Top-down mass spectrometry and assigning internal fragments for determining disulfide bond positions in proteins.

Disulfide bonds in proteins have a substantial impact on protein structure, stability, and biological activity. Localizing disulfide bonds is critical for understanding protein folding and higher-order structure. Conventional top-down mass spectrometry (TD-MS), where only terminal fragments are assigned for disulfide-intact proteins, can access disulfide information, but suffers from low fragmentation efficiency, thereby limiting sequence coverage. Here, we show that assigning internal fragments generated from TD-MS enhances the sequence coverage of disulfide-intact proteins by 20-60% by returning information from the interior of the protein sequence, which cannot be obtained by terminal fragments alone. The inclusion of internal fragments can extend the sequence information of disulfide-intact proteins to near complete sequence coverage. Importantly, the enhanced sequence information that arise from the assignment of internal fragments can be used to determine the relative position of disulfide bonds and the exact disulfide connectivity between cysteines. The data presented here demonstrates the benefits of incorporating internal fragment analysis into the TD-MS workflow for analyzing disulfide-intact proteins, which would be valuable for characterizing biotherapeutic proteins such as monoclonal antibodies and antibody-drug conjugates.

A Multiplatform Approach for the Discovery of Novel Drug-Induced Kidney Injury Biomarkers.

Drug-induced kidney injury (DIKI) is a common toxicity observed in pharmaceutical development. We demonstrated the use of label-free liquid chromatography-mass spectrometry (LC-MS) and multiplex liquid chromatography-single reaction monitoring (LC-SRM) as practical extensions of standard immunoassay based safety biomarker assessments for identification of new toxicity marker candidates and for improved mechanistic understanding. Two different anticancer drugs, doxorubicin (DOX) and cisplatin (cis-diamminedichloridoplatinum, CDDP), were chosen as the toxicants due to their different modes of nephrotoxicity. Analyses of urine samples from toxicant treated and untreated rats were compared to identify biochemical analytes that changed in response to toxicant exposure. A discovery (label-free LC-MS) and targeted proteomics (multiplex LC-SRM) approach was used in combination with well established immunoassay experiments for the identification of a panel of urinary protein markers related to drug induced nephrotoxicity in rats. The initial generation of an expanded set of markers was accomplished using the label-free LC-MS discovery screen and ELISA based analysis of six nephrotoxicity biomarker proteins. Diagnostic performance of the expanded analyte set was statistically compared to conventional nephrotoxicity biomarkers. False discovery rate (FDR) analysis revealed 18 and 28 proteins from the CDDP and DOX groups, respectively, exhibiting significant differences between the vehicle and treated groups. Multiplex SRM assays were constructed to more precisely quantify candidate markers selected from the discovery screen and immunoassay experiments. To evaluate the sensitivity and specificity for each of the candidate biomarkers, histopathology severity scores were used as a benchmark for renal injury followed by receiver-operating characteristic (ROC) curve analysis on selected biomarkers. Further examination of the best performing analytes revealed relevant biological significance after consideration of anatomical localization and functional roles. In summary, the inclusion of mass spectrometry together with conventional ELISA based assays resulted in the identification of an expanded set of biomarkers with a realistic potential for providing additional beneficial information in mechanistic investigations of drug induced kidney injury and with similar responsiveness to conventionally applied indicators of renal injury.

Proteomics Coupled with Metabolite and Cell Wall Profiling Reveal Metabolic Processes of a Developing Rice Stem Internode.

