A general method for the covalent labeling of fusion proteins with small molecules in vivo

Characterizing the movement, interactions, and chemical microenvironment of a protein inside the living cell is crucial to a detailed understanding of its function. Most strategies aimed at realizing this objective are based on genetically fusing the protein of interest to a reporter protein that monitors changes in the environment of the coupled protein. Examples include fusions with fluorescent proteins, the yeast two-hybrid system, and split ubiquitin. However, these techniques have various limitations, and considerable effort is being devoted to specific labeling of proteins in vivo with small synthetic molecules capable of probing and modulating their function. These approaches are currently based on the noncovalent binding of a small molecule to a protein, the formation of stable complexes between biarsenical compounds and peptides containing cysteines, or the use of biotin acceptor domains. Here we describe a general method for the covalent labeling of fusion proteins in vivo that complements existing methods for noncovalent labeling of proteins and that may open up new ways of studying proteins in living cells.     

Fluorescent labeling of proteins. A new methodology

A new reagent, 2-methoxy-2,4-diphenyl-3(2$\text{H}$)-furanone (MDPF) has been utilized for the fluorescent labeling of proteins. MDPF, which is nonfluorescent, reacts with primary amino groups to form fluorescent $\text{N}$-substituted 3,5-diphenyl-5-hydroxy-2-pyrrolin-4-ones. Antibodies labeled with MDPF afford intense immunofluorescent staining.     

Site-specific covalent labeling of proteins inside live cells using small molecule probes

The study of dynamic movement and interactions of proteins inside living cells in real time is critical for a better understanding of cellular mechanisms and functions in molecular detail. Genetically encoded fusions to fluorescent protein(s) (FP) have been widely used for this purpose [Annu. Rev. Biochem. 1998, 67, 509-544]. To obviate some of the drawbacks associated with the use of FPs [Curr. Opin. Biotechnol. 2005, 16, 1-6; Nat. Methods2006, 3, 591-596], we report a small molecule-based approach that exploits the unique reactivity between the cysteine residue at the N-terminus of a target protein and cell-permeable, thioester-based small molecule probes resulting in site-specific, covalent tagging of proteins. This approach has been demonstrated by the in vivo labeling of proteins in both bacterial and mammalian systems thereby making it potentially useful for future bioimaging applications.     

A new assay for proteases using fluorescent labeling of proteins

A simple assay for proteases based on the fluorescent labeling of insoluble proteins (fibrin) or of soluble casein by 2-methoxy-2,4-diphenyl-3(2H)furanone has been developed. Fluorescence of the liberated peptide-fluorophors resulting from enzymatic hydrolysis is easily measured in the supernatant after separation of the unreacted fluorescent fibrin by centrifugation or from unreacted casein-fluorophor by acid precipitation. Nanogram quantities of trypsin, chymotrypsin, and elastase can be measured.     

A new view of water dynamics in immobilized proteins.

The inflection frequency of the deuteron magnetic relaxation dispersion from water in rotationally immobilized protein samples has recently been found to be essentially independent of temperature and protein structure. This remarkable invariance has been interpreted in terms of a universal residence time of 1 microseconds for protein-associated water molecules. We demonstrate here that this interpretation is an artifact of the conventional perturbation theory of spin relaxation, which is not valid for rotationally immobile proteins. Using a newly developed non-perturbative, stochastic theory of spin relaxation, we identify the apparent correlation time of 1 microseconds with the inverse of the nuclear quadrupole frequency, thus explaining its invariance. The observed dispersion profiles are consistent with a broad distribution of residence times, spanning the microseconds range. Furthermore, we argue that the deuteron dispersion is due to buried water molecules rather than to the traditional surface hydration previously invoked, and that the contribution from rapidly exchanging protein hydrogens cannot be neglected. The conclusions of the present work are also relevant to proton relaxation in immobilized protein samples and to magnetic resonance imaging of soft tissue.     

Development of a small peptide tag for covalent labeling of proteins

A 21-mer peptide that can be used to covalently introduce synthetic molecules into proteins has been developed. Phage-displayed peptide libraries were subjected to reaction-based selection with 1,3-diketones. The peptide was further evolved by addition of a randomized region and reselection for improved binding. The resulting 21-mer peptide had a reactive amino group that formed an enaminone with 1,3-diketone and was used as a tag for labeling of maltose binding protein. Using this peptide tag and 1,3-diketone derivatives, a variety of molecules such as reporter probes and functionalities may be covalently introduced into proteins of interest.     

