Elsevier

Analytica Chimica Acta

Volume 618, Issue 2, 23 June 2008, Pages 168-183
Analytica Chimica Acta

Mass spectrometric identification of formaldehyde-induced peptide modifications under in vivo protein cross-linking conditions

https://doi.org/10.1016/j.aca.2008.04.049Get rights and content

Abstract

Formaldehyde cross-linking of proteins is emerging as a novel approach to study protein–protein interactions in living cells. It has been shown to be compatible with standard techniques used in functional proteomics such as affinity-based protein enrichment, enzymatic digestion, and mass spectrometric protein identification. So far, the lack of knowledge on formaldehyde-induced protein modifications and suitable mass spectrometric methods for their targeted detection has impeded the identification of the different types of cross-linked peptides in these samples. In particular, it has remained unclear whether in vitro studies that identified a multitude of amino acid residues reacting with formaldehyde over the course of several days are suitable substitutes for the much shorter reaction times of 10–20 min used in cross-linking experiments in living cells. The current study on model peptides identifies amino-termini as well as lysine, tryptophan, and cysteine side chains, i.e. a small subset of those modified after several days, as the major reactive sites under such conditions, and suggests relative position in the peptide sequence as well as sequence microenvironment to be important factors that govern reactivity. Using MALDI-MS, mass increases of 12 Da on amino groups and 30 Da on cysteines were detected as the major reaction products, while peptide fragment ion analysis by tandem mass spectrometry was used to localize the actual modification sites on a peptide. Non-specific cross-linking was absent, and could only be detected with low yield at elevated peptide concentrations. The detailed knowledge on the constraints and products of the formaldehyde reaction with peptides after short incubation times presented in this study is expected to facilitate the targeted mass spectrometric analysis of proteins after in vivo formaldehyde cross-linking.

Introduction

The study of protein–protein interactions in living cells is an important aspect of understanding the function of individual proteins and their role within the cellular context. Affinity-based enrichment of proteins and their interaction partners from cell or tissue lysate is a commonly applied technique to isolate protein complexes and to identify and characterize its components by subsequent mass spectrometric analysis. Moreover, it is at the heart of several large-scale approaches to map the interactome of model organisms such as yeast [1], [2], [3] and E. coli[4], or signaling pathways in human cell lines [5], [6].

A major limitation of this general strategy is the loss of cellular context prior to the affinity enrichment step. As cell lysis converts the highly ordered protein assemblies inside a cell into disordered and dilute protein solutions, it eliminates the spatial and temporal constraints that govern protein interactions in cells. This process thus exposes the proteins under investigation to abundant cellular proteins they would not necessarily be in contact with otherwise. Moreover, it disturbs the equilibria of individual protein interactions and introduces molecular diffusion as an additional factor. As a consequence, interactions that show higher dissociation constants are lost, i.e. those that are of transient and weak nature and result in high turnover and/or low stoichiometry. Together, these two processes lead to an increase in false positive and a decrease in true positive interactions over time. Several ways of combating this phenomenon have been introduced, which include the use of tandem affinity tags [7] to eliminate non-specific binding and cryolysis [8] to reduce the protein association and dissociation rates.

Formaldehyde-mediated protein cross-linking has recently emerged as an additional means to preserve cellular protein interactions during affinity enrichment. It has been shown to be compatible with: affinity enrichment based on single [9] or tandem affinity tags [10], [11] as well as the endogenous protein [12], [13]; incorporation of stable isotope labels [10], [11], [13]; the use of non-denaturing [9], [12], [13] as well as denaturing [10], [11] purification conditions; and the application in cultured cells [9], [10], [11] as well as animal tissue [12], [13]. In all of the aforementioned cases, interacting proteins were identified by tandem mass spectrometry, whereas the actual cross-linking sites were not determined. This is in stark contrast to in vitro cross-linking strategies that utilize homo- or hetero-bifunctional cross-linking reagents with characteristic linker lengths for the analysis of protein interaction geometries [14], [15], [16]. These are based on well-known reactions that generate chemical structures with predictable changes in peptide masses, which facilitates their analysis and identification by mass spectrometry. Conversely, the chemistry of the formaldehyde cross-linking is presumed to be considerably more complex, and the resulting mixture of products to be much more heterogeneous and difficult to assess by mass spectrometry. Two recent studies involving model peptides [17] and proteins [18] have shed some light onto the reaction products after formaldehyde treatment. Extended length of formaldehyde exposure over several days has been shown to result in the modification of a multitude of different amino acid residues (amino-termini, lysine, arginine, histidine, cysteine, tyrosine, tryptophan, and phenylalanine), seemingly confirming the complexity of this reaction. Whether the results of these studies are applicable to conditions typically used for the cellular protein interaction studies outlined above, in particular with respect to their much shorter reaction time of 10–20 min that are required to prevent extensive protein loss, has not been addressed.

