Rational Design of Cyanine-Based Fluorogenic Dimers to Reduce Nonspecific Interactions with Albumin and Lipid Bilayers: Application to Highly Sensitive Imaging of GPCRs in Living Cells.

Fluorogenic dimers with polarity-sensitive folding are powerful probes for live-cell bioimaging. They switch on their fluorescence only after interacting with their targets, thus leading to a high signal-to-noise ratio in wash-free bioimaging. We previously reported the first near-infrared fluorogenic dimers derived from cyanine 5.5 dyes for the optical detection of G protein-coupled receptors. Owing to their hydrophobic character, these dimers are prone to form nonspecific interactions with proteins such as albumin and with the lipid bilayer of the cell membrane resulting in a residual background fluorescence in complex biological media. Herein, we report the rational design of new fluorogenic dimers derived from cyanine 5. By modulating the chemical structure of the cyanine units, we discovered that the two asymmetric cyanine 5.25 dyes were able to form intramolecular H-aggregates and self-quenched in aqueous media. Moreover, the resulting original dimeric probes enabled a significant reduction of the nonspecific interactions with bovine serum albumin and lipid bilayers compared with the first generation of cyanine 5.5 dimers. Finally, the optimized asymmetric fluorogenic dimer was grafted to carbetocin for the specific imaging of the oxytocin receptor under no-wash conditions directly in cell culture media, notably improving the signal-to-background ratio compared with the previous generation of cyanine 5.5 dimers.

Rapid and Highly Selective Fluorescent Labeling of Peptides via a Thia-Diels-Alder Cycloaddition: Application to Apelin.

Herein, we describe a catalyst-free thia-Diels-Alder cycloaddition for the chemoselective labeling of fully deprotected phosphonodithioester-peptides in solution with fluorophores functionalized with an exocyclic diene. The reaction was optimized on the model tripeptide 1 containing a lysine residue, which enabled its rapid and straightforward labeling with three different fluorophores (fluorescein, lissamine rhodamine B, and squaraine) in very mild conditions (H2O/iPrOH, 37 °C, 1 h). The reaction was then successfully applied to the chemoselective labeling of fully deprotected apelin-13 with squaraine dye. The resulting fluorescent ligand 18 exhibited a high affinity (0.17 ± 0.03 nM) for apelinR. It enabled the development of time-resolved FRET-based competition assays for high-throughput screening and drug discovery. Thanks to its fluorogenic properties, ligand 18 was also successfully involved in the live-cell optical imaging of apelinR in no-wash conditions.

Target-Focused Library Design by Pocket-Applied Computer Vision and Fragment Deep Generative Linking.

We here describe a computational approach (POEM: Pocket Oriented Elaboration of Molecules) to drive the generation of target-focused libraries while taking advantage of all publicly available structural information on protein-ligand complexes. A collection of 31 384 PDB-derived images with key shapes and pharmacophoric properties, describing fragment-bound microenvironments, is first aligned to the query target cavity by a computer vision method. The fragments of the most similar PDB subpockets are then directly positioned in the query cavity using the corresponding image transformation matrices. Lastly, suitable connectable atoms of oriented fragment pairs are linked by a deep generative model to yield fully connected molecules. POEM was applied to generate a library of 1.5 million potential cyclin-dependent kinase 8 inhibitors. By synthesizing and testing as few as 43 compounds, a few nanomolar inhibitors were quickly obtained with limited resources in just two iterative cycles.

Nile Red-Based GPCR Ligands as Ultrasensitive Probes of the Local Lipid Microenvironment of the Receptor.

