A phenotypic and functional analysis of long-lived B and T lymphocytes

The lymphocyte composition of spleen, lymph nodes, bone marrow, and thymus of mice submitted to hydroxyurea treatments for four consecutive days was studied. The treatment selects for small lymphocyte populations that represent between 4 and 20% of control numbers in the various organs. Spleen and bone marrow contain the same B cell population with a low IgM, high IgD, low I-E phenotype, which respond to LPS at control clonal frequencies. The T cell compartment is equally depleted, and the lymphocytes remaining contain frequencies of clonable cells in response to mitogens and IL-2 that are comparable to those detected in normal spleen cells. Overall, the results suggest that only a minor fraction of all lymphocytes in a normal young adult mouse have life spans longer than 4 days.     

Activation and growth requirements for cytotoxic and noncytotoxic T lymphocytes

The number and nature of the "signals" required for lymphocyte activation have been so repetitively and academically discussed over the last 15 years that both the readers and the authors appear exhausted by such exercises. Yet, what may be considered to be the essential question, the basis for self-nonself discrimination, remains to be clarified. Since it has been established that clonal expansion and maturation to effector functions are brought about by polyclonally ("immunologically nonspecific") active factors, it is obvious that the crucial "step" in this context is the initial interaction of antigen with specific receptors of immunocompetent lymphocytes. This initial discriminatory event appears to proceed differently on the various cell subsets. We first deal with the mechanism of induction and growth of cytotoxic-T-lymphocyte precursors, and then discuss the inductive requirements leading to proliferation of T helper cells.     

Characterization of B lymphocytes present in the demyelinating lesions induced by Theiler's virus

Theiler's virus, a murine picornavirus, persists in the central nervous system of SJL/J mice and causes inflammation and demyelination in the white matter of spinal cord. We isolated inflammatory cells from the central nervous system of infected animals and studied their functions in vitro. Flow microfluorimetry analysis showed the presence of all major lymphocyte subsets, namely CD4+ and CD8+ T cells as well as B lymphocytes. B lymphocytes were activated in vitro and the antigenic specificity of secreted Ig was determined by immunoblotting. Secreted Ig reacted strongly with viral capsid proteins VP1 and VP2 and had neutralizing activity. They reacted also with two nonviral white matter components which were present only in infected animals. Therefore, it is likely that Igs secreted at the site of infection play a role in limiting virus spread. It is also possible that virus induced autoreactive antibodies participate in demyelination.     

Voluntary control of autonomic responses: A case for a dialogue between individual and group experimental methodologies

This paper discusses the general methodological controversy between individual and group research approaches by comparing the main characteristics of these two methods as applied to the specific context of basic research on voluntary heart rate control. A review of the literature published over the past 19 years in this area of study shows an imbalance in the frequency of utilization of these two methods that strongly favors short-term group designs. Implications of this research tendency are discussed. The relevance and the advantages of applying the individual approach to voluntary autonomic control research are outlined. This area is particularly amenable to the individual approach because the phenomena under study seem to be characterized by, among other things, a smaller intrasubject than intersubject variability. It is suggested that the present imbalanced tendency in the choice of a research method be corrected and that researchers adopt a more flexible attitude in the choice of the best method for studying each specific problem.     

Neutron-encoded Signatures Enable Product Ion Annotation From Tandem Mass Spectra

