JOURNAL OF MASS SPECTROMETRYJ. Mass Spectrom. 35, 612–617 (2000)

Identification of non-covalent structure inapocytochromec by hydrogen exchange andmass spectrometry

Ashraf S. Raza, Kuruppu Dharmasiri and David L. Smith*Department of Chemistry, University of Nebraska–Lincoln, Lincoln, Nebraska 68588-0304, USA

Apocytochrome c, the in vivo precursor to active cytochrome c, was analyzed by amide hydrogen exchangeand mass spectrometry to search for fixed, non-covalent structure. The protein was incubated in H2O atpH 3.3 or 6.7 for various times, then exposed to D2O to initiate isotope labeling of unfolded regions. Followingacid quenching of hydrogen exchange, the labeled apocytochromec was digested with pepsin into fragmentsthat were analyzed by directly coupled high-performance liquid chromatography/electrospray ionization massspectrometry. The intermolecular distribution of deuterium and the deuterium levels in structurally distinctivepopulations were determined from the mass spectra of the peptic fragments. Spectra of peptic fragments derivedfrom apocytochrome c incubated at pH 3.3 had single envelopes of isotope peaks with masses indicating thatall of the amide hydrogens had been replaced with deuterium. These results showed that apocytochromec atpH 3.3 offered little resistance to hydrogen exchange, indicating that it was unfolded with little fixed structure.However, mass spectra of peptic fragments including residues 81–94 of apocytochromec incubated at pH 6.7had two envelopes of isotope peaks, indicating that one population was unfolded and the other population washighly structured in this region. Mass spectra of peptic fragments including residuesN -terminal to residue 81indicated that this region of the protein remained unfolded with little fixed structure at pH 6.7. Copyright 2000 John Wiley & Sons, Ltd.

KEYWORDS: hydrogen exchange; mass spectrometry; apocytochromec; non-covalent structure; proteins

INTRODUCTION

Cytochromec, which is normally found in the inter-membrane space of mitochondria, consists of a singlepolypeptide chain with approximately 100 residues andheme covalently attached to two cysteinyl residues. Theprocesses through which apocytochromec (cytochromec without the heme) is synthesized in the cytoplasm,translocated through the outer mitochondrial membraneand joined to the heme have been studied extensively.Early in vitro studies of apocytochromec showed thatit is highly disordered with properties characteristic ofunfolded proteins.1–3 More recent studies have shownthat the interaction of apocytochromec with the outermitochondrial membrane causes the protein to undergopartial folding.4,5 At some point in the translocation pro-cess, the heme is attached and the protein folds to formactive cytochromec.6,7 Whether apocytochromec hassome residual structure prior to interaction with the outermitochondrial membrane is important to our understand-ing of the general problem of transporting proteins acrosslipid membranes.8–13

* Correspondence to: D. L. Smith, Department of Chemistry, Uni-versity of Nebraska–Lincoln, Lincoln, Nebraska 68588-0304, USA.E-mail: dsmith7@unl.edu

Contract/grant sponsor: National Institutes of Health;Contract/grant number: R01 40384.

Contract/grant sponsor: Nebraska Center for Mass Spectrometry.

The present study was undertaken to determine whetheramide hydrogen exchange and mass spectrometry givesuseful information about possible residual structurein apocytochromec. The rates at which hydrogenlocated at peptide amide linkages undergoes isotopicexchange are highly sensitive to protein structure.14–16

Although nuclear magnetic resonance (NMR) has beenthe primary method used to determine amide hydrogenexchange rates for two decades, the advantages ofmass spectrometry for such measurements have becomeincreasingly evident.17–20 As demonstrated in the presentstudy, mass spectrometry is particularly useful fordetermining the intermolecular distributions of deuteriumfrom which structural heterogeneity in proteins maybe determined.19,21–23 The present hydrogen exchangeresults indicate that apocytochromec is unfolded in acid,exhibiting no detectable resistance to hydrogen exchange.However, when apocytochromec is incubated in neutralsolutions, folded structure in a region including¾14residues near theC-terminus leads to two populationsof molecules. One offers no resistance to exchangewhereas the other is highly resistant to exchange.

