Structure Determination of Membrane Proteins Using X-Ray Crystallography

Abstract

Membrane proteins serve essential roles in all aspects of life and make up roughly one-third of all genomes from prokaryotes to eukaryotes. Their responsibilities include mediating cell signaling, nutrient import, waste export, cellular communication, trafficking, and immunity. For their critical role in many cellular processes, membrane proteins serve as targets for up to 50% of drugs currently on the market and remain primary targets in new therapeutics being developed. Despite their importance and abundance in nature, only ~1% of structures in the Protein Data Bank are of transmembrane proteins. This discrepancy can be directly attributed to the biochemical properties of membrane proteins and the difficulty in producing sufficient yields for structural studies or the difficulty in growing well-ordered crystals. Here, we present methods from our work that outline our general pipeline from cloning to structure determination of membrane proteins, with a focus on using X-ray crystallography, which still yields ~90% of all structures being deposited into the Protein Data Bank.

1. Introduction

Transmembrane proteins can have either an α-helical or β-barrel membrane domain [1–7]. Both types serve similar roles in the cell, however, an overarching difference between the two types is that while α-helical membrane proteins can be found in all cellular membranes from prokaryotes to eukaryotes, β-barrel membrane proteins are only found in the outer membranes of Gram-negative bacteria, mitochondria, and plastids such as chloroplasts [3,5,8]. This chapter will be mostly relevant to membrane proteins expressed in bacteria with a focus on the structure determination of β-barrel membrane proteins, however, the methods can be easily adapted to α-helical membrane proteins and for other expression systems (Figure 1).

Figure 1. The structure determination pipeline for membrane proteins using X-ray crystallography.

A. Examples of α-helical and β-barrel membrane proteins. B. Highlighted are the two bottlenecks in the pipeline (1) expression/purification of the protein target and (2) growth of well diffracting crystals. A ‘feedback loop’ would be necessary when crystallization fails to produce well-ordered crystals and a new construct must be explored.

Many aspects of protein expression, purification, and crystallization may be addressed by considering them in the initial cloning of the protein (Figure 2). Cloning choices that affect the expression of the protein include the construct size, the presence of tags or fusions, whether a native or denatured state is desired, and how the expression will be induced [9–11]. In instances where the entire protein cannot be successfully expressed or crystallized, truncations of the protein, such as those that omit a disordered portion of the protein, may yield better results. Other truncations or fragments simply generate more manageable pieces of the protein specifically for structure determination.

Figure 2. Considerations and strategies for cloning membrane proteins for structure determination.

A. Like with soluble proteins (top), there are many factors to consider when studying membrane proteins. Additional factors (bottom) must also be considered that further complicate production of these families of proteins. B. For cloning, one must carefully consider the expression system(s) that will be tested and an appropriate vector for delivering the gene and induction system. Each system has advantages and disadvantages that must be factored into the decision-making process.

The types of affinity tags, localization signal sequences, and protease sites encoded in a vector may affect both the expression and purification of the protein. Affinity tags may be moved from the N-terminus to the C-terminus if needed and can sometimes make a drastic difference in expression levels and stability. Commonly used affinity tags include 6X-or 10X-histidine, GST, and Strep. To aid its crystallization, a protein may be cloned as a fusion with another, easily crystalized protein fragment, such as lysozyme or maltose binding protein [12,10]. Additionally, mutations can be engineered to reduce flexibility of the protein, increasing its thermostability [13,14]. When working with membrane proteins, homologs and orthologs of the protein of interest are often screened for levels of expression, thermostability, mobility within a detergent or lipid, and initial crystallization screening to determine the most promising candidate(s).