We report the use of neutron-encoded (NeuCode) stable isotope labeling of amino acids in cell culture for the purpose of C-terminal product ion annotation. Two NeuCode labeling isotopologues of lysine, ¹³C₆¹⁵N₂ and ²H₈, which differ by 36 mDa, were metabolically embedded in a sample proteome, and the resultant labeled proteins were combined, digested, and analyzed via liquid chromatography and mass spectrometry. With MS/MS scan resolving powers of ∼50,000 or higher, product ions containing the C terminus (i.e. lysine) appear as a doublet spaced by exactly 36 mDa, whereas N-terminal fragments exist as a single m/z peak. Through theory and experiment, we demonstrate that over 90% of all y-type product ions have detectable doublets. We report on an algorithm that can extract these neutron signatures with high sensitivity and specificity. In other words, of 15,503 y-type product ion peaks, the y-type ion identification algorithm correctly identified 14,552 (93.2%) based on detection of the NeuCode doublet; 6.8% were misclassified (i.e. other ion types that were assigned as y-type products). Searching NeuCode labeled yeast with PepNovo⁺ resulted in a 34% increase in correct de novo identifications relative to searching through MS/MS only. We use this tool to simplify spectra prior to database searching, to sort unmatched tandem mass spectra for spectral richness, for correlation of co-fragmented ions to their parent precursor, and for de novo sequence identification.The ability to make de novo sequence identifications directly from tandem mass spectra has long been a holy grail of the proteomic community. Such a capability would wean the field from its reliance upon sequenced genome databases. Even for organisms with fully annotated genomes, events such as single nucleotide polymorphisms, alternative splicing, gene fusion, and a host of other genomic transformations can result in altered proteomes. These alterations can vary from cell to cell and individual to individual. Thus, one could argue that the most valuable proteomic information, the individual and cellular proteome variation from the genome, remains elusive (1). This problem has received considerable attention; that said, it is not easy to de novo correlate spectrum to sequence in a large-scale, automated fashion (2–6). Improvements in mass accuracy have helped, but routine, reliable de novo sequencing without database assistance is not standard (7–10).A primary means to facilitate de novo spectral interpretation is the simple annotation of m/z peaks in tandem mass spectra as either N- or C-terminal. We and others have investigated this seemingly simple first step. Real-world spectra, however, are complex. Difficulties often arise in determining the charge state of the fragment or in differentiating between fragment ions and peaks arising from neutral loss, internal fragmentation, or spectral noise, both electronic and chemical. Several strategies have focused on product ion annotation. These approaches have included manipulation of the N-terminus basicity combined with electron transfer dissociation (ETD)¹ (11–13). This approach can yield mostly N-terminal fragments for peptides having only two charges. However, it requires both ETD and the protease LysN. Other methods have used differential labeling of N- and C-terminal peptides to shift either one or the other product ion series, by either metabolic or chemical means (14–18). Metabolic incorporation of amino acids is an efficient method of introducing distinctive labels that eliminates in vitro labeling, but this method requires that the sample be amenable to cell culture (19, 20). Additionally, it may be difficult to achieve complete labeling in complex systems. Several other approaches used to introduce heavy isotopes onto one terminus have been investigated, including trypsin digestion in ¹⁸O water (21–23), differential isotopic esterification (24, 25), derivatization of the C-terminal carboxylate by p-bromophenethylamine (8, 26), N-terminal derivatization with sulfonic acid groups (27, 28), and formaldehyde labeling via reductive amination (29–31). These chemical modifications are introduced after cell lysis, often immediately prior to analysis. Although chemical labeling strategies can be used with a variety of samples, difficulties can arise from differences in labeling efficiency between samples, and often a clean-up step is required following labeling, which may lead to sample loss. No matter the labeling method, in this regime, the two precursors must be separately isolated, fragmented, and analyzed either together or separately. The recognition and selection of the broadly spaced doublet in real time also are necessary. These requirements have limited the utility of these approaches. Our own laboratory discovered that the c- and ^●z-type product ions generated from either electron capture dissociation or ETD have distinct chemical formulae and therefore can always be distinguished based on accurate mass alone (32). The problem with this approach is that extremely high mass accuracy (<500 ppb) is required in order to distinguish these product ion types above ∼600 Da in mass. Thus, the majority of the product ions within a spectrum cannot be readily mapped to either terminus with high confidence.Despite these difficulties, we assert that robust de novo sequencing methodology would benefit greatly from a simple method that could be used to distinguish N- and C-terminal product ions with high accuracy and precision. Ideally, the approach would work regardless of the choice of proteolytic enzyme or dissociation method. Recently, we described a new technology for protein quantification called neutron encoding (NeuCode) (33). NeuCode embeds millidalton mass differences into peptides and proteins by exploiting the mass defect induced by differences in the nuclear binding energies of the various stable isotopes of common elements such as C, N, H, and O. For example, consider the amino acid lysine, which has eight additional neutrons (+8 Da). One way to synthesize this amino acid is to add six ¹³C atoms and two ¹⁵N atoms (+8.0142 Da). Another isotopologue could be constructed by adding eight ²H atoms (+8.0502). These two isotopologues differ by only 36 mDa; peptide precursors containing both of these amino acids would appear as a single, unresolved precursor m/z peak at a mass resolving power of less than ∼100,000. However, under high resolving powers (i.e. greater than ∼100,000 at m/z 400), this doublet is resolved. We first developed this NeuCode concept in the context of metabolic labeling, akin to stable isotope labeling with amino acids in cell culture (SILAC), except that instead of the precursor partners being separated by 4 to 8 Da, they are separated by only 6 to 40 mDa. For quantitative purposes, NeuCode promises to deliver ultraplexed SILAC (>12) without increasing spectral complexity.We reasoned that the isotopologues of Lys that permit NeuCode SILAC would generate a distinct fingerprint on C-terminal product ions. Specifically, peptides that have been labeled with NeuCode SILAC and digested with LysC uniformly contain Lys at the C terminus. Upon MS/MS, all C-terminal product ions should present as doublets (with duplex NeuCode), whereas N-terminal products will be detected as a single m/z peak. The very close m/z spacing of the NeuCode SILAC partners will ensure that each partner is always co-isolated and that the signatures are visible only upon high-resolving-power mass analysis. Here we investigate the combination of NeuCode SILAC and high-resolving-power MS/MS analysis to allow the straightforward identification of C-terminal product ions.