EXPERIMENTAL

Preparation of apocytochromec

The heme, which was covalently attached to cytochromec, was removed from holocytochromec as described

Copyright 2000 John Wiley & Sons, Ltd. Received 4 October 1999Accepted 19 January 2000

IDENTIFICATION OF NON-COVALENT STRUCTURE 613

previously.2 Stock solutions of horse heart cytochromec (0.39 mM) and silver sulfate (26 mM in 9% acetic acid)were combined (1 : 10), incubated in the dark for 4 h at40°C and centrifuged to precipitate heme. The supernatantwas applied to a Sephadex G-25 column (28ð 100 mm)equilibrated in 0.1M acetic acid to remove non-volatilesalts. Protein collected from the column was dried in aSpeed Vac concentrator and dissolved in 2 ml of 0.05Mammonium acetate solution (pH 5.0, 6M guanidinehydrochloride and 1M dithiothreitol). This solution wasincubated in the dark for 2 h to cleave the silvermercaptide bond. Apocytochromec was isolated usinga Sephadex G-50 column (28ð 100 mm, equilibrated in0.05 M ammonium acetate), followed by reversed-phasehigh-performance liquid chromatography (RP-HPLC)(acetonitrile–water–0.1% trifluoroacetic acid). Analysisof the apocytochromec by electrospray ionization massspectrometry (ESIMS) confirmed that the heme had beenremoved without altering the cytochromec backbone (Mrfound, 11 741.9; calculated, 11 743.6 Da.).

Apocytochrome c labeling in D2O

The non-covalent structure of apocytochromec atpH 3.3 or 6.7 was investigated by amide hydrogenexchange/mass spectrometry using general methodsdescribed previously.20,24 The experimental procedureused for this study is illustrated in Fig. 1. Apocytochromec (0.6 mM) was incubated in aqueous phosphate buffer(pH 3.3 or 6.7) for various times. Following theincubation period, the solution was diluted 20-fold withD2O (0.1 M phosphate, pD 6.8, 20°C) to initiate thelabeling process. Isotope exchange was quenched after10–120 s by decreasing the pH to 2.4 and the temperatureto 0°C. Samples were either analyzed immediately orstored at�80°C.

Quantification of deuterium by HPLC/MS

The deuterium levels in short segments of the protein weredetermined by fragmenting the labeled protein with pepsin(E : S 1 : 1, 5 min, 0°C) and analyzing the digest by RP-HPLC/ESIMS. The H2O : D2O ratio during digestion was1 : 1. The HPLC column (150ð 0.32 mm i.d.) was packedwith POROS 10 R2 perfusive media (PerSeptive Biosys-tems). Peptides were eluted following a 3 min desaltingperiod by a gradient of 0–60% acetonitrile (0.05% triflu-oroacetic acid) in 5 min. The mobile phase flow-rate was50 µl/min�1. To minimize artifactual isotope exchangeduring HPLC, the injector and column were submergedin an ice-bath.

The molecular masses of partially deuterated peptideswere determined by monitoring the HPLC effluent witha Micromass Autospec high-resolution mass spectrometerequipped with an electrospray ionization source and focalplane detector. The magnet was scanned through them/zrange 400–1600 in 10 steps where¾18% of the centralm/z was recorded in each step. Each scan required¾3 sfor completion. Mass spectra were recorded in the profilemode with an OPUS data system (Micromass).

Figure 1. Schematic representation of the pulse labeling pro-cedure used to determine the deuterium levels of the pepticfragments of apocytochrome c when equilibrated in phosphatebuffer of different pH.

RESULTS AND DISCUSSION

The rates at which hydrogens located at peptide amidelinkages undergo isotopic exchange are generally sensitiveto the non-covalent structures of proteins.14–16 To searchfor non-covalent structural features in apocytochromecunder acidic (pH 3.3) and neutral (pH 6.7) conditions,the protein was incubated in buffered H2O solutions forvarious times, then exposed to D2O to label unfoldedregions. The isotopically labeled protein was fragmentedinto peptides with pepsin and analyzed by directly coupledHPLC/ESIMS under conditions where 70–80% of thedeuterium at peptide amide linkages was retained. Sincedeuterium located in side-chains was lost during the HPLCstep,20,25,26 the molecular masses of the peptides reflectedthe total number of deuteriums at peptide amide linkages.

