|Home | About | Journals | Submit | Contact Us | Français|
Alpha-cyano-4-hydroxycinnamic acid (α-CHCA) as a matrix facilitates the ionization of proteins and peptides in a matrix-assisted laser desorption/ionization (MALDI) time-of-flight (TOF) mass spectrometer. The matrix itself also ionizes and so do its sodium and potassium adducts. Matrix clusters and metal ion adducts interfere with peptide ionization and peptide mass spectrum interpretation. These matrix adducts are significantly reduced with addition of ammonium monobasic phosphate or ammonium dibasic citrate to the matrix and sample deposited onto the MALDI target. The reduction of matrix adducts results in the increase of peptide intensity and signal-to-noise ratio as well as in improvement of peptide ionization for samples deposited onto the target at levels of 10 fmol or below. These improvements were particularly significant in the detection of peptides at amol levels when reduced amounts of matrix were also used.
Alpha-cyano-4-hydroxycinnamic acid (α-CHCA) has been widely used as matrix to facilitate the ionization of proteins and peptides in matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS).1,2 However, sodium and potassium ions induce formation of α-CHCA adducts. These elements are present in solvents and buffers, and are extracted from many plastics that are used for sample preparation and storage. Matrix adducts are ubiquitous in MALDI-TOF mass spectra, and are particularly evident at low sample concentrations. Thus, detection of low abundance, low mass analytes often becomes problematic. To remedy this situation matrix additives3 can be used to scavenge or exchange metal ions. In this work, reduced matrix adducts were observed in mass spectra by addition of ammonium monobasic phosphate (NH4H2PO4) or ammonium dibasic citrate [(NH4)2C6H6O7] to the matrix/sample.* Another observed benefit from the addition of ammonium monobasic phosphate to samples was an increase in the intensity and the signal-to-noise ratio of peptide peaks in MALDI-TOF mass spectra.
Bovine serum albumin (BSA) and Escherichia coli (E.coli) β-galactosidase (both from Sigma Chemical Company, Milwaukee, WI) were digested with bovine trypsin (Sigma). These samples were diluted with 50:50 water–acetonitrile with 0.1% trifluoroacetic acid (TFA) to a series of concentrations from 0.2 fmol/mL to 1 pmol/mL. Solutions of various concentrations of ammonium monobasic phosphate and ammonium dibasic citrate (Sigma) were prepared by dissolving each salt in deionized water. A 5-mg/mL α-CHCA MALDI matrix was prepared by dissolving recrystallized α-CHCA (Sigma) in 50:50 water–acetonitrile with 0.1% TFA which contained 0–50 mM ammonium phosphate or 0–50 mM ammonium citrate. The MALDI samples were prepared by mixing one portion of protein digest sample with nine portions of α-CHCA solution before spotting on the MALDI plate. A 2-mg/mL α-CHCA matrix was prepared in the same way as the 5-mg/mL α-CHCA matrix, which contained 10 mM ammonium phosphate.
MS and tandem mass spectrometry (MS/MS) data were acquired with the Applied Biosystems 4700 Proteomics Analyzer with TOF/TOF optics.5 For MS/MS spectra, the collision energy was 1 keV and the collision gas was air. The interpretation of both the MS and MS/MS data was carried out using the GPS Explorer software (Applied Biosystems, Framingham, MA).
The α-CHCA matrix adducts are typically most abundant in the range of m/z 800–1100 of a MALDI mass spectrum, which is part of the mass range used for protein identification by both peptide mass fingerprinting (PMF) and MS/MS peptide sequencing. These MALDI matrix adduct signals are generally not dominant in relatively concentrated samples (e.g., 1 pmol deposited onto the MALDI target). However, they become quite dominant at analyte concentrations in the low femtomole levels or below. This results in complicated mass spectra and in difficulty with data interpretation, especially when the latter must be carried out in automated mode. An example of this effect was in the MALDI mass spectrum (Fig. 1A1A)) of 500 amol E. coli β-galactosidase digest prepared in the standard manner by mixing with α-CHCA. The matrix adduct ions were reduced substantially with the addition of ammonium phosphate to the matrix, as shown in Figure 1B1B.. In the latter spectrum several β-galactosidase peptides were observed that had not been detected in the former spectrum, which suggests that the addition of ammonium phosphate facilitated peptide ionization. These mass spectra were interpreted with GPS Explorer software using the MASCOT search engine for protein identification by PMF. The MASCOT score from the sample with ammonium phosphate was higher than that from the sample without ammonium phosphate (Fig. 22),), thus increasing the confidence in the correct identification of the proteins.