Sample Preparation

Saccharomyces cerevisiae strain BY4741 Lys1Δ was grown in defined synthetic complete (SC, Sunrise Science, San Diego, CA) drop-out media with either heavy 6C¹³/2N¹⁵ lysine (+8.0142 Da, Cambridge Isotopes, Tewksbury, MA), or heavy 8D (+8.0502 Da, Cambridge Isotopes). Cells were propagated to a minimum of 10 doublings. At mid-log phase, cells were harvested via centrifugation at 3,000 × g for 3 min and then washed three times with chilled double distilled H₂O. Cell pellets were resuspended in 5 ml lysis buffer (50 mm Tris pH 8, 8 m urea, 75 mm sodium chloride, 100 mm sodium butyrate, 1 mm sodium orthovanadate, protease and phosphatase inhibitor tablet), and protein was extracted via glass bead milling (Retsch, Haan, Germany). Protein concentration was measured via BCA (Pierce). Cysteines in the yeast lysate were reduced with 5 m dithiothreitol at ambient temperature for 30 min, alkylated with 15 mm iodoacetamide in the dark at ambient temperature for 30 min, and then quenched with 5 mm dithiothreitol. 50 mm tris (pH 8.0) was used to dilute the urea concentration to 4 m. Proteins were digested with LysC (1:50 enzyme:protein ratio) at ambient temperature for 16 h. The digestion was quenched with TFA and desalted with a tC18 Sep-Pak (Waters, Etten-Leur, The Netherlands). Samples were prepared by mixing 6C¹³/2N¹⁵ (+8.0412 Da) and 8D (+8.0502 Da) labeled peptides in 1:1 ratios by mass. For strong cation exchange fractionation, peptides were dissolved in 400 μl of strong cation exchange buffer A (5 mm KH₂PO₄ and 30% acetonitrile; pH 2.65) and injected onto a polysulfoethylaspartamide column (9.4 mm × 200 mm; PolyLC) attached to a Surveyor LC quarternary pump (Thermo Electron, West Chester, PA) operating at 3 ml/min. Peptides were detected by photodiode array detector (Thermo Electron, West Chester, PA). Fractions were collected every 2 min starting at 10 min into the following gradient: 0–2 min at 100% buffer A, 2–5 min at 0%–15% buffer B (5 mm KH₂PO₄, 30% acetonitrile, and 350 mm KCl (pH 2.65)), and 5–35 min at 15%–100% buffer B. Buffer B was held at 100% for 10 min. Finally, the column was washed with buffer C (50 mm KH₂PO₄ and 500 mm KCl (pH 7.5)) and water before recalibration. Fractions were collected by hand every 2 to 3 min starting at 10 min into the gradient and were lyophilized and desalted with a tC18 Sep-Pak (Waters).

