In Vitro Production of Perdeuterated Proteins in H2O for Biomolecular NMR Studies

Enabling site-specific NMR investigations of therapeutic Fab using a cell-free based isotopic labeling approach: application to anti-LAMP1 Fab

Monoclonal antibodies (mAbs) are biotherapeutics that have achieved outstanding success in treating many life-threatening and chronic diseases. The recognition of an antigen is mediated by the fragment antigen binding (Fab) regions composed by four different disulfide bridge-linked immunoglobulin domains. NMR is a powerful method to assess the integrity, the structure and interaction of Fabs, but site specific analysis has been so far hampered by the size of the Fabs and the lack of approaches to produce isotopically labeled samples. We proposed here an efficient in vitro method to produce [15N, 13C, 2H]-labeled Fabs enabling high resolution NMR investigations of these powerful therapeutics. As an open system, the cell-free expression mode enables fine-tuned control of the redox potential in presence of disulfide bond isomerase to enhance the formation of native disulfide bonds. Moreover, inhibition of transaminases in the S30 cell-free extract offers the opportunity to produce perdeuterated Fab samples directly in 1H2O medium, without the need for a time-consuming and inefficient refolding process. This specific protocol was applied to produce an optimally labeled sample of a therapeutic Fab, enabling the sequential assignment of 1HN, 15N, 13C', 13Cα, 13Cβ resonances of a full-length Fab. 90% of the backbone resonances of a Fab domain directed against the human LAMP1 glycoprotein were assigned successfully, opening new opportunities to study, at atomic resolution, Fabs' higher order structures, dynamics and interactions, using solution-state NMR.

Efficient and economic protein labeling for NMR in mammalian expression systems: Application to a preT-cell and T-cell receptor protein

Protein nuclear magnetic resonance (NMR) spectroscopy relies on the ability to isotopically label polypeptides, which is achieved through heterologous expression in various host organisms. Most commonly, Escherichia coli is employed by leveraging isotopically substituted ammonium and glucose to uniformly label proteins with 15N and 13C, respectively. Moreover, E. coli can grow and express proteins in uniformly deuterium-substituted water (D2O), a strategy useful for experiments targeting high molecular weight proteins. Unfortunately, many proteins, particularly those requiring specific posttranslational modifications like disulfide bonding or glycosylation for proper folding and/or function, cannot be readily expressed in their functional forms using E. coli-based expression systems. One such class of proteins includes T-cell receptors and their related preT-cell receptors. In this study, we present an expression system for isotopic labeling of proteins using a nonadherent human embryonic kidney cell line, Expi293F, and a specially designed media. We demonstrate the application of this platform to the β subunit common to both receptors. In addition, we show that this expression system and media can be used to specifically label amino acids Phe, Ile, Val, and Leu in this system, utilizing an amino acid-specific labeling protocol that allows targeted incorporation at high efficiency without significant isotopic scrambling. We demonstrate that this system can also be used to express proteins with fluorinated amino acids. We were routinely able to obtain an NMR sample with a concentration of 200 μM from 30 mL of culture media, utilizing less than 20 mg of the labeled amino acids.

Protein deuteration via algal amino acids to circumvent proton back-exchange for 1H-detected solid-state NMR

With perdeuteration, solid-state NMR spectroscopy of large proteins suffers from incomplete amide-proton back-exchange. Using a 72 kDa micro-crystalline protein, we show that deuteration exclusively via deuterated amino acids, well-established in solution to suppress sidechain protonation without proton back-exchange obstacles, provides spectral resolution comparable to perdeuterated preparations at intermediate spinning frequencies.

Site-Specific Introduction of Alanines for the Nuclear Magnetic Resonance Investigation of Low-Complexity Regions and Large Biomolecular Assemblies

Nuclear magnetic resonance (NMR) studies of large biomolecular machines and highly repetitive proteins remain challenging due to the difficulty of assigning frequencies to individual nuclei. Here, we present an efficient strategy to address this challenge by engineering a Pyrococcus horikoshii tRNA/alanyl-tRNA synthetase pair that enables the incorporation of up to three isotopically labeled alanine residues in a site-specific manner using in vitro protein expression. The general applicability of this approach for NMR assignment has been demonstrated by introducing isotopically labeled alanines into four distinct proteins: huntingtin exon-1, HMA8 ATPase, the 300 kDa molecular chaperone ClpP, and the alanine-rich Phox2B transcription factor. For large protein assemblies, our labeling approach enabled unambiguous assignments while avoiding potential artifacts induced by site-specific mutations. When applied to Phox2B, which contains two poly-alanine tracts of nine and twenty alanines, we observed that the helical stability is strongly dependent on the homorepeat length. The capacity to selectively introduce alanines with distinct labeling patterns is a powerful tool to probe structure and dynamics of challenging biomolecular systems.

