Zhixin Tian

Carbanions in the gas phase.

ion) and produce a new radical anion along with water, however, is what makes it a valuable reagent. Dihydrogen abstraction from the same carbon, adjacent sites, and remote centers all have been reported, but the inability to control or even predict the structure of the resulting radical anion, if one forms at all, is the major limitation of this approach. Given the lack of selectivity often observed in the formation of radical anions with O− •, Squires et al. developed the use of molecular fluorine to overcome this limitation. In this methodology, a bis(trimethylsilyl) derivative is reacted with fluoride anion to generate a (M − SiMe3) carbanion. Reaction of this ion with F2 undoubtedly occurs via electron transfer to afford F2 − • or F− (and a fluorine atom), which then removes the remaining trimethylsilyl group to give the desired radical anion. For example, metaand para-bis(trimethylsilyl)benzene were reacted with fluoride anion and then F2 to afford the radical anions of metaand para-benzyne as illustrated for the meta isomer (eq 101). This synthetic strategy is predictable, regioselective, and capable of producing the large ion currents needed for photoelectron spectroscopy. Its major drawback is the need to use molecular fluorine as a neutral reagent given that it is exceedingly reactive, hard to handle, and harsh on vacuum systems. Fluorine is also difficult to introduce at low flows and pressures as needed for a FTMS; consequently, a third route was developed by the Kass group that makes use of multiply charged anions. Electrospray ionization of dicarboxylic acids typically affords strong signals of the corresponding dicarboxylate dianions. Fragmentation of these species has been reported to violate the even electron rule and lead to the expulsion of carbon dioxide and an electron to afford a radical anion as illustrated for naphthalene-2,6-dicarboxylate (eq 102). Distonic ions of this sort are quite interesting and have been extensively investigated by the Blanksby group because they provide a platform for studying radical reactivity with a mass spectrometer. A second CID step, however, leads to the loss of a second molecule of carbon dioxide and affords a new radical anion where the original carboxyl groups have been replaced by an unpaired electron and an anionic site (eq 103). This synthetic strategy is also predictable and regioselective, but it requires an instrument outfitted with an ESI source and capable of carrying out two CID steps. Additional methods that are more limited in scope have been used to produce radical anions because these species can be used to obtain valuable thermodynamic data on reactive intermediates such as biradicals, carbenes, and other fleetingly stable compounds. As this area overlaps with another review in this issue, and has been the subject of several more recent reviews, only some selected data from the past 10 years are presented in Table 11. Chemical Reviews Review dx.doi.org/10.1021/cr4000896 | Chem. Rev. XXXX, XXX, XXX−XXX T Reactivity studies on radical anions are fairly limited. Zhong et al. reported that electron ionization (EI) of trifluoromethyl acetyl compounds (CF3COR, R = OH, OSiMe3, OMe, CH2COCH3, CH2COCF3, CH3, and 3,5-(CF3)2C6H3CH2) affords a series of radical anions resulting from the addition of an electron and the loss of hydrogen fluoride (eqs 104 and 105).These biradical anions react like closed-shell species and abstract acidic hydrogens, but they also afford products that were attributed to radical addition−eliminations. For example, CF2CO2 − undergoes CF2 − • transfer upon reaction with SO2 (eq 106); in a separate study, CH2CO2 − was found to transfer CH2 − • to sulfur dioxide. In a similar manner, CF2COCHAr − was found to undergo aryl and acetyl transfers (eqs 107 and 108). These reactions presumably occur via a radical addition to the benzene ring or one of the carbonyl groups. Radical abstraction of bromine atom from molecular bromine was also noted with these ions. More recently, the reactivity of chlorocarbene anion (CHCl− •) with CHnCl4−n (n = 0−3) was investigated. 