LC-MS/MS

Samples were loaded onto a 15-cm-long, 75-μm capillary column packed with 5 μm Magic C18 (Michrom, Auburn, CA) particles in mobile phase A (0.2% formic acid in water). Peptides were eluted with mobile phase B (0.2% formic acid in acetonitrile) over a 120-min gradient at a flow rate of 300 nl/min. Eluted peptides were analyzed by an Orbitrap Elite mass spectrometer. For the nonfractionated samples, mass spectrometer instrument methods comprised one MS¹ scan followed by data-dependent MS² scans of the five most intense precursors. A survey MS¹ scan was performed by the Orbitrap at 30,000 resolving power to identify precursors to sample for tandem mass spectrometry, and this was followed by an additional MS¹ scan at 480,000 resolving power (at m/z 400; actual mass resolving power of 470,700). Data-dependent tandem mass spectrometry was performed via beam-type collisional activated dissociation (HCD) in the Orbitrap at a resolving power of 15,000, 60,000, 120,000, or 240,000 and a collision energy of 30. Preview mode was enabled, and precursors of unknown charge or with a charge of +1 were excluded from MS² sampling. For experiments comparing the duty cycle and resolving power required in order to distinguish y-ion doublets, MS¹ and MS² target ion accumulation values were set to 5 × 10⁵ and 5 × 10⁴, respectively. For all other experiments, MS¹ target accumulation values were set to 1 × 10⁶ and MS² accumulation values were set to 4 × 10⁵. Dynamic exclusion was set to 30 s for −0.55 m/z and +2.55 m/z of selected precursors. For ETD analysis, data-dependent top-five mass spectrometry was performed at a resolving power of 240,000 (m/z 400; actual MS² mass resolving power of 271,000) (34). ETD accumulation values were set to 1 × 10⁶ for MS¹ target accumulation and 4 × 10⁵ for MS² target accumulation. The fluoranthene reaction time was set to 100 ms. For the high-pH strong cation exchange fractions, data-dependent tandem mass spectrometry was performed via HCD at a resolving power of either 60,000 or 120,000 and a collision energy of 30. Preview mode was enabled, and precursors of unknown charge or with a charge of +1 were excluded from MS² sampling. MS¹ targets were set to 1 × 10⁶, and MS² accumulation values were set to 4 × 10⁵. Dynamic exclusion was set to 45 s for −0.55 m/z and +2.55 m/z of selected precursors. Analysis by use of a wide isolation window was performed on an Orbitrap Fusion. MS¹ analysis was performed at 450,000 resolving power (m/z 200), and MS² analysis was performed at 120,000 resolving power (m/z 400). Data-dependent top-N mass spectrometry was performed, with precursors selected from sequential 25-Da windows. HCD was performed twice on the same precursor, first by use of a quadrupole isolation width of 0.7 m/z for peptide identification, and then using 25 m/z quadrupole isolation. Fragment ions were analyzed in the Orbitrap at a mass resolving power of 120,000 (m/z 400). MS¹ and MS² target accumulation values were set to 2 × 10⁵ and 5 × 10⁴, respectively.