Internodes of grass stems function in mechanical support, transport, and, in some species, are a major sink organ for carbon in the form of cell wall polymers. This study reports cell wall composition, proteomic, and metabolite analyses of the rice elongating internode. Cellulose, lignin, and xylose increase as a percentage of cell wall material along eight segments of the second rice internode (internode II) at booting stage, from the younger to the older internode segments, indicating active cell wall synthesis. Liquid-chromatography tandem mass spectrometry (LC-MS/MS) of trypsin-digested proteins from this internode at booting reveals 2,547 proteins with at least two unique peptides in two biological replicates. The dataset includes many glycosyltransferases, acyltransferases, glycosyl hydrolases, cell wall-localized proteins, and protein kinases that have or may have functions in cell wall biosynthesis or remodeling. Phospho-enrichment of internode II peptides identified 21 unique phosphopeptides belonging to 20 phosphoproteins including a leucine rich repeat-III family receptor like kinase. GO over-representation and KEGG pathway analyses highlight the abundances of proteins involved in biosynthetic processes, especially the synthesis of secondary metabolites such as phenylpropanoids and flavonoids. LC-MS/MS of hot methanol-extracted secondary metabolites from internode II at four stages (booting/elongation, early mature, mature, and post mature) indicates that internode secondary metabolites are distinct from those of roots and leaves, and differ across stem maturation. This work fills a void of in-depth proteomics and metabolomics data for grass stems, specifically for rice, and provides baseline knowledge for more detailed studies of cell wall synthesis and other biological processes characteristic of internode development, toward improving grass agronomic properties.

Multi-mode acquisition (MMA): An MS/MS acquisition strategy for maximizing selectivity, specificity and sensitivity of DIA product ion spectra.

In proteomics studies, it is generally accepted that depth of coverage and dynamic range is limited in data-directed acquisitions. The serial nature of the method limits both sensitivity and the number of precursor ions that can be sampled. To that end, a number of data-independent acquisition (DIA) strategies have been introduced with these methods, for the most part, immune to the sampling issue; nevertheless, some do have other limitations with respect to sensitivity. The major limitation with DIA approaches is interference, i.e., MS/MS spectra are highly chimeric and often incapable of being identified using conventional database search engines. Utilizing each available dimension of separation prior to ion detection, we present a new multi-mode acquisition (MMA) strategy multiplexing both narrowband and wideband DIA acquisitions in a single analytical workflow. The iterative nature of the MMA workflow limits the adverse effects of interference with minimal loss in sensitivity. Qualitative identification can be performed by selected ion chromatograms or conventional database search strategies.

Rapid label-free profiling of oral cancer biomarker proteins using nano-UPLC-Q-TOF ion mobility mass spectrometry.

PURPOSE: It has become quite clear that single cancer biomarkers cannot in general provide high sensitivity and specificity for reliable clinical cancer diagnostics. This paper explores the feasibility of rapid detection of multiple biomarker proteins in model oral cancer samples using label-free protein relative quantitation. EXPERIMENTAL DESIGN: MS-based label-free quantitative proteomics offer a rapid alternative that bypasses the need for stable isotope containing compounds to chemically bind and label proteins. Total protein content in oral cancer cell culture conditioned media was precipitated, subjected to proteolytic digestion, and then analyzed using a nano-UPLC (where UPLC is ultra-performance liquid chromatography) coupled to a hybrid Q-Tof ion-mobility mass spectrometry (MS). RESULTS: Rapid, simultaneous identification and quantification of multiple possible cancer biomarker proteins was achieved. In a comparative study between cancer and noncancer samples, approximately 952 proteins were identified using a high-throughput 1D ion mobility assisted data independent acquisition (IM-DIA) approach. As we previously demonstrated that interleukin-8 (IL-8) and vascular endothelial growth factor A (VEGF-A) were readily detected in oral cancer cell conditioned media(1), we targeted these biomarker proteins to validate our approach. Target biomarker protein IL-8 was found between 3.5 and 8.8 fmol, while VEGF-A was found at 1.45 fmol in the cancer cell media. CONCLUSIONS AND CLINICAL RELEVANCE: Overall, our data suggest that the nano-UPLC-IM-DIA bioassay is a feasible approach to identify and quantify proteins in complex samples without the need for stable isotope labeling. These results have significant implications for rapid tumor diagnostics and prognostics by monitoring proteins such as IL-8 and VEGF-A implicated in cancer development and progression. The analysis in tissue or plasma is not possible at this time, but the subsequent work would be needed for validation.

Enhanced Detection of Low-Abundance Host Cell Protein Impurities in High-Purity Monoclonal Antibodies Down to 1 ppm Using Ion Mobility Mass Spectrometry Coupled with Multidimensional Liquid Chromatography.

