Skip to main content
NCBI home page
Protein Science : A Publication of the Protein Society logoLink to Protein Science : A Publication of the Protein Society
. 2015 Aug 27;24(11):1756–1763. doi: 10.1002/pro.2766

Large-scale identification of membrane proteins with properties favorable for crystallization

Jared Kim 1, Allison Kagawa 1, Kellie Kurasaki 1, Niloufar Ataie 1, Il Kyu Cho 2, Qing X Li 2, Ho Leung Ng 1,3,*
PMCID: PMC4622209  PMID: 26257393

Abstract

Membrane protein crystallography is notoriously difficult due to challenges in protein expression and issues of degradation and structural stability. We have developed a novel method for large-scale screening of native sources for integral membrane proteins that have intrinsic biochemical properties favorable for crystallization. Highly expressed membrane proteins that are thermally stable and nonaggregating in detergent solutions were identified by mass spectrometry from Escherichia coli, Saccharomyces cerevisiae, and Sus scrofa cerebrum. Many of the membrane proteins identified had been crystallized previously, supporting the promise of the approach. Most identified proteins have known functions and include high-value targets such as transporters and ATPases. To validate the method, we recombinantly expressed and purified the yeast protein, Yop1, which is responsible for endoplasmic reticulum curvature. We demonstrate that Yop1 can be purified with the detergent dodecylmaltoside without aggregating.

Keywords: membrane proteins, crystallization, proteomics, mass spectrometry

Introduction

Membrane proteins are underrepresented in X-ray crystallography structures. Less than 600 unique membrane protein structures had been solved as of 2015.1 In contrast, there are over 100,000 crystal structures of soluble proteins in the Protein Data Bank. Less than 25% of the membrane proteins that have been crystallized are from eukaryotes.2 Progress has been slow despite enormous interest in membrane proteins, which constitute more than 30% of proteomes and 60% of known drug targets.3 Multiple bottlenecks exist in the membrane protein crystallography pipeline: recombinant expression, solubilization, and purification in detergent and crystallization.4,5 State-of-the-art, high-throughput membrane protein crystallography pipelines report attrition rates well over 90%.6,7 Because of this attrition, it is underappreciated that a significant number of membrane proteins are biochemically well behaved and relatively straightforward to crystallize. Examples from E. coli include aquaporin Z,8 multidrug efflux pump AcrB,9,10 outer membrane protein OmpF,11 and palmitoyltransferase PagP.12 In fact, AcrB has been recognized as a highly crystallizable contaminant in multiple membrane protein crystallography projects.

Membrane proteins with high expression levels and stability and monodispersity in detergent solutions are strongly associated with higher structure elucidation success rates.13–17 Moreover, highly abundant membrane proteins are good candidates for extraction from native sources if recombinant expression is unsuccessful. We report a protocol for identification of membrane proteins with intrinsic biochemical properties that correlate with crystallizability (Fig. 1). The protocol, which involves detergent solubilization, heat precipitation, and mass spectrometry, was used to identify native membrane proteins from E. coli, S. cerevisiae, and porcine (Sus scrofa) cerebral tissue that are promising crystallization candidates. Previous studies have identified thermostable cytosolic proteins from different organisms18–22; this is the first report identifying thermostable membrane proteins.

Figure 1.

Figure 1

Open in a new tab

Overview of approach for isolating and identifying membrane proteins with properties that favor crystallization.

Results

Cells or tissues were lysed, and membrane proteins were extracted into buffer containing 1% n-dodecyl β-D-maltoside (DDM). DDM was chosen because it is the single most effective detergent used in membrane protein crystallization.23 Samples containing solubilized membrane proteins were heat precipitated at 50–60°C for 10 min. The supernatant fraction containing soluble membrane proteins was passed through a 1,000,000-Da filter to remove heavily aggregated protein but retain oligomers.24,25 Membrane protein aggregation is well known to inhibit crystallization.13,16,25,26 DDM was removed from the filtrate by a combination of methanol–chloroform extraction and detergent affinity resins. Samples were then denatured with urea and digested with trypsin, and peptides were identified using a liquid chromatography mass spectrometry (LC-MS/MS) method on an LTQ Orbitrap XL using Mascot software.

We screened heat precipitation temperatures up to 70°C. Compared to 60° and 65°C, far more proteins from E. coli and S. cerevisiae were removed by heat precipitation at 70°C (Supporting Information, Fig. 1), a quality we considered undesirable. Given the ability of high-resolution LC-MS/MS to identify hundreds to thousands of proteins from a complex mixture, we chose heat precipitation temperatures of 50–60°C to retain more proteins.

After heat precipitation and filtration to remove aggregated proteins, 354 proteins from E. coli, 272 proteins from S. cerevisiae, and 565 proteins from porcine cerebrum were identified with a Mascot expectation value of <0.05. Many identified proteins are not membrane proteins but are, instead, highly abundant, soluble proteins commonly observed in proteomics studies such as ribosomal, heat shock, and cytoskeletal proteins.27–30 Not surprisingly, we also identified many peripheral, membrane-associated proteins in addition to integral membrane proteins (IMPs). The IMPs, which were of interest in this study, with the highest Exponentially Modified Protein Abundance Index (emPAI) scores, which correlate with abundance,31 from each sample are displayed in Tables3. Full lists of identified proteins are provided in the Supporting Information.

