They Scan, They SCORE - American Chemical Society

---

currents

N ew M ethod forProtein Quantitation The traditional approach for measuring relative quantities of proteins is comparing spot intensities on two independent electrophoresis gels. Recently, less time-consuming and more accurate isotope labeling procedures for quantitation based on MS peptide maps have been investigated for proteomics applications. ICAT reagents and H218O are among those that have shown potential in a number of cases, but there is still a need for simple, less expensive alternatives. Salvatore Sechi from the National Institutes of Health (Bethesda, MD) has developed a method that not only can be used for MS-based quantification but also improves the quality of the same MS data for protein identification. Previously, Sechi reacted a protein mixture with acrylamide, a well-known alkylating agent of the amino acid cysteine, immediately preceding gel electrophoresis. In the subsequent peptide map, a number of alkylated ions derived from cysteine-containing peptides were observed that when unalkylated were not large enough to be detected (Anal. Chem. 1998, 70, 5150–5158). This increased peptide coverage enhances the effectiveness of database matching. In the same report, a protein mixture was combined with both protonated and deuterated acrylamide. Because cysteine would react with both, the isotopic distribution in the peptide map indicates the number of cysteines in each peptide. Knowledge of cysteine content reduces the number of possible proteins, which increases the accuracy of the database-driven identification. Recently, to determine relative protein concentrations, Sechi treated a one-femtomolar protein mixture with protonated acrylamide and a second mixture with deuterated acrylamide, prior to their combined gel electrophoresis (Rapid Commun. Mass Spectrom. 2002, 16, 1416–1424). The relative protein quantities could be determined by measuring the

Peptide–Disaccharide Interactions Sugars play a large but undervalued role in cell functions such as biological recognition and signaling. Therefore, an understanding of how proteins and sugars interact is critical. To address this problem, Hans-Joachim Gabius and colleagues at various institutes recently undertook a study of the minimal structural requirements for sugar binding by polypeptides (Biochemistry 2002, 41, 9707–9717). Animal lectins are typically composed of more than 100 amino acids, complicating the analysis of their car© 2002 American Chemical Society

bohydrate-binding behavior. Recent phage-display experiments with a library of pentadecapeptides, however, identified three peptides (see figure) with high affinity for Galβ1,3GalNAcα1,R, the Thomsen–Friedenreich (TF) antigen. The small size of the oligopeptides provides for potentially favorable pharmacokinetic properties and therefore might facilitate their use in biomedical applications. Gabius and colleagues tested the peptides in both the absence and presence of TF antigen and maltose (a disaccharide control) by NMR spectroscopy, ESI-MS,

ratios of the MS isotope patterns of the ions derived from cysteine-containing peptides. When the mixtures had protonated:deuterated protein ratios of 1:1 to 2:1 (or 1:2), the quantitation error was within 10%, but errors that resulted from more disparate ratios (5:1 or 10:1) were higher (up to 40%). In all cases, however, the prevalence of the deuterated or protonated protein was correctly determined. The error values might be improved, Sechi indicates, by using acrylamide labeled with both deuterium and 13C and by obtaining MS/MS data for some notable cysteine-containing peptides in the spectra. However, inherent limitations to this technique, such as the potential for coincidental overlap and cases where no cysteine-containing peptide is present, would still remain.

How m any cysteines? A MALDI-TOF spectrum of a peptide mixture treated with protonated and deuterated acrylamide, highlighting the isotopic envelopes of peptides containing one cysteine (A) and two cysteines (B, C). (Adapted with permission from Sechi, S. Rapid Commun. Mass Spectrom. 2002,16, 1416–1424. Copyright 2002 John Wiley & Sons, Ltd.)

and molecular modeling to determine how the molecules interact. The NMR studies suggested that the peptides alone were predominantly unfolded and that the aromatic residues long thought to form a hydrophobic core for folding did not do so. This unfolded state remained largely unchanged when TF antigen was added, although there were specific interactions between the sugar and peptides that did not occur with maltose. There was some concern that the aromatic residues might prompt the peptides to aggregate, but although the ESI-MS experiments showed

the presence of peptide homo- or heterodimers, larger complexes were absent. Furthermore, when the TF antigen was added to the mixture, it only bound to the monomeric form of the peptides. Finally, the molecular modeling experiments confirmed the results of the biophysical methods in that there was no evidence of peptide secondary structure formation in the presence of the TF antigen. Although the aromatic residues were unable to form a hydrophobic core for a folded secondary structure, the researchers were enthusiastic about the possibility of

JournalofProteom e Research • Vol. 1, No. 5, 2002 395

currents introducing two cysteine residues into the peptides in such a way that a disulfide bridge could be formed, as is the case in several other lectins. This, the authors argue, might reduce the conformational freedom of the peptides and thereby allow the peptides to take a form that will more tightly and selectively bind to the TF antigen.