The local lipid microenvironment of transmembrane receptors is an essential factor in G protein coupled receptor (GPCR) signaling. However, tools are currently missing for studying endogenously expressed GPCRs in primary cells and tissues. Here, we introduce fluorescent environment-sensitive GPCR ligands for probing the microenvironment of the receptor in living cells using fluorescence microscopy under no-wash conditions. We designed and synthesized antagonist ligands of the oxytocin receptor (OTR) by conjugating a high-affinity nonpeptidic OTR ligand PF-3274167 to the environment-sensitive fluorescent dye Nile Red. The length of the polar PEG spacer between the pharmacophore and the fluorophore was adjusted to lower the nonspecific interactions of the probe while preserving a strong fluorogenic response. We demonstrated that the new probes embed into the lipid bilayer in the vicinity of the receptor and convey information about the local polarity and the lipid order via the wavelength-shifting emission of the Nile Red fluorophore.

LIT-001, the First Nonpeptide Oxytocin Receptor Agonist that Improves Social Interaction in a Mouse Model of Autism.

Oxytocin (OT) and its receptor (OT-R) are implicated in the etiology of autism spectrum disorders (ASD), and OT-R is a potential target for therapeutic intervention. Very few nonpeptide oxytocin agonists have currently been reported. Their molecular and in vivo pharmacology remain to be clarified, and none of them has been shown to be efficient in improving social interaction in animal models relevant to ASD. In an attempt to rationalize the design of centrally active nonpeptide full agonists, we studied in a systematic way the structural determinants of the affinity and efficacy of representative ligands of the V1a and V2 vasopressin receptor subtypes (V1a-R and V2-R) and of the oxytocin receptor. Our results confirm the subtlety of the structure-affinity and structure-efficacy relationships around vasopressin/oxytocin receptor ligands and lead however to the first nonpeptide OT receptor agonist active in a mouse model of ASD after peripheral ip administration.

A Time-Resolved FRET Cell-Based Binding Assay for the Apelin Receptor.

Analogues of apelin-13 carrying diverse spacers and an ad hoc DY647-derived fluorophore were designed and synthesized by chemoselective acylation of α-hydrazinopeptides. The resulting probes retain very high affinity and efficacy for both the wild-type and SNAP-tagged apelin receptor (ApelinR). They give a time-resolved FRET (TR-FRET) signal with rare-earth lanthanides used as donor fluorophores grafted onto the SNAP-tagged receptor. This specific signal allowed the validation of a binding assay with a high signal-to-noise ratio. In such an assay, the most potent sub-nanomolar fluorescent probe was found to be competitively displaced by the endogenous apelin peptides with binding constants similar to those obtained in a classical radioligand assay. We have thus validated the first TR-FRET cell-based binding assay for ApelinR with potential high-throughput screening applications.

Fluorogenic squaraine dimers with polarity-sensitive folding as bright far-red probes for background-free bioimaging.

Polarity-sensitive fluorogenic dyes raised considerable attention because they can turn on their fluorescence after binding to biological targets, allowing background-free imaging. However, their brightness is limited, and they do not operate in the far-red region. Here, we present a new concept of fluorogenic dye based on a squaraine dimer that unfolds on changing environment from aqueous to organic and thus turns on its fluorescence. In aqueous media, all three newly synthesized dimers displayed a short wavelength band characteristic of an H-aggregate that was practically nonfluorescent, whereas in organic media, they displayed a strong fluorescence similar to that of the squaraine monomer. For the best dimer, which contained a pegylated squaraine core, we obtained a very high turn-on response (organic vs aqueous) up to 82-fold. Time-resolved studies confirmed the presence of nonfluorescent intramolecular H-aggregates that increased with the water content. To apply these fluorogenic dimers for targeted imaging, we grafted them to carbetocin, a ligand of the oxytocin G protein-coupled receptor. A strong receptor-specific signal was observed for all three conjugates at nanomolar concentrations. The probe derived from the core-pegylated squaraine showed the highest specificity to the target receptor together with minimal nonspecific binding to serum and lipid membranes. The obtained dimers can be considered as the brightest polarity-sensitive fluorogenic molecules reported to date, having ∼660,000 M(-1) cm(-1) extinction coefficient and up to 40% quantum yield, whereas far-red operation region enables both in vitro and in vivo applications. The proposed concept can be extended to other dye families and other membrane receptors, opening the route to new ultrabright fluorogenic dyes.