Table 3.

Twenty Most Abundant Porcine Cerebral Proteins with Favorable Properties for Crystallization

emPAI Exp. value Protein name UniProt accession MW Structure?
15.48 2.00E-29 Vesicle-associated membrane protein 2 P63026 12641 N
12.65 1.50E-99 Syntaxin-1B P61266 33224 N
6.31 0 Sodium/potassium-transporting ATPase subunit alpha-3 P13637 111677 N
4.9 0 Sodium/potassium-transporting ATPase subunit alpha-2 D2WKD8 112137 N
4.32 1.80E-22 Membrane-associated progesterone receptor component 1 Q95250 21596 N
3.91 4.30E-19 Cell cycle exit and neuronal differentiation protein 1 Q29026 13959 N
3.37 0 Sodium/potassium-transporting ATPase subunit alpha-1 Q9N0Z6 112922 N
2.91 1.60E-31 Myelin proteolipid protein Q8HXW7 30124 N
2.83 9.70E-22 Vesicle-associated membrane protein-associated protein B A5GFS8 27036 N
1.48 4.70E-19 Sodium/potassium-transporting ATPase subunit beta-1 P05027 35136 Y
1.4 1.70E-09 Neuronal membrane glycoprotein M6-a Q0VD07 31188 N
1.4 0.00017 Cytochrome b5 P00171 15320 N
1.17 1.30E-13 Cell adhesion molecule 2 Q8N3J6 47524 N
1.16 2.20E-36 Excitatory amino acid transporter 1 P46411 59553 N
1.06 2.30E-25 Neuroplastin P97300 44345 N
0.99 2.90E-16 Syntaxin-1A P32850 33071 N
0.96 2.20E-50 V-type proton ATPase 116 kDa subunit a isoform 1 Q9Z1G4 96404 N
0.92 1.20E-08 B-cell receptor-associated protein 31 Q5R8H3 27930 N
0.9 4.20E-63 Plasma membrane calcium-transporting ATPase 2 Q01814 136789 N
0.9 1.20E-18 Cell adhesion molecule 4 Q8R464 42697 N
Open in a new tab

Table 1.

Twenty Most Abundant E. coli Membrane Proteins with Favorable Properties for Crystallization

emPAI Exp. value Protein Name UniProt Accession MW Structure?
5479.9 0 ATP synthase subunit b P0ABA0 17253 N
116.63 3.70E-63 Inner membrane protein YhcB P0ADW3 14952 N
115.74 6.40E-93 Uncharacterized protein YqjD P64581 11045 N
56.99 0 Outer membrane protein A P0A910 37178 Y
43.63 0 Cytochrome bd-I ubiquinol oxidase subunit 1 P0ABJ9 58167 N
33.67 2.50E-46 NADH-quinone oxidoreductase subunit A P0AFC3 16447 N
27.45 1.30E-25 UPF0092 membrane protein YajC P0ADZ7 11879 Y
22.74 0 Outer membrane protein TolC P02930 53708 Y
17.3 1.40E-90 Cytoskeleton protein RodZ P27434 36169 N
14.48 9.20E-123 ATP-dependent zinc metalloprotease FtsH P0AAI3 70663 N
11.22 0 Peptidyl-prolyl cis-trans isomerase D P0ADY1 68108 N
10.38 1.00E-54 Modulator of FtsH protease HflC P0ABC3 37626 N
9.34 9.10E-18 Protein ElaB P0AEH5 11299 N
8.76 0 Magnesium-transporting ATPase, P-type 1 P0ABB8 99403 N
7.65 2.30E-46 Uncharacterized protein YgiM P0ADT8 23062 Y
7.31 2.10E-62 Modulator of FtsH protease HflK P0ABC7 45517 N
5.33 5.30E-27 Probable phospholipid ABC transporter-binding protein MlaD P64604 19564 N
5.11 3.10E-20 Sec-independent protein translocase protein TatA P69428 9658 N
4.11 3.00E-36 UPF0070 protein YfgM P76576 22162 N
4.04 8.70E-10 NADH-quinone oxidoreductase subunit K P0AFE4 10838 Y
Open in a new tab

Across samples, 165 out of 354 E. coli, 50 out of 272 S. cerevisiae, and 181 out of 565 porcine protein identifications were membrane-associated proteins (both integral and peripheral membrane proteins). The S. cerevisiae sample contained proportionally fewer membrane-associated proteins than the E. coli and porcine samples. We attribute the difference to the increased difficulty of lysing yeast cells and isolating their membrane fractions, which led to contamination of yeast samples with cytosolic proteins. Only 18 proteins from the S. cerevisiae sample were conclusively IMPs.

Crystal structures are available for five of the 20 most abundant E. coli membrane proteins and five of the 18 most abundant S. cerevisiae membrane proteins. The high percentage of crystallized membrane proteins identified strongly supports the predictive power of this method. In contrast, a crystal structure is available for only one of the 20 abundant porcine cerebral membrane proteins identified; this protein is a sodium/potassium-transporting ATPase. The lower percentage is presumably due to less crystallography research attempted with mammalian membrane proteins.