Axonal transport of synaptic vesicle proteins in the rat optic nerve

The optic nerve, as a part of the central nervous system (CNS), has been used to study axonal transport for decades. The present study has concentrated on the axonal transport of synaptic vesicle proteins in the optic nerve, using the “stop-flow/nerve crush” method. After blocking fast axonal transport, distinct accumulations of synaptic vesicle proteins developed during the first hour after crush-operation and marked increases were observed up to 8 h postoperative. Semiquantitative analysis, using cytofluorimetric scanning (CFS) of immunoincubated sections, revealed that the ratio between distal accumulations (organelles in retrograde transport) and proximal accumulations (organelles in anterograde transport) was much higher (up to 80–90%) for the transmembrane proteins than that for surface adsorbed proteins (only 10–20%). The pattern of axonal transport in the optic nerve was comparable to that in the sciatic nerve. However, clathrin and Rab3a immunoreactivities were accumulated in much lower amounts than that in the sciatic nerve. Most synaptic vesicle proteins were colocalized in the axons proximal to the crush. A differential distribution of synaptobrevin I and II, however, was observed in the optic nerve axons; synaptobrevin I was present in large-sized axons, while synaptobrevin II immunoreactivity was present in most axons, including the large ones. The two isoforms were, thus, partially colocalized. The results demonstrate that (1) cytofluorimetric scanning techniques could be successfully used to study axonal transport not only in peripheral nerves, but also in the CNS; (2) synaptic vesicles are transported with fast axonal transport in this nerve; and (3) some differences were noted compared with the sciatic nerve, especially for Rab3a and clathrin. © 1997 John Wiley & Sons, Inc. J Neurobiol 32: 237–250, 1997.     

Axonal transport—I. Diffusion

The phenomenon of axonal transport of material has been well documented (Ochs, 1971; Lasek, 1970; and Grafstein, 1967). This report seeks to establish the role of diffusion—if any—in such a transport process. We report that diffusion cannot account for the observed build-up of material as reported in the literature.     

Axonal transport—II. Convection

The phenomenon of axonal transport has been well documented (Ochs, 971; Lasek, 1970; and Grafstein, 1967). In a previous paper, we showed how diffusion alone could not account for this process. In this report we show that convection or convection with diffusion can account for the observed build-up of material. By including a first-order catabolic sequestration term, we are able to offer an understanding of the several apparent rates of transport with the same underlying velocity and variable sequestration.     

A new view of convective-diffusive transport processes in the arterial intima

In this paper a new theoretical framework is presented for analyzing the filtration and macromolecular convective-diffusive transport processes in the intimal region of an artery wall with widely dispersed macromolecular cellular leakage sites, as proposed in the leaky junction-cell turnover hypothesis of Weinbaum et al. In contrast to existing convection-diffusive models, which assume that the transport is either 1-D, or convection is primarily in a direction normal to the endothelial surface, the present model considers for the first time the nonuniform subendothelial pressure field that arises from the different hydraulic resistances of normal and leaky endothelial clefts and the special role of the internal elastic lamina (IEL) in modulating the horizontal transport of macromolecules after they have passed through the leaky clefts of cells that are either in mitosis or demonstrate IgG labeling. The new theory is able to quantitatively explain the growing body of recent experiments in which an unexpectedly rapid early-time growth of the leakage spot has been observed and the longer time asymptotic behavior in which the leakage spot appears to approach an equilibrium diameter. The new theory also predicts the observed doubling in macromolecular permeability between EBA labeled blue and white areas when the frequency of leakage sites is doubled. This frequency for doubling of permeability, however, is an order of magnitude smaller than predicted by the author's previous model, Tzeghai et al., in which only convection normal to the endothelial surface was considered and the pressure was uniform in the intima. The longer time model predictions are used to explain the time scale for the formation of liposomes in subendothelial tissue matrix in animal feeding experiments where it has been observed that the extracellular lipid concentration rises sharply prior to the entry of monocytes into the intima.     

N-terminal labeling of proteins using initiator tRNA

Methodology based on tRNA mediated protein engineering is described for the introduction of fluorophores and other labels at the N-terminus of proteins produced in cell-free translation systems. One method for low-level (trace) N-terminal labeling is based on the use of an Escherichia coli initiator tRNA(fMet) misaminoacylated with methionine modified at the alpha-amino group. In addition to the normal formyl group, the protein translational machinery incorporates the fluorophore BODIPY-FL and the affinity tag biotin at an N-terminal end of the nascent protein. A second method for higher N-terminal labeling uses a chemically aminoacylated amber initiator suppressor tRNA and a DNA template which contains a complementary amber (UAG) codon instead of the normal initiation (AUG) codon. This more versatile approach is demonstrated using a variety of N-terminal markers including fluorescein, biotin, PC-biotin, and a novel dual marker conjugate (Biotin/BODIPY-FL).     

Axonal transport of phospholipids in rat visual system.