The matrix adducts in the range of m/z 800–1100 are apparently formed from α-CHCA tetramers or pentamers, sodium and potassium salts, and hydrates, the likely compositions of which and corresponding m/z values are listed in Table 11.* One possible explanation for the observed reduction of matrix adducts by ammonium phosphate is that these adducts are dissociated upon addition of this salt to the matrix–sample solution. This is illustrated by comparing the signal-to-noise ratios for the ions of two matrix adducts and two β-galactosidase tryptic peptides as a function of the ammonium phosphate concentration, as shown in Figure 3A3A.. Ammonium dibasic citrate has a similar effect to that of ammonium monobasic phosphate, though ammonium phosphate can be used over a broader concentration range (1–50 mM) than ammonium citrate (0.5–10 mM) for the reduction of matrix adducts. However, the matrix adducts were not completely removed at ammonium citrate concentrations of 2 mM or lower, whereas peptide signals were dramatically decreased with the addition of ammonium citrate at concentrations of 5 mM or higher. The decrease of peptide signals, paradoxically, may be due to the lower solubility of ammonium monobasic phosphate in water compared with ammonium dibasic citrate. Thus, ammonium monobasic phosphate crystallizes and precipitates with the matrix–sample solution, but at concentrations of 20 mM or higher, a significant portion of ammonium dibasic citrate precipitates on top of matrix–sample crystals as the solvent evaporates. This supposition was confirmed by visual inspection of the samples on MALDI plates. Overall, ammonium citrate addition causes reduced peptide intensities and lower signal-to-noise ratios at concentrations of 5 mM or higher (Fig. 3B3B).). By contrast, peptide signal-to-noise ratios increased 40–70% along with absolute peptide ion intensities with addition of 1–20 mM ammonium monobasic phosphate (Fig. 3B3B).). Even though the peptide signal-to-noise ratio increased with addition of ammonium citrate at 0.5–2 mM concentration, the matrix adducts were not reduced dramatically.
Matrix adduct-related fragment ions are observed in a peptide MS/MS spectrum (Fig. 4A4A)) if the matrix adduct ion m/z is close to the peptide ion m/z (Fig. 4B4B),), which complicates the spectrum interpretation. However, this MS/MS spectrum interference is reduced with the addition of ammonium phosphate to the matrix–sample mixture (Fig. 5A5A),), because the matrix adduct ion abundance is significantly reduced in the MS spectrum (Fig. 5B5B).). This, in turn, improves the confidence in the sequence derived from the peptide MS/MS spectrum, whether the spectrum is interpreted de novo or used for database searching.
For very low concentration sample (e.g., 200 amol deposited on the MALDI target), along with ammonium phosphate addition, reduced matrix concentration further improves peptide ionization and peptide fragmentation. Three more peptides from a 200 amol BSA trypsin digest were observed with 2 mg/mL α-CHCA (Fig. 6B6B)) than with 5 mg/mL α-CHCA (Fig. 6A6A);); ammonium phosphate had been added to both matrix preparations. By database searching using Mascot,6 BSA was identified as the top protein hit for both matrix concentrations. However, the score returned by Mascot was below the significant threshold for the sample analyzed in 5 mg/mL matrix (43 vs 62), whereas for the sample analyzed in 2 mg/mL matrix the Mascot score was well above the significant threshold (73 vs 62), and so the confidence in the correct protein identification obtained from the latter results is significantly greater. Also, more complete peptide sequence information, due to the presence of more abundant and numerous MS/MS fragment ions from these peptides, was obtained from the MS/MS spectra than from the sample analyzed in 2 mg/mL matrix, allowing for positive confirmation of the protein identification results (Fig. 77).
Addition of ammonium monobasic phosphate to protein digests deposited on MALDI targets can be complemented by the use of this salt as a modifier in reversed-phase high performance liquid chromatography (HPLC) mobile phases during fractionation of protein digests. Ammonium phosphate can facilitate the separation of peptides eluted from a reversed phase HPLC column.7 With this approach, HPLC fractions, mixed with matrix and deposited directly onto MALDI plates for mass spectrometric analysis, may benefit from the better separation of peptides and the effect of ammonium phosphate in reducing or eliminating interfering matrix clusters with sodium and potassium ions which are often present in samples separated by HPLC.
Both ammonium monobasic phosphate and ammonium dibasic citrate reduce formation of α-CHCA adducts and increase the peptide signal-to-noise ratio when added to the matrix–sample mixture. Ammonium monobasic phosphate can be used in a wider concentration range than ammonium citrate, benefitting both, and it is our preferred matrix additive. These effects are enhanced by the use of α-CHCA matrix at a reduced concentration for peptide samples deposited onto the MALDI target at the amol level, which further improves sensitivity and, apparently, also enhances the fragmentation of peptides.
*Ammonium dibasic citrate has been used as an additive for the analysis of phosphopeptides by MALDI-TOF-MS.4
*It is worth noting that the fractional mass of the matrix adducts is generally smaller than that of a typical isobaric peptide. This mass deficiency is due to the high oxygen content of these adducts.