In our chapter here, we will use an example protocol for the cloning to structure determination of BamA from Neisseria gonorrhoeae [15], however, the protocols can be easily adapted to other membrane proteins of interest. Here, the pET20b vector was modified to include an N-terminal 10xHis-tag followed by a TEV protease site prior to the commercial NcoI site. The TEV protease site may be replaced with an enterokinase or thrombin site, if preferred [16]. A short list of common vectors commonly used for the production of membrane proteins include the pET family of vectors (specifically pET20b, pET-DUET, and pET26), pRSF-1b, pCDF-1b, pBAD, pTrc, pMAL, pGEX, pASG, and pHIS2-parallel. Primers may be designed to incorporate various restriction enzyme sequences onto the ends of the construct to accommodate the process of restriction enzyme digestion and ligation. Alternatively, other cloning strategies may be used including Golden Gate Assembly, Gibson Assembly, Gateway Cloning, and ligation-independent cloning.

Once an expression system has been chosen and a suitable clone or library of clones have been prepared, the next step is to transform the plasmids into our desired cell type which we will first use to verify expression of the gene in small-scale assays. Once expression is confirmed, we can proceed with large-scale expression. There are many factors affecting the conditions for optimal large-scale expression yields [9,17,18] (Figure 3). Many of these include the cell type, medium being used, and expression temperature, all of which can be optimized during small-scale expression tests. There are other considerations that will be made in the interest of not only maximizing overall protein expression levels, but also maximizing the final yield of active, natively folded protein. With these interests in mind, large-scale expression of NgBamA was performed at low temperature (20°C) without IPTG induction [15,19]. Cell lines such as BL21(DE3) often exhibit ‘leaky’ expression of the recombinant gene, meaning the T7 polymerase is expressed at low yields even in the absence of IPTG yielding low levels of recombinant protein. Both the low temperature and the lack of induction have the effect of slowing the production of the protein of interest and in the case of membrane proteins, this can be advantageous for allowing sufficient time to complete transit and folding into the membrane. Alternatively, the use of autoinduction media is also commonly used in the expression of membrane proteins [20,21]. Media can also be an important consideration for achieving high protein yields [17,22]. For the expression of NgBamA, Terrific Broth (TB) was used since it can achieve very high cell densities ranging from an OD₆₀₀ of 10-20 and yielding 30-40 g of cells per liter of medium. Whereas less rich media like 2XYT will produce ~5 g of cells per liter of medium. In general, it may be advantageous to grow in richer media to increase protein production, but it is also important to consider that richer media contains a greater variety of nutrients which may activate the expression of a greater variety of genes. This may interfere with the production of your protein or cause difficulty with protein purification due to the presence of additional impurities.

Figure 3. Protein expression flow diagram.

The diagram can be used to guide the decision-making process when expression membrane proteins, particularly in bacterial systems. Colony screening is sometimes necessary for difficult targets while easily expressed targets behave much like well-behaved soluble protein samples.

We will describe the following procedure using BL21(DE3) cells which utilize the T7 expression system allowing IPTG-mediated inducible expression [23]. This cell type also lacks proteases such as OmpT and is therefore sufficient for expressing non-toxic genes. However, in the event of low or no expression, we suggest trying alternate expression systems. Rosetta or CodonPlus cells can be used to express genes with rare codons. C41 or C43 cells can be used to express toxic genes or many genes from other organisms [24,25]. In the event that expression in several alternate cell lines fails, we suggest revisiting the vector selection.

The purification of membrane proteins is uniquely distinguished from the purification of soluble proteins and careful consideration must be applied through the process [19,17,11,24]. In our example, NgBamA was cloned into the pET20b-HT vector which contains an N-terminal pelB signal sequence, routing it across the inner membrane where is then folded and inserted into the outer membrane by the β-barrel assembly machinery [26]. Therefore, to purify membrane-embedded proteins such as NgBamA, the cellular membranes must be isolated, and the protein solubilized (released). The most obvious and important distinction from soluble protein purification is the requirement of a detergent or detergent/lipid mix to stabilize the hydrophobic, transmembrane regions of the protein. Most commonly, membranes are isolated by centrifugation and the membranes are then solubilized in a buffer supplemented with detergent (Figure 4). Throughout the purification process, it is important to keep the detergent above the critical micelle concentration (CMC), so the protein does not aggregate or precipitate out of solution [27–29].

Figure 4. Solubilization and purification of membrane proteins.