The identified proteins are structurally diverse. Four of the membrane proteins identified have or are predicted to have a beta barrel structure: OmpA and TolC from E. coli and mitochondrial porin 1 and ECM33 from yeast. Less-abundant E. coli membrane proteins identified also include previously crystallized proteins AcrB and OmpF (Supplementary Data Spreadsheet 1).9–11 In contrast, none of the proteins identified from S. crofa is predicted to have a beta barrel structure. Surprisingly, molecular weights of the most abundant IMPs identified vary widely across the three samples, ranging from 8,375 Da for yeast V-type proton ATPase subunit e to 170,970 Da for yeast tricalbin-3. Many of the identified proteins are subunits of well-characterized membrane protein complexes such as ATP synthase, cytochrome bd-I ubiquinol oxidase, TatA protein translocase, and modulator of FtsH protease from E. coli and sodium/potassium-transporting ATPase and syntaxin 1 from S. scrofa. A large fraction of the identified proteins are single-pass membrane proteins: 55% from E. coli, 56% from yeast, and 55% from pig. Of all known membrane proteins, 62% are predicted to cross the membrane only once.32 Interestingly, only 10% of all unique, membrane protein structures are single-pass membrane proteins.1 This suggests that single-pass membrane proteins may be harder to express or crystallize than other membrane proteins despite their abundant identification with this method. Alternatively, it is possible that single-pass membrane protein structures are underrepresented because research is primarily conducted on truncated nontransmembrane domains.

The majority of identified proteins in all three samples are involved in metabolic processes as defined by UniProt Gene Ontology terms33; of the identified proteins, 58.2% of E. coli, 66.6% of S. cerevisiae, and 48.1% of S. scrofa proteins are involved in metabolism. ATP synthase subunits, electron transport chain proteins, and small-molecule transporters were among the most abundant transmembrane proteins identified using our protocol, which is consistent with results from other membrane proteomics studies.29 Our method successfully identified metabolic enzymes which are potential drug targets. Notably, no G-protein-coupled receptors or kinase receptors were identified from the eukaryote samples, presumably due to low levels of expression or instability. Only 5 out of the top 20 E. coli and 6 of the top 18 S. cerevisiae membrane proteins identified have no known functions. Surprisingly, all of the top 20 S. scrofa membrane proteins identified have characterized functions.

To confirm that the membrane proteins identified by our method are not prone to aggregation, we recombinantly expressed in E. coli and purified Yop1, which was identified as a lower scoring hit by our screen. It does not appear in Table2 but does appear in the complete data included as Supporting Information. We chose to study Yop1 because of our interest in its role in generating membrane curvature in the endoplasmic reticulum.34,35 We purified His6-tagged Yop1 by nickel affinity chromatography in buffer containing 0.1% DDM. The size exclusion chromatogram for purified Yop1 revealed two main peaks that correspond to particles with molecular weights of 404 kDa and 42 kDa (Fig. 2). Very little eluted at the void volume, indicating minimal aggregation of Yop1 in DDM. The large particle with an estimated mass of 404 kDa may correspond to an oligomerized form of Yop1. It remains to be seen whether an oligomerized form of Yop1 is biologically relevant. This species is not heterogeneously aggregated given the symmetric elution profile and significant resolution from the exclusion volume elution point. We will describe detailed characterization of purified Yop1 in a separate publication.

Table 2.

Eighteen Most Abundant S. cerevisiae Membrane Proteins with Favorable Properties for Crystallization

emPAI Exp. value Protein name UniProt accession MW Structure?
2.62 1.20E-35 Reticulon-like protein 1 Q04947 32895 N
1.76 0.0015 V-type proton ATPase subunit e Q3E7B6 8375 N
1.12 5.30E-28 Om45p E7LVL3 44602 N
0.95 1.30E-14 ATP synthase subunit 4, mitochondrial P05626 26993 Y
0.9 1.10E-18 Sso1p E7QA36 33127 Y
0.85 8.4E-06 Pet9p E7Q0S8 34273 Y
0.81 3.00E-17 Mitochondrial outer membrane protein porin 1 P04840 30410 N
0.57 2.10E-07 Synaptobrevin homolog 2 P33328 12963 Y
0.45 8.50E-10 Protein transport protein SFT2 P38166 24258 N
0.42 0.0021 Prohibitin 2 P50085 34886 N
0.39 7.10E-24 Kar2p C8ZBH9 74406 N
0.38 2.50E-08 Rotenone-insensitive NADH-ubiquinone oxidoreductase, mitochondrial P32340 57214 Y
0.35 8.50E-26 Plasma membrane ATPase 2 P19657 102093 N
0.29 3.50E-13 Cell wall protein ECM33 P38248 48278 N
0.27 6.30E-11 Golgi SNAP receptor complex member 1 P38736 25379 N
0.19 5.90E-07 Mitochondrial import receptor subunit TOM70 P07213 70165 N
0.08 0.0028 SKG6p P32900 81799 N
0.06 1.10E-08 Tricalbin-3 Q03640 170970 N
Open in a new tab

Figure 2.

Figure 2

Open in a new tab

Size exclusion chromatogram for purified Yop1 (yeast) in 0.1% DDM.