Abstract— Seventeen day old rats were injected intraocularly with a phospholipid precursor, [³²P]phosphate, and a glycoprotein precursor, [³H]fucose. Animals were killed between 1 h and 21 days later, and structures of the visual pathway (retina, optic nerve, optic tract, lateral geniculate body, and superior colliculus) were dissected. Radioactivity in phospholipids ([³²P] in solvent-extracted material) and in glycoproteins ([³H] in solvent-extracted residue) was determined. Incorporation of [³H]fucose into retinal glycoproteins peaked at 6–8 h. Labelled glycoproteins were present in superior colliculus by 2h after injection, indicating a rapid rate of transport; maximal labelling was at 8–10 h after injection. Incorporation of [³²P]phosphate into retinal phospholipids peaked at 1 day after injection. Phospholipids were also rapidly transported since label was present in the superior colliculus by 3 h after injection: however, maximal labelling did not occur until 5–6 days. These results indicate that newly synthesized phospholipids enter a preexisting pool, part of which is later committed to transport at a rapid rate. Transported phospholipids were catabolized at the nerve endings with a maximum half-life of several days; there was minimal recycling of precursor label. Lipids were fractionated by thin-layer chromatography, and radioactivity in individual phospholipid classes determined. Choline and ethanolamine phosphoglycerides were the major transported phospholipids, together accounting for approx 85% of the total transported lipid radioactivity. At early time points, the ratio of radioactivity in choline phosphoglycerides to that in ethanolamine phosphoglycerides increased in structures progressively removed from the site of synthesis (retina) but by 2 days approached a constant value. In each structure, choline phosphoglyceride-ethanolamine phosphoglyceride radioactivity ratios decreased with time, rapidly at first, but plateaued by 2 days. These results indicate that choline phosphoglycerides are committed to transport sooner than ethanolamine phosphoglycerides. Some experiments were also conducted using [2-³H]glycerol as a phospholipid precursor. Results concerning incorporation of this precursor into individual phospholipid classes and their subsequent axonal transport were comparable to those obtained using [³²P]phosphate, with the following exceptions: (a) incorporation of [2-³H]glycerol into retinal phospholipids was relatively rapid (near-maximal levels at 1 h after injection) although transport to the superior colliculus showed an extended time course very similar to [³²P]-labelled lipids; (b) [2-³H]glycerol was somewhat less efficient than [³²P]phosphate in labelling lipids committed to transport relative to labelling those which remained in the retina; and (c) [2-³H]glycerol did not label plasmalogens.     

Tumor labeling in vivo using cyanine-conjugated monoclonal antibodies

Far-red-emitting cyanine fluorochromes have many properties desirable for in vivo imaging: absorption and emission at wavelengths where blood and tissue are relatively transparent, high quantum yields, and good solubility even at high molar ratios of fluorochrome to antibody. Potentially, conjugation by multiple linkages should minimize hydrolysis in vivo. We conjugated two tumor-targeting monoclonal antibodies: anti-SSEA-1 (IgM, ) at ratios of 1.2–35 mol dye/mol antibody and 9.2.27 (IgG_2a, ) at 0.6–6 mol dye/mol antibody, using the cyanine fluorochromes Cy3.18, Cy5.18, and Cy5.5.18. Nude mice were inoculated using the SSEA-1-expressing MH-15 teratocarcinoma or the 9.2.27 antigen-expressing SK-MEL-2 melanoma to give tumors at several sites. Conjugated antibody was injected, and mice were imaged immediately after injection and at appropriate intervals thereafter using a standard camera lens, dissecting microscope, or endoscopes. Images were acquired using either an image-intensified video camera or cooled CCD cameras. Immediately after injection, major blood vessels and the heart, liver, and kidneys were readily visualized. After 1 day, tumor-targeting antibody conjugates were concentrated in tumors and there was little circulating conjugate; however, the bladder and kidneys were still visible. Tumors labeled by specific antibody were the most fluorescent tissues at 2 days after injection, but non-specific antibody conjugates did not concentrate in the tumors. The small intestine was weakly visualized by both specific and non-specific antibody conjugates. These data support the possibility of visualizing tumor metastasis by optical means, including currently available endoscopes.This work was supported by funds of the Division of Urologic Surgery, Department of Surgery, University of Pittsburgh; by an NSF Center Grant (MCB-8920118) to the Center for Light Microscope Imaging and Biotechnology, Carnegie-Mellon University; and by intramural funds of the VAMC     

HAM: A new epitope-tag for in vivo protein labeling

The ability to metabolically label proteins with ³⁵S-methionine is critical for the analysis of protein synthesis and turnover. Despite the importance of this approach, however, efficient labeling of proteins in vivo is often limited by a low number of available methionine residues, or by deleterious side-effects associated with protein overexpression. To overcome these limitations, we have created a methionine-rich variant of the widely used HA tag, called HAM, for use with ectopically expressed proteins. Here we describe the development of a series of vectors, and corresponding antisera, for the expression and detection of HAM-tagged proteins in mammalian cells. We show that the HAM tag dramatically improves the sensitivity of ³⁵S-methionine labeling, and permits the analysis of Myc oncoprotein turnover even when HAM-tagged Myc is expressed at levels comparable to that of the endogenous protein. Because of the improved sensitivity provided by the HAM tag, the vectors and antisera described here should be useful for the analysis of protein synthesis and destruction at physiological levels of protein expression.     