The elements comprising most materials have multipleisotopes that are distributed randomly among the entire

Copyright 2000 John Wiley & Sons, Ltd. J. Mass Spectrom. 35, 612–617 (2000)

614 A. S. RAZA, K. DHARMASIRI AND D. L. SMITH

population of molecules comprising a sample. Molecularmass measurements are normally directed to determiningthe mass of the lightest population (monoisotopic mass)27

or to the mass averaged over the entire population ofmolecules. However, for hydrogen exchange studies ofproteins, the intermolecular distribution of deuterium maynot be random. In such cases, the distribution of deu-terium may be as important as the level of deuterium. Forexample, the intermolecular distribution of deuterium forproteins of uniform structure is random, giving a singleenvelope of isotope peaks in the mass spectra of a proteinor its peptic fragments. However, if one population of theprotein is folded and another is unfolded, the intermolec-ular distribution of deuterium following exchange will bebimodal. Mass spectra of these samples will have two ormore envelopes of isotope peaks. The ability to detectthe intermolecular distribution of deuterium in proteins isan important advantage of mass spectrometry over NMRwhen hydrogen exchange is used to investigate structuralheterogeneity in proteins.19,21–23 In the present study, themass spectra of apocytochromec fragments were ana-lyzed in terms of both the intermolecular distributions ofdeuterium and the average deuterium levels in structurallysimilar populations of the protein.

Immediately before analysis by HPLC/ESIMS, thelabeled apocytochromec was digested with pepsin underconditions required to minimize isotope exchange at pep-tide amide linkages. Digestion gave¾20 fragments,28

from which the deuterium level along the entire backboneof the protein, and also the intermolecular distributionof deuterium, could be determined. This information wasused to search for structure and structural heterogeneity inapocytochromec. To assess quantitatively the deuteriumlevels in the peptic fragments, two reference samplesrepresenting unlabeled.Href/ and completely exchanged.Dref/ apocytochromec were analyzed. Mass spectra of

the C2 charge states of three peptic fragments includ-ing residues 67–82, 81–94 and 95–104 are presented inFig. 2. The top row of spectra, designated Href, indicatethe deuterium level expected for unfolded apocytochromec that was subjected only to quench and HPLC/MS anal-ysis. The deuterium levels in these ions are slightly ele-vated above natural abundance because a small amountof deuterium exchanged into the apocytochromec back-bone during digestion. The deuterium level found in thecompletely deuterated sample was 70–80% of the levelexpected if all of the amide linkages were deuterated.Since exchange was presumably complete prior to analy-sis, the differences between the expected and found levelsof deuterium in Dref fragments indicate the amounts ofdeuterium lost during digestion and HPLC. Losses of20–30% in the present experimental procedure are consis-tent with losses expected based on the exchange behaviorof model peptides.26 The molecular masses of peptides inthe Href and Dref samples indicate the minimum and max-imum molecular masses expected for the same segmentsderived from partially deuterated apocytochromec.

Mass spectra of the same fragments derived fromapocytochromec incubated at pH 3.3 for 1, 2 and 4 hand labeled in D2O for 30 s are also presented inFig. 2. These fragments give single envelopes of iso-tope peaks, indicating that all of the apocytochromecmolecules were structurally homogeneous in the back-bone regions sampled by these peptic fragments. Fur-thermore, the deuterium levels in these three peptic frag-ments matched withinš0.3 Da the deuterium levels foundin the same fragments when derived from completelyexchanged apocytochromec. Similar results were foundfor peptides derived from all other regions of the apoc-ytochromec backbone. Finding all segments completelydeuterated indicates that apocytochromec incubated atpH 3.3 has no non-covalent structure that significantly

Figure 2. Mass spectra of peptides representing three segments of the C-terminal region of apocytochrome c. Residue numbers areshown on the top of the spectra. Apocytochrome c was equilibrated in aqueous phosphate buffer at pH 3.3 for 1 4 h prior to 30 sincubation in D2O (pD 6.8). Results obtained for the same peptides derived from apocytochrome c that was completely protonated orcompletely deuterated are indicated by Href and Dref, respectively.

Copyright 2000JohnWiley & Sons,Ltd. J. MassSpectrom. 35, 612–617 (2000)

IDENTIFICATION OF NON-COVALENT STRUCTURE 615

affects the rates of isotope exchange at the peptide amidelinkages.

