Introduction

Hydrogen/deuterium exchange (HDX) combined with mass spectrometry is an important tool in structural proteomics. The principle of the method is that backbone amide hydrogen atoms can exchange with deuterium upon exposure of the protein to a D2O-based buffer at physiological pH. The rate of exchange depends on the structural context of each amide bond. The peptide bond amide hydrogen atoms that are involved in hydrogen bonding, such as those in secondary structural elements, undergo slow exchange, whereas non-bonded atoms, such as those in disordered regions of the protein, exchange more rapidly. Owing to the mass difference between hydrogen and deuterium, this exchange can conveniently be measured by mass spectrometry.

Traditionally, characterization of the HDX level by mass spectrometry has been done by digestion of the protein at low pH (where the exchange is minimized) with proteolytic enzymes, usually pepsin, and then by determination of the deuteration levels of the resulting peptides. Alternatively, the level of HDX can be assessed by using our recently developed top-down approach [1, 2]. In this method, deuterated protein is directly infused into a mass spectrometer and fragmented using an ECD fast fragmentation technique. Fragmentation results in a series of c- and z-ions with preserved, “non-scrambled” deuteration patterns. Comparing deuteration levels of consecutive fragments in the series allows determination of the deuteration status of nearly every residue in the protein sequence. This approach relies on the determination of the deuteration status of fragments that differ by one exchanged hydrogen atom, so data processing is time-consuming and challenging—particularly if it is done manually by centroiding low-intensity isotopic envelopes. In contrast, the fitting of theoretically predicted deuterated isotopic envelopes is thought to be less subject to errors resulting from missing peaks in the envelopes and other spectral imperfections ([3] and references therein).

Numerous approaches have been developed mainly for the processing of bottom-up HDX data, ranging from simple peak processing utilities [4] to fully integrated instrument-software platform solutions [3, 5–7]. Arguably, the most advanced approaches for determining deuteration levels combine the identification of the deuterated peptide (thereby determining its elemental composition), calculation of its deuterated isotopic distribution, and fitting it to the experimental data. Predicting the theoretical isotopic distributions is also advantageous for the identification of the high-mass species that do not have clearly distinguishable monoisotopic peaks in their mass spectra. This problem often occurs in top-down protein fragmentation. We have, therefore, developed this method for the identification of the protein fragments by mass in high resolution non-HDX MS/MS spectra and for the determination of deuteration levels of the corresponding fragments in HDX MS/MS spectra.

Methods

Data were acquired as described previously [2]. Briefly, the experimental set-up was as follows: the protein solution and the D2O buffer solution were fed from syringes and continuously mixed at a 1:4 ratio in a HDX capillary (80% D2O). The effluent was then continuously mixed in a 1:1 ratio with a low-pH quenching solution of formic acid/acetonitrile/80% D2O (i.e., 80% D2O in the aqueous portion of the solution), and infused directly into the mass spectrometer.

The program was written in Microsoft Visual Basic .NET, 2008 Express Edition. Downloadable files are posted on www.proteincentre.com and at www.creativemolecules.com/CM_Software.htm. Executing the downloaded programs requires the Microsoft .NET Framework. The program is primarily oriented towards Bruker ESI-FTMS-ECD MS/MS data, but can be used with any tab-separated mass-intensity lists as input and, therefore, is instrument-independent. All results are saved in text files as tab-delimited values and can, therefore, be easily opened as Excel spreadsheets.

Results

Algorithm

The algorithm of the fragment deuteration determination is based on predicting the isotopic distribution of any particular deuterated c- or z-ion fragment and fitting it to the experimental isotopic envelope with the fragment deuteration level as a variable. For every fragment ion, the following assumptions are made:

$$ {\mathrm{H}}_{\mathrm{tot}}={\mathrm{H}}_{\mathrm{nexch}}+{\mathrm{H}}_{\mathrm{exch}\ \max}\kern0.5em \mathrm{and}\kern.5em {\mathrm{H}}_{\mathrm{exch}\ \max }={\mathrm{H}}_{\mathrm{exch}}+{\mathrm{H}}_{\mathrm{prot}}, $$

where Htot is the total number of hydrogen atoms in the fragment, Hnexch is number of non-exchangeable hydrogens, Hexch max is the maximum number of exchangeable hydrogens, Hexch and Hprot are the numbers of exchanged and protected hydrogen atoms, respectively.