They Scan,They SCORE Current efforts to elucidate the protein-encoding components of the genome tell only half of the story, because within their native context, these genes cannot express their products without the strict control of the regulatory sequences that surround them. To fully address this genomic blueprint, James Posakony and colleagues at the University of California at San Diego developed a computational method for identifying clusters of sites where regulatory proteins bind in the genome (Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 9888– 9893). The method, called SCORE (site clustering over random expectation), is an algorithm designed to detect and statistically evaluate cisregulatory sequences within whole-genome data to identify previously unrecognized gene expression enhancers. To test SCORE, the researchers searched the Drosophila genome for sites that match the binding consensus sequence for the transcription factor Suppressor of Hairless [Su(H)]. The identified sites were then clustered according to the number of adjacent sites within specificlength windows (100−5000 basepairs). The researchers then performed Monte Carlo simulations to determine the random incidence of Su(H) binding site clusters and compared these values with the real data to determine which clusters were likely to

Binding site clusters.The purities of clusters of Suppressor of Hairless [Su(H)] binding sites are plotted against the sizes of the Drosophila DNA fragments screened and the number of Su(H) sites found within the sequence. (Adapted with permission from Rebeiz, M.; et al. Proc. Natl. Acad. Sci. U.S.A. 2002,99, 9888–9893. Copyright 2002 National Academy of Sciences, U.S.A.)

be biologically relevant—in other words, their purity (see figure). Using SCORE, the researchers were able to identify several clusters near genes that were known or suspected to be regulated by Su(H). They were also able to identify novel genes, one of which they experimentally determined to be controlled by Su(H) in situ. On the basis of this test case, the researchers wrote, “SCORE and other similar techniques will no doubt prove increasingly valuable as tools for reading the regulatory genome.”

Enzym atic M icroreactoron-a-Chip Two challenges to improving the utility of peptide mapping for high-throughput proteomics are 1) making it a faster, more automated process, and 2) avoiding the occurrence of autodigestion of the proteolytic enzyme (e.g., trypsin), which complicates protein assignments. In a recent study, Jean M. J. Frechet and colleagues at the Univer-

tions within the microfluidic chips themselves, in which the optimized porous properties lead to very low back pressures, enabling the use of simple mechanical pumping for inducing flow. Trypsin solution was pumped through the monolith micromolds and immobilization took place for a period of 1–8 h, depending on the porosity. The activity of the immobilized enzyme was probed with three different substrates. First benzoylL-arginine ethyl ester, a standard trypsin assay, was pumped through the microreactor. Absorbance spectra were used to calculate the different enzyme kinetics with the varying monoliths and trypsin concentrations. Myoglobin solution was then run through the channels over 11.7 s and was digested approximately 95%, according to a comparison between the myoglobin MS peaks before and after proteolysis. The MS peptide map actually showed 67% coverage of all the myoglobin peptides, which is a comparable result to those yielded by macroanalysis methods that could take from 5 to 15 min. The digestion activity of the microreactor was further illustrated with fluorescently labeled casein, which was pumped at different flow rates through the chip. The fluorescence of the eluted material increased as the flow rate increased. Since the

sity of California, Berkeley, developed an enzymatic microreactor-on-a-chip that combines microfluidics for greater analysis speed with immobilization of trypsin on a solid support during proteolysis to eliminate unwanted autodigestion (Anal. Chem. 2002, 74, 4081–4088). The researchers used highly porous polymer monoliths molded into microfluidic channels, which provided extensive surface area for trypsin molecules to bind. The monoliths were copolymers of 2-vinyl-4,4-dimethylazlactone, which reacts readily with the amine and thiol groups of enzymes by photoactivation. Initially the monoliths were produced in bulk in large-volume molds using different solvent systems, copolymer ratios, and reaction times so that the researchers could control the pore Before and after.Mass spectrum of fluoresproperties (volume, dicently tagged casein before (A) and after (B) ameter, etc.). The scien- digestion in a trypsin-containing monolith. tists subsequently carried (Adapted from Peterson, D. S.; et al. Anal. out the photopolymeriza- Chem. 2002,74, 4081–4088.) JournalofProteom e Research • Vol. 1, No. 5, 2002 397