Stabilization of a DNA aptamer by ligand binding

G-rich aptamers such as AS1411 are small oligonucleotides that present several benefits comparatively to monoclonal antibodies, since they are easier to manufacture and store, have small size and do not stimulate an immune response. We analyzed AT11-B1, a modified sequence of AT11 (itself a modified version of AS1411), in which one thymine was removed from the bulge region. We studied G-quadruplex (G4) formation/stabilization using PhenDC3, PDS, BRACO-19, TMPyP4 and 360A ligands by different biophysical techniques, namely circular dichroism (CD), Förster resonance energy transfer (FRET-melting) and nuclear magnetic resonance (NMR). The CD spectra showed that AT11-B1 adopts a predominant G4 of parallel topology when the buffer contains KCl or when ligands are added. PhenDC3 induced a ΔT_m of 30 °C or more of the G4 structure as shown by CD- and FRET-melting experiments. The ligands demonstrate high affinity for AT11-B1 G4 and the NMR studies revealed that the AT11-B1 G4 involves four G-tetrad layers. The in silico studies suggest that all ligands bind AT11-B1 G4, namely, by stacking interactions, except PDS that may bind to the loop/groove interface. In addition, molecular dynamics simulations revealed that nucleolin (NCL) interacts with the AT11-B1 G4 structure through the RNA binding domain (RBD) 2 and the 12-residue linker between RBD1,2. Moreover, AT11-B1 G4 was internalized into a NCL-positive tongue squamous cell carcinoma cell line. In a nutshell, this study may help the identification of the ligands scaffolds to bind and stabilize AT11-B1, improving the targeting towards NCL that is overexpressed in cancer cells.

Pre-miRNA-149 G-quadruplex as a molecular agent to capture nucleolin

One of the most significant challenges in capturing and detecting biomarkers is the choice of an appropriate biomolecular receptor. Recently, RNA G-quadruplexes emerged as plausible receptors due to their ability to recognize with high-affinity proteins. Herein, we have unveiled and characterized the capability of the precursor microRNA 149 to form a G-quadruplex structure and determined the role that some ligands may have in its folding and binding capacity to nucleolin. The G-quadruplex formation was induced by K+ ions and stabilized by ligands, as demonstrated by nuclear magnetic resonance and circular dichroism experiments. Surface plasmon resonance measurements showed a binding affinity of precursor microRNA 149 towards ligands in the micromolar range (10-5-10-6 M) and a strong binding affinity to nucleolin RNA-binding domains 1 and 2 (8.38 × 10-10 M). Even in the presence of the ligand PhenDC3, the binding remains almost identical and in the same order of magnitude (4.46 × 10-10 M). The molecular interactions of the RNA G-quadruplex motif found in precursor miRNA 149 (5'-GGGAGGGAGGGACGGG- 3') and nucleolin RNA-binding domains 1 and 2 were explored by means of molecular docking and molecular dynamics studies. The results showed that RNA G-quadruplex binds to a cavity between domains 1 and 2 of the protein. Then, complex formation was also evaluated through polyacrylamide gel electrophoresis. The results suggest that precursor microRNA 149/ligands and precursor microRNA 149/nucleolin RNA-binding domains 1 and 2 form stable molecular complexes. The in vitro co-localization of precursor microRNA 149 and nucleolin in PC3 cells was demonstrated using confocal microscopy. Finally, a rapid and straightforward microfluidic strategy was employed to check the ability of precursor microRNA 149 to capture nucleolin RNA-binding domains 1 and 2. The results revealed that precursor microRNA 149 can capture nucleolin RNA-binding domains 1 and 2 labeled with Fluorescein 5-isothiocyanate in a concentration-dependent manner, but PhenDC3 complexation seems to decrease the ability of precursor microRNA 149 to capture the protein. Overall, our results proved the formation of the G-quadruplex structure in the precursor microRNA 149 and the ability to recognize and detect nucleolin. This proof-of-concept study could open up a new framework for developing new strategies to design improved molecular receptors for capture and detection of nucleolin in complex biological samples.

Determination of the two-component systems regulatory network reveals core and accessory regulations across Pseudomonas aeruginosa lineages

Pseudomonas aeruginosa possesses one of the most complex bacterial regulatory networks, which largely contributes to its success as a pathogen. However, most of its transcription factors (TFs) are still uncharacterized and the potential intra-species variability in regulatory networks has been mostly ignored so far. Here, we used DAP-seq to map the genome-wide binding sites of all 55 DNA-binding two-component systems (TCSs) response regulators (RRs) across the three major P. aeruginosa lineages. The resulting networks encompass about 40% of all genes in each strain and contain numerous new regulatory interactions across most major physiological processes. Strikingly, about half of the detected targets are specific to only one or two strains, revealing a previously unknown large functional diversity of TFs within a single species. Three main mechanisms were found to drive this diversity, including differences in accessory genome content, as exemplified by the strain-specific plasmid in IHMA87 outlier strain which harbors numerous binding sites of conserved chromosomally-encoded RRs. Additionally, most RRs display potential auto-regulation or RR-RR cross-regulation, bringing to light the vast complexity of this network. Overall, we provide the first complete delineation of the TCSs regulatory network in P. aeruginosa that will represent an important resource for future studies on this pathogen.