Discussion

From three different species, we identified hundreds of membrane proteins that are naturally abundant and possess physical properties that predict successful crystallization. We considered the actual quantity of recovered protein more important than that relative to the baseline without heat treatment, which has already been extensively studied.36,37 We reasoned that even if the majority of an expressed membrane protein is aggregated, researchers are most concerned with the actual amount recovered. For example, some insoluble cytosolic proteins can be crystallized after low yield refolding.38,39

It remains to be seen whether good membrane protein crystallization candidates can be predicted in silico, as has been done with some success for soluble proteins40–42; however, a significant number of the identified proteins have already been crystallized. Of the 4,288 predicted open reading frames in E. coli,36 1,151 are predicted to be integral membrane proteins.43,44 Of these, the structures of 63, or 5.5% of 1,151, have been determined by X-ray crystallography and deposited in the PDB. Of the 20 most abundant membrane proteins identified by our screen, five have crystal structures in the PDB, delivering an enrichment factor of 4.5. The low numbers described here preclude discussion of statistical significance. Most of the membrane proteins identified in our screen have known functions, with many members of important protein families such as metabolic enzymes, ATPases, and transporters.

The proteins identified by experimental screening should also be analyzed with traditional bioinformatics tools for prioritization for crystallization attempts. Features that correlate with successful crystallization of soluble proteins, such as lack of predicted disordered regions, molecular weight, oligomerization state, and hydrophobicity, are also likely to affect membrane protein crystallization. Future analysis of results from large-scale membrane protein structural genomics projects will likely provide new insights into physical features associated with crystallization.45,46

The current method has several limitations. Identification of membrane proteins was not comprehensive given variability in membrane fraction isolation, natural protein abundance, and peptide ionization and detection by mass spectrometry. More specialized sample preparation techniques may allow identification of additional high-value, low-abundance membrane proteins, as has been demonstrated in high coverage proteomics studies.47 Finally, thermal stability and monodispersity of membrane proteins depend on the detergents and lipids used. Testing for membrane protein stability in the presence of detergents and lipids other than DDM, which was used in our screen, such as n-octyl-β-d-glucopyranoside (β-OG), lauryldimethylamine-oxide (LDAO), and cholesterol, may identify additional candidates for crystallography.

Materials and Methods

Sample preparation and lysis

A 2.5-L shake flask containing 1 L of Terrific Broth media was inoculated with Rosetta 2 (DE3) (Novagen) Escherichia coli, and a second 2.5-L shake flask containing 1 L of yeast extract peptone dextrose media was inoculated with Saccharomyces cerevisiae from frozen glycerol stocks. The flasks were incubated overnight at 37°C and 30°C, respectively. Cells were harvested by centrifugation at 3,000g for 45 min and washed three times by resuspension in 35 mL of lysis buffer (0.3M NaCl, 50 mM HEPES, pH 7.5, 0.1 mM tris(2-carboxyethyl)phosphine (TCEP)), followed by centrifugation at 26,200g for 15 min. The E. coli pellet was resuspended in 35 mL of lysis buffer with 1 μL of DNase I (25 ng/mL) and was then sonicated on ice for 2 min. The S. cerevisiae pellet was resuspended in in 35 mL of lysis buffer with 1 μL of DNase I, mixed with an equal volume of 0.4-mm silica beads, and lysed in a Biospec Bead-Beater under ice for a total of 3 min in 30 s pulses. Cerebral tissue was dissected from fresh porcine brain; meninges, membranes, and vasculature were removed. Aliquots were stored in phosphate buffered saline at −70°C. To extract membrane proteins, 3.5 g of porcine cerebrum were suspended in 35 mL of lysis buffer, and lysed in a 55 mL Teflon homogenizer until all visible solid particles were dispersed. Lysates from E. coli, yeast, and porcine brain were centrifuged at 10,000g for 10 min, and the pellets were discarded before being centrifuged at 56,700g for 30 min. The supernatants were discarded. The S. cerevisiae pellet was stored at −80°C. The porcine and the E. coli pellets were each resuspended in 35 mL of lysis buffer containing 1M NaCl to remove loosely associated proteins. The porcine and E. coli lysates were centrifuged a second time at 56,700g for 30 min. The pellets were stored at −80°C.

Detergent extraction and heat precipitation

The 8.5 mg E. coli pellet was suspended in 100 μL of lysis buffer containing 1% DDM detergent and 0.3 μL of DNase I (25 ng/mL). The porcine sample (7.1 mg) was suspended in 100 μL of lysis buffer with DDM. For S. cerevisiae, the 4 mg pellet was suspended in 40 μL of lysis buffer with DDM. The samples were placed on a rotating inverter at 25°C for 2 h before centrifugation at 24,000g for 15 min. The supernatant fractions were collected and heat precipitated: the S. cerevisiae sample was heated at 60°C for 10 min, and the E. coli and porcine samples heat treated at 50°C for 10 min. The samples were centrifuged at 24,000g for 5 min before being transferred to Vivaspin 500 centrifugal concentrators with 1-MDa molecular weight cutoff, and centrifuged at 24,000g for 30 min at 25°C. The filtrate was retained.

