Here, we present a robust and optimized protocol for the production of milligram quantities of native, tag-free monomers and fibrils of the exon1 of the Huntingtin protein (Httex1) based on the transient fusion of small ubiquitin related modifier (SUMO).
Huntington's Disease (HD) is an inherited fatal neurodegenerative disease caused by a CAG expansion (≥36) in the first exon of the HD gene, resulting in the expression of the Huntingtin protein (Htt) or N-terminal fragments thereof with an expanded polyglutamine (polyQ) stretch.
The exon1 of the Huntingtin protein (Httex1) is the smallest Htt fragment that recapitulates many of the features of HD in cellular and animal models and is one of the most widely studied fragments of Htt. The small size of Httex1 makes it experimentally more amenable to biophysical characterization using standard and high-resolution techniques in comparison to longer fragments or full-length Htt. However, the high aggregation propensity of mutant Httex1 (mHttex1) with increased polyQ content (≥42) has made it difficult to develop efficient expression and purification systems to produce these proteins in sufficient quantities and make them accessible to scientists from different disciplines without the use of fusion proteins or other strategies that alter the native sequence of the protein. We present here a robust and optimized method for the production of milligram quantities of native, tag-free Httex1 based on the transient fusion of small ubiquitin related modifier (SUMO). The simplicity and efficiency of the strategy will eliminate the need to use non-native sequences of Httex1, thus making this protein more accessible to researchers and improving the reproducibility of experiments across different laboratories. We believe that these advances will also facilitate future studies aimed at elucidating the structure-function relationship of Htt as well as developing novel diagnostic tools and therapies to treat or slow the progression of HD.
Htt is a 348 kDa protein and has been implicated in several physiological functions1. When Htt contains an expanded polyQ region of more than 36 residues in its N-terminus, it causes HD2,3. HD pathology is characterized by cellular inclusions in the striatum and cortex, which leads to neuronal death and atrophy of the affected tissues4,5. Several N-terminal Htt fragments that contain the polyQ repeat tract have been detected in post-mortem brains from HD patients and are thought to be generated by proteolytic processing of the huntingtin protein6. Recent studies suggest that Httex1 could also be formed due to aberrant mRNA splicing. Httex1 contains the pathological polyQ mutation and its overexpression in animals can recapitulate many of the key features of HD7, thus highlighting a possible central role of this fragment in HD pathology and disease progression6,8,9.
Due to the high aggregation propensity of mutant Httex1 (mHttex1) with expanded polyQ tract, the majority of existing expression systems are based on the transient fusion of Httex1 to proteins (such as glutathione-S-transferase (GST), thioredoxin (TRX) or maltose-binding-protein (MBP) and/or peptides (poly-histidine) that differentially improve its expression, stability, purification and/or solubility10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28. The fusion partner is linked to Httex1 with a short sequence containing a cleavage site for proteases such as trypsin, tabacco etch virus (TEV) protease or PreScission to allow for the cleavage and release of Httex1 prior to the initiation of aggregation or purification. Shortcomings of these methods include the possibility of leaving additional residues due to non-traceless cleavage and the creation of truncated fragments due to miscleavage within the sequence of Httex1, in addition to heterogeneity due to incomplete cleavage (see Vieweg et al. for more in-depth discussion on the advantages and limitations of this approach)10. To address these limitations, we recently developed an expression strategy enabling the generation of tag-free native Httex1 for the first time by utilizing a transient N-terminal fusion of the Synechocystis sp. (Ssp) DnaB intein to Httex110. While the intein cleavage is traceless and specific and yields mg quantity of proteins, it still suffers two drawbacks that could reduce the yield: namely, premature cleavage of the intein which can occur during the expression, and the fact that cleavage occurs over several hours, which could lead to loss of protein due to aggregation, especially for Httex1 with expanded polyQ repeats.
To address these limitations and to refine our strategy for the production of native, tag-free Httex1, we developed a new expression system based on the transient fusion of SUMO, more exactly the yeast homolog Smt3 to Httex1. The application of the SUMO system for the production of recombinant proteins was first published in 200429, where an increased rate of expression and solubility of SUMO fusion protein was demonstrated. The SUMO tag can be cleaved by the ubiquitin like protein-specific protease 1 (ULP1), which does not require a recognition site, but recognizes the tertiary structure of SUMO and practically eliminates the possibility of miscleavage30. Furthermore, the ULP1-mediated cleavage is fast and traceless and does not leave additional residues behind. The premature cleavage of the fusion tag, as observed with the autocatalytic intein10, is completely avoided by the requirement of an external protease. While the SUMO strategy is nowadays widely used for recombinant protein production31,32,33, we demonstrate in this paper that it is especially useful for the generation of an intrinsically disordered, aggregation-prone, amyloidogenic protein such as Httex1. We believe that the simplicity, efficiency and robustness of our SUMO-fusion-based method will make native, tag-free Httex1 more accessible to researchers from different disciplines and eliminate the need to use non-native sequences of Httex1 in vitro. This is an important advance that will facilitate future studies to elucidate the structure-function relationship of Httex1.