Red fluorescent turn-on ligands for imaging and quantifying G protein-coupled receptors in living cells.

Classical fluorescence-based approaches to monitor ligand-protein interactions are generally hampered by the background signal of unbound ligand, which must be removed by tedious washing steps. To overcome this major limitation, we report here the first red fluorescent turn-on probes for a G protein-coupled receptor (oxytocin receptor) at the surface of living cells. The peptide ligand carbetocin was conjugated to one of the best solvatochromic (fluorogenic) dyes, Nile Red, which turns on emission when reaching the hydrophobic environment of the receptor. We showed that the incorporation of hydrophilic octa(ethylene glycol) linker between the pharmacophore and the dye minimized nonspecific interaction of the probe with serum proteins and lipid membranes, thus ensuring receptor-specific turn-on response. The new ligand was successfully applied for background-free imaging and quantification of oxytocin receptors in living cells.

High-throughput automated refolding screening of inclusion bodies.

One of the main stumbling blocks encountered when attempting to express foreign proteins in Escherichia coli is the occurrence of amorphous aggregates of misfolded proteins, called inclusion bodies (IB). Developing efficient protein native structure recovery procedures based on IB refolding is therefore an important challenge. Unfortunately, there is no "universal" refolding buffer: Experience shows that refolding buffer composition varies from one protein to another. In addition, the methods developed so far for finding a suitable refolding buffer suffer from a number of weaknesses. These include the small number of refolding formulations, which often leads to negative results, solubility assays incompatible with high-throughput, and experiment formatting not suitable for automation. To overcome these problems, it was proposed in the present study to address some of these limitations. This resulted in the first completely automated IB refolding screening procedure to be developed using a 96-well format. The 96 refolding buffers were obtained using a fractional factorial approach. The screening procedure is potentially applicable to any nonmembrane protein, and was validated with 24 proteins in the framework of two Structural Genomics projects. The tests used for this purpose included the use of quality control methods such as circular dichroism, dynamic light scattering, and crystallogenesis. Out of the 24 proteins, 17 remained soluble in at least one of the 96 refolding buffers, 15 passed large-scale purification tests, and five gave crystals.

Structure of Escherichia coli YhdH, a putative quinone oxidoreductase.

As part of a structural genomics project on bacterial gene products of unknown function, the crystal structures of YhdH, a putative quinone oxidoreductase, and its complex with NADP have been determined at 2.25 and 2.6 A resolution, respectively. The overall fold of YhdH is very similar to that of alcohol dehydrogenases and quinone reductases despite its low sequence identity. The absence of any Zn ion indicates that YdhH is a putative quinone oxidoreductase. YhdH forms a homodimer, with each subunit composed of two domains: a catalytic domain and a coenzyme-binding domain. NADP is bound in a deep cleft formed between the two domains. Large conformational changes occur upon NADP binding, with the two domains closing up to each other and narrowing the NADP-binding cleft. Comparisons of the YdhH active site with those of the quinone oxidoreductases from Escherichia coli and Thermus thermophilus made it possible to identify essential conserved residues as being Asn41, Asp43, Asp64 and Arg318. The active-site size is very narrow and unless an induced fit occurs is accessible only to reagents the size of benzoquinone.

Crystal structure and kinetics identify Escherichia coli YdcW gene product as a medium-chain aldehyde dehydrogenase.