Mass spectrometry sample preparation

The E. coli and porcine samples were further extracted with chloroform/methanol to remove detergent.48 The methanol extracted E. coli and porcine protein samples were resuspended in 15 μL of 50 mM ammonium bicarbonate. All samples were then resuspended in 6M urea and 1 mM fresh dithiothreitol and incubated for 30 min. Proteins were then alkylated by incubation in 5.5 mM fresh iodoacetamide for 20 min. We then added 60 ng trypsin (Amresco, proteomics grade) to each sample and incubated overnight.49 Detergent was removed with Pierce 87776 detergent removal spin columns.50 Samples were desalted with Thermo Scientific C18 StageTips and resuspended in 20 μL of 0.1% formic acid before mass spectrometry analysis.

Mass spectrometry and data analysis

Samples were analyzed with a linear ion trap (LTQ) Orbitrap equipped with a Waters nanoAcquity UPLC system, a Waters Symmetry C18 180 μm × 20 mm trap column, and a 1.7 μm × 75 μm × 250 mm nanoACQUITY UPLC column (35°C) for peptide separation. The flow rate was 300 nL/min. A linear gradient (201 min) was run with 5% buffer B (100% acetonitrile, 0.075% formic acid) and 95% buffer A (100% water, 0.1% formic acid) at initial conditions, 50% B at 200 min, and 85% B at 201 min. MS (m/z range = 300–2,000 at average resolving power of 60,000) was acquired in the Orbitrap using one microscan in parallel with six data-dependent MS/MS acquisitions (based on the top six most intense MS peaks) in the LTQ. Peaks targeted for MS/MS fragmentation by collision-induced dissociation were first isolated with a 2-Da window followed by normalized collision energy of 35%. Dynamic exclusion was activated when former target ions were excluded for 60 s. Each MS Orbitrap scan required 1.4 s to acquire, whereas up to six MS/MS scans were acquired over an average time of 1 s. The total cycle time for both MS and MS/MS acquisition was 2.4 s.

Database search

All raw LC-MS/MS spectral data were searched with the Mascot algorithm (Matrix Science, v. 2.5) with the Mascot Distiller program used to generate compatible files. The Mascot Distiller program combines and centroids sequential MS/MS scans from profile data with the same precursor ion. Ions with charge state with +7 or less were preferentially located, and a peak list was generated for database searching. Data was searched against SwissProt_2013_12.fasta tax:Escherichia coli, NCBInr_20130805.fasta tax:Saccharomyces cerevisiae, and SwissProt_2013_12.fasta tax:Mammalia databases for the E. coli, S. cerevisiae, and porcine samples, respectively. The parameters were as follows: (i) the number of allowed miscleavages was set to four, (ii) error tolerances were set with peptide mass tolerance of 10 ppm or less and fragment mass tolerance of 0.5 Da or less, (iii) variable settings were used for oxidation of methionine and propanimide modification and carboamidomethylation of cysteine, and (iv) preferential charge state of ≤7+ was used. A protein was considered identified when Mascot listed it as a significant match/score (P < 0.05) with trypsin cleavage sites. The Mammalia database was chosen for identification of porcine proteins as the S. scrofa genome database is not well annotated. For protein identification we also required that the matched protein (based on accession number) had MS/MS fragmentation showing at least three consensus amino acid sequences.

Annotation of results

Proteins identified by Mascot were sorted by emPAI scores. Proteins were verified as integral membrane proteins by (i) inclusion of the terms, “single-pass membrane protein” or “multi-pass membrane protein” in UniProt, (ii) annotation of transmembrane topology in Uniprot, (iii) prediction of transmembrane alpha helices by TMHMM Server v.2.0, or (iv) prediction of transmembrane beta barrels by PRED-TMBB.51–53 Both TMHMM and PRED-TMBB utilize hidden Markov models and are among the most effective software packages for detecting transmembrane regions.

Expression, purification, and characterization of Yop1

Yop1 was expressed from plasmid pMCSG28 (Midwest Center for Structural Genomics) with a C-terminal His6-tag from E. coli Lemo21 cells (New England Biolabs). Expression was induced at 18°C with 0.1 mM IPTG and 0.5M rhamnose. Cells were harvested 16 h after expression was induced and were lysed by sonication. Membrane fractions were isolated by centrifugation and solubilized by gentle inversion for 2 h with buffer containing 0.1M HEPES, pH 7.4, 0.1 mM TCEP, 80 mM NaCl, 10% glycerol, and 1% DDM. Yop1 was purified by nickel affinity chromatography using HisPur NTA resin (Thermo Scientific) and eluted in buffer containing 0.1M HEPES, pH 7.4, 0.1 mM TCEP, 80 mM NaCl, 0.2M imidazole, and 0.1% DDM. Purified protein was analyzed by size-exclusion chromatography on an Akta Purifier with a Superdex 200 10/300 GL column with flow rate 0.5 mL/min. The elution profile of the column was calibrated against protein molecular weight standards from Sigma-Aldrich.

Acknowledgments

We thank Tu Kiet Lam and the staff of the Yale School of Medicine Keck Proteomics lab for performing the mass spectrometry analysis.

Supporting Information

Additional Supporting Information may be found in the online version of this article.