In vivo covalent binding of [14C]trinitrotoluene to proteins in the rat.

When a single dose of [14C]trinitrotoluene was administered intraperitoneally (i.p.) to rats at 1, 10 or 50 mg/kg of body weight, covalently bound radioactivity was detected in globin, plasma proteins and proteins in the liver and kidney. The extent of covalent binding was dose dependent and was highest in plasma and renal proteins at all times up to 4 h after dosing. Covalent adduct levels in globin, however, decline slower than others. At a dose of 50 mg/kg of body weight, globin covalent adduct levels peaked at 1 h after dosing at 182 pmol/mg protein and subsequently decreased to approximately 50 pmol/mg protein between days 1 and 8. Of the covalent adduct levels in liver and kidney, those in the 10,000 x g and microsomal fractions were found to be higher than that in the cytosolic fraction. Radioactivity covalently bound to globin and the hepatic proteins was susceptible to dilute acid hydrolysis from which 2-amino-4,6-dinitrotoluene (2A) and 4-amino 2,6-dinitrotoluene (4A) were the major products recovered by solvent extraction. Upon acetylation, the hydrolysate gave rise to derivatives identified as the acetates of 2A and 4A on the basis of mass spectrometry and HPLC cochromatography with authentic samples. Four hours after an i.p. dose of [14C]TNT at 50 mg/kg of body weight about 0.4% of the dose was found as bound adducts to hemoglobin, of which approximately 48% was recovered as solvent extractable radioactivity after acid hydrolysis. About 2% of the radioactive dose was in the liver, of which approximately 30% was covalently bound to hepatic proteins, and approximately 49% of that was convertible to solvent extractable radioactivity upon acid hydrolysis. In vitro incubation of [14C]TNT with blood showed that there was a linear increase of covalent adducts in globin during the first 2 h of incubation; the concentration of covalent adducts was slightly higher than that with plasma proteins. The major compounds recovered from the hydrolysate of the globin adducts were also 2A and 4A as obtained from globin in the in vivo studies. On the basis of the in vitro and in vivo study results, we have confirmed the formation of protein adducts following a single i.p. administration of [14C]TNT at 1, 10 or 50 mg/kg of body weight to the rat or by in vitro incubation with blood.(ABSTRACT TRUNCATED AT 400 WORDS)     

Evaluation of a new reagent for covalent attachment of polyethylene glycol to proteins.

Methoxypolyethylene glycol of molecular weight 5000 was converted to a reactive succinimidyl carbonate form (SC-PEG). The usefulness of this new polymeric reagent for the covalent attachment of polyethylene glycol to proteins was evaluated. SC-PEG was found to be sufficiently reactive to produce extensively modified proteins under mild conditions within 30 min, showing the highest reactivity around pH 9.3. The commonly used succinimidyl succinate derivative of methoxypolyethylene glycol (SS-PEG) served as a reference standard to which the new reagent was compared. The stability of the polymer-protein linkages, studied on a series of PEG-modified bovine serum albumins, provided the single most important difference between the two activated polymers. Urethane-linked PEG-proteins obtained through the use of SC-PEG showed considerably higher chemical stability than SS-PEG-derived conjugates. The measured rate constants of aminolysis (using N alpha-acetyllysine) and hydrolysis showed that SC-PEG is slightly less reactive yet more selective of the two reagents. Hydrolysis of the active groups on SC-PEG was on average twofold slower than that on SS-PEG. The differences in the rates of aminolysis were even smaller than those in hydrolysis. PEG-trypsin conjugates produced by both activated polymers showed similar properties: they had no proteolytic activity, well-preserved esterolytic activity, and enhanced activity toward p-nitroanilide substrates. Michaelis-Menten constants of the modified enzymes were determined using N alpha-benzyloxycarbonyl-L-arginine p-nitroanilide. These measurements indicated that the attachment of PEG to trypsin caused an increase in both the rate of turnover of the substrate and its affinity toward the modified enzymes. Through a series of experiments involving the appropriate polymeric and low-molecular-weight model compounds, it was demonstrated that these increases in amidolytic activity were unrelated to tyrosyl residues acylation by either one of the activated polymers.