Data Analysis

Thermo.raw files were converted to searchable DTA text files using the Coon OMSSA Proteomic Analysis Software Suite (COMPASS) (35). DTA files containing exclusively y-ions were generated using an in-house algorithm. DTA files were searched against the UniProt yeast database (version 132) with Lys-C specificity using the Open Mass Spectrometry Search Algorithm (OMSSA), version 2.1.9 (36). Methionine oxidation was searched as a variable modification. Cysteine carbamidomethylation and the mass shift imparted by the lysine isotopolgues were searched as fixed modifications. For MS² scans performed at a resolving power of 60,000, 120,000, or 240,000, a shift of +8.0142, representing the mass shift of the ¹³C₆¹⁵N₂ isotopologue, was searched. For MS² scans performed at 15,000 resolving power, the average shift of the ¹³C₆¹⁵N₂ and ⁸H₂ isotopologues (+8.0322) was searched. For all analyses, the precursor mass was obtained from the 480,000 MS scan. The precursor mass tolerance was defined as 50 ppm, and the fragment ion mass tolerance was set to 0.01 Da. A histogram of precursor mass error at different search tolerances is presented in supplemental Fig. S1. Using the COMPASS software suite, obtained search results were filtered to 1% FDR based on E-values. y-ion doublets were extracted from raw files using an in-house algorithm explained in the supplemental information. Briefly, an ensemble of three different machine learning models was used to score each MS/MS spectral peak for C-terminal product ion prediction. To train our ensemble learner to correctly distinguish C-terminal product ion peaks from N-terminal product ion peaks and noise peaks within our experimental MS/MS spectra, we generated a representative training set of spectral data. Instances used for training and test sets were peaks acquired only from MS/MS spectra associated with a peptide identification. Peaks with a signal-to-noise value of less than 5 were not used. Feature information for each training/testing instance was extracted from raw spectral data. Seven MS/MS spectral features were selected to generate training and test set data: (1) “has doublet” (evaluated as “true” only if a spectral peak could be found at the predicted m/z of the peak''s “heavy” partner), (2) “signal-to-noise” (discretized using a scale of 1–5 based on the peak''s signal-to-noise value), (3) “is isotope,” (4) “is neutral loss,” (5) “number of isotopes,” (6) “number of doublet isotopes,” and (7) “has neutral loss.”To evaluate NeuCode SILAC labeling for automated de novo sequencing, PepNovo⁺ (8) was benchmarked on y-ion predicted spectra. First, a set of identified spectra from 13,832 unique peptides (>7,400 per precursor charge 2–3) was produced to train PepNovo⁺ so it could learn features such as the relative peak height ranks of b/y-ions and the probability of noise at each mass interval. These training spectra were acquired under the 11 NeuCode yeast strong cation exchange fractions prepared as described above. Thermo raw files were converted into mzXML format using ProteoWizard v2.2.2828 (with peak-picking turned on) and identified by MS-GF⁺ v9358 (37) at a 1% spectrum-level FDR against the UniProt yeast database (plus isoforms), v20110729. A fixed modification of K⁺8.0142 was imposed along with variable modifications of oxidized Met and deamidated Asn/Gln. All MS/MS scans were searched with a 50-ppm precursor mass tolerance, the high-accuracy LTQ instrument setting, the HCD fragmentation setting, and one allowed missed Lys-C cleavage.Thermo.raw files were also converted into DTA spectra as before, except the in-house algorithm for selecting y-ion doublets was slightly altered to boost the peak height of predicted y-ions above that of other peaks (the cumulative peak height was equal to the sum of the monoisotopic doublet peaks, all isotopic doublet peaks, and two times the peak height of the base peak) and to convert their m/z to charge one. Remaining peaks not predicted to be y-ions were converted to charge one by a previously described MS/MS deconvolution tool (38). Deconvoluted DTA spectra that originated from identified MS/MS scans were then paired with the MSGF⁺ peptide IDs and passed to PepNovo⁺ for training. The resulting PepNovo⁺ scoring model lacked the rank-boosting component (39), which requires identified spectra from >100,000 unique peptides per precursor charge state and extensive modification of the PepNovo⁺ source code to train. Still, the model was sufficient to perform de novo peptide sequencing on the y-ion predicted spectra. PepNovo⁺ was also run on the raw MS/MS scans (mzXML spectra converted to MGF with all MS/MS peaks converted to charge one) by use of a previously trained HCD scoring model that also lacks the rank-boosting component (40). The following PepNovo⁺ parameters were set at all stages of training and benchmarking: fixed modification of K⁺8.0142; variable modifications of oxidized Met and deamidated Asn; 0.01-Da fragment mass tolerance; use of spectrum precursor charge; and use of spectrum precursor m/z.     

Origins of bone repair in the armour of fossil fish: response to a deep wound by cells depositing dentine instead of dermal bone

The outer armour of fossil jawless fishes (Heterostraci) is, predominantly, a bone with a superficial ornament of dentine tubercles surrounded by pores leading to flask-shaped crypts (ampullae). However, despite the extensive bone present in these early dermal skeletons, damage was repaired almost exclusively with dentine. Consolidation of bone, by dentine invading and filling the vascular spaces, was previously recognized in Psammolepis and other heterostracans but was associated with ageing and dermal shield wear (reparative). Here, we describe wound repair by deposition of dentine directly onto a bony scaffold of fragmented bone. An extensive wound response occurred from massive deposition of dentine (reactionary), traced from tubercle pulp cavities and surrounding ampullae. These structures may provide the cells to make reparative and reactionary dentine, as in mammalian teeth today in response to stimuli (functional wear or damage). We suggest in Psammolepis, repair involved mobilization of these cells in response to a local stimulatory mechanism, for example, predator damage. By comparison, almost no new bone is detected in repair of the Psammolepis shield. Dentine infilling bone vascular tissue spaces of both abraded dentine and wounded bone suggests that recruitment of this process has been evolutionarily conserved over 380 Myr and precedes osteogenic skeletal repair.     

