the loop in the domain I of the EF-P, located close to the PTC (Fig. 1).

Fig. 1
figure 1

Structure of elongation factor P (left) and conservative region of EF-P loop from domain I (right) [2]. Side chain is shown as sticks for highly conservative residue Lys32 which can carry a special post-translational modification

Full size image

There are four known types of possible EF-P PTM of highly conservative loop I of domain I: b-lysinilation [11], rhamnosilation [12], 5-aminopentilation [13] and hypusine in case of eIF5A [14]. Approximately one-third of all bacteria possess modifying enzymes corresponding to the known modification [1]. Modification status of other bacteria is unknown. Highly possible modification of S.aureus EF-P could be 5-aminopentanolitation of 34 K, the same type modification as for B.subtillus EF-P [15].

Bacteria with mutant forms of the EF-P loop, with the absence of EF-P or the absence of EF-P modifying enzymes have phenotypes that can be correlated with specific cell processes such as stress resistance, cell motility and virulence—that could be due to involvement in this processes proteins containing stalling motifs, regulated by EF-P [2]. Proteins with such sites are often involved in secretion processes, including secretion pathogenicity factors of microorganisms. Last study of EF-P complex with stalled ribosome made by CryoEM [16] proposed some structural mechanism of B-lysinilation modification action for polyproline antistalling, but there are still few details how this modification could work. Also there is no understanding how another types of modification could function and what is the reason for their structural difference.

Therefore the knowledge about dynamics of the loop I could be crucial for understanding of EF-P interaction process with P-tRNA near PTC and the function of modifications. Recently we analyzed a dynamic of SaEF-P backbone amide protons by high-resolution NMR spectroscopy [17] and due to high mobility of the loop between β-sheets β2 and β3 (residues 30-34) the amide protons have not been detected in 15N-1H heteronuclear NMR spectra due to fast proton exchange leading to the absence of the corresponding amide resonances. Site-directed spin labeling in concert with electron paramagnetic resonance (EPR) spectroscopy is a powerful method for studying the nature of proteins in solution which has been especially useful for examining proteins that are large, flexible or highly dynamic [18,19,20]. One of the most used spin labeling is achieved by chemically modifying engineered Cys residues with a nitroxide moiety provided by small molecules such as MTSL (Fig. 2). In present paper by EPR spectroscopy we investigated a spinlabeled mutated (K32C) EF-P (20,5 kDa) from Staphylococcus aureus (SaEF-P)—a pathogenic bacterium which cause various human diseases and despite the presence of different type of antibiotics is still a significant threat for human health, including a nosocomial infections. The structure of S.aureus EF-P is still absent. The approbation of this method with SaEF-P alone will allow us to check the dynamics of SaEF-P in complex with S.aureus ribosome and, in general, with any other protein that binds to the ribosome.

Fig. 2
figure 2

Chemical structure of nitroxide spin label MTSL

Full size image

2 Materials and methods

2.1 Synthesis and purification of EF-P‑SL

Elongation Factor P from S. aureus (SaEF-P) protein with mutation K32C and histidine tag at its C-terminus was expressed in Escherichia coli BL21star(DE3) transformed with a pGS21A plasmid. The cells were dissolved in the lysis buffer (20 mM Tris–HCl, 200 mM NH4Cl, pH 7.6) with addition of PIC and PMSF. Lisate was clarified by centrifugation at 13.000 rpm, 4 °C for 30 min (Beckman 25.50 rotor). SaEF-P-HIS was purified by gravity flow MAC chromatography (Ni–NTA). After a final step of a gel-filtration chromatography on Superdex 75 10/300, the purified protein was dissolved in phosphate buffer 50 mM phosphate buffer pH 7.4, 250 mM NH4Cl).

2.2 MTSL labeling of EF-P

MTSL spin label, 2,2,5,5-tetramethyl-1-oxyl-3-methyl methanethiosulfonate (O875000, from Toronto Research Chemicals) was dissolved at 10 mg/mL concentration in DMSO. For spin labeling, the SaEF-P(K32C) solution was mixed with ×10 ammount of MTSL and incubated 18 h at 4 °C. The excess spin label was removed by gel exclusion chromatography (gel filtration). Final sample of SaEF-P labeled protein was concentrated to 1 mM with Amicon Ultra Centrifugal Filter Device (10,000 Da cut off) to final volume equal to 500 μl.

2.3 Electron paramagnetic resonance spectroscopy

EPR measurements were done in continuous wave mode by using the abilities of Bruker X-band Elexsys 580 spectrometer at room temperature (Centre of the shared facilities at Kazan Federal University). The solutions were placed into the 0.8 mm inner diameter quartz tubes. Registration parameters were chosen to be 100 µW for the microwave power and 0.2 G at 100 kHz modulation to avoid saturation and overmodulation effects. EPR parameters (isotropic g-factor and hyperfine constant A) and rotational correlation times (by using the values of g- and A- components for MTSL in water listed in [21] were extracted from the EPR spectra fitting in Easyspin module for Matlab [22].

3 Results and discussion

EPR spectrum of MTSL in buffer solution contains 3 lines of hyperfine splitting typical for nitroxide radical in the low-viscous environments caused by the interaction of S = ½ and I = 1 for 14N nuclei (Fig. 3). From experiment parameters Aiso = 16.1 G, giso = 2.0059 in correspondence with the literature data for water solution [21], individual components line widths \(\Delta B_{pp }^{ + 1} = 1.13 G\), \(\Delta B_{pp}^{0} = 1.12 G\), \(\Delta B_{pp}^{ - 1} = 1.15 G\), can be estimated as well as a correlation time τR = 4·10−11 s (40 ps).

Fig. 3
figure 3

EPR spectrum of free MTSL spin label in phosphate buffer along with the corresponding simulation with the parameters given in the text and τR = 4·10−11 s

Full size image

For spin labeling, the SaEF-P(K32C) solution was mixed with ×10 amount of MTSL and the excess spin label was removed by Gel Exclusion Chromatography and final sample was analyzed by EPR. The results are presented in Fig. 4. The lines broadening due to immobilization, i.e. incomplete averaging by the motion of the g-and A component (\(\Delta B_{pp }^{ + 1} = 1.62 G\), \(\Delta B_{pp}^{0} = 1.41 G\), \(\Delta B_{pp}^{ - 1} = 2.54 G\)) was observed which means that spin label was bound to the protein. Correlation time of MTSL label bound to SaEF-P was estimated as τR = 8·10−10 s.

Fig. 4
figure 4

EPR spectrum MTSL spin label covalently attached to 32Cys with the corresponding simulation with τR = 8·10−10 s

Full size image

4 Conclusion

Addition of MTSL label covalently bound to 32Cys in the highly conservative part of EF-P loop in the domain I of the EF-P located near the CCA-end of P-tRNA and the peptidyl transferase center (PTC) allows us to analyze protein dynamic by EPR spectroscopy of the high mobility region which was not observed in NMR spectra due to fast proton exchange leading to the absence of the corresponding amide resonances. We believe that our result could be extended in future for the analysis of EF-P—ribosome complex formation process by EPR spectroscopy.