Supporting Information

pro0024-1756-sd1.docx (2.2MB, docx)

References

  1. White S. Membrane proteins of known 3D structure. Available from: http://blanco.biomol.uci.edu/mpstruc.
  2. Lee JK, Stroud RM. Unlocking the eukaryotic membrane protein structural proteome. Curr Opin Struct Biol. 2010;20:464–470. doi: 10.1016/j.sbi.2010.05.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Russell RB, Eggleston DS. New roles for structure in biology and drug discovery. Nat Struct Biol. 2000;7(Suppl):928–930. doi: 10.1038/80691. [DOI] [PubMed] [Google Scholar]
  4. Newby ZER, O'Connell JD, Gruswitz F, Hays FA, Harries WEC, Harwood IM, Ho JD, Lee JK, Savage DF, Miercke LJW. A general protocol for the crystallization of membrane proteins for X-ray structural investigation. Nat Protoc. 2009;4:619–637. doi: 10.1038/nprot.2009.27. , et al. ( [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Wiener MC. A pedestrian guide to membrane protein crystallization. Methods. 2004;34:364–372. doi: 10.1016/j.ymeth.2004.03.025. [DOI] [PubMed] [Google Scholar]
  6. Li M, Hays FA, Roe-Zurz Z, Vuong L, Kelly L, Ho C-M, Robbins RM, Pieper U, O'Connell JD, III, Miercke LJW. Selecting optimum eukaryotic integral membrane proteins for structure determination by rapid expression and solubilization screening. J Mol Biol. 2009;385:820–830. doi: 10.1016/j.jmb.2008.11.021. , et al. ( [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Carpenter EP, Beis K, Cameron AD, Iwata S. Overcoming the challenges of membrane protein crystallography. Curr Opin Struct Biol. 2008;18:581–586. doi: 10.1016/j.sbi.2008.07.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Savage DF, Egea PF, Robles-Colmenares Y, O'Connell JD, Stroud RM. Architecture and selectivity in aquaporins: 2.5 a X-ray structure of aquaporin Z. PLoS Biol. 2003;1:E72. doi: 10.1371/journal.pbio.0000072. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Veesler D, Blangy S, Cambillau C, Sciara G. There is a baby in the bath water: AcrB contamination is a major problem in membrane-protein crystallization. Acta Crystallogr Sect F Struct Biol Cryst Commun. 2008;64:880–885. doi: 10.1107/S1744309108028248. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Psakis G, Polaczek J, Essen L-O. AcrB et al.: obstinate contaminants in a picogram scale. One more bottleneck in the membrane protein structure pipeline. J Struct Biol. 2009;166:107–111. doi: 10.1016/j.jsb.2008.12.007. [DOI] [PubMed] [Google Scholar]
  11. Cowan SW, Garavito RM, Jansonius JN, Jenkins JA, Karlsson R, König N, Pai EF, Pauptit RA, Rizkallah PJ, Rosenbusch JP. The structure of OmpF porin in a tetragonal crystal form. Structure. 1995;3:1041–1050. doi: 10.1016/s0969-2126(01)00240-4. [DOI] [PubMed] [Google Scholar]
  12. Cuesta-Seijo JA, Neale C, Khan MA, Moktar J, Tran CD, Bishop RE, Pomès R, Privé GG. PagP crystallized from SDS/cosolvent reveals the route for phospholipid access to the hydrocarbon ruler. Structure. 2010;18:1210–1219. doi: 10.1016/j.str.2010.06.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Hattori M, Hibbs RE, Gouaux E. A fluorescence-detection size-exclusion chromatography-based thermostability assay for membrane protein precrystallization screening. Structure. 2012;20:1293–1299. doi: 10.1016/j.str.2012.06.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Fan J, Huang B, Wang X, Zhang XC. Thermal precipitation fluorescence assay for protein stability screening. J Struct Biol. 2011;175:465–468. doi: 10.1016/j.jsb.2011.05.003. [DOI] [PubMed] [Google Scholar]
  15. Tomasiak TM, Pedersen BP, Chaudhary S, Rodriguez A, Colmanares YR, Roe-Zurz Z, Thamminana S, Tessema M, Stroud RM. General qPCR and plate reader methods for rapid optimization of membrane protein purification and crystallization using thermostability assays. Current Protocols in Protein Science. 2001;77:29.11:29.11.1–29.11.14. doi: 10.1002/0471140864.ps2911s77. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Löw C, Moberg P, Quistgaard EM, Hedrén M, Guettou F, Frauenfeld J, Haneskog L, Nordlund P. High-throughput analytical gel filtration screening of integral membrane proteins for structural studies. Biochim Biophys Acta. 2013;1830:3497–3508. doi: 10.1016/j.bbagen.2013.02.001. [DOI] [PubMed] [Google Scholar]
  17. Cornvik T, Dahlroth S-L, Magnusdottir A, Herman MD, Knaust R, Ekberg M, Nordlund P. Colony filtration blot: a new screening method for soluble protein expression in Escherichia coli. Nat Methods. 2005;2:507–509. doi: 10.1038/nmeth767. [DOI] [PubMed] [Google Scholar]