Here we present a detailed analysis of the reaction of formaldehyde with model peptides, which demonstrates that only a subset of the previously reported amino acid residues show significant reactivity within the first 10 minutes. Moreover, consistent with previous studies we find that the relative position of the residue in the peptide sequence is an additional constraint of its reactivity, while we determine the microenvironment of a residue, i.e. the peptide sequence, as another important influence. Peptide cross-linking was negligible under the conditions in this study, and previously reported peptide mass increases of 30 Da were only observed as amino group intermediates or as thiohemiacetal product of cysteine. In contrast, the major reaction products observed as increases by multiples of 12 Da were associated with and mostly located on the peptides’ amino-termini as well as lysine and tryptophan residues. Taken together, these results indicate that reactivity towards formaldehyde over short incubation times requires not only reactive residues but also favorable local environments to be present in a peptide.

Section snippets

Chemicals

Model peptides were purchased from Sigma (St. Louis, MO), Peptide 2.0 Inc. (Chantilly, VA), or synthesized in-house. α-Cyano-4-hydroxycinnamic acid (CHCA) was also obtained from Sigma. Paraformaldehyde (PFA), formic acid (FA, 88%) and acetonitrile (ACN, HPLC grade) were purchased from Fisher (Fair Lawn, NJ). C18 extraction tips were either obtained from Millipore (Bedford, MA) or Varian (Lake Forest, CA). 0.22 μm filters were purchased from Pall Corporation (Ann Arbor, MI). Deionized water (18  

Results and discussion

In order to determine whether the results of extended formaldehyde treatment over several days [17], [18] can be used as a model to understand the formaldehyde cross-linking reaction in living cells, 14 model peptides that differed in length and amino acid composition were treated with formaldehyde at reaction conditions (100 μM peptide, 83 mM (0.25% (w/v)) formaldehyde, 10–20 min, 37 °C) that are similar to those typically used for in vivo cross-linking. To qualify for these experiments, each

Conclusions

This study was undertaken to determine whether previously identified amino acid residues that react with formaldehyde over the course of several days, i.e. amino-termini, lysine, arginine, histidine, cysteine, tyrosine, tryptophan, and phenylalanine are identical to those involved in formaldehyde cross-linking in living cells. From the results presented here, it is evident that these two conditions lead to clearly distinct outcomes. Although the residues that have reacted with formaldehyde to a

Acknowledgments

The authors would like to thank Brent Sutherland for critical reading of the manuscript. This work was funded by an operating grant from the Natural Sciences and Engineering Research Council of Canada (NSERC), and infrastructure grants by the Michael Smith Foundation for Health Research (MSFHR) and the Canada Foundation for Innovation (CFI).

References (19)

  • I.M. Cristea et al.

    Mol. Cell. Proteomics

    (2005)
  • C. Tagwerker et al.

    Mol. Cell. Proteomics

    (2006)
  • C. Guerrero et al.

    Mol. Cell. Proteomics

    (2006)
  • Y. Bai et al.

    Mol. Cell. Proteomics

    (2008)
  • B. Metz et al.

    J. Biol. Chem.

    (2004)
  • A.C. Gavin et al.

    Nature

    (2002)
  • Y. Ho et al.

    Nature

    (2002)
  • A.C. Gavin et al.

    Nature

    (2006)
  • G. Butland et al.

    Nature

    (2005)
There are more references available in the full text version of this article.

Cited by (0)

View full text