The membrane is then disrupted when the cells are lysed. The insoluble cell membranes containing the protein of interest can then be isolated by centrifugation, separating them from the soluble portion of the lysate. In order to purify the protein, these isolated membranes must be solubilized in a detergent to put the protein into solution. To accomplish this, the collected membranes can be readily mixed with a buffer containing a solubilization detergent in a dounce homogenizer. This solution can then be left to stir overnight. Afterwards, the membrane protein will be solubilized in a detergent micelle. Thus, it can be separated by subsequent centrifugation to separate it from insoluble debris in the supernatant. The solubilized protein of interest is then ready for the proceeding steps of affinity purification and size exclusion chromatography.

If the membrane protein is expressed as inclusion bodies, it may be refolded similar to methods used for soluble proteins [30–32]. However, in order to stabilize the protein during the refolding process, detergent must be present in the refolding buffers. Upon purification (and even during refolding), additives such as lipids, cholesterol, and other small molecules can often improve overall protein stability, which would assist in functional characterization and even structural studies.

Typical yields for the purification of natively folded membrane proteins are on average much lower when compared to soluble proteins, however, in some cases, yields can be greatly improved if the target protein is expressed as inclusion bodies and refolded. Producing sufficiently large yields of highly pure membrane proteins is most often the largest hurdle for successful structure determination. Over the past few decades, several strategies have been developed to improve the purity of membrane proteins. For example, the target membrane protein can be expressed using higher affinity, more specific tags or even duel affinity tags, such as 6-10xHIS-tag on the N-terminus and a Strep-tag on the C-terminus. An ideal sample for crystallization would produce a single, symmetric, and monodisperse peak from SEC and would be >90% pure when visualized using SDS-PAGE. This is indicative of a stable and homogeneous protein sample, ideal for crystallization. One must also keep in mind that not all detergents are stabilizing to all membrane proteins, therefore, our suggestion is to try up to five different detergents (DDM, LDAO, OG, C₈E₄, LMNG) and consider even supplementing with cholesterol or lipids, if necessary [33]. And even once a suitable detergent is found, one must also carefully consider the increase in detergent concentration, as the protein is concentrated to 10 mg/mL for initial crystallization, and how this may affect crystallization screening [27].

Membrane proteins have large hydrophobic domains that mediate their embedment within the core of the membrane, therefore, once removed from the membrane bilayer, the exposed hydrophobic surfaces must be stabilized by detergents/lipids for in-vitro studies. The presence of the highly dynamic detergent micelles occupying a large portion of the protein surface makes crystallization even more challenging than normally expected. Crystallization of membrane proteins follows a parallel workflow to that of soluble proteins and variables like temperature, pH, type of buffer, precipitant concentration, salt concentration, and sample concentration should be optimized. However, there are additional considerations and methods specifically tailored for membrane proteins. For example, there are commercially available crystallization screens that are marketed specifically for membrane proteins, as well as, detergent additive screens and the use of hydrophobic crystallization sheets for hanging drop crystallization. Further, detergent screening, bicelle crystallization, and lipidic cubic phase (LCP) crystallization are all methods specifically for the crystallization of membrane proteins (Figure 5). For the structure of NgBamA, only bicelle crystallization produced well-ordered crystals despite detergent screening yielding several lead conditions with very large, yet poorly diffraction, crystals.

Figure 5. Crystallization methods for membrane proteins.

Outline of crystallization methods specifically for membrane proteins including detergent screening, bicelles crystallization, and lipidic cubic phase (LCP) crystallization. Note that detergent screening and bicelle crystallization can be performed with the same instrumentation as with soluble proteins, however, LCP crystallization will require additional tools and robotics for manual and automated methods.

The presence of detergents in the sample buffer, however, does pose a challenge though since detergents significantly reduce the surface tension of the crystallization drops, making sitting drop crystallization and hanging drop crystallization with hydrophilic crystallization cover sheets nearly impossible for membrane proteins. Therefore, high-throughput screening requires the use of hydrophobic crystallization cover sheets to ensure a well-formed drop is maintained throughout the incubation period of the crystallization trails. One advantage here though is that most often, the hydrophobic cover sheets are UV transparent, making it convenient when identifying protein crystals from salt using UV imaging.

