Electron-capture mass spectrometry: recent advances.

Electron-capture (EC) is a sensitive and selective ionization technique for mass spectrometry (MS). In the most familiar form of EC, a susceptible analyte (electrophore) is detected after eluting from a gas chromatography (GC) column, where a low attomole detection limit for standards is routine. High-performance liquid chromatography can facilitate sample cleanup prior to detection by GC-EC-MS, but carryover and shifts in retention time for the "invisible" analyte can be difficulties. Solid-phase extraction avoids these difficulties, but the degree of cleanup and recovery can be problems. Alternative electrophoric derivatizing reagents are available to help deal with interferences, and new reagents such as "AMACE1" are emerging. Releasable forms of electrophores can be used as tags for labeling macromolecules, motivated by the desire to multiplex ligand-type assays. The conventional, gas-phase ion source for EC is not well-understood, especially the role of wall reactions. Using an electron monochromator to tune the electron energy adds to the selectivity and information provided by EC-MS. High-resolution and tandem EC-MS measurements are emerging. Electron-capture dissociation is a new technique to sequence small- to medium-sized peptides, having the advantage of providing more extensive sequence information relative to other MS techniques. Particle-beam EC-MS tends to be less sensitive than GC-EC-MS, but not always. Recently it was demonstrated that EC-MS can be accomplished on an ordinary laser desorption time-of-flight mass spectrometer, and also by using atmospheric pressure chemical ionization. Two applications are discussed here in detail: bile acids and oxidized phenylalanine. EC-MS is well-established as a useful technique for trace analysis in special cases, and the scope of its usefulness is broadening (qualitative analysis and detection of more polar and larger molecules), based on advances in both the chemical and instrumental aspects of this technique.

AMACE1: versatile aminoacetamide electrophore reagent.

Acetamide, 2-amino-N-[[3,5-bis(trifluoromethyl)phenyl]-methyl]-N-methyl-, monohydrochloride, which we have named AMACE1, was synthesized in three steps starting from N-tritylglycine. AMACE1 was coupled via its primary amine group (pKa 8.2) under aqueous conditions to four model analytes for oxidative sugar damage to DNA: glycolate, 3-hydroxy-2-butanone, 3-phenylbutyraldehyde, and alpha-hydroxy-gamma-butyrolactone, relying on cyanoborohydride for coupling to a keto function and a water-soluble carbodiimide for coupling to a carboxyl function. Further reaction with butyric anhydride led to products that could be detected by gas chromatography/electron capture mass spectrometry when 1 microL of ethyl acetate containing essentially 20 amol of each product was injected, on the basis of selected ion monitoring of the analyte characteristic anion fragment from dissociative loss of the 3,5-bis-(trifluoromethyl)phenylmethyl moiety: m/z 215, 289, 299, and 329, respectively. Since many small, organic analytes contain a keto or carboxylic acid group, AMACE1 should be useful in general in the area of trace organic analysis.

Matrix-assisted laser desorption/ionization mass spectrometry of deoxynucleotides labeled with an IMI dye.

The four major deoxynucleotides of DNA, and adduct mixtures resulting from separate reactions of 5'-dAMP and 5'-dGMP with benzo[a]pyrene diolepoxide (BPDE), were analyzed by matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) after labeling of their phosphate group with an IMI dye. The latter reagent comprises an imidazole functional group attached to a BODIPY (4,4-difluoro-5,7-dimethyl-4-bora-3a,4a-diaza-s-indacene+ ++) fluorophore. Good sensitivity was observed in the detection of the IMI-labeled products by MALDI-MS: 300-500 fmol in the laser spot (1% of the 30-50 pmol sample on the target) gave a signal-to-noise (S/N) of >/=30 from 20-30 superimposed laser shots. The BPDE reaction products, after the IMI labeling, were also subjected to capillary electrophoresis with laser-induced fluorescence detection, which revealed a complex mixture of products. Overall the results encourage the further development of this 'IMI-postlabeling' methodology as an alternative to (32)P-postlabeling for the detection of DNA adducts.

Preparation of an IMI dye (imidazole functional group) containing a 4-(N,N-dimethylaminosulfonyl)-2,1,3-benzoxadiazole fluorophore for labeling of phosphomonoesters.