The influence of fire frequency on the structure and botanical composition of savanna ecosystems

Savannas cover 60% of the land surface in Southern Africa, with fires and herbivory playing a key role in their ecology. The Limpopo National Park (LNP) is a 10,000 km² conservation area in southern Mozambique and key to protecting savannas in the region. Fire is an important factor in LNP's landscapes, but little is known about its role in the park's ecology. In this study, we explored the interaction between fire frequency (FF), landscape type, and vegetation. To assess the FF, we analyzed ten years of the Moderate resolution Imaging Spectroradiometer (MODIS) burned area product (2003–2013). A stratified random sampling approach was used to assess biodiversity across three dominant landscapes (Nwambia Sandveld‐NS, Lebombo North‐LN, and Shrubveld Mopane on Calcrete‐C) and two FF levels (low—twice or less; and high—3 times or more, during 10 years). Six ha were sampled in each stratum, except for the LN versus high FF in which low accessibility allowed only 3 ha sampling. FF was higher in NS and LN landscapes, where 25% and 34% of the area, respectively, burned more than three times in 10 years. The landscape type was the main determinant of grass composition and biomass. However, in the sandy NS biomass was higher under high FF. The three landscapes supported three different tree/shrub communities, but FF resulted in compositional variations in NS and LN. Fire frequency had no marked influence on woody structural parameters (height, density, and phytomass). We concluded that the savannas in LNP are mainly driven by landscape type (geology), but FF may impose specific modifications. We recommend a fire laissez‐faire management system for most of the park and a long‐term monitoring system of vegetation to address vegetation changes related to fire. Fire management should be coordinated with the neighboring Kruger National Park, given its long history of fire management. Synthesis: This study revealed that grass and tree/shrub density, biomass, and composition in LNP are determined by the landscape type, but FF determines some important modifications. We conclude that at the current levels FF is not dramatically affecting the savanna ecosystem in the LNP (Figure 1). However, an increase in FF may drive key ecosystem changes in grass biomass and tree/shrub species composition, height, phytomass, and density.     

CD25+ CD4+ T cells regulate the expansion of peripheral CD4 T cells through the production of IL-10

The mechanisms by which the immune system achieves constant T cell numbers throughout life, thereby controlling autoaggressive cell expansions, are to date not completely understood. Here, we show that the CD25(+) subpopulation of naturally activated (CD45RB(low)) CD4 T cells, but not CD25(-) CD45RB(low) CD4 T cells, inhibits the accumulation of cotransferred CD45RB(high) CD4 T cells in lymphocyte-deficient mice. However, both CD25(+) and CD25(-) CD45RB(low) CD4 T cell subpopulations contain regulatory cells, since they can prevent naive CD4 T cell-induced wasting disease. In the absence of a correlation between disease and the number of recovered CD4(+) cells, we conclude that expansion control and disease prevention are largely independent processes. CD25(+) CD45RB(low) CD4 T cells from IL-10-deficient mice do not protect from disease. They accumulate to a higher cell number and cannot prevent the expansion of CD45RB(high) CD4 T cells upon transfer compared with their wild-type counterparts. Although CD25(+) CD45RB(low) CD4 T cells are capable of expanding when transferred in vivo, they reach a homeostatic equilibrium at lower cell numbers than CD25(-) CD45RB(low) or CD45RB(high) CD4 T cells. We conclude that CD25(+) CD45RB(low) CD4 T cells from nonmanipulated mice control the number of peripheral CD4 T cells through a mechanism involving the production of IL-10 by regulatory T cells.     

The Generating Function of CID,ETD, and CID/ETD Pairs of Tandem Mass Spectra: Applications to Database Search