  18. Galea CA, High A, Obenauer JC, Mishra A, Park C-G, Punta M, Schlessinger A, Ma J, Rost B, Slaughter CA. Large-scale analysis of thermo-stable, mammalian proteins provides insights into the intrinsically disordered proteome. J Proteome Res. 2009;8:211–226. doi: 10.1021/pr800308v. , et al. ( [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kwon S, Jung Y, Lim D. Proteomic analysis of heat-stable proteins in Escherichia coli. BMB Rep. 2008;41:108–111. doi: 10.5483/bmbrep.2008.41.2.108. [DOI] [PubMed] [Google Scholar]
  20. Prosinecki V, Botelho HM, Francese S, Mastrobuoni G, Moneti G, Urich T, Kletzin A, Gomes CM. A proteomic approach toward the selection of proteins with enhanced intrinsic conformational stability. J Proteome Res. 2006;5:2720–2726. doi: 10.1021/pr0602491. [DOI] [PubMed] [Google Scholar]
  21. Wang W, Wu X, Xiong E, Tai F. Improving gel-based proteome analysis of soluble protein extracts by heat prefractionation. Proteomics. 2012;12:938–943. doi: 10.1002/pmic.201100475. [DOI] [PubMed] [Google Scholar]
  22. Yun S-H, Choi C-W, Kwon SO, Lee YG, Chung Y-H, Jung HJ, Kim Y-J, Lee J-H, Choi J-S, Kim S. Enrichment and proteome analysis of a hyperthermostable protein set of archaeon Thermococcus onnurineus NA1. Extremophiles. 2011;15:451–461. doi: 10.1007/s00792-011-0376-1. , et al. ( [DOI] [PubMed] [Google Scholar]
  23. Privé GG. Detergents for the stabilization and crystallization of membrane proteins. Methods. 2007;41:388–397. doi: 10.1016/j.ymeth.2007.01.007. [DOI] [PubMed] [Google Scholar]
  24. Bondos SE, Bicknell A. Detection and prevention of protein aggregation before, during, and after purification. Anal Biochem. 2003;316:223–231. doi: 10.1016/s0003-2697(03)00059-9. [DOI] [PubMed] [Google Scholar]
  25. Vergis JM, Purdy MD, Wiener MC. A high-throughput differential filtration assay to screen and select detergents for membrane proteins. Anal Biochem. 2010;407:1–11. doi: 10.1016/j.ab.2010.07.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Niegowski D, Hedrén M, Nordlund P, Eshaghi S. A simple strategy towards membrane protein purification and crystallization. Int J Biol Macromol. 2006;39:83–87. doi: 10.1016/j.ijbiomac.2006.02.011. [DOI] [PubMed] [Google Scholar]
  27. Mellacheruvu D, Wright Z, Couzens AL, Lambert J-P, St-Denis NA, Li T, Miteva YV, Hauri S, Sardiu ME, Low TY. The CRAPome: a contaminant repository for affinity purification-mass spectrometry data. Nat Methods. 2013;10:730–736. doi: 10.1038/nmeth.2557. , et al. ( [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Wang P, Bouwman FG, Mariman ECM. Generally detected proteins in comparative proteomics – a matter of cellular stress response? Proteomics. 2009;9:2955–2966. doi: 10.1002/pmic.200800826. [DOI] [PubMed] [Google Scholar]
  29. Daley DO, Rapp M, Granseth E, Melén K, Drew D, von Heijne G. Global topology analysis of the Escherichia coli inner membrane proteome. Science. 2005;308:1321–1323. doi: 10.1126/science.1109730. [DOI] [PubMed] [Google Scholar]
  30. Weiner JH, Li L. Proteome of the Escherichia coli envelope and technological challenges in membrane proteome analysis. Biochim Biophys Acta. 2008;1778:1698–1713. doi: 10.1016/j.bbamem.2007.07.020. [DOI] [PubMed] [Google Scholar]
  31. Shinoda K, Tomita M, Ishihama Y. emPAI Calc—for the estimation of protein abundance from large-scale identification data by liquid chromatography-tandem mass spectrometry. Bioinformatics. 2010;26:576–577. doi: 10.1093/bioinformatics/btp700. [DOI] [PubMed] [Google Scholar]
  32. Kanapin A, Batalov S, Davis MJ, Gough J, Grimmond S, Kawaji H, Magrane M, Matsuda H, Schönbach C, Teasdale RD. Mouse proteome analysis. Genome Res. 2003;13:1335–1344. doi: 10.1101/gr.978703. , et al. ( [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, Davis AP, Dolinski K, Dwight SS, Eppig JT. Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet. 2000;25:25–29. doi: 10.1038/75556. , et al. ( [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Hu J, Shibata Y, Voss C, Shemesh T, Li Z, Coughlin M, Kozlov MM, Rapoport TA, Prinz WA. Membrane proteins of the endoplasmic reticulum induce high-curvature tubules. Science. 2008;319:1247–1250. doi: 10.1126/science.1153634. [DOI] [PubMed] [Google Scholar]