We are studying dye-imidazole conjugates ("IMI dyes") as reagents for labeling phosphomonoesters such as nucleotides. Previously we have employed a BODIPY dye in our IMI reagents, and analyzed the labeled products by capillary electrophoresis with laser-induced fluorescence detection (CE-LIF) involving an argon ion laser. (The BODIPY fluorophore is a 4,4-difluoro-4-bora-3a,4a-diaza-s-indacene). Here we broaden the technology by preparing a DBD-IMI dye [DBD = 4-(N,N-dimethylaminosulfonyl)-2,1,3-benzoxadiazole], and using a helium-cadmium laser. While DBD-IMI (IMI3) is about 50x more stable photolytically than a BODIPY-IMI dye (IMI2, a conjugate of a BODIPY dye with histamine, was tested), the detection limit for IMI2 (5.10(-11) M; S/N = 5, CE-LIF with an argon ion laser) is tenfold better than that for IMI3 (5.10(-10) M, S/N = 5, helium-cadmium laser). IMI3 conjugates of the four major DNA nucleotides were prepared and detected by CE-LIF.

Contrasting behavior of pentafluorophenoxyacetone and pentafluorobenzyloxyacetone in electron impact and electron capture mass spectrometry.

Towards the goal of finding new ketone electrophores suitable as molecular labels for electrophoric release tags, pentafluorophenoxyacetone (1) and pentafluorobenzyloxyacetone (2) were prepared. Both ketones were evaluated by electron capture (EC) and electron impact (EI) modes of mass spectrometry (MS). By EC-MS, 1 nearly gave a single ion (as desired), whereas 2 gave many ions. This behavior was completely reversed in EI-MS. To account for certain ion fragments in the EC mass spectrum of 2, an anion radical McLafferty-type rearrangement and loss of a carbene neutral were postulated. Electron impact of 1 gave an abundant ion at m/z 117 (C5F3+), which was suggested to be a diyne cation.

Phosphate-specific fluorescence labeling with BO-IMI: reaction details.

Previously we reported than BO-IMI, a reagent which contains a BODIPY fluorophore linked to an imidazole group, can be used to covalently label a phosphomonoester in a single step under aqueous conditions [P. Wang, R.W. Giese, Anal. Chem. 65 (1993) 3518]. The reaction was conducted in the presence of a water-soluble carbodiimide 1-ethyl-3-(3'-N,N'-dimethylaminopropyl) carbodiimide [EDC] to activate the phosphomonoester, and the coupling took place onto both the N1 and N3 imidazole nitrogens of BO-IMI. Whether the two BO-IMI-phosphomonoester regioisomers migrated separately or together during capillary electrophoresis depended on the pH, due to a difference in their pKa values. Since then, we have studied the reaction in more detail leading to the information reported here. First, we have learned that the regioisomer ratio changes during the course of the reaction, and found that the mechanism involves both spontaneous and BO-IMI-catalyzed hydrolysis of the less stable isomer. Second, there is a background reaction in which BO-IMI becomes attached to EDC. Third, the BO-IMI-phosphomonoester product (a mixture of two isomers), that is observed by capillary electrophoresis at an alkaline pH, is found to no longer contain the two fluorine atoms present in the starting BO-IMI reagent. This is because they are placed by hydroxy groups at high pH. Finally, an event was discovered which complicates the detection of less than about 60 fmol of a phosphomonoester with BO-IMI: hydrolysis of a tiny fraction of the BO-IMI takes place during the coupling reaction, which leads to chemical noise in the capillary electropherogram.

N-hydroxysuccinimide ester labeling 5'-aminoalkyl DNA oligomers: reaction conditions and purification.

Difficulties were encountered in labeling 5'-aminoalkyl DNA oligomers with glycolketo electrophore N-hydroxysuccinimide esters in aqueous sodium bicarbonate (a common base for this purpose), followed by C18-silica reversed-phase high-performance liquid chromatography (HPLC) to achieve purification. The electrophore-labeled oligomers were not separated readily either from the hydrolyzed electrophore or from the starting oligomer. This problem was overcome by conducting the reaction with triethylamine as a base, organic washing the reaction mixtures after evaporation, and separating on a C18-poly(styrene-divinylbenzene) HPLC packing.

Electrophore mass tag dideoxy DNA sequencing.

Toward a goal of dideoxy sequencing DNA utilizing electrophore labels, we prepared four electrophore-labeled DNA oligonucleotide primers. Each primer has a different electrophore and DNA sequence but a common glycol keto (alpha,beta-dihydroxyketo) release group. Cleavage of this latter group by either periodate oxidation or a thermal retroaldol reaction releases the electrophores for detection by mass spectrometry. Successful sequencing data with these primers was obtained by capillary electrophoresis on an ABI Model 310 after fluorescence dideoxy terminator cycle sequencing reactions were conducted. In a separate experiment, it was demonstrated that a cocktail of the four electrophore DNA primers could be detected as a dried sample spot by CO2 laser desorption/capillary collection/gas chromatography electron capture mass spectrometry. These results establish some feasibility for our long-term goal of high-speed multiplex electrophore mass tag dideoxy DNA sequencing. Ultimately we plan to use a higher number of electrophore mass tags and to rely on direct detection of the desorbed electrophores by electron capture time-of-flight mass spectrometry.

