Letter pubs.acs.org/ac
Cite This: Anal. Chem. XXXX, XXX, XXX-XXX
Engineering Nanodisc Scaffold Proteins for Native Mass Spectrometry Deseree J. Reid,† James E. Keener,† Andrew P. Wheeler,† Dane Evan Zambrano,† Jessica M. Diesing,† Maria Reinhardt-Szyba,‡ Alexander Makarov,‡ and Michael T. Marty*,† †
Department of Chemistry and Biochemistry, University of Arizona, Tucson, Arizona 85721, United States Thermo Fisher Scientific, 28199 Bremen, Germany
‡
S Supporting Information *
ABSTRACT: Lipoprotein nanodiscs are ideally suited for native mass spectrometry because they provide a relatively monodisperse nanoscale lipid bilayer environment for delivering membrane proteins into the gas phase. However, native mass spectrometry of nanodiscs produces complex spectra that can be challenging to assign unambiguously. To simplify interpretation of nanodisc spectra, we engineered a series of mutant membrane scaffold proteins (MSP) that do not affect nanodisc formation but shift the masses of nanodiscs in a controllable way, eliminating isobaric interference from the lipids. Moreover, by mixing two different belts before assembly, the stoichiometry of MSP is encoded in the peak shape, which allows the stoichiometry to be assigned unambiguously from a single spectrum. Finally, we demonstrate the use of mixed belt nanodiscs with embedded membrane proteins to confirm the dissociation of MSP prior to desolvation.
N
Despite their utility, native MS analysis of nanodiscs is limited by potential challenges with isobaric species that can complicate data analysis and interpretation. For example, MSP1D1(−) has a mass of 22 044 Da, which is only 2 Da larger than 29 molecules of 1-palmitoyl-2-oleoyl-sn-glycero-3phosphocholine (POPC) (22 042 Da). Although nanodiscs are more monodisperse than alternative membrane mimetics,25 they still have an intrinsic distribution in the number of lipids per complex of at least ±5 lipids.12 When membrane proteins are ejected from nanodiscs in the gas phase under gradual dissociation, they can retain a large number of bound lipids in an even broader distribution.4 At current instrumental resolving power, it can be impossible to assign isobaric peaks where contributions could come from a combination of lipids, membrane protein subunits, and MSP belts. To solve this problem, we previously collected spectra from nanodiscs with different MSP belts and lipids, thereby shifting the masses of the MSP and lipids so that they were no longer isobaric.4 However, this approach presents several challenges. It consumes triple the material and assumes that changing the lipids and MSP belts will not perturb the system. It also relies on having suitable nonisobaric combinations of MSP, lipid, and membrane protein available. Because the lipid and membrane protein are the critical components of the system, the most straightforward way to achieve a nonisobaric combination is to change the MSP, but existing constructs are designed for different functions without concern for their mass. Here, we
ative or noncovalent mass spectrometry (MS) has emerged as a powerful technique for characterizing the oligomeric state and interactions for soluble and membrane protein complexes. Because of the high precision of native MS analysis, the stoichiometry of the complex can often be determined unambiguously from the mass of the intact complex. However, isobaric masses present the potential for unassignable spectra and unresolvable species. Solution and gas phase separation techniques such as ion mobility may help to resolve isobaric species, but these are not always available or suitable. Thus, it remains challenging to assign complex spectra with ambiguous masses. Nanodiscs are a promising platform for native MS of membrane proteins due to their size, homogeneity, and relative monodispersity.1−19 To complement conventional detergentbased methods, nanodiscs allow membrane proteins to be delivered for native MS in a lipid bilayer environment, which may be a better model of the physiological environment of membrane proteins. Nanodiscs are nanoscale lipoprotein particles consisting of a lipid bilayer surrounded by two membrane scaffold protein (MSP) belts. There are a number of different MSP constructs available, but they are all derived from ApoAI, the primary component of high-density lipoprotein particles.20−24 Klassen and co-workers have employed nanodiscs with catch-and-release ESI to investigate glycolipid interactions with soluble proteins.1,3,5,10,11,13−15,17 We showed that nanodisc complexes can be preserved intact for native MS,12,18 and Robinson and co-workers later showed that nanodiscs can be used to deliver membrane protein complexes into the gas phase.4,16 © XXXX American Chemical Society
Received: August 31, 2017 Accepted: October 19, 2017 Published: October 19, 2017 A
DOI: 10.1021/acs.analchem.7b03569 Anal. Chem. XXXX, XXX, XXX−XXX
Letter
Analytical Chemistry
voltage was ramped in 20 V increments from 0 to 200 V, and spectra were deconvolved using UniDec9 as previously described.6 Representative spectra at 120 V are shown in Figure S2A with their deconvolved zero-charge mass spectra in Figure 2A. By shifting the mass, the isobaric overlap between MSP and POPC lipids are eliminated.