The protocol describes the purification of Httex1 from 12 L of bacterial culture, but the protocol could be easily adapted for smaller or larger scale productions. The protocol describes the production of wild type Httex1 (wtHttex1) with a polyQ repeat length below (23Q) and mutant Httex1 (mHttex1) with a polyQ repeat length above (43Q) the pathogenic threshold (36Q).
1. Expression of Recombinant Httex1 23Q and 43Q
2. Cell Lysis and Purification of His6-SUMO Httex1 Fusion Protein by Immobilized Metal Affinity Chromatography (IMAC)
3. Cleavage of the His6-SUMO-tag and HPLC Purification
Caution: Trifluoroacetic acid (TFA) is a volatile liquid and can cause severe burns so handle with care. Carry out all handling in a fume hood and wear adequate personal protective equipment (i.e., disposable nitrile gloves, safety glasses and a lab coat).
4. Disaggregation and Resolubilization of Httex1 Proteins
Caution: TFA is a volatile liquid and can cause severe burns so handle with care. Carry out all handling in a fume hood and wear adequate personal protective equipment (i.e. disposable nitrile gloves, safety glasses and a lab coat).
5. Monitoring of the Aggregation Kinetics of Httex1 43Q using UPLC and circular Dichroism (CD) Spectroscopy and Characterization of the Aggregates by Transmission Electron Microscopy (TEM)
Httex1 is expressed in E. coli with an N-terminal His6-SUMO tag. The representative results of the expression and purification of the fusion protein are summarized in Figure 1. The sequence of Httex1 consists of the residues 2-90 of Htt and starts with Ala2, because Met1 is fully cleaved in vivo42. The numbering of the amino acids refers to the 23Q variant, the complete sequence of the expressed fusion protein is shown in Figure 1A. The plasmids will be deposited at Addgene in the near future to be shared with the community. A schematic of the plasmid and the expressed fusion protein is shown in Figure 1B. His6-SUMO Httex1 expresses at a medium level (Figure 1C) and most of the fusion protein is present in the soluble fraction after lysis, both for the 23Q and the 43Q variant. The fusion protein migrates higher than expected, based on the molecular weight. This is partly due to the strong fold of SUMO but mostly due to the unusual sequence composition of Httex1, containing mainly glutamine and proline residues. Both the wildtype (23Q) and the mutant (43Q) fusion protein can be enriched to ~80% purity by IMAC (Figure 1D) The presence of co-purifying host protein can be explained by the comparatively low expression level of Httex1 and the big sample volume applied to the column.
The cleavage of the His6-SUMO tag and the purification of Httex1 is shown in Figure 2A. UPLC is an efficient tool to monitor the cleavage of the His6-SUMO tag (Figure 2B). The original peak of the fusion protein is consumed and two new and well-separated peaks corresponding the His6-SUMO tag and Httex1 appear. The cleavage reaction is finished in 10-20 min. The Western Blot (WB) is too slow to monitor the cleavage reaction efficiently, but it has been included in the figure for reference and to demonstrate the completeness of the SUMO cleavage. Both Httex1 23Q and 43Q can be separated well from the His6-SUMO tag by RP-HPLC (Figure 2C) and were obtained in high purity as shown by UPLC, MS and SDS-PAGE analysis (Figure 2D).
To illustrate that the Httex1 proteins prepared by this method retain the expected aggregation properties of Httex1, we assessed the fibrillization kinetics of mutant Httex1 at 37 °C by a sedimentation assay, monitored the changes in secondary structure by CD spectroscopy, and characterized the morphology of the aggregates by TEM. A representative data set of the aggregation kinetics of mHttex1 fibril formation as determined by a sedimentation assay is shown inFigure 3A. The loss of soluble Httex1 43Q over time, due to fibril formation was quantified by UPLC. We observe a complete depletion of soluble protein after 48 hours of incubation. Additionally, we determined the secondary structure of the protein by CD spectroscopy (Figure 3B). Httex1 43Q shifts from unstructured (λmin 205 nm) to mainly β-sheet rich conformation (λmin 215 nm) after 48 hours of incubation. This structural change is accompanied by the formation of long fibrillar aggregates as observable by TEM at 48 hours (Figure 3C).
Figure 1. Expression and purification of the His6-SUMO Httex1 fusion protein.