  35. Shibata Y, Voss C, Rist JM, Hu J, Rapoport TA, Prinz WA, Voeltz GK. The Reticulon and Dp1/Yop1p proteins form immobile oligomers in the tubular endoplasmic reticulum. J Biol Chem. 2008;283:18892–18904. doi: 10.1074/jbc.M800986200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Blattner FR, Plunkett G, Bloch CA, Perna NT, Burland V, Riley M, Collado-Vides J, Glasner JD, Rode CK, Mayhew GF. The complete genome sequence of Escherichia coli K-12. Science. 1997;277:1453–1462. doi: 10.1126/science.277.5331.1453. , et al. ( [DOI] [PubMed] [Google Scholar]
  37. Wu F, Sun D, Wang N, Gong Y, Li L. Comparison of surfactant-assisted shotgun methods using acid-labile surfactants and sodium dodecyl sulfate for membrane proteome analysis. Anal Chim Acta. 2011;698:36–43. doi: 10.1016/j.aca.2011.04.039. [DOI] [PubMed] [Google Scholar]
  38. Cowieson NP, Wensley B, Listwan P, Hume DA, Kobe B, Martin JL. An automatable screen for the rapid identification of proteins amenable to refolding. Proteomics. 2006;6:1750–1757. doi: 10.1002/pmic.200500056. [DOI] [PubMed] [Google Scholar]
  39. Oganesyan N, Kim S-H, Kim R. On-column protein refolding for crystallization. J Struct Funct Genomics. 2005;6:177–182. doi: 10.1007/s10969-005-2827-3. [DOI] [PubMed] [Google Scholar]
  40. Slabinski L, Jaroszewski L, Rychlewski L, Wilson IA, Lesley SA, Godzik A. XtalPred: a web server for prediction of protein crystallizability. Bioinformatics. 2007;23:3403–3405. doi: 10.1093/bioinformatics/btm477. [DOI] [PubMed] [Google Scholar]
  41. Overton IM, van Niekerk CAJ, Barton GJ. XANNpred: Neural nets that predict the propensity of a protein to yield diffraction-quality crystals. Proteins Struct Funct Bioinforma. 2011;79:1027–1033. doi: 10.1002/prot.22914. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Mizianty MJ, Kurgan L. Meta prediction of protein crystallization propensity. Biochem Biophys Res Commun. 2009;390:10–15. doi: 10.1016/j.bbrc.2009.09.036. [DOI] [PubMed] [Google Scholar]
  43. Marani P, Wagner S, Baars L, Genevaux P, de Gier J-W, Nilsson I, Casadio R, von Heijne G. New Escherichia coli outer membrane proteins identified through prediction and experimental verification. Protein Sci. 2006;15:884–889. doi: 10.1110/ps.051889506. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Papanastasiou M, Orfanoudaki G, Koukaki M, Kountourakis N, Sardis MF, Aivaliotis M, Karamanou S, Economou A. The Escherichia coli peripheral inner membrane proteome. Mol Cell Proteomics. 2013;12:599–610. doi: 10.1074/mcp.M112.024711. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Lewinson O, Lee AT, Rees DC. The funnel approach to the precrystallization production of membrane proteins. J Mol Biol. 2008;377:62–73. doi: 10.1016/j.jmb.2007.12.059. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Loll PJ. Membrane proteins, detergents and crystals: what is the state of the art? Acta Crystallogr Sect F Struct Biol Commun. 2014;70:1576–1583. doi: 10.1107/S2053230X14025035. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Hebert AS, Richards AL, Bailey DJ, Ulbrich A, Coughlin EE, Westphall MS, Coon JJ. The one hour yeast proteome. Mol Cell Proteomics. 2014;13:339–347. doi: 10.1074/mcp.M113.034769. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Wessel D, Flügge UI. A method for the quantitative recovery of protein in dilute solution in the presence of detergents and lipids. Anal Biochem. 1984;138:141–143. doi: 10.1016/0003-2697(84)90782-6. [DOI] [PubMed] [Google Scholar]
  49. Wiśniewski JR, Zougman A, Nagaraj N, Mann M. Universal sample preparation method for proteome analysis. Nat. Methods. 2009;6:359–362. doi: 10.1038/nmeth.1322. [DOI] [PubMed] [Google Scholar]
  50. Antharavally BS, Mallia KA, Rosenblatt MM, Salunkhe AM, Rogers JC, Haney P, Haghdoost N. Efficient removal of detergents from proteins and peptides in a spin column format. Anal Biochem. 2011;416:39–44. doi: 10.1016/j.ab.2011.05.013. [DOI] [PubMed] [Google Scholar]
  51. Consortium TU. UniProt: a hub for protein information. Nucleic Acids Res. 2015;43:D204–D212. doi: 10.1093/nar/gku989. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Bagos PG, Liakopoulos TD, Spyropoulos IC, Hamodrakas SJ. PRED-TMBB: a web server for predicting the topology of beta-barrel outer membrane proteins. Nucleic Acids Res. 2004;32:W400–W404. doi: 10.1093/nar/gkh417. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Krogh A, Larsson B, von Heijne G, Sonnhammer EL. Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes. J Mol Biol. 2001;305:567–580. doi: 10.1006/jmbi.2000.4315. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting Information

pro0024-1756-sd1.docx (2.2MB, docx)

Articles from Protein Science : A Publication of the Protein Society are provided here courtesy of The Protein Society

RESOURCES