161 This ion is known from its photoelectron spectrum to have a singly occupied π orbital, so carbene behavior, in principle, can be observed. Proton transfer and attack at chlorine are the major pathways in the reaction with chloroform, but a small amount (∼5%) of Cl2 • and Cl− was observed. Surprisingly, their isotope distributions were found to depend on the reactant chlorine isotopomer. That is, CH37Cl− • led to both Cl− and Cl2 − • with enhanced amounts of Cl, while CH35Cl− • led to greater incorporation of the lighter chlorine isotope. These observations indicate that chlorine is scrambled between the reactant ion and chloroform. This result was accounted for via a carbene-like insertion−elimination mechanism (eq 109). The measured isotope ratios also led Villano et al. to deduce that C−Cl insertion takes place in preference to Table 11. Thermochemical Data for Selected Biradicals, Carbenes, and a Triradical Data are from refs 153−160. Positive values are for ground-state singlets. Chemical Reviews Review dx.doi.org/10.1021/cr4000896 | Chem. Rev. XXXX, XXX, XXX−XXX U attack at the C−H bond. This contrasts with the behavior of carbenes in solution, but the presence of an additional electron increases the nucleophilicity of the carbene anion. Additional reactions of this carbene anion and related ones (i.e., CHBr− •, CF2 − •, CCl2 − •, and CBrCl− •) with CO, CO2, COS, CS2, N2O, and O2 were also investigated. 162 The Wenthold group has reported on the preparation and characterization of an interesting triradical anion (i.e., 5dehydro-m-xylylene anion). In the process, it was found to undergo three sequential oxidation steps upon reaction with molecular oxygen to afford 3,5-bis(formyl)-phenoxide anion (eq 110). An electron-promoted Cope cyclization of several 1,5-hexadiene radical anions was also described (eq 111). 4. CONCLUSIONS AND OUTLOOK Carbanions are common synthetic intermediates and play an indispensable role in synthesis. Complications in solution due to solvation, counterion effects, and aggregation have hindered our understanding of these species. As a result, the intrinsic reactivity and properties of isolated and truly free carbanions in the gas phase are of interest. Versatile synthetic strategies have been developed for the preparation of a wide variety of gaseous carbanions, and it seems at this time as if almost any target species of interest (or a simple derivative of it) can be formed. This has enabled the reactivities and thermodynamic properties of many carbanions to be experimentally determined. Moreover, because the electron can be viewed as the simplest protecting group, the reactivities and energetics of neutral species such as radicals, biradicals, carbenes, and other fleetingly stable species also can be explored. Some of the most reliable quantitative data on these species (e.g., bond dissociation energies, heats of formation, singlet−triplet gaps) have been obtained in this way. New insights often result and provide an invaluable testing ground for theoretical models and computational methods. For example, based upon the proton affinities of allyl anions derived from cyclobutene, cyclopentene, cyclohexene, and propene, it is now apparent that cyclopropenyl anion follows a trend and, consequently, is best thought of as a nonaromatic ion. Continuing instrumental improvements in mass spectrometry and the recent explosion of gas-phase spectroscopic studies on ions all but ensure that new insights into carbanions will be obtained for years to come. In this regard, the spectroscopic characterization of Grignard reagents, organolithiums, and other metal-containing carbanions of synthetic and catalytic importance is eagerly awaited. AUTHOR INFORMATION Corresponding Author *E-mail: kass@umn.edu (S.R.K.), zhixintian@dicp.ac.cn (Z.T.).