Protein crystals are fragile and harvesting them must be done with care. This is particularly true for membrane protein crystals which are grown in detergents and lipids, often making them even more fragile and more sensitive to even subtle temperature and buffer changes. As with all protein crystals, careful screening of cryo-protectants must be performed in order to find optimal conditions that preserve or enhance the diffraction quality of the crystals [34,35]. There is often a reduced solvent content observed within membrane protein crystals due to the detergent, which can assist in cryoprotection, however, the LCP matrix, along with the precipitant, often serve as a sufficient cryoprotectant for these crystals [36].

Once crystals have been grown, structure determination of membrane proteins parallels that of soluble proteins with some additional considerations, which we will summarize here. But generally, the methods and instrumentation are shared, except for when LCP crystallization is being performed [37,38,33]. While crystals grown in detergent and bicelles [39,40] are often similar in size to soluble proteins, LCP crystals are typically much smaller and more difficult to harvest, particularly from glass sandwich plates [36].

While earlier data collection on LCP crystals required multiple crystals for a complete dataset, advances in synchrotron instrumentation such as automated sample changing robotics, beam optics, and detectors can enable complete data collection on single LCP crystals. Microbeams ranging in size from 5 – 20 μm have enabled not only improved data collection from LCP harvested crystals, but also in finding well diffracting regions that are more ordered compared to other regions on larger non-uniformly packed crystals [41,42]. Both ultrafast goniometers and high-speed hybrid photon counting detectors with improved signal ratio have complemented the use of microbeams by allowing for superfast data collection to alleviate the effects of radiation damage without sacrificing data quality or resolution [43]. Rastering methods have also boosted data collection for membrane proteins by automating the search for ‘sweet spots’ on larger crystals and the location of crystals within loops containing an LCP matrix [44,45]. Lastly, SONICC technology has improved the visualization of not only crystals within an LCP matrix, but also for the identification of nanocrystals which may be too small for UV imaging [46,47]; however, its usefulness as a routine method for identifying protein crystals for X-ray or electron crystallography remains to be determined.

Crystals of membrane proteins often do not diffract to as high resolution as soluble protein crystals; a summary of all reported membrane protein structures can be found at the following website maintained by the White Laboratory at UC Irvine: https://blanco.biomol.uci.edu/mpstruc/. Therefore, data collection strategies and expectations must be adjusted accordingly, such that collecting a nice 3 Å complete dataset is the goal and anything higher resolution being a luxury, with anything lower resolution being the reality of working on membrane proteins. Of course, there are always some families of membrane proteins that are more stable and are more easily crystallized and routinely produce higher resolution diffraction, but this is not the norm. It is more likely that many crystals, sometime hundreds, will need to be screened in order to find a few that produce good enough diffraction quality for structure determination.

Once a dataset is collected and processed, the next step is structure determination using either molecular replacement or experimental phasing methods. Both phasing methods require knowing how many molecules are in the asymmetric unit to improve the likelihood of success. However, the need to stabilize membrane proteins with detergents or lipids leads to an increase in the average Matthew’s coefficient (packing parameter, Vm) and an increase in the expected range of acceptable solvent contents; information used to estimate the number of molecules in the asymmetric unit [48]. A correct solution for a membrane protein might have a Vm of 3.9 and a solvent content of 70% for a resolution of 3 Å, while this would likely not be the case for a closely packed soluble protein (Vm ranges from 1.7 – 3.5) of similar size and resolution. Therefore, during initial phasing, if a correct number of molecules per asymmetric unit cannot be determined with high confidence, one should run jobs in parallel for each of the possible options determined from analysis of the cell contents.

2. Materials

2.1. Cloning

2.1.1. Polymerase Chain Reaction (PCR)

1. dH₂O (DNase and RNase-free).

