|Home | About | Journals | Submit | Contact Us | Français|
Advancement in proteomics research relies on the development of new, innovative tools for identifying and characterizing proteins. Here, we describe a protocol for analyzing peptides and proteins on a chromatographic timescale by coupling nanoflow reverse-phase (RP) liquid chromatography (LC) to electron-transfer dissociation (ETD) mass spectrometry. For this protocol, proteins can be proteolytically digested before ETD analysis, although digestion is not necessary for all applications. Proteins ≤30 kDa can be analyzed intact, particularly if the objective is protein identification. Peptides or proteins are loaded onto a RP column and are gradient-eluted into an ETD-enabled mass spectrometer. ETD tandem mass spectrometry (MS/MS) provides extensive sequence information required for the unambiguous identification of peptides and proteins and for characterization of posttranslational modifications. ETD is a powerful MS/MS technique and does not compromise the sensitivity and speed necessary for online chromatographic separations. The described procedure for sample preparation, column packing, sample loading and ETD analysis can be implemented in 5–15 h.
A comprehensive and sensitive analysis of peptides using liquid chromatography (LC) online with tandem mass spectrometry (MS/ MS) has, until recently, mainly utilized collision-activated dissociation (CAD) for peptide fragmentation. In a traditional LC-MS/MS analysis of peptides, CAD is employed to dissociate small peptides derived from proteins by proteases, such as trypsin1,2. The peptide mixture is commonly separated by reverse-phase (RP) LC and introduced online into a mass spectrometer via electrospray ionization (ESI)3. Following mass analysis of the fragment ions resulting from peptide dissociation, the experimental data are usually searched in silico against theoretical peptides created from predicted protein sequences using database-searching algorithms4,5. Peptide identifications are then correlated to protein sequences in the database to identify the source proteins.
Although CAD has been widely adopted for peptide characterization, it has several shortcomings. These include the following: (i) CAD often promotes the loss of labile posttranslational modifications (PTMs) (e.g., phosphorylation), making PTM site-mapping difficult; (ii) CAD fails to generate random cleavage along the backbone of peptides that contain multiple basic residues and (iii) CAD provides limited sequence information for large (>30 amino acids), highly charged peptides and intact proteins6.
Electron-transfer dissociation (ETD) is a powerful fragmentation technique that overcomes these limitations. ETD was first introduced by Syka et al. in 2004 (ref. 6) and is the ion/ion analogue of electron-capture dissociation (ECD)7. ECD was first demonstrated on a Fourier transform ion cyclotron resonance mass spectrometer. ETD is performed on relatively inexpensive quadrupole ion trap mass spectrometers, requires short ion/ion reaction times (tens of milliseconds, short duty cycles) and promotes extensive fragmentation of the protein backbone within the time frame of a single spectrum acquisition6,8. Because ETD is highly efficient and takes place on a millisecond timescale, it is compatible with online chromatographic separations and can be used to analyze samples at the low femtomole level6,9.
ETD results when multiply charged peptide cations in the gas phase are allowed to react with radical anions of a polyaromatic hydrocarbon such as fluoranthene6,8 (Fig. 1). During the reaction, the radical anion transfers an electron to the multiply charged peptide cation.
This process is exothermic and triggers highly selective fragmentation of N-Cα bonds along the peptide/protein amide backbone, and fragment ions of type c′ and z′• result6 (Fig. 1). ETD is likely a nonergodic process, not a threshold energy dissociation process such as CAD. Peptides and proteins undergo highly efficient fragmentation, and labile PTMs remain intact upon peptide/protein dissociation6,7,10. Shown in Figure 2 is the ETD MS/MS spectrum of phosphorylated peptide, SGDpSDEELIRTVR. The observed c′-and z′•-type fragment ions are labeled above and below the peptide sequence, respectively. These ions facilitate complete sequence analysis of the phosphorylated peptide, including the unambiguous assignment of the modified serine residue.