In the context of a medium-scaled structural genomics program aiming at solving the structures of as many as possible bacterial unknown open reading frame products from Escherichia coli (Y prefix), we have solved the structure of YdcW at 2.1A resolution, using molecular replacement. According to its sequence identity, YdcW has been classified into the betaine aldehyde dehydrogenases family (EC 1.2.1.8), catalysing the oxidation of betaine aldehyde into glycine betaine. The structure of YdcW resembles that of other aldehyde dehydrogenases: it is tetrameric and binds a NADH molecule in each monomer. The NADH molecules, bound in the active site by soaking, are revealed to be in the "hydrolysis position". Activities experiments demonstrate that YdcW is more active on medium-chains aldehyde than on betaine aldehyde. However, soaking of betaine into YdcW crystals revealed its presence in one of the subunits, in two positions, a putative resting position and a hydride transfer ready position. Analysis of kinetics data and of the active site shape suggest an optimum binding of n-alkyl aldehydes up to seven to eight carbon atoms, possibly followed by a bulky cyclic or aromatic group.

Crystal structure of E.coli alcohol dehydrogenase YqhD: evidence of a covalently modified NADP coenzyme.

In the course of a structural genomics program aiming at solving the structures of Escherichia coli open reading frame (ORF) products of unknown function, we have determined the structure of YqhD at 2.0A resolution using the single wavelength anomalous diffraction method at the Pt edge. The crystal structure of YqhD reveals that it is an NADP-dependent dehydrogenase, a result confirmed by activity measurements with several alcohols. The current interpretation of our findings is that YqhD is an alcohol dehydrogenase (ADH) with preference for alcohols longer than C(3). YqhD is a dimer of 2x387 residues, each monomer being composed of two domains, a Rossmann-type fold and an alpha-helical domain. The crystals contain two dimers in the asymmetric unit. While one of the dimers contains a cofactor in both subunits, only one of the subunits in the second dimer contains it, making it possible to compare bound and unbound active sites. The active site contains a Zn atom, as verified by EXAFS on the crystals. The electron density maps of NADP revealed modifications of the nicotinamide ring by oxygen atoms at positions 5 and 6. Further analysis by electrospray mass spectrometry and comparison with the mass spectra of NADP and NADPH revealed the nature of the modification and the incorporation of two hydroxyl moieties at the 5 and 6 position in the nicotinamide ring, yielding NADPH(OH)(2). These modifications might be due to oxygen stress on an enzyme, which would functionally work under anaerobic conditions.

The crystal structure of the Escherichia coli lipocalin Blc suggests a possible role in phospholipid binding.

Lipocalins form a large multifunctional family of small proteins (15-25 kDa) first discovered in eukaryotes. More recently, several types of bacterial lipocalins have been reported, among which Blc from Escherichia coli is an outer membrane lipoprotein. As part of our structural genomics effort on proteins from E. coli, we have expressed, crystallized and solved the structure of Blc at 1.8 A resolution using remote SAD with xenon. The structure of Blc, the first of a bacterial lipocalin, exhibits a classical fold formed by a beta-barrel and a alpha-helix similar to that of the moth bilin binding protein. Its empty and open cavity, however, is too narrow to accommodate bilin, while the alkyl chains of two fatty acids or of a phospholipid could be readily modeled inside the cavity. Blc was reported to be expressed under stress conditions such as starvation or high osmolarity, during which the cell envelope suffers and requires maintenance. These data, together with our structural interpretation, suggest a role for Blc in storage or transport of lipids necessary for membrane repair or maintenance.

The Escherichia coli YadB gene product reveals a novel aminoacyl-tRNA synthetase like activity.

In the course of a structural genomics program aiming at solving the structures of Escherichia coli open reading frame products of unknown function, we have determined the structure of YadB at 1.5A using molecular replacement. The YadB protein is 298 amino acid residues long and displays 34% sequence identity with E.coli glutamyl-tRNA synthetase (GluRS). It is much shorter than GluRS, which contains 468 residues, and lacks the complete domain interacting with the tRNA anticodon loop. As E.coli GluRS, YadB possesses a Zn2+ located in the putative tRNA acceptor stem-binding domain. The YadB cluster uses cysteine residues as the first three zinc ligands, but has a weaker tyrosine ligand at the fourth position. It shares with canonical amino acid RNA synthetases a major functional feature, namely activation of the amino acid (here glutamate). It differs, however, from GluRSs by the fact that the activation step is tRNA-independent and that it does not catalyze attachment of the activated glutamate to E.coli tRNAGlu, but to another, as yet unknown tRNA. These results suggest thus a novel function, distinct from that of GluRSs, for the yadB gene family.