2. Phusion HF buffer (5X, ThermoFisher).

3. Dimethyl sulfoxide (DMSO) (100%).

4. Deoxynucleotide triphosphate (dNTP) mix (10 mM of each nucleotide).

5. Phusion HF DNA polymerase (ThermoFisher).

6. NgBamA genomic or optimized template DNA sequence (50-100 ng/mL concentration).

7. Forward and reverse primers (1000 ng/mL concentration each) for each construct.

8. Thermocycler (e.g. Eppendorf Mastercycler EP Gradient S).

2.1.2. Agarose Gel Electrophoresis

1. Agarose.

2. Tris-borate-EDTA (TBE, 1X) buffer: dissolve 10.8 g tris and 5.5 g boric acid in 900 mL of water, add 4 mL Na₂EDTA (0.5 M, pH 8.0) and adjust volume to 1 L.

3. Ethidium bromide (10 mg/mL).

4. DNA gel loading dye (e.g. Thermo Scientific DNA gel loading dye 6X).

5. DNA ladder (e.g. New England Biolabs 1-kb DNA ladder).

6. DNA gel casting system.

7. DNA gel electrophoresis apparatus (e.g. Biorad Minisub Cell GT Cell + Powerpac).

8. UV light/imager (e.g. Biorad UView Mini Transilluminator).

2.1.3. DNA Gel Extraction/PCR Cleanup

1. DNA gel extraction/cleanup kit (e.g. IBI Scientific Gel/PCR DNA Fragments Extraction Kit).

2. NgBamA PCR insert.

3. Benchtop centrifuge (e.g. Thermo Scientific Sorvall Legend Micro 17).

4. Microvolume spectrophotometer (e.g. ThermoFisher NanoDrop, BioTek Epoch 2).

2.1.4. Restriction Enzyme Digestion

1. Purified NgBamA PCR insert.

2. Purified plasmid expression vector (e.g. pET20bHT).

3. Relevant restriction enzymes (e.g. New England Biolabs NcoI and XhoI).

4. Cutsmart buffer (10X, New England Biolabs).

5. 37°C incubator.

2.1.5. Ligation of Insert and Vector

1. Purified, digested NgBamA PCR insert.

2. Purified, digested pET20bHT.

3. T4 DNA ligase (New England Biolabs).

4. T4 DNA ligase buffer (10X, New England Biolabs).

2.1.6. Transformation of Ligation Reaction into Chemically Competent DH5α

1. DH5α cells.

2. Ligation reaction.

3. 42°C heat block or water bath.

4. LB agar plates with carbenicillin (50 μg/mL) (LB/carb).

5. SOC outgrowth medium: 2 % tryptone, 0.5 % yeast extract, 10 mM NaCl, 2.5 mM KCl, 10 mM MgCl₂, 10 mM MgSO₄, 20 mM glucose; autoclave.

2.1.7. Overnight Cultures

1. Round-bottom 14 mL culture tubes.

2. Bacterial growth medium (e.g. 2XYT or LB).

3. Ampicillin: stock solution prepared at 50 mg/mL in water and stored in 500 μL aliquots at −20°C.

4. LB/carb agar plates with transformed DH5α colonies.

5. 37°C shaker.

2.1.8. DNA Plasmid Minipreps

1. Overnight cultures of transformed DH5α.

2. DNA plasmid isolation kit (e.g. IBI High-speed Plasmid Mini Kit).

2.1.9. DNA Sequencing Analysis

1. Purified DNA plasmids.

2. Sequencing primers.

3. Computer.

4. Vector analysis software or BLAST tool (e.g. Vector NTI, SnapGene).

2.2. Expression

2.2.1. Transformation of clone into BL21(DE3) cells

1. LB agar plates with carbenicillin (50 μg/mL) (LB/carb).

2. Incubator (37°C).

3. BL21(DE3) chemically competent cells.

4. Sequence verified plasmid.

5. Heat block (42°C).

6. LB medium: dissolve 5 g NaCl, 10 g peptone, and 5 g yeast extract into 800 mL of water. Adjust the volume to 1 L and autoclave.

7. Thermomixer (37°C, 1000 RPM) (e.g. Eppendorf Thermomixer C).

8. Inoculation loop or L spreader.

2.2.2. Small-scale expression test

1. Round-bottom 14 mL culture tubes.

2. LB medium: dissolve 5 g NaCl, 10 g peptone, and 5 g yeast extract into 800 mL of water. Adjust the volume to 1 L and autoclave.