The enormous dynamic range of proteinaceous species present in protein biotherapeutics poses a significant challenge for current mass spectrometry (MS)-based methods to detect low-abundance HCP impurities. Previously, an HCP assay based on two-dimensional chromatographic separation (high pH/low pH) coupled to high-resolution quadrupole time-of-flight (QTOF) mass spectrometry and developed in the author's laboratory has been shown to achieve a detection limit of about 50 ppm (parts per milion) for the identification and quantification of HCPs present in monoclonal antibodies following Protein A purification.1 To improve the HCP detection limit we have explored the utility of several new analytical techniques for HCP analysis and thereby developed an improved liquid chromatography-mass spectrometry (LC-MS) methodology for enhanced detection of HCPs. The new method includes (1) the use of a new charge-surface-modified (CSH) C18 stationary phase to mitigate the challenges of column saturation, peak tailing, and distortion that are commonly observed in the HCP analysis; (2) the incorporation of traveling-wave ion mobility (TWIM) separation of coeluting peptide precursors, and (3) the improvement of fragmentation efficiency of low-abundance HCP peptides by correlating the collision energy used for precursor fragmentation with their mobility drift time. As a result of these improvements, the detection limit of the new methodology was greatly improved, and HCPs present at a concentration as low as 1 ppm (1 ng HCP/mg mAb) were successfully identified and quantified. The newly developed method was applied to analyze two high-purity mAbs (NIST mAb and Infliximab) expressed in a murine cell line. For both samples, low-abundance HCPs (down to 1 ppm) were confidently identified, and the identities of the HCPs were further confirmed by targeted MS/MS experiments. In addition, the performance of the assay was evaluated by an interlaboratory study in which three independent laboratories performed the same HCP assay on the mAb sample. The reproducibility of this assay is also discussed.

Negative ion fragmentation of cysteic acid containing peptides: cysteic acid as a fixed negative charge.

We present here a study of the collision induced dissociation (CID) of deprotonated cysteic acid containing peptides produced by MALDI. The effect of cysteic acid (C(ox)) position is interrogated by considering the positional isomers, C(ox)LVINVLSQG, LVINVLSQGC(ox), and LVINVC(ox)LSQG. Although considerable variation between the CID spectra is observed, the mechanistic picture that emerges involves charge retention at the deprotonated cysteic acid side chain. Fragmentation occurs in the proximity of the cysteic acid group by charge directed mechanisms as well as remote from this group to form ions, which may be rationalized by charge remote mechanisms. Additionally, the formation of the SO(3)(-•) ion is observed in all cases. Fragmentation of C(ox)LVINVLSQC(ox) provides both N- and C-terminal, y and b ions, respectively indicating that the negative charge may be retained at either of the cysteic acids; however, there is some evidence that charge retention at the C-terminal cysteic acid may be preferred. Fragmentation of tryptic type peptides containing a C-terminal arginine or lysine residue is considered through comparison of three peptides C(ox)LVINKLSQG, C(ox)LVINVLSQK, and C(ox)LVINVLSQR. Lastly, we rationalize the formation of b(n-1)+ H(2)O and a(n-1) ions through a mechanism involving rearrangement of the C-terminal residue to form a mixed anhydride intermediate.

Studies of histidine as a suitable isoelectric buffer for tryptic digestion and isoelectric trapping fractionation followed by capillary electrophoresis-mass spectrometry for proteomic analysis.

The use of histidine as a protein digestion buffer followed by isoelectric trapping separations using "membrane separated wells for isoelectric focusing and trapping" (MSWIFT) and mass spectrometry (MS) analysis is described. Tryptic digestion of bovine serum albumin (BSA) performed in histidine buffered solutions yields similar amino acid sequence coverage values to those obtained using ammonium bicarbonate buffer. Time course studies suggest that histidine buffers provide faster migration of peptides from the loading compartment compared to digestions prepared in ammonium bicarbonate due to differences in conductivities of the two buffers. In addition, this sample preparation method and MSWIFT separations have been coupled with capillary electrophoresis (CE) and matrix-assisted laser desorption ionization-mass spectrometry (MALDI-MS) as an alternative separation approach for proteomic studies. Tryptic peptides of ribosomal proteins in histidine are fractionated using MSWIFT followed by CE-MALDI-MS, which further illustrates the ability to couple fractions from a pI based separation device to CE-MS. Specifically, two-dimensional CE-MS plots provide a direct correlation between the numbers of basic residues within the peptide sequence displayed in charge-state trend lines. Combining MSWIFT and CE-MS provides added information regarding peptide sequence, specifically pI and in-solution charge state. Post-translational modifications can also be identified using this method.