currents Screening Proteins w ith Electroporation Most methods used in the study of protein expression and function rely on the removal of the proteins from their cellular environment. Unfortunately, this eliminates much of the contextual information about these proteins, and therefore, in situ analytical methods are needed. But most proteins reside within cells, and most protein labels, drugs, and transfection agents are membrane-impermeable. To address this problem, Owe Orwar and colleagues at Göteborg University and the Chalmers University of Technology (Göteborg, Sweden) developed an electroporation method whereby fluorogenic substrates can be delivered into cells using an electrolyte-filled capillary (EFC) with few effects on cell viability. Fluorescence microscopy then follows the behavior of these substrates as they interact with their protein targets (Anal. Chem. 2002, 74, 4300–4305). Using a high-graduation micromanipulator, the EFC is positioned 20–40 µm above the cell. The EFC is then lowered toward the cell until the capillary comes into focus microscopically. The cell-bathing medium is grounded with a platinum wire, and a short electrical burst opens small transient pores in the cell membrane, allowing the entry of the electrolyte and fluorescently labeled ligand. The researchers tested their system by fluorescently tracing the Ca2+ efflux that followed electroporation of cells with 1,4,5-inositol triphosphate (IP3). They also examined a similar efflux following the introduction of cyclic adenosine diphosphate ribose (cADPr) to ryanodine receptors, and then showed that this efflux could be interrupted by the coelectroporation of ruthenium red with cADPr. Therefore, this assay might be useful in screening compounds that inhibit or block original solution had a very low level of fluorescence, this response demonstrates the occurrence of rapid cleavage of this high-molecularweight protein. The researchers expect these results to open up new avenues for the design of high-throughput protein mapping systems.

Low -Abundance Peptide DREAM S Cell and organism proteomes are complex in both protein variety and abundance levels, 398

Detecting alkaline phosphatase and proteases.The substrate fluorescein diphosphate was electroporated into a single cell, and the resulting fluorescence from its cleavage by alkaline phosphatase was measured (B). Images A and C are the same cells before and after electroporation, respectively. The fluorescence resulting from the cleavage of casein BIODIPY FL by cellular proteases was similarly detected (E). Images D and F are the same cells before and after electroporation, respectively. (Adapted from Nolkrantz, K.; et al. Anal. Chem. 2002,74, 4300–4305.)

specific cellular functions. The scientists then detected alkaline phosphatase and proteases by measuring the behaviors of fluorescein diphosphate and casein BIODIPY FL, respectively, when electroporated into the cells (see figure). Finally, by growing and electroporating cells in a microwell format, the researchers demonstrated that their method was capable of highthroughput screening. “Multiplexed EFC electroporation of a plurality of cells grown in arrays,” wrote the authors, “might become a useful tool in the search for agents such as drugs, genes, antisense oligonucleotides, etc., that affect intracellular chemistry and for the detection of pathophysiological states using specific markers.”

with the latter potentially ranging over 6 orders of magnitude. And it is this large range that complicates the identification and analysis of low-abundance proteins by techniques such as Fourier transform ion cyclotron resonance (FTICR) MS, where the detector can be swamped by signals from high-abundance peptides. This fact typically limits standard MS to a dynamic range of 103. To surmount this problem, researchers have combined FTICR MS with high-resolution capillary

JournalofProteom e Research • Vol. 1, No. 5, 2002

LC, but even this addition only improves the range to 104–105. As an improvement over these methods, Richard Smith and his colleagues at the Pacific Northwest National Laboratory (Richland, WA) developed a new method of FTICR MS that offers an expanded dynamic range. They call their system dynamic range enhancement applied to MS, or DREAMS, and recently coupled this system to a highefficiency capillary reversephase liquid chromatogram

( J. Am. Soc. Mass Spectrom. 2002, 13, 954–963). In DREAMS, a standard spectral acquisition event is followed by a second acquisition step where the most abundant peptide species are selectively removed through the superposition of a resonant waveform in the selection quadrupole. Their selection conditions eliminated the 10 most abundant peptides before initiating a prolonged accumulation event. This step allowed the accumulation quadrupole to be filled for a longer period with the remaining low-abundance peptides. Thus, DREAMS offers both a standard peptide spectrum (first acquisition) and a low-abundance peptide spectrum (second acquisition). The researchers tested DREAMS on a global trypsin digest of proteins from 14N- and 15N-labeled mouse cells. The first spectrum resulted in the detection of 9896 14N/15N peptide pairs, while the second spectrum revealed 8856 pairs, of which only 939 were found in the first spectrum—an increase of 80%. Similarly, the researchers examined peptide samples from 14N- and 15N-labeled D. radiodurans cells and compared their results with proteins identified in a D. radiodurans database. Using DREAMS, they were able to identify 835 (48% improvement) unique peptides, which represented 548 (62% improvement) potential protein coding regions not otherwise found in the database. Even though these results show that DREAMS is useful in expanding proteome coverage, the researchers are still hoping to make improvements, such as increasing the resolution of the ion ejection step and reducing the throughput time, but they still see their method as being a critical next step in proteomic analysis.