3. Ampicillin: stock solution prepared at 50 mg/mL in water and stored in 500 μL aliquots at −20°C.

4. Inoculation loop.

5. Temperature controlled shaker (e.g. New Brunswick Innova 44R, ThermoFisher MaxQ 8000).

6. UV-vis spectrometer (e.g. Eppendorf Biophotometer Plus 6132).

7. Isopropyl β-d-1-thiogalactopyranoside (IPTG): stock solution prepared at 200 mM in water and stored in 500 μL aliquots at −20°C.

2.2.3. Prepare a glycerol stock of the transformed cells

1. Microcentrifuge tubes (1.5 mL).

2. Overnight culture from LB/carb agar plate.

3. Sterile 60% glycerol/LB solution.

2.2.4. Preparing cells for SDS-PAGE analysis

1. Benchtop microcentrifuge (e.g. Eppendorf 5424, ThermoFisher AccuSpin Micro 17).

2. 2x SDS-Loading Buffer.

3. Heat block.

4. SDS-PAGE gel running apparatus (e.g. BioRad PROTEAN system).

5. Pre-cast SDS-PAGE gel (e.g. BioRad 4-15% Mini-PROTEAN TGX).

6. SDS-PAGE running buffers.

7. Gel stain (e.g. ThermoFisher PageBlue Stain).

2.2.5. Prepare a 5 mL overnight starter culture

1. Round-bottom 14 mL culture tubes.

2. LB medium: dissolve 5 g NaCl, 10 g peptone, and 5 g yeast extract in 800 mL of water. Adjust the volume to 1 L and autoclave.

3. Ampicillin: stock solution prepared at 50 mg/mL in water and stored in 500 μL aliquots at −20°C.

4. Expression-verified cells.

5. Temperature-controlled shaker (e.g. New Brunswick Innova 44R, ThermoFisher MaxQ 8000).

2.2.6. Large-scale expression

1. Overnight culture from LB/carb agar plate.

2. 2 L flasks.

3. TB medium: dissolve 20 g tryptone, 24 g yeast extract, 4 mL glycerol, and 100 mL of phosphate buffer (0.17 M, KH₂PO₄, 0.72 K₂HPO₄) into 800 mL of water. Adjust the volume to 1 L and autoclave.

4. Ampicillin: stock solution prepared at 50 mg/mL in water and stored in 500 μL aliquots at −20°C.

5. Temperature-controlled shaker (e.g. New Brunswick Innova 44R, ThermoFisher MaxQ 8000).

2.3. Purification

2.3.1. Preparation of Lysate

1. 1xPBS (12 mM phosphate buffer, pH 7.4, 137 mM NaCl, 2.7 mM KCl).

2. DNase I: prepare a stock solution at 10 mg/ml in 10 mM NaCl and stored at −20°C. Phenylmethylsulfonyl fluoride (PMSF): prepare a stock solution at 200 mM by dissolving solid PMSF in 100% ethanol and store in 0.5 mL aliquots at −20°C.

3. Apparatus for large scale cell lysis (e.g. an Avestin Emulsiflex C3 high-pressure homogenizer).

2.3.2. Solubilization from Whole Cell Lysate

1. High-speed centrifuge (e.g. ThermoFisher Lynx 6000).

2. n-Dodecyl β-D-maltoside (DDM): prepare a stock solution at 20% in water and store at 4°C.

2.3.3. Solubilization from Isolated Membranes

1. High-speed centrifuge (e.g. ThermoFisher Lynx 6000).

2. Ultracentrifuge (e.g. Beckman-Coulter XPN90).

3. 40 mL glass dounce homogenizer (Wheaton).

4. n-Dodecyl β-D-maltoside (DDM): prepare a stock solution at 20% in water and store at 4°C.

5. Elugent: prepare a stock solution at 25% in water.

2.3.4. Isolation of Inclusion Bodies and Refolding

1. High-speed centrifuge (e.g. ThermoFisher Lynx 6000).

2. n-Dodecyl β-D-maltoside (DDM): prepare a stock solution at 20% in water and store at 4°C.