Results illustrating isotopic exchange in the same seg-ments in apocytochromec incubated at pH 6.7 for 5 minto 72 h are presented in Fig. 3. Reference spectra labeledHref and Dref are as described above. The molecularmasses and isotope patterns obtained for fragment 67–82incubated at pH 3.3 and 6.7 are similar. Similar resultswere found for all segments derivedN-terminal to thissegment. In contrast, mass spectra of the segment includ-ing residues 81–94 derived from apocytochromec incu-bated at pH 6.7 for more than¾20 min showed twoenvelopes of isotope peaks (i.e. a bimodal distribution).The average molecular masses of segments comprising thelow- and high-mass envelopes are equal to the molecularmasses of the same segments in reference samples Hrefand Dref, respectively. The low level of exchange in thepopulation represented by the low-mass envelope showsthat this population has considerable non-covalent struc-ture in the region including residues 81–94. Likewise,complete exchange in the population represented by thehigh-mass envelope indicates that these molecules havevery little non-covalent structure in this region. The rela-tive abundances of the two populations are given by theareas of the two envelopes of isotope peaks. The change inthe relative abundances of the structured and unstructuredforms with incubation time at pH 6.7 indicates that steady-state concentrations of 35 and 65%, respectively, wereestablished after¾3 h. Mass spectra of three additionalfragments (residues 81–96, 83–94 and 83–96) exhib-ited features and behaviors similar to those described forresidues 81–94 (data not shown).

The isotope pattern of the 95–104 segment is asym-metric, indicating that it was comprised of two envelopesof isotope peaks. Incomplete separation of these envelopesshows that only some of the amide hydrogens in the struc-tured population were protected from hydrogen exchange.

The mass of the low-mass envelope suggests that approx-imately half of this segment in the structured populationhad non-covalent, fixed structure. The change in the rela-tive abundances of the two envelopes with incubation timeis similar to that found for residues 81–94. These resultsshow that incubation of apocytochromec at pH 6.7 formore than 3 h leads to the formation of substantial non-covalent structure in the region that starts near residue 81,extends to residue 94 and likely ends near residue 99.

The nature (i.e. tightness) of the non-covalent structurein this region was assessed by varying the time theprotein was labeled in D2O. Mass spectra for the segmentincluding residues 81–94 derived from apocytochromec incubated at pH 6.7 for 4 h and labeled in D2O for10–120 s are presented in Fig. 4. These results show thatthe deuterium levels in the low- and high-mass envelopesof the 81–94 segment did not change. Finding some(¾65%) of the peptides from this segment fully deuteratedafter only 10 s exposure to D2O suggests a protectionfactor of <10.16,29 Similarly, finding no deuterium inthe remaining peptides derived from this segment, evenafter incubation in D2O for 120 s, suggests a protectionfactor for these molecules of>100. The results presentedin Fig. 5 are useful for comparing the structure of thefolded population with that of native cytochromec, in thebackbone region including residues 81–94. The deuteriumlevel in the low-mass envelope (middle panel) is muchless than the deuterium level in the same segment derivedfrom native cytochromec (top panel). Both forms of theprotein were labeled under the same conditions. Theseresults show that the structure in residues 81–99 of thefolded population of apocytochromec is much tighter thanthe same region in native cytochromec.

This study shows that apocytochrome incubated atpH 3.3 is unstructured, as sensed by amide hydrogenexchange. However, incubation at pH 6.7 leads to twopopulations, one unstructured and one highly structured.The region that exhibits this phenomenon includes residues

Figure 3. Mass spectra of peptides representing the same three C-terminal segments as depicted in Fig. 2. Apocytochrome c wasequilibrated in aqueous phosphate buffer at pH 6.7 for 5 min 72 h prior to 30 s incubation in D2O (pD 6.8). Href and Dref as in Fig. 2.

Copyright 2000JohnWiley & Sons,Ltd. J. MassSpectrom. 35, 612–617 (2000)

616 A. S. RAZA, K. DHARMASIRI AND D. L. SMITH

Figure 4. Electrospray ionization mass spectra of the 81 94segment of apocytochrome c following incubation in aqueousphosphate buffer of pH 6.7 for 4 h and labeled in D2O for10 120 s. Href and Dref as in Fig. 2.

81–94 and approximatelyhalf of the 95–104 segment.Protectionagainsthydrogenexchangein this region issubstantiallygreaterthan in the sameregion of nativecytochromec, suggestingthat this region is highly struc-tured.Intramolecularhydrogenbonding(e.g.˛-helicesand

Figure 5. Mass spectra of the segment including residues 81 94derived from folded cytochrome c (top), apocytochrome c atpH 6.7 (middle) and apocytochrome c at pH 3.3 (bottom). In eachcase, the intact protein was labeled by incubating in D2O for 30 s.