The crystal structure of the Escherichia coli YfdW gene product reveals a new fold of two interlaced rings identifying a wide family of CoA transferases.

Because of its toxicity, oxalate accumulation from amino acid catabolism leads to acute disorders in mammals. Gut microflora are therefore pivotal in maintaining a safe intestinal oxalate balance through oxalate degradation. Oxalate catabolism was first identified in Oxalobacter formigenes, a specialized, strictly anaerobic bacterium. Oxalate degradation was found to be performed successively by two enzymes, a formyl-CoA transferase (frc) and an oxalate decarboxylase (oxc). These two genes are present in several bacterial genomes including that of Escherichia coli. The frc ortholog in E. coli is yfdW, with which it shares 61% sequence identity. We have expressed the YfdW open reading frame product and solved its crystal structure in the apo-form and in complex with acetyl-CoA and with a mixture of acetyl-CoA and oxalate. YfdW exhibits a novel and spectacular fold in which two monomers assemble as interlaced rings, defining the CoA binding site at their interface. From the structure of the complex with acetyl-CoA and oxalate, we propose a putative formyl/oxalate transfer mechanism involving the conserved catalytic residue Asp169. The similarity of yfdW with bacterial orthologs (approximately 60% identity) and paralogs (approximately 20-30% identity) suggests that this new fold and parts of the CoA transfer mechanism are likely to be the hallmarks of a wide family of CoA transferases.

A medium-throughput crystallization approach.

The first results of a medium-scale structural genomics program clearly demonstrate the value of using a medium-throughput crystallization approach based on a two-step procedure: a large screening step employing robotics, followed by manual or automated optimization of the crystallization conditions. The structural genomics program was based on cloning in the Gateway vectors pDEST17, introducing a long 21-residue tail at the N-terminus. So far, this tail has not appeared to hamper crystallization. In ten months, 25 proteins were subjected to crystallization; 13 yielded crystals, of which ten led to usable data sets and five to structures. Furthermore, the results using a robot dispensing 50-200 nl drops indicate that smaller protein samples can be used for crystallization. These still partial results might indicate present and future directions for those who have to make crucial choices concerning their crystallization platform in structural genomics programs.

Control of domain swapping in bovine odorant-binding protein.

As revealed by the X-ray structure, bovine odorant-binding protein (OBPb) is a domain swapped dimer [Tegoni, Ramoni, Bignetti, Spinelli and Cambillau (1996) Nat. Struct. Biol. 3, 863-867; Bianchet, Bains, Petosi, Pevsner, Snyder, Monaco and Amzel (1996) Nat. Struct. Biol. 3, 934-939]. This contrasts with all known mammalian OBPs, which are monomers, and in particular with porcine OBP (OBPp), sharing 42.3% identity with OBPb. By the mechanism of domain swapping, monomers are proposed to evolve into dimers and oligomers, as observed in human prion. Comparison of bovine and porcine OBP sequences pointed at OBPp glycine 121, in the hinge linking the beta-barrel to the alpha-helix. The absence of this residue in OBPb might explain why the normal lipocalin beta-turn is not formed. In order to decipher the domain swapping determinants we have produced a mutant of OBPb in which a glycine residue was inserted after position 121, and a mutant of OBPp in which glycine 121 was deleted. The latter mutation did not result in dimerization, while OBPb-121Gly+ became monomeric, suggesting that domain swapping was reversed. Careful structural analysis revealed that besides the presence of a glycine in the hinge, the dimer interface formed by the C-termini and by the presence of the lipocalins conserved disulphide bridge may also control domain swapping.