3. 40 mL dounce homogenizer (Wheaton).

4. Triton X-100: prepare a 20% stock solution in water.

5. 8 M urea: dissolve 480.5 g urea into 500 mL of hot water or 1X PBS; adjust the volume to 1 L.

2.3.5. Purification of Solubilized Membrane Proteins

1. FPLC column (5 mL) packed with HisPur Ni-NTA resin (Thermofisher).

2. FPLC automated purification system (e.g. GE Healthcare Pure 25 L).

3. Buffer A (1X PBS, 0.05% DDM).

4. Buffer B (1X PBS, 0.05% DDM, 1 M imidazole).

5. Dialysis tubing (Millipore Sigma).

6. Amicon Ultra Centrifugal filter units (30 and 100 MWCO) (Millipore Sigma).

7. Sephacryl S-300 HR 16/60 Hi-Prep column (GE Healthcare).

8. SEC buffer (25 mM Tris, pH 7.5, 150 mM NaCl, and 0.6% C₈E₄).

2.4. Crystallization

2.4.1. Aliquoting Commercial Crystallization Broad-matrix Screens (Hanging Drop)

1. Commercial broad matrix screens in 96-well blocks (e.g. Shot Gun, Crystal Screen, MCSG Suite).

2. Plate dispensing/stamping instrument (e.g. Rainin Liquidator-96, ThermoFisher Versette, Beckman Biomek).

3. Automated microplate heat sealer (e.g. ThermoFisher ALPS 3000); or manual plate seals.

4. Round-bottom 96-well polypropylene or polystyrene plates (e.g. Greiner Bio-One #650201).

5. Microplate sealing film roller.

2.4.2. Preparing 96-well Optimization Screens

1. Crystallization buffer components (from lead conditions in broad-matrix screening).

2. Automated liquid handling robot (e.g. SPT Dragonfly; Formulatrix Formulator).

3. Microplate mixer.

2.4.3. Performing Broad-matrix and Optimization Crystallization Screening Using an Automated Crystallization Robot (Hanging and Sandwich Drop)

1. Broad-matrix crystallization screens in 96-well plates.

2. Optimization crystallization screens in 96-well plates.

3. Hydrophobic hanging drop crystallization sheets (e.g. Grace Bio-Labs ProCrystal #875238, SPT Labtech Viewdrop III).

4. LCP sandwich plates (e.g. Hampton Research LCP Sandwich Set, Molecular Dimensions Laminex Plates).

5. Bicelles (e.g. MemX Biosciences MemMagic Bicelle Screen kit).

6. Additives (e.g. Hampton Research Additive Screen).

7. Multichannel pipetter (12-channel; 0.5 – 10 μL).

8. Automated crystallization robot (e.g. SPT Labtech mosquito LCP, Art Robbins Crystal Gryphon LCP).

9. V-bottom 96 well plates.

10. Centrifuge with microplate holders (e.g. ThermoFisher Lynx 4000, Beckman Allegra 15R).

2.4.4. Performing Optimization Screening Manually

1. Large format 24-well plates with sealant (e.g. VDXm, VDX, EasyXtal).

2. LCP sandwich plates (e.g. Hampton Research LCP Sandwich Set, Molecular Dimensions Laminex Plates).

3. Siliconized crystallization coverslips (18 mm round and square).

4. Repeating dispenser for Hamilton syringes.

5. Crystallization buffer components (from lead conditions in broad-matrix screening).

6. Additives.

2.4.5. Visualizing and Assessing Crystallization Leads

1. Automated crystallization imaging robot (e.g. Formulatrix Rock Imager, Art Robbins CrysCam, JanSi UVEX).

2. UV imaging microscope (e.g. Korima PRS-1000).

3. Stereomicroscope (e.g. Zeiss SteREO Discovery.V8).

4. Micro tools (e.g. Hampton Research Micro Tools, Mitegen MicroTools Kit).

5. Scalpel or razor blade.