address these challenges by engineering MSP1D1 with a series of nonfunctional mutations to have a unique mass signature that encodes the stoichiometry of the MSP belts in the MS peak shape. Designing Mutant Membrane Scaffold Proteins. We designed MSP mutants to have specific masses by adding additional amino acids near the N-terminus, immediately after the TEV protease cleavage site. Because the presence or absence of the polyhistidine purification tag at the N-terminus does not adversely affect nanodisc formation, we predicted that modifications at this site would not hinder nanodisc formation. Using MSP1D1 as a template, we inserted G, T, R, or TT residues after the starting GST that remained after TEV cleavage (Figure 1). These amino acid additions were chosen to
Figure 2. (A) The deconvolved masses of nanodiscs with mutant belts show a predictable shift to greater mass. (B) Mixing two different belts creates a triplet pattern caused by random mixing of the two belts. Figure 1. (Top) Sequence of MSP mutants and associated mass values as determined from (Bottom) ESI MS of purified MSP. For convenience, the name of conventional MSP1D1(−) is shortened to D1T0. Minor peaks in the ESI mass spectrum are sodium adducts.
The stoichiometry of the shifted MSP belts can be assigned directly using macromolecular mass defect analysis (Figure S3A) as described previously.4 Briefly, the measured masses are divided by the mass of the repeating monomer unit, in this case POPC. Because the mass defect of POPC is zero by definition, the mass defect reports on the stoichiometry of nonlipid components. Because MSP1D1(−) is nearly isobaric with POPC, its mass defect is also nearly zero, so we cannot determine the stoichiometry of MSP1D1(−) from the mass defect. However, the additional mass of our mutant MSP belts creates nonzero mass defects. Because the mass defect shift of the resulting nanodiscs is twice the predicted mass defect of the additional residues, we can confirm that each nanodisc contains two belts as expected. After demonstrating that the novel mutants form nanodiscs with the intended mass shifts, we created nanodiscs containing a roughly equimolar mixture of two different MSP mutants. Each MSP was expressed and purified independently and mixed prior to assembly. Because detergent is added to the mixture prior to assembly, the two belts mix and assemble into nanodiscs randomly. Similar to NMR, the statistical distribution of peaks allows the stoichiometry of the MSP belts (n) to be derived from the peak shape pattern using the n+1 rule based on Pascal’s Triangle. Because nanodiscs contain two MSP belts, native MS analysis of the mixed belt nanodiscs showed a triplet peak shape corresponding to light-light, heavy-light/light-heavy, and heavy-heavy peaks (Figure 2B and Figure S3B). Using the high resolving power of the Orbitrap mass spectrometer,27 we observed the triplet peak shape for all mutants larger than the D1T0 + D1T1 mixture, which has a separation between peaks
shift the masses by 57, 101, 156, and 202 Da, respectively, to create a series of mutants separated by roughly 50 Da. Each of these expressed well in E. coli and was purified with immobilized metal affinity chromatography according to conventional MSP purification protocols.26 We performed ESI mass spectrometry on the purified MSP to confirm that each protein construct contained the correct mass shift (Figure 1). We also generated A, Q, and TTT mutants, but the Q and TTT mutants failed to express in reasonable quantities. The A mutant expressed but was too similar to the G mutant to prove useful at current instrumental resolving power. Detailed experimental methods are provided in the Supporting Information. Next, we created POPC nanodiscs using each of these purified mutant MSP belts using conventional reconstitution protocols.26 The resulting nanodiscs were characterized by size exclusion chromatography (SEC) as shown in Figure S1. The mutant belt nanodiscs did not show any substantial deviations from the control MSP1D1(−) nanodiscs, indicating that the mutations did not affect nanodisc formation. These results demonstrate that our novel mutants can be substituted directly for conventional MSP. Native MS of Mutant Nanodiscs. We performed native MS using an Ultra-High Mass Range (UHMR) research option Q-Exactive HF mass spectrometer27 on each of the purified mutant nanodisc forms to observe the mass shifts. The collision B