Recent studies indicate that ETD can also be employed to characterize PTMs on large peptides, providing a means for identifying long-range, combinatorial modifications11,12. When the primary sequence of the protein being studied is known, digestion of the protein can be tailored to generate large peptides. These peptides are generally more amenable to fragmentation via ETD and are often produced by proteases other than trypsin such as endoproteinases Lys-C, Asp-N and Glu-C12-14. If the primary sequence of the protein is not known, then Lys-C is often an appropriate choice for protein digestion as it both generates large peptides and ensures the presence of a C-terminal lysine which facilitates interpretation of the resulting ETD spectra6,13,15. If the peptide or protein exists in multiple charge states, the best sequence coverage (optimal formation of c′- and z′•-type fragment ions) is likely to be obtained by acquiring an ETD spectrum on one of the higher rather than lower charge states. Competitive with the production of c′-and z′•-type ions is the process of charge reduction. In this latter pathway, the multiply charged peptide or protein accepts an electron from the fluroanthene radical anion and undergoes backbone cleavage, but the resulting c′ and z′• fragments fail to dissociate because they are held together by multiple hydrogen bonds or salt bridge interactions. In Figure 2, charge reduction of the [M + 3H]+3 ion generates the [M + 3H]+2• and [M + 3H]+1•• ions at m/z 779 and m/z 1,558, respectively. Supplemental activation can be employed to facilitate selective dissociation of these ions, to improve the yield of c′- and z′•-type fragment ions6,16 while preserving labile PTMs.
Using the method detailed in this protocol, we often interrogate large peptides and intact proteins. ETD of these large, highly charged species commonly results in complex MS/MS spectra containing fragment ions in a variety of charge states. To simplify spectra, another ion/ion reaction termed proton-transfer reaction (PTR) is employed sequentially to deprotonate multiply charged fragment ions8. PTR is currently available from another instrument vendor (i.e., Bruker Daltonics), but is not yet available on the Thermo Scientific ETD-enabled linear ion trap (LTQ XL). Investigators are also utilizing extended ETD reaction times in order to simplify spectra and ease spectral interpretation17.
ETD shows great promise in the field of proteomics, and here we present a protocol for implementing ETD for dissociation of both peptides and proteins on a chromatographic timescale. We demonstrate the utility of this method on a peptide containing a labile PTM, a large peptide (90 residues in length), and an intact protein (21 kDa) that were successfully interrogated via ETD.
RP chromatography gradient; Step 9:
|Time interval (min)||Gradient (% B, vol/vol)|
A sensitive analysis is achieved using small i.d. columns (50–75-μm diameters) and 50–100 nl min−1 flow rates. We prefer using an Agilent 1100 binary pump delivering solvent at 0.2 ml min−1 and splitting the flow prior to the column as previously described18. We apply a spray voltage of 2 kV for all online nanoflow experiments. We prefer acetic acid as the ion-pairing agent when separating peptides for online LC-MS/MS. Acetic acid or formic acid can be utilized for analyzing intact proteins via LC-MS/ MS. You can also use higher organic content (>70% acetonitrile (vol/vol)) for solvent B when analyzing intact proteins. Under the ESI operating conditions outlined above, we observe that acetic and formic acids afford ion currents that are more than five times higher than those obtained with stronger ion-pairing agents.
|Reagent AGC target (radical anions)||2 × 105 ion counts|
|Full AGC target (MS1 precursor cations)||1 × 104 ion counts|
|MSn AGC target (MS2 isolated precursor cations)||2 × 104 ion counts|
|Isolation window||3 m/z|
|MS1 scan range||300–2,000 m/z|
|Reaction time||100 ms*|
|Repeat duration||20 s|
|Exclusion duration||30 s|
The table above lists the parameters that are appropriate for a data-dependent MS/MS analysis. We recommend using a method where a full-scan mass spectrum is acquired followed by six full-scan MS/MS spectra acquired sequentially on the six most abundant ions detected in the initial full-scan. The time to complete a duty cycle is lengthened as the number of data-dependent scans increases, and a significant increase in duty cycle time will negatively affect the dynamic range. We prefer acquiring six data-dependent scans as the average duty cycle is ~2 s. Sequence information from peptides and proteins can typically be obtained routinely at the level of 10 and 100 fmol, respectively. To evaluate instrument performance, two peptides, angiotensin I and vasoactive intestinal peptide, are employed as internal standards at the 100 fmol level.