(A) The amino acid sequence of the His6-SUMO-Httex1-QN fusion constructs (His6-SUMO in green and Httex1-QN in orange); (B) Schematic overview of the expression and purification of the fusion protein; (C) SDS-PAGE analysis of the expression and the solubility of the fusion protein after lysis; (D) Chromatogram of the IMAC purification of the fusion protein and analysis of the fractions by SDS-PAGE (red bar: unbound fraction, blue bar: wash fraction, black bar: fractions containing the elution peak); Please click here to view a larger version of this figure.
Figure 2. Cleavage of the His6 SUMO tag and purification of tag-free Httex1-QN proteins.
(A) schematic overview; (B) analysis of the cleavage of the SUMO tag with ULP1 by UPLC (blue: before addition of ULP1; black: 20 min (23Q), respectively 10 min (43Q) after addition of ULP1) and WB (MAB5492 1:2000, secondary goat anti mouse antibody 1:5000); (C) Chromatogram of the preparative RP-HPLC purification of Httex1; D: analysis of the purified Httex1 by UPLC, SDS-PAGE and ESI-MS; the expected molecular weight is 9943 Da (23Q) and 12506 Da (43Q) respectively. Please click here to view a larger version of this figure.
Figure 3. Aggregation of Httex1-43Q: (A) Sedimentation assay based on UPLC. (B) CD spectra of the secondary structure at 0 h and 48 h. (C) TEM micrographs of the aggregates at 48 h (scale bars are 200 nm). Please click here to view a larger version of this figure.
In this protocol, we have outlined an efficient procedure for obtaining milligram quantities of native, untagged Httex1 containing 23 or 43 glutamine residues. This was achieved by expressing Httex1 as a C-terminal fusion to a His6-SUMO tag, which is used to isolate the fusion protein from the cell lysate by IMAC and is cleaved prior to HPLC purification of Httex1. While the SUMO strategy has been used in the production of several other proteins, our method shows that the unique properties SUMO could also be used to generate intrinsically disordered, aggregation-prone, amyloidogenic protein that have previously proved to be extremely difficult to handle and produce43,44. We present a protocol that is straightforward, easy to use and comparable to a protocol for the generation of a "well-behaved" protein. The SUMO fusion solubilizes and stabilizes Httex1 during expression and the IMAC purification step. Premature cleavage of the tag, as observed with the intein strategy10 and aggregation were no longer an issue.
Intrinsically disordered proteins are especially vulnerable to degradation. While N-terminal degradation in the N17 region is not an issue using this protocol, truncations in the PRD of Httex1 can occur. As the truncated proteins are very similar to Httex1 in hydrophobicity, charge and size, removing them by chromatographic means is challenging, thus it is best to prevent their formation in the first place. Sticking closely to the protocol, always working on ice and using a sufficient amount of protease inhibitor should help keep the level of observed truncation very low. Applying a fusion tag at the C-terminus of Httex1 could remove truncations in the PRD easily as the truncated protein would lose the affinity tag as well. However, if the native sequence needs to be maintained this option cannot be applied as Httex1 ends with proline and to the best of our knowledge there are no C-terminal fusion tags that are known to induce traceless and efficient cleavage after proline.
The most critical part of the protocol is the handling of the Httex1 liberated after cleavage of the SUMO tag by ULP1. The protein should be purified immediately by RP-HPLC. Fortunately, this is an efficient and fast reaction that is usually completed in 10-20 min at 4 °C. In contrast, the intein strategy required several hours for complete cleavage of the intein, thus requiring a trade-off between incomplete cleavage and beginning aggregation in order to maximize the yield. A fast workup is required for mutant Httex1, as it will start to aggregate at the comparatively high concentration present in the cleavage reaction, whereas the 23Q variant is stable for a longer time. During the RP-HPLC purification, another advantage of SUMO becomes apparent: While the Ssp DnaB intein is hydrophobic and sticks strongly to the column, SUMO is more hydrophilic and elutes completely from the C4 reversed-phase column. Although commercial ULP1 is quite costly, the protein can be easily produced in high yield following previously published protocols29.
The critical importance of applying a disaggregation protocol prior to using Httex1 cannot be stressed enough. Lyophilized polyQ proteins such as Httex1 are stable and can be stored long periods, but are not completely soluble in water and buffers. The presence of preformed oligomers or fibrils could have a significant impact on aggregation kinetics and biophysical properties of the protein45. The disaggregation protocol described here allows the disaggregation of the protein, removal of preformed aggregates and generation of a solution of monomeric Httex1 from a lyophilized sample. We observed similar aggregation kinetics and fibril morphology for Httex1 obtained with the SUMO and the intein strategy.