Recent emergence of new mass spectrometry techniques (e.g. electron transfer dissociation, ETD) and improved availability of additional proteases (e.g. Lys-N) for protein digestion in high-throughput experiments raised the challenge of designing new algorithms for interpreting the resulting new types of tandem mass (MS/MS) spectra. Traditional MS/MS database search algorithms such as SEQUEST and Mascot were originally designed for collision induced dissociation (CID) of tryptic peptides and are largely based on expert knowledge about fragmentation of tryptic peptides (rather than machine learning techniques) to design CID-specific scoring functions. As a result, the performance of these algorithms is suboptimal for new mass spectrometry technologies or nontryptic peptides. We recently proposed the generating function approach (MS-GF) for CID spectra of tryptic peptides. In this study, we extend MS-GF to automatically derive scoring parameters from a set of annotated MS/MS spectra of any type (e.g. CID, ETD, etc.), and present a new database search tool MS-GFDB based on MS-GF. We show that MS-GFDB outperforms Mascot for ETD spectra or peptides digested with Lys-N. For example, in the case of ETD spectra, the number of tryptic and Lys-N peptides identified by MS-GFDB increased by a factor of 2.7 and 2.6 as compared with Mascot. Moreover, even following a decade of Mascot developments for analyzing CID spectra of tryptic peptides, MS-GFDB (that is not particularly tailored for CID spectra or tryptic peptides) resulted in 28% increase over Mascot in the number of peptide identifications. Finally, we propose a statistical framework for analyzing multiple spectra from the same precursor (e.g. CID/ETD spectral pairs) and assigning p values to peptide-spectrum-spectrum matches.Since the introduction of electron capture dissociation (ECD)1 in 1998 (1), electron-based peptide dissociation technologies have played an important role in analyzing intact proteins and post-translational modifications (2). However, until recently, this research-grade technology was available only to a small number of laboratories because it was commercially unavailable, required experience for operation, and could be implemented only with expensive FT-ICR instruments. The discovery of electron-transfer dissociation (ETD) (3) enabled an ECD-like technology to be implemented in (relatively cheap) ion-trap instruments. Nowadays, many researchers are employing the ETD technology for tandem mass spectra generation (4–9).Although the hardware technologies to generate ETD spectra are maturing rapidly, software technologies to analyze ETD spectra are still in infancy. There are two major approaches to analyzing tandem mass spectra: de novo sequencing and database search. Both approaches find the best-scoring peptide either among all possible peptides (de novo sequencing) or among all peptides in a protein database (database search). Although de novo sequencing is emerging as an alternative to database search, database search remains a more accurate (and thus preferred) method of spectral interpretation, so here we focus on the database search approach.Numerous database search engines are currently available, including SEQUEST (10), Mascot (11), OMSSA (12), X!Tandem (13), and InsPecT (14). However, most of them are inadequate for the analysis of ETD spectra because they are optimized for collision induced dissociation (CID) spectra that show different fragmentation propensities than those of ETD spectra. Additionally, the existing tandem mass spectrometry (MS/MS) tools are biased toward the analysis of tryptic peptides because trypsin is usually used for CID, and thus not suitable for the analysis of nontryptic peptides that are common for ETD. Therefore, even though some database search engines support the analysis of ETD spectra (e.g. SEQUEST, Mascot, and OMSSA), their performance remains suboptimal when it comes to analyzing ETD spectra. Recently, an ETD-specific database search tool (Z-Core) was developed; however it does not significantly improve over OMSSA (15).We present a new database search tool (MS-GFDB) that significantly outperforms existing database search engines in the analysis of ETD spectra, and performs equally well on nontryptic peptides. MS-GFDB employs the generating function approach (MS-GF) that computes rigorous p values of peptide-spectrum matches (PSMs) based on the spectrum-specific score histogram of all peptides (16).² MS-GF p values are dependent only on the PSM (and not on the database), thus can be used as an alternative scoring function for the database search.Computing p values requires a scoring model evaluating qualities of PSMs. MS-GF adopts a probabilistic scoring model (MS-Dictionary scoring model) described in Kim et al., 2009 (17), considering multiple features including product ion types, peak intensities and mass errors. To define the parameters of this scoring model, MS-GF only needs a set of training PSMs.