DOI: 10.1021/acs.analchem.7b03569 Anal. Chem. XXXX, XXX, XXX−XXX
Letter
Analytical Chemistry
S7). The doublet indicates that these peaks arise from monomeric MSP stripped from the nanodisc complex with several lipids still bound. This assignment is confirmed by the deconvolved mass and mass defect. Because the doublet ratio agrees with the triplet ratios, we do not observe any preference in the ejection of heavier or lighter MSP. Previously, multiple nanodisc samples were required to unambiguously assign the complex spectra from membrane protein nanodiscs, but the unique signatures (or lack thereof) from mixed-belt nanodiscs allow definitive assignment from a single spectrum. Thus, encoding the stoichiometry of MSP in the peak shape solves the three-body problem of using the mass to assign the stoichiometry of MSP, lipids, and membrane proteins inside the nanodisc.
of 101 Da. The D1T0 + D1T2 mixture was baseline resolved with a separation between peaks of 202 Da. In contrast, the resolving power was too low to resolve the 56 Da differences for the triplet peak of the D1T0 + D1G1 (Figure S4). By encoding the stoichiometry of one component of the system in the peak shape, the mass of the MSP can be subtracted to assign the stoichiometry of the embedded lipids or proteins. This is relatively trivial in the case of a twocomponent system like nanodiscs with only MSP and lipids, but it is essential with a three-component system like nanodiscs with embedded membrane protein oligomers. Membrane Protein Nanodiscs. To demonstrate the utility of mixed-belt nanodiscs for membrane proteins, we assembled trimeric AmtB into POPC nanodiscs with several different combinations of MSP mutants. SEC analysis confirmed that the mutant nanodiscs do not affect membrane protein assembly into nanodiscs (Figure S5). Native MS of mixed-belt AmtB nanodiscs continued to show a single peak, and spectra were very similar to those previously described for MSP1D1(−) (Figure 3 and Figures S3C and S6).4 The absence of a triplet or doublet peak confirms that the MSP belt is lost during desolvation to present the AmtB trimer surrounded only by lipids. To observe the AmtB trimer with bound lipids, we tuned the instrument for maximum signal for high mass species. When the instrument is adjusted to transmit more low mass species, several charge states with doublet peaks are observed (Figure
■
CONCLUSIONS We have engineered several new types of MSP with subtle shifts in mass. Mixing these mass mutants allows the stoichiometry of MSP to be determined from the peak shape, which simplifies interpretation and assignment of complex spectra. We demonstrated that empty nanodiscs ionize as intact complexes with two belts, but membrane protein nanodiscs lose the MSP belt before they can be resolved. We expect that these new mutant forms of MSP will prove critical for future native MS analysis of membrane protein-nanodisc systems. We have deposited the sequences with Addgene (see below) to make them easily available for the growing nanodisc MS community. The use of small, nonfunctional mutations to encode stoichiometry in peak shape may prove useful for solving the oligomeric state of complex assemblies or for biophysics experiments. A similar approach has been used previously for subunit exchange with either a polyhistidine tag28 or with stable isotope labeling.29 Our approach of single amino acid additions may be less disruptive than a larger tag, and it avoids costly isotopic labeling. As native mass spectrometry instruments increase in resolving power, there are a number of exciting possibilities for new types of quantitative analysis of complex bimolecular systems.