Accurate and Efficient Resolution of Overlapping Isotopic Envelopes in Protein Tandem Mass Spectra.

It has long been an analytical challenge to accurately and efficiently resolve extremely dense overlapping isotopic envelopes (OIEs) in protein tandem mass spectra to confidently identify proteins. Here, we report a computationally efficient method, called OIE_CARE, to resolve OIEs by calculating the relative deviation between the ideal and observed experimental abundance. In the OIE_CARE method, the ideal experimental abundance of a particular overlapping isotopic peak (OIP) is first calculated for all the OIEs sharing this OIP. The relative deviation (RD) of the overall observed experimental abundance of this OIP relative to the summed ideal value is then calculated. The final individual abundance of the OIP for each OIE is the individual ideal experimental abundance multiplied by 1 + RD. Initial studies were performed using higher-energy collisional dissociation tandem mass spectra on myoglobin (with direct infusion) and the intact E. coli proteome (with liquid chromatographic separation). Comprehensive data at the protein and proteome levels, high confidence and good reproducibility were achieved. The resolving method reported here can, in principle, be extended to resolve any envelope-type overlapping data for which the corresponding theoretical reference values are available.

Carbon–Hydrogen Bond Dissociation Energies: The Curious Case of Cyclopropene

The ionization energy (IE) of the 3-cyclopropenyl radical (6.00 ± 0.17 eV) was measured in the gas phase by reacting 3-cyclopropenium cation (c-C3H3(+)) with a series of reference reagents of known IEs. This result was combined in a thermodynamic cycle to obtain the heat of formation of c-C3H3(•) (118.9 ± 4.0 kcal mol(-1)) and the allylic C-H bond dissociation energy (BDE) of cyclopropene (104.4 ± 4.0 kcal mol(-1)). These experimental values are well reproduced by high level G3 and W1 computations and reveal that the BDE is similar to that for cyclopropane and the vinyl position of cyclopropene. This is unprecedented and is a reflection of the unusual nature of cyclopropene.

Are neutral loss and internal product ions useful for top-down protein identification?

Neutral loss and internal product ions have been found to be significant in both peptide and protein tandem mass spectra and they have been proposed to be included in database search and for protein identification. In addition to common canonical b/y ions in collision-based dissociation or c/z ions in electron-based dissociation, inclusion of neutral loss and internal product ions would certainly make better use of tandem mass spectra data; however, their ultimate utility for protein identification with false discovery rate control remains unclear. Here we report our proteome-level utility benchmarking of neutral loss and internal product ions with tandem mass spectra of intact E. coli proteome. Utility of internal product ions was further evaluated at the protein level using selected tandem mass spectra of individual E. coli proteins. We found that both neutral loss and internal products ions do not have direct utility for protein identification when they were used for scoring of P Score; but they do have indirect utility for provision of more canonical b/y ions when they are included in the database search and overlapping ions between different ion types are resolved. BIOLOGICAL SIGNIFICANCE Tandem mass spectrometry has evolved to be a state-of-the-art method for characterization of protein primary structures (including amino acid sequence, post-translational modifications (PTMs) as well as their site location), where full study and utilization tandem mass spectra and product ions are indispensable. This primary structure information is essential for higher order structure and eventual function study of proteins.

Top-down characterization of histone H4 proteoforms with ProteinGoggle 2.0

Top-down characterization of combinatorial and dense post-translational modifications (PTMs) on core histones has long been a big analytical challenge because of enormous putative proteoforms for identification and simultaneously enormous putative sites of each individual PTM for localization. ProteinGoggle 2.0, as implemented with the isotopic mass-to-charge ratio and envelope fingerprinting algorithm, has multiple unique strengths for top-down characterization of histone PTMs together with high-resolution tandem mass spectrometry. Here we report our database search and proteoform identification of HeLa core histone H4 using ProteinGoggle 2.0. The Theoretical database containing all putative proteoforms was created with shotgun annotation from the human H4 flat text downloaded from UniProt; information including the amino acid sequence, putative PTMs (methylation, di-methylation, tri-methylation, acetylation and phosphorylation) and amino acid variation (A77 to P) was adopted from the flat text file. A total of 426 proteoforms were confidently identified with a spectrum level false discovery rate of less than 1%, which represents the most comprehensive H4 proteoforms reported so far. Side-by-side comparison of these proteoforms with those identified by ProSightPC 2.0 was also made. By and large, ProteinGoggle 2.0 can be adopted for database search and proteoform identification of proteins with multiple combinatorial PTMs as well as amino acid variation.

Mass measurement accuracy of the Orbitrap in intact proteome analysis.