Htot and Hexch max can be predicted from the sequence of the fragments and the charge state. The theoretical isotopic distribution for the maximum deuteration level with Hexch max is then calculated for each fragment using Kubinyi algorithm [8]. We have found that rather than changing the deuterium content for all hydrogen atoms non-selectively, better results for the distribution calculations are obtained if we use the natural % deuterium for non-exchanged hydrogens and the % deuterium in the exchange/quenching solutions as the % deuterium of the exchanged hydrogen atoms (Figure 1). Hexch is then gradually decreased from Hexch max and the corresponding distributions are calculated, normalized, and fitted to the normalized experimental isotopic distributions, with quality of the fit given by the R2 value. The best fit for the theoretical distribution to the experimental distribution thus gives the number of protected hydrogen atoms, Hprot, for any particular fragment.

Figure 1
figure 1

Theoretical distributions for the SADTT- c5 1+ fragment. The elemental formula for this fragment is C18H33N6O10. White, experimental non-deuterated distribution; blue, deuterated distribution calculated for 33 total hydrogens with deuterium content of 27% (the value calculated from the maximum number of exchangeable protons in 80% D2O); red, deuterated distribution with 21 non-exchangeable hydrogens and a deuterium content of 0.0156% and 12 exchangeable hydrogens with deuterium content of 80%; black, normalized experimental deuterated distribution, obtained using 80% D2O in the experiment. A model using 12 exchanged protons with a deuterium content of 80% (i.e., the same as the D2O% concentration used in the exchange and quench solutions), and 21 non-exchanged protons with a deuterium content of 0.0156% (i.e., the same as the naturally-occurring % deuterium), gives a better fit to the experimental data

Full size image

Data Analysis Workflow

The mass-intensity lists are generated from the ECD-FT-MS/MS spectra of the non-deuterated and deuterated samples and, optionally, an additional deuterated sample under different experimental conditions. The best results are produced with recalibrated spectra (i.e., spectra that have been recalibrated using the predicted monoisotopic masses for the identified fragments in the non-HDX spectra and using masses deduced from these monoisotopic masses with additions of 1.0063 Da for the identified fragments in the HDX spectra). The c- and z-ion series can be then predicted for the selected mass and charge state ranges. Each fragment ion can be then located in the non-HDX spectrum and the corresponding deuterated fragment ion can then be located in HDX spectra by varying the Hexch parameter, starting from the Hexch max value. The results can be saved as text files and used later for generating the deuteration plots of the protein sequence.

Graphical User Interface

The graphical user interface of the program consists of single page and is designed for the efficient practical interrogation of the HDX datasets. It contains all the necessary parameters for the analysis (Table 1), zoomable views of the spectra (non-HDX and two HDX spectra panels for the concurrent HDX analysis of two different protein states), and boxes for the input and output data (Figure 2). When a match is found in the spectrum, the theoretical isotopic envelope is plotted in red.

Table 1 Control Parameters of the HDX Match Program
Full size table
Figure 2
figure 2

Graphical user interface of the HDX Match program. In a single window, the c- and z-ion fragment masses can be predicted and located in the non-HDX and HDX spectra. If a match for the selected fragment is found in non-HDX spectrum, the theoretical distribution of the selected fragment is plotted in red in the top (non-HDX) spectrum panel. Maximum number of the exchangeable protons is calculated based on the elemental composition for the selected fragment. The isotopic distributions of the deuterated form of the corresponding fragment are plotted in red in the middle or bottom (HDX) spectrum panels. The best fit to the experimental data (i.e., the fit with the maximum R2 value) is found by gradually reducing the number of exchanged protons (increasing number of protected protons) in the calculation of the theoretical deuterated isotopic distribution of the fragment ion

Full size image

Conclusions

HDX Match software is a practical tool for rapid fragment assignment and deuteration level determination from top-down ECD-FTMS HDX experiments. Using this software dramatically reduces the time required (from weeks to hours) for analysis of the data from such experiments and, as a result, will help to promote and to spread use of this promising approach for protein structural analysis.