■
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.7b03569. Supplemental methods, table of primers designed for SLIM insertion of additional amino acids into the Nterminal region of MSP1D1, size exclusion chromatograms, raw data of nanodiscs, normalized mass defects of nanodiscs, spectra and deconvolved masses of nanodiscs, deconvolved charge state peaks, mass spectra, and supplemental references (PDF)
■
AUTHOR INFORMATION
Corresponding Author
Figure 3. Native MS spectra of nanodiscs containing AmtB and mixed MSP belts. AmtB nanodiscs do not show a triplet or doublet peak shape, which indicates that MSP is ejected during desolvation and is not present in the complex at higher collision energy. One apparent triplet cluster is present in the complex pattern, but this is simply three adjacent charge states as shown in Figure S6. A positive control spectrum without AmtB (black) shows a clear triplet on each of the peaks.
*E-mail:
[email protected]. ORCID
Michael T. Marty: 0000-0001-8115-1772 Notes
The authors declare the following competing financial interest(s): M.T.M. has a disclosed financial interest in C
DOI: 10.1021/acs.analchem.7b03569 Anal. Chem. XXXX, XXX, XXX−XXX
Letter
Analytical Chemistry
Membrane Proteins. In Membrane Protein Structure and Function Characterization: Methods and Protocols; Lacapere, J.-J., Ed.; Springer New York: New York, 2017; pp 205−232. (20) Nasr, M. L.; Baptista, D.; Strauss, M.; Sun, Z.-Y. J.; Grigoriu, S.; Huser, S.; Pluckthun, A.; Hagn, F.; Walz, T.; Hogle, J. M.; Wagner, G. Nat. Methods 2017, 14, 49−52. (21) Denisov, I. G.; Sligar, S. G. Chem. Rev. 2017, 117, 4669−4713. (22) Hagn, F.; Etzkorn, M.; Raschle, T.; Wagner, G. J. Am. Chem. Soc. 2013, 135, 1919−1925. (23) Denisov, I. G.; Grinkova, Y. V.; Lazarides, A. A.; Sligar, S. G. J. Am. Chem. Soc. 2004, 126, 3477−3487. (24) Bayburt, T. H.; Grinkova, Y. V.; Sligar, S. G. Nano Lett. 2002, 2, 853−856. (25) Marty, M. T.; Hoi, K. K.; Robinson, C. V. Acc. Chem. Res. 2016, 49, 2459−2467. (26) Ritchie, T. K.; Grinkova, Y. V.; Bayburt, T. H.; Denisov, I. G.; Zolnerciks, J. K.; Atkins, W. M.; Sligar, S. G. Reconstitution of Membrane Proteins in Phospholipid Bilayer Nanodiscs. In Methods in Enzymology; Nejat, D., Ed.; Academic Press: San Diego, CA, 2009; Vol. 464; pp 211−231. (27) van de Waterbeemd, M.; Fort, K. L.; Boll, D.; Reinhardt-Szyba, M.; Routh, A.; Makarov, A.; Heck, A. J. Nat. Methods 2017, 14, 283− 286. (28) Gupta, K.; Donlan, J. A. C.; Hopper, J. T. S.; Uzdavinys, P.; Landreh, M.; Struwe, W. B.; Drew, D.; Baldwin, A. J.; Stansfeld, P. J.; Robinson, C. V. Nature 2017, 541, 421−424. (29) Baldwin, A. J.; Lioe, H.; Robinson, C. V.; Kay, L. E.; Benesch, J. L. P. J. Mol. Biol. 2011, 413, 297−309.
University of Oxford that had no involvement in the work reported here. The terms of this arrangement have been properly disclosed to the University of Arizona and reviewed by the Institutional Review Committee in accordance with its conflict of interest policies. M.R.-S. and A.M. are employees of Thermo Fisher Scientific, the manufacturer and supplier of Orbitrap-based mass spectrometers. Plasmids for MSP1D1G1, MSP1D1T1, MSP1D1R1, and MSP1D1T2 are available from Addgene (https://www. addgene.org/Michael_Marty/). UniDec software used for data processing is available at http://unidec.chem.ox.ac.uk.
■
ACKNOWLEDGMENTS The authors thank Dr. Jani Bolla and Dr. Carol Robinson for initial support with mutagenesis and Chelsie Hurst for assistance with production of MSP. The template plasmid, pMSP1D1, was a gift from Stephen Sligar (Addgene plasmid no. 20061). This work was funded by an American Cancer Society Institutional Research Grant (Grant IRG-16-124-37IRG) and the Bisgrove Scholar Award from Science Foundation Arizona.