RATIONALE The mass measurement accuracy (MMA) of Orbitrap mass spectrometers is 1-5 ppm according to the manufacturers specification; yet, up to 50 ppm has been used as mass tolerance to interpret Orbitrap data in the literature. A systematic evaluation of MMA is thus necessary to find the optimal mass tolerance to be used. METHODS Reversed-phase liquid chromatography/tandem mass spectrometry (RPLC/MS/MS) analyses of the intact E. coli proteome were carried out on a Q Exactive Orbitrap mass spectrometer coupled to a Dionex UltiMate 3000 RSLCnano system. The analysis included three technical replicates each day and was repeated for six continuous days right after a mass calibration. The obtained raw datasets were searched using ProteinGoggle 2.0 under four different mass tolerances of 5, 10, 15, and 20 ppm. RESULTS With both forward and random database searches and FDR ≤1% at the spectrum level, the most protein spectrum matches and protein IDs were obtained at a mass tolerance of 15 ppm. The average mass accuracy of both precursor and product ions from three representative high, medium, and low abundance proteins as well as the common proteins identified in all the 18 replicate runs was found to be 0-4 ppm; and no significant drift of measured mass accuracy was observed within the calibration period of 1 week. CONCLUSIONS Despite the mass measurement accuracy of 1-5 ppm of the Orbitrap stated by the manufacturer, the optimal mass tolerance for protein identification was found to be 15 ppm for both the precursor and product ions. Weekly mass calibration is appropriate because no significant drift in MMA was found within the 6-day period. Copyright

Selective fragmentation of the N-glycan moiety and protein backbone of ribonuclease B on an Orbitrap Fusion Lumos Tribrid Mass Spectrometer

RATIONALE The functional study and application of an intact glycoprotein require the structural characterization of both the protein backbone and the glycan moiety; the former has been successfully demonstrated with selective fragmentation of the protein backbone in CID and ExD; whether the latter can be achieved with selective fragmentation of the glycan moiety remains to be explored. METHODS RNase B solution was electrosprayed and its intact glycoforms of GlcNAc2 Mann (n = 5-9) with the highest abundance (charge state z = 16) were isolated individually and fragmented using CID, ETD, HCD, ETciD, and EThcD on the Orbitrap Fusion Lumos Tribrid mass spectrometer; the dissociation parameters were optimized for selective fragmentation of the N-glycan moiety and protein backbone as well as high sequence coverage. The obtained spectra were interpreted using the protein and N-glycan database search engines ProteinGoggle and GlySeeker, respectively. RESULTS With exploration of different dissociation parameters for all the five methods, selective fragmentation of the N-glycan moiety (the protein backbone staying intact) was observed in both HCD and EThcD at low collisional energies, but only a few matched product ions were observed; more comprehensive fragmentation was observed at high collisional energies (the protein backbone lost). Selective protein backbone fragmentation was observed in all the five dissociation methods. CONCLUSIONS For comprehensive structural characterization of intact N-glycoproteins using tandem mass spectrometry, the composition and topology of the N-glycan moiety can be identified using HCD and EThcD complementarily at low and high energies; while the amino acid sequence and glycosite can be identified using CID, ETD, HCD, ETciD, and EThcD with their optimal dissociation parameters.

Top-down characterization of chicken core histones

Core histones and their PTMs play important roles in regulating gene transcription and other DNA-related processes. The study of core histones PTMs, their cross-talk and functional roles is not only of broad biological significance but also of wide pathological and clinical relevance. Having the strength of comprehensive proteoform identification with 100% amino acid sequence coverage and combinatorial PTMs, top-down proteomics has become the state-of-the-art analytical tool for combinatorial PTM characterization of core histones. In this study, we report our top-down characterization of chicken (Gallus gallus domesticus) core histones, which have been widely used as models for chromosome re-construction among others because of easy availability and not-so-dense PTMs. With nanoRPLC-MS/MS analysis and ProteinGoggle database search, a total of 58 proteoforms were identified for the core histone families of H4, H2B, H2A, and H3.