The reaction time, denoted with an asterisk in the table above, is an important parameter that needs to be customized for each experiment. For options, see options A–C for peptide analysis with ETD, peptide analysis with ETD and supplemental activation, and peptide/protein analysis using ETD/PTR or extended ETD, respectively.
If a C8 column was used for LC-MS/MS analysis and poor chromatography resulted from the intact proteins having excessive retention times, then a C4 column may be more appropriate for separation.
Upon data analysis in Step 11, you may observe ETD MS/MS spectra wherein charge-reduced ions constitute the majority of the ion current. Supplemental activation can be employed in Step 9 to disrupt intramolecular interactions (e.g., hydrogen bonds) present within these charge-reduced species. Supplemental activation increases the relative abundances of c′-and z′•-type ions and results in more extensive sequence information. These features of supplemental activation are shown in Figure 4.
Figure 2 shows the ETD MS/MS spectrum recorded on [M + 3H]+3 ions derived from the phosphorylated peptide, SGDpSDEELIRTVR. This peptide was derived from HIV type 1 protein, Rev (regulator of expression of virion products)25,26. Product ions of type c′ and z′• allow for extensive sequence coverage of the peptide and enable assignment of the modified serine residue. Identification of the phosphorylated residue was achieved following gradient elution of the Rev peptide from a C18 RP column. The ETD reaction time employed was 60 ms, and the total time required to obtain the MS/MS spectrum was <350 ms.
Figure 4 displays MS/MS spectra recorded from ETD, with and without supplemental activation, on [M + 3H]+3 ions of the 18–39 peptide of ACTH (18–39). An extended ETD reaction time of 400 ms was used to ensure the [M + 3H]+2• reduced-charge species was the most abundant peak in the MS/MS spectra before implementing supplemental activation. The MS/MS spectrum acquired after ETD followed by supplemental activation is shown in Figure 4b,c. The [M + 3H]+2• and [M + 3H]+1•• charge-reduced products were activated. Note that although the [M + 3H]+1•• species is outside the scanned mass range, it is retained in the LTQ, and, because the m/z of the [M + 3H]+1•• is <2,500, it will be activated by the supplemental activation procedure. Following supplemental activation, ETD product ions are increased in relative abundance, and, furthermore, product ions are present that were previously not detected without supplemental activation. As seen in ECD27, activation of charge-reduced product ions resulting from electron transfer can yield ions of type c′ and z′• that have undergone radical transfer reactions; a z′•-type radical ion abstracts a hydrogen radical from the c′-type ion with which it is in complex. Therefore, observed masses for z′•-type ions are increased by 1 Da, and c′-type ions are decreased by 1 Da.
Figure 5 displays the ETD/PTR MS/MS spectrum of an Asp-N-generated, 90-amino acid peptide derived from the RNA-binding protein Sam68 (Src-associated in mitosis of 68 kDa)28,29. Because the sequence of the protein was known, Asp-N was selected for digestion to generate the 90mer peptide harboring a C-terminal lysine, making the peptide amenable to dissociation via ETD. To obtain this spectrum, the Sam68 peptide was gradient-eluted from a C18 RP column into the mass spectrometer. The total scan time (including both ETD and PTR times) was <3 s. Two arginine methylation motifs are present within this large peptide, and the coverage obtained with doubly charged c′-type ions facilitated identification of the dimethylated arginine residue indicated in the spectrum.
Figure 6 displays the ETD/PTR MS/MS spectrum of the 21-kDa protein subunit, p21, of the Arp2/3 (actin related protein 2/3) complex30. Identification of p21 was achieved following gradient elution of the intact protein from a C4 RP column. Less than 2 s was required to dissociate p21, resulting in extensive sequence information of the N- and C-termini of the intact protein. Nanoflow RP chromatography coupled to ETD MS/MS enabled the identification of p21 from a mixture of all the intact Arp2/3 protein subunits.
The authors thank Marie-Louise Hammarskjold, David Rekosh, Yukiko Misawa and Emily Sloan for providing the Sam68 and Rev protein samples and Dorothy Schafer and Tatyana Kotova for providing the Arp2/3 protein sample. This work was supported by grants from the National Institutes of Health (GM37537 to D.F. Hunt).