Hydrazide as a ligand moiety in immobilized metal ion affinity chromatography. Separation of BO-IMI and BODIPY-hydrazide.

BODIPY hydrazide (BO-HZ, a commercially available fluorescent dye) and BO-IMI (obtained by coupling the hydrazide moiety of BODIPY to the carboxyl group of N-acetylhistidine) were separated on three forms of a Sepharose-iminodiacetic acid column: Cu(II), Ni(II) and Zn(II). Whereas BO-IMI eluted first on the Cu(II) and Ni(II) columns (a pH gradient from 7.0 to 2.0 was applied), it eluted last on the Zn(II) column. BO-HZ eluted from the Zn(II) column without displacing this metal. The explanation suggested for these results is that BODIPY hydrazide undergoes strong, bidendate binding only to the Cu(II) and Ni(II) columns.

Electron-capture detection: difluorobenzyl and related electrophores.

Difluorobenzyl derivatives (several isomers were tested) of 4-hydroxyacetophenone were synthesized and found to have similar properties (retention and response) by both reversed-phase HPLC and GC-ECD relative to each other, and also relative to that of a corresponding conventional pentafluorobenzyl derivative. The same was true for a representative difluorobenzyl derivative of thymine and 1-naphthoic acid. Overall, the responses by GC-ECD for the same core structure were only about two- to four-fold lower for a difluorobenzyl compared to a corresponding pentafluorobenzyl derivative. This makes a difluorobenzyl derivative attractive as an HPLC-UV retention marker, and sometimes as a substitute for a pentafluorobenzyl derivative (to help overcome an interference) in a method based on detection by electron capture. We also observed, somewhat as an aside, that the GC-ECD response of the benzyl derivative of 4-hydroxyacetophenone was only seven-fold lower than that of the corresponding pentafluorobenzyl derivative, and that this former benzyl derivative gave a 2.10(4) higher response than acetophenone. Thus, replacing the ring hydrogen atoms of a benzyl group with fluorine atoms had a relatively small impact on both the hydrophobicity and electron capture properties of the compounds tested here.

Detection via laser desorption and mass spectrometry of multiplex electrophore-labeled albumin.

Albumin was reacted with a mixture of six electrophore N-hydroxysuccinimide esters, each of which possessed an interior glycolketo linkage. The purpose of this linkage is to release the attached electrophore as a ketone when heated, due to a thermal retro-aldol reaction. The multiplex electrophore-labeled albumin was detected as a dried spot deposited on a polyimide membrane by laser desorption/capillary collection (of the released ketone electrophores)/off-line gas chromatography/electron capture-mass spectrometry. This encourages further study of such electrophore labels in immunoassays and related techniques, where there is a need to make advances in multi-analyte detection.

General method for determining ethylene oxide and related N7-guanine DNA adducts by gas chromatography-electron capture mass spectrometry.

A 112-micrograms sample of DNA was spiked with 103 pg of N7-(2'-hydroxyethyl)guanine and 100 pg of N7-(2'-hydroxyethyl-d4)guanine, the internal standard. The sample was subjected to the following sequence of steps: heating at 100 degrees C, precipitation of the DNA with HCl, reaction with nitrous acid to form the corresponding xanthines, reaction twice with pentafluorobenzyl bromide (first to derivatize NH, then OH), solid-phase extraction on silica and detection by gas chromatography-electron capture mass spectrometry. The absolute, overall yield of final product for both the analyte and internal standard was 9.7%. Conveniently, the three chemical reactions are conducted sequentially in the same vial and, aside from a washing step, are separated only by evaporations. Corresponding N7-guanine methyl, phenyl and styrene oxide adducts were detected at about the 50-ng level by the procedure, to indicate the generality of the method.

Phosphate-specific fluorescence labeling of pepsin by BO-IMI.