ˇ-sheets)andlimited accessto the solventarethe princi-pal factorsthat reduceisotopeexchangeat peptideamidelinkages.Approximately3 h wererequiredto achievethesteady-statepopulationsof structuredand unstructuredformsandtheseformsdo not interconvert,evenaftersev-eral days. It is not possibleto determineunequivocallyfrom the presentresultswhetherthis structureis due tomonomersor multimers of apocytochromec. However,the high protectionagainsthydrogenexchangesuggeststhat the structuredform may involve multimersof apoc-ytochromec. Experimentsperformedat differentconcen-trationsof proteingavedifferentpopulationsof structuredandunstructuredforms,asexpectedfor an intermolecularprocess.However,thereproducibilityof theseexperimentswasnotsufficient to point to aspecificaggregationmecha-nism.It is notedthatresultsof atime-resolvedfluorescencestudysuggestthatapocytochromec maybea dimerat theconcentrationusedin thepresentstudy.8

Acknowledgements

This work was supportedby a grant from the National Institutesof Health (GM RO1 40384) and the NebraskaCenter for MassSpectrometry.

REFERENCES

1. Stellwagen E, Rysavy R, Babul G. J. Biol. Chem. 1972; 247:8074.

2. Fisher WR, Taniuchi H, Anfinsen CB. J. Biol. Chem. 1973;248: 3188.

3. Cohen JS, Fisher WR, Schechter AN. J. Biol. Chem. 1974;249: 1113.

4. Jordi W, Zhou L, Pilon M, Demel RA, De Kruijff B. J. Biol.Chem. 1989; 264: 2292.

5. Bryson E, Rankin S, Carey M, Watts A, Pinheiro T. Biochem.1999; 38: 9758.

6. Smith M, Leung DW, Gillam S, Astell CR, Montgomery DL,Hall BD. Cell 1979; 16: 753.

7. Pfanner N, Hartl FU, Guiard B, Neupert W. Eur. J. Biochem.1987; 169: 289.

8. Vincent M, Brochon JC, Merola F, Jordi W, Gallay J. Bio-chemistry 1988; 27: 8752.

9. Hamada D, Hoshino M, Kataoka M, Fink AL, Goto Y. Bio-chemistry 1993; 32: 10351.

10. Kuroda Y. Biochemistry 1993; 32: 1219.

Copyright 2000JohnWiley & Sons,Ltd. J. MassSpectrom. 35, 612–617 (2000)

IDENTIFICATION OF NON-COVALENT STRUCTURE 617

11. Snel MME, de Kruijff B, Marsh D. Biochemistry 1994; 33:7146.

12. Tong J, Zhu L, Yang F. Chin. Sci. Bull. 1996; 41: 438.13. Rankin SE, Watts A, Pinheiro TJT. Biochemistry 1998; 37:

12588.14. Hvidt A, Nielsen SO. Adv. Protein Chem. 1966; 21: 287.15. Woodward C, Simon I, Tuchsen E. Mol. Cell. Biochem. 1982;

48: 135.16. Englander SW, Kallenbach NR. Q. Rev. Biophys. 1984; 16:

521.17. Katta V, Chait BT. Rapid Commun. Mass Spectrom. 1991; 5:

214.18. Thevenon-Emeric G, Kozlowski J, Zhang Z, Smith DL. Anal.

Chem. 1992; 64: 2456.19. Miranker A, Robinson CV, Radford SE, Aplin RT, Dobson

CM. Science 1993; 262: 896.

20. Smith DL, Deng Y, Zhang Z. J. Mass Spectrom. 1997; 32:135.

21. Zhang Z, Post CB, Smith DL. Biochemistry 1996; 35: 779.22. Yang H, Smith DL. Biochemistry 1997; 36: 14992.23. Heidary DK, Gross LA, Roy M, Jennings PA. Nature Struct.

Biol. 1997; 4: 725.24. Zhang Z, Smith DL. Protein Sci. 1993; 2: 522.25. Englander JJ, Rogero JR, Englander SW. Anal. Biochem.

1985; 147: 234.26. Bai Y, Milne JS, Mayne L, Englander SW. Proteins: Struct.

Funct. Genet. 1993; 17: 75.27. Yergey J, Heller D, Cotter RJ, Fenselau C. Anal. Chem. 1983;

55: 353.28. Dharmasiri K, Smith DL. Anal. Chem. 1996; 68: 2340.29. Jeng M-F, Englander SW, Elove GA, Wand AJ, Roder H.

Biochemistry 1990; 29: 10433.

Copyright 2000 John Wiley & Sons, Ltd. J. Mass Spectrom. 35, 612–617 (2000)