Effect of cysteic acid position on the negative ion fragmentation of proteolytic derived peptides.

A study on the effect of cysteic acid position on the types of fragment ions formed by collision-induced dissociation (CID) of [M - H](-) ions is presented. Of particular note is the observation of d-type fragment ions for peptides that contain an N-terminal cysteic acid (fixed negative charge) and cleavable amino acid side chains possessing a ß-γ carbon-carbon bond. For example, the CID mass spectrum of oxidized cys-kemptide (C(ox)LRRASLG) [M - H + O(3)](-) ions contains abundant series of d-type fragment ions, and similar results are observed for oxidized cysteine-containing ribonuclease A proteolytic peptides. The d(i) fragment ions are assumed to arise by a charge-remote and/or charge-assisted fragmentation mechanism, which both occur at high collision energies and involve consecutive reactions (i.e., the formation of a(i) ions followed by the elimination of the side chain to form d(i) ions).

High-throughput method for on-target performic acid oxidation of MALDI-deposited samples.

An information-rich on-target performic acid oxidation method, which is compatible with alkylation for differentiation of free cysteine versus disulfide-containing peptides, is described. On-target oxidation is achieved using performic acid vapor to oxidize disulfide-containing peptides and/or small proteins on the matrix-assisted laser desorption/ionization (MALDI) sample deposits. The on-target oxidation method is preferred over solution-phase oxidation methods because (1) less sample handling is required, (2) oxidation throughput is drastically increased and (3) ion suppression effects are reduced because performic acid is not added directly to the MALDI spot. The utility of this method is demonstrated by simultaneous oxidation of multiple MALDI sample deposits containing model disulfide-linked peptides, intact bovine insulin and a bovine ribonuclease A proteolytic digest.

Hepatic phenotype of liver fatty acid binding protein gene-ablated mice.

Although the function of liver fatty acid binding protein in hepatic fatty acid metabolism has been extensively studied, its potential role in hepatic cholesterol homeostasis is less clear. Although hepatic cholesterol accumulation was initially reported in L-FABP-null female mice, that study was performed with early N2 backcross generation mice. To resolve whether the hepatic cholesterol phenotype in these L-FABP(-/-) mice was attributable to genetic inhomogeneity, these L-FABP(-/-) mice were further backcrossed to C57Bl/6 mice up to the N10 (99.9% homogeneity) generation. Hepatic total cholesterol accumulation was observed in female, but not male, L-FABP(-/-) mice at all (N2, N4, N6, N10) backcross generations examined. The greater total cholesterol was due to increased hepatic levels of both unesterified (free) cholesterol and esterified cholesterol. Altered hepatic cholesterol accumulation correlated directly with L-FABP's ability to bind cholesterol with high affinity as shown by direct L-FABP binding of fluorescent cholesterol analogs (NBD-cholesterol, dansyl-cholesterol), a photoactivatable cholesterol analog [free cholesterol benzophenone (FCBP)], and free cholesterol (circular dichroism, isothermal titration microcalorimetry). One mole of fluorescent sterol was bound per mole of L-FABP. This was confirmed by photo-cross-linking studies with the photoactivatable cholesterol analog FCBP and by isothermal titration calorimetry with free cholesterol, which showed that L-FABP bound only one sterol molecule per L-FABP molecule. In contrast, the hepatic phenotype of male, but not female, L-FABP(-/-) mice was characterized by decreased hepatic triacylglycerol levels at all backcross generations examined. Taken together, these data support the hypothesis that L-FABP plays a role in physiological regulation of not only hepatic fatty acid metabolism, but also that of hepatic cholesterol.