■
REFERENCES
(1) Li, J.; Richards, M. R.; Kitova, E. N.; Klassen, J. S. J. Am. Soc. Mass Spectrom. 2017, 28, 2054−2065. (2) Henrich, E.; Peetz, O.; Hein, C.; Laguerre, A.; Hoffmann, B.; Hoffmann, J.; Dötsch, V.; Bernhard, F.; Morgner, N. eLife 2017, 6, e20954. (3) Han, L.; Morales, L. C.; Richards, M. R.; Kitova, E. N.; Sipione, S.; Klassen, J. S. Anal. Chem. 2017, 89, 9330−9338. (4) Marty, M. T.; Hoi, K. K.; Gault, J.; Robinson, C. V. Angew. Chem., Int. Ed. 2016, 55, 550−554. (5) Li, J.; Fan, X.; Kitova, E. N.; Zou, C.; Cairo, C. W.; Eugenio, L.; Ng, K. K. S.; Xiong, Z. J.; Prive, G. G.; Klassen, J. S. Anal. Chem. 2016, 88, 4742−4750. (6) Hoi, K. K.; Robinson, C. V.; Marty, M. T. Anal. Chem. 2016, 88, 6199−6204. (7) Cleary, S. P.; Thompson, A. M.; Prell, J. S. Anal. Chem. 2016, 88, 6205−6213. (8) Campuzano, I. D. G.; Li, H. L.; Bagal, D.; Lippens, J. L.; Svitel, J.; Kurzeja, R. J. M.; Xu, H.; Schnier, P. D.; Loo, J. A. Anal. Chem. 2016, 88, 12427−12436. (9) Marty, M. T.; Baldwin, A. J.; Marklund, E. G.; Hochberg, G. K.; Benesch, J. L.; Robinson, C. V. Anal. Chem. 2015, 87, 4370−4376. (10) Leney, A. C.; Rezaei Darestani, R.; Li, J.; Nikjah, S.; Kitova, E. N.; Zou, C.; Cairo, C. W.; Xiong, Z. J.; Prive, G. G.; Klassen, J. S. Anal. Chem. 2015, 87, 4402−4408. (11) Han, L.; Kitova, E. N.; Li, J.; Nikjah, S.; Lin, H.; Pluvinage, B.; Boraston, A. B.; Klassen, J. S. Anal. Chem. 2015, 87, 4888−4896. (12) Marty, M. T.; Zhang, H.; Cui, W.; Gross, M. L.; Sligar, S. G. J. Am. Soc. Mass Spectrom. 2014, 25, 269−277. (13) Leney, A. C.; Fan, X.; Kitova, E. N.; Klassen, J. S. Anal. Chem. 2014, 86, 5271−5277. (14) Zhang, Y.; Deng, L.; Kitova, E. N.; Klassen, J. S. J. Am. Soc. Mass Spectrom. 2013, 24, 1573−1583. (15) Lin, H.; Kitova, E. N.; Klassen, J. S. Anal. Chem. 2013, 85, 8919−8922. (16) Hopper, J. T.; Yu, Y. T.; Li, D.; Raymond, A.; Bostock, M.; Liko, I.; Mikhailov, V.; Laganowsky, A.; Benesch, J. L.; Caffrey, M.; Nietlispach, D.; Robinson, C. V. Nat. Methods 2013, 10, 1206−1208. (17) Zhang, Y.; Liu, L.; Daneshfar, R.; Kitova, E. N.; Li, C.; Jia, F.; Cairo, C. W.; Klassen, J. S. Anal. Chem. 2012, 84, 7618−7621. (18) Marty, M. T.; Zhang, H.; Cui, W.; Blankenship, R. E.; Gross, M. L.; Sligar, S. G. Anal. Chem. 2012, 84, 8957−8960. (19) van Dyck, J. F.; Konijnenberg, A.; Sobott, F. Native Mass Spectrometry for the Characterization of Structure and Interactions of D
DOI: 10.1021/acs.analchem.7b03569 Anal. Chem. XXXX, XXX, XXX−XXX