Pepsin (3.6 nmol) was detected by the following three-step procedure: (i) reaction with a 20-fold molar excess of BO-IMI (a fluorophore containing a reactive imidazole group) in the presence of a 150-fold molar excess of 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC) and 2% sodium dodecyl sulfate; (ii) gel filtration (spin column) to remove most of the residual BO-IMI; and (iii) capillary electrophoresis with laser-induced fluorescence detection. For the latter step, 8.5 x 10(-7) of the original sample was injected. BO-IMI/EDC targets phosphomonoesters and does not label albumin (prior knowledge). Progressive dephosphorylation of pepsin with acid phosphatase reduced its labeling with BO-IMI. Thus, the BO-IMI, as intended, labels the phosphate group on pepsin. Such BO-IMI labeling should be useful in general for studying phosphoproteins and phosphopeptides.

Masking as a mechanism for evaporative loss of trace analyte, especially after solid-phase extraction.

Using a Pasteur pipette plugged with silanized glass wool and packed with C18-silica particles, we attempted to remove K2CO3 from an aqueous acetonitrile solution. In spite of extensive washing of the column with water after the sample was applied, elution with acetonitrile followed by evaporation gave a visible, white residue. It was found that the residue was derived from both the sample and the packing, including particles from the latter. Substitution of a plastic column/polyethylene frit for the Pasteur pipette/glass wool gave a more consistent residue, apparently because this improved the retention of particles. Subsequent experiments were conducted in the plastic hardware. The amount of the residue was observed to vary as much as 19-fold when C18-silica particles were tested from different manufacturers, and the residue could be reduced in amount as much as 9-fold when a column was prepared in the laboratory vs. the use of a comparable, pre-packed column. The water itself contributed some of the residue: even the "purest" water routinely available left a visible residue when 1.0 ml was appropriately evaporated (e.g. on Saran Wrap in a microwave oven). The recovery of an arbitrary trace analyte and internal standard (pentafluorobenzylated nucleobases at the low pg level) was 32% less when they were evaporated in acetonitrile that had been passed through an acetonitrile and water-washed cartridge containing C18-Si vs. evaporation in untreated acetonitrile. Collectively these results reveal that an evaporation can risk some loss of an analyte from masking by even subtle solvent contaminants.(ABSTRACT TRUNCATED AT 250 WORDS)

Chemical transformation/derivatization of O6-methyl- and O6-(hydroxyethyl)guanine for detection by GC-EC/MS.

In this project we set out to make an important class of DNA adducts, comprising O6-alkyl and O6-(hydroxyalkyl)guanines, susceptible to sensitive detection by GC-EC/MS. While existing literature indicated that pentafluorobenzylation would be useful for the ring NH site on these compounds, how to best overcome the polarity of the exocyclic NH2 and OH groups, without losing the O6-alkyl moiety, was less clear. Working with O6-methylguanine and O6-(2'-hydroxyethyl)guanine as representative analytes, we found that the NH2 group could be converted into fluoro without loss of the O6 substituent. For the OH group, a comparison of several derivatives (OR') led to R' = tert-butyl as the best choice at this stage. The latter work, especially via NMR, also allowed exact structural assignments to be made for the N7 and N9 pentafluorobenzyl isomeric derivatives that formed. Of these R' derivatives, the N7 isomers migrated slower on silica-TLC, had higher GC retention times, had lower responses by GC-EC/MS, and were preferentially destroyed as the GC column aged. However, the N9 isomer was slower on TLC when the OH was not derivatized. This behavior was rationalized using a concept of "polar footprint" for the derivatives. The concept also seemed to explain the puzzling GC-EC/MS behavior of some related compounds in our laboratory. Apparently the polar footprint should be minimized in designing derivatives for trace detection by GC-EC/MS.

Methods development toward the measurement of polyaromatic hydrocarbon-DNA adducts by mass spectrometry.

The measurement of DNA adducts in human samples is at an early stage. The accuracy of some of the current measurements is not defined, the structures are unknown for a significant number of the adducts that have been detected, and there is little information about how many adducts remain to be discovered. This is due largely to the trace amounts of human DNA adducts in any sample. A consequence of this is that the true potential of DNA adducts as indicators of exposure and risk in human toxicology is far from realized. Mass spectrometry, a powerful technique for organic analysis, is the key to exploiting fully the usefulness of human DNA adducts as biomarkers of human exposure and risk. Mass spectrometry can make accurate measurements, discover unknown compounds, and determine the structures of these unknown compounds. However, the trace (very small) amounts of human DNA adducts have limited mass spectrometry's usefulness in analyzing such samples. This project focused on increasing the sensitivity of mass spectrometry for measuring human DNA adducts. Advances in sensitivity have been achieved for two modes of mass spectrometry applied to standards related to DNA adducts: gas chromatography with electron-capture negative ion mass spectrometry, and fast-atom-bombardment mass spectrometry. These advances involve both sample preparation and instrument conditions.