Structure and function of the virulence-associated high-temperature requirement A of Mycobacterium tuberculosis.

The high-temperature requirement A (HtrA) family of serine proteases has been shown to play an important role in the environmental and cellular stress damage control system in Escherichia coli. Mycobacterium tuberculosis ( Mtb) has three putative HtrA-like proteases, HtrA1, HtrA2, and HtrA3. The deletion of htrA2 gives attenuated virulence in a mouse model of TB. Biochemical analysis reveals that HtrA2 can function both as a protease and as a chaperone. The three-dimensional structure of HtrA2 determined at 2.0 A resolution shows that the protease domains form the central core of the trimer and the PDZ domains extend to the periphery. Unlike E. coli DegS and DegP, the protease is naturally active due to the formation of the serine protease-like catalytic triad and its uniquely designed oxyanion hole. Both protease and PDZ binding pockets of each HtrA2 molecule are occupied by autoproteolytic peptide products and reveal clues for a novel autoregulatory mechanism that might have significant importance in HtrA-associated virulence of Mtb.

Structure and function of the sterol carrier protein-2 N-terminal presequence.

Although sterol carrier protein-2 (SCP-2) is encoded as a precursor protein (proSCP-2), little is known regarding the structure and function of the 20-amino acid N-terminal presequence. As shown herein, the presequence contains significant secondary structure and alters SCP-2: (i) secondary structure (CD), (ii) tertiary structure (aqueous exposure of Trp shown by UV absorbance, fluorescence, and fluorescence quenching), (iii) ligand binding site [Trp response to ligands, peptide cross-linked by photoactivatable free cholesterol (FCBP)], (iv) selectivity for interaction with anionic phospholipid-rich membranes, (v) interaction with a peroxisomal import protein [FRET studies of Pex5p(C) binding], the N-terminal presequence increased SCP-2's affinity for Pex5p(C) by 10-fold, and (vi) intracellular targeting in living and fixed cells (confocal microscopy). Nearly 5-fold more SCP-2 than proSCP-2 colocalized with plasma membrane lipid rafts and caveolae (AF488-CTB); 2.8-fold more SCP-2 than proSCP-2 colocalized with a mitochondrial marker (Mitotracker), but nearly 2-fold less SCP-2 than proSCP-2 colocalized with peroxisomes (AF488 antibody to PMP70). These data indicate the importance of the N-terminal presequence in regulating SCP-2 structure, cholesterol localization within the ligand binding site, membrane association, and, potentially, intracellular targeting.

Utility of CE-MS data in protein identification.

A new method for displaying CE-MALDI-MS data for proteolytic digests is described. This data display mode yields distinct charge-based trends for plots of mass-to-charge (m/z) versus CE migration time. These trends arise owing to the in-solution charge state of the peptides, and this interpretation was confirmed by using empirical peptide electromigration models and peptide standards as charge-state markers. These charge-state specific trends exhibit analytical utility by providing additional chemical information about the peptides, which increases the confidence level of peptide identification and provides a rapid means for screening for posttranslationally modified peptides.

Amino acid profiling in plant cell cultures: an inter-laboratory comparison of CE-MS and GC-MS.

A CE-MS method for metabolic profiling of amino acids was developed and used in an integrated functional genomics project to study the response of Medicago truncatula liquid suspension cell cultures to stress. This project required the analysis of more than 500 root cell culture extracts. The CE-MS method profiled 20 biologically important amino acids. The CE-MS method required no sample derivatization prior to injection and used minimal sample preparation. The method is described in terms of CE and MS operational parameters, reproducibility of migration times and response ratios, sample preparation, sample throughput, and reliability. This method was then compared with a previously published report that used GC-MS metabolic profiling for the same tissues. The data reveal a high level of similarity between the CE-MS and GC-MS amino acid profiling methods, thus supporting these as complementary technologies for metabolomics. We conclude that CE-MS is a valid alternative to GC-MS for targeted profiling of metabolites, such as amino acids, and possesses some significant advantages over GC-MS.