³ This set of PSMs can be obtained in a variety of ways: for example, one can generate CID/ETD pairs and use peptides identified by CID to form PSMs for ETD. Alternatively, one can generate spectra from a purified protein (when PSMs can be inferred from the accurate parent mass alone) or use a previously developed (not necessary optimal) tool to generate training PSMs. From these training PSMs, MS-GF automatically derives scoring parameters without assuming any prior knowledge about the specifics of a particular peptide fragmentation method (e.g. ETD, CID, etc.) and/or proteolytic origin of the peptides. MS-GF was originally designed for the analysis of CID spectra, but now it has been extended to other types of spectra generated by various fragmentation techniques and/or various enzymes. We show that MS-GF can be successfully applied to novel types of spectra (e.g. ETD of Lys-N peptides (18, 19)) by simply retraining scoring parameters without any modification. Note that although the same scoring model is used for different types of spectra, the parameters derived to score different types of spectra are dissimilar.We compared the performance of MS-GFDB with Mascot on a large ETD data set and found that it generated many more peptide identifications for the same false discovery rates (FDR). For example, at 1% peptide level FDR, MS-GFDB identified 9450 unique peptides from 81,864 ETD spectra of Lys-N peptides whereas Mascot only identified 3672 unique peptides, ≈160% increase in the number of peptide identifications (a similar improvement is observed for ETD spectra of tryptic peptides).⁴ MS-GFDB also showed a significant 28% improvement in the number of identified peptides from CID spectra of tryptic peptides (16,203 peptides as compared with 12,658 peptides identified by Mascot).The ETD technology complements rather than replaces CID because both technologies have some advantages: CID for smaller peptides with small charges, ETD for larger and multiply charged peptides (20, 21). An alternative way to utilize ETD is to use it in conjunction with CID because CID and ETD generate complementary sequence information (20, 22, 23). ETD-enabled instruments often support generating both CID and ETD spectra (CID/ETD pairs) for the same peptide. Although the CID/ETD pairs promise a great improvement in peptide identification, the full potential of such pairs has not been fully realized yet. In the case of de novo sequencing, de novo sequencing tools utilizing CID/ETD pairs indeed result in more accurate de novo peptide sequencing than traditional CID-based algorithms (23, 24, 25). However, in the case of database search, the argument that the use of CID/ETD pairs improves peptide identifications remains poorly substantiated. A few tools are developed to use CID/ETD (or CID/ECD) pairs for the database search but they are limited to preprocessing/postprocessing of the spectral data before or following running a traditional database search tool (26, 27). Nielsen et al., 2005 (22) pioneered the combined use of CID and ECD for the database search. Given a CID/ECD pair, they generated a combined spectrum comprised only of complementary pairs of peaks, and searched it with Mascot.⁵ However, this approach is hard to generalize to less accurate CID/ETD pairs generated by ion-trap instruments because there is a higher chance that the identified complementary pairs of peaks are spurious. More importantly, using traditional MS/MS tools (such as Mascot) for the database search of the combined spectrum is inappropriate, because they are not optimized for analyzing such combined spectra; a better approach would be to develop a new database search tool tailored for the combined spectrum. Recently, Molina et al., 2008 (26) studied database search of CID/ETD pairs using Spectrum Mill (Agilent Technologies, Santa Clara, CA) and came to a counterintuitive conclusion that using only CID spectra identifies 12% more unique peptides than using CID/ETD pairs. We believe that it is an acknowledgment of limitations of the traditional MS/MS database search tools for the analysis of multiple spectra generated from a single peptide.In this paper, we modify the generating function approach for interpreting CID/ETD pairs and further apply it to improve the database search with CID/ETD pairs. In contrast to previous approaches, our scoring is specially designed to interpret CID/ETD pairs and can be generalized to analyzing any type of multiple spectra generated from a single peptide. When CID/ETD pairs from trypsin digests are used, MS-GFDB identified 13% and 27% more peptides compared with the case when only CID spectra and only ETD spectra are used, respectively. The difference was even more prominent when CID/ETD pairs from Lys-N digests were used, with 41% and 33% improvement over CID only and ETD only, respectively.Assigning a p value to a PSM greatly helped researchers to evaluate the quality of peptide identifications. We now turn to the problem of assigning a p value to a peptide-spectrum-spectrum match (PS²M) when two spectra in PS²M are generated by different fragmentation technologies (e.g. ETD and CID). We argue that assigning statistical significance to a PS²M (or even PSⁿM) is a prerequisite for rigorous CID/ETD analyses. To our knowledge, MS-GFDB is the first tool to generate statistically rigorous p values of PSⁿMs.The MS-GFDB executable and source code is available at the website of Center for Computational Mass Spectrometry at UCSD (http://proteomics.ucsd.edu). It takes a set of spectra (CID, ETD, or CID/ETD pairs) and a protein database as an input and outputs peptide matches. If the input is a set of CID/ETD pairs, it outputs the best scoring peptide matches and their p values (1) using only CID spectra, (2) using only ETD spectra, and (3) using combined spectra of CID/ETD pairs.