Compared to previous methods for producing Httex1, the SUMO strategy described here offers several advantages and expands the range of possible studies to investigate the structure and functional properties of this protein in health and disease. The SUMO-Httex1 fusion protein is easy to handle, it can be frozen and stored or kept in solution for 24 h at ambient temperature, while the free mHttex1 would aggregate quickly. The stability and high solubility of the SUMO-Httex1 fusion proteins provide greater flexibility to manipulate the protein and/or introduce enzymatic and chemical modifications into mHttex1 that would otherwise not be possible after cleavage. This includes the introduction of post-translational modifications, fluorophores, spin labels, biotin tags, etc. The advances presented here should 1) facilitate future studies to elucidate structure-function relationships of Httex1; 2) generate new tools to investigate Htt aggregation and pathology spreading; 3) enable the development of new assays to identify molecules that stabilize mutant Httex1 and prevent its aggregation; and 4) encourage scientists from other fields to bring to work on this protein and join our quest to find cures for Huntington's disease.
The authors have nothing to disclose.
This work was funded primarily by grants from the CHDI foundation and the Swiss National Science Foundation. We thank Dr. Sophie Vieweg for useful discussions during the development of this new expression system and other members of the Lashuel group for sharing their experience with this expression system and for their input and valuable feedback. We also thank Prof. Oliver Hantschel for providing the ULP1 plasmid. The authors thank Dr. John B. Warner and Dr. Senthil K. Thangaraj for critical review of the manuscript
Uranyl formate (UO2(CHO2)2) | EMS | 22450 | |
Formvar/carbon 200 mesh, Cu 50 grids | EMS | FCF200-Cu-50 | |
High Precision Cell made of Quartz SUPRASIL 1 mm light path from Hellma Analytics | HellmaAnalytics | 110-1-40 | |
Buffer Substance Dulbecco's (PBS w/o Ca and Mg) ancinne ref 47302 (RT) SERVA | Witech | SVA4730203 | |
Ampicillin | AxonLab | A0839.0100 | |
Luria Broth (Miller's LB Broth) | Chemie Brunschwig | 1551.05 | |
Isopropyl β-D-1-thiogalactopyranoside (IPTG) | AxonLab | A1008.0025 | |
E. coli B ER2566 | NEB | NEB# E4130 | |
Imidazole | Sigma | 56750-500G | |
cOmplete Protease Inhibitor Cocktail | Roche | 4693116001 | |
Anti-Huntingtin Antibody, a.a. 1-82 | Merck Millipore Corporation | MAB5492 | |
IRDye 680RD Goat anti-Mouse IgG (H + L) | Licor | 925-68070 | |
PMSF | AxonLab | A0999.0005 | |
HisPrep 16/10 column | GE Healthcare | 28936551 | |
C4 HPLC column | Phenomenex | 00G-4168.P0 | 10 µm C4 300 Å, LC Column 250 x 21.2 mm, Phenomenex, 19×10 mm guard column, not temperature jacketed |
Acetonitrile HPLC | MachereyNagel | C2502 | |
Filtre seringue Filtropur S 0,45 ul sans prefiltre sterile | Sarstedt AG | 83.1826 | |
Spectrophotometer semi-micro cuvette | Reactolab S.A. | 2534 | |
Superloop, 1/16" fittings (ÄKTAdesign), 50 ml | GE Healthcare | 18111382 | |
Trifluoroacetic acid | Sigma | 302031 | |
GREINER Tubes fo FPLC 16 x 100 mm, cap. 12.0 ml | Greiner Bio-One | 7.160 102 | |
100 kD Microcon fast flow filters | Merck Millipore Corporation | MRCF0R100 | |
Vibra-cell VCX130 ultrasonic liquid processor | Sonics | ||
Äkta 900 equipped with a fraction collector | GE Healthcare | ||
Jasco J-815 Circular Dichroism | Jasco | ||
Waters UPLC system | Waters | C8 BEH acquity 2.1×100 mm 1.7 micron column , preheated column (40 °C), flow rate of 0.6 mL/min, injection volume of 4 μL | |
waters HPLC system | Waters | 2489 UV detector and 2535 quaternary gradient module, 20 mL loop in a FlexInject housing | |
ESI-MS: Finnigan LTQ | Thermo Fisher Scientific | ||
lyophylizer instrument | FreeZone 2.5 Plus | ||
Tecnai Spirit BioTWIN | FEI | electron microscope equipped with a LaB6 gun and a 4K x 4K FEI Eagle CCD camera (FEI) and operated at 80 kV | |
37 °C shaking incubator | Infors HT multitron Standard | ||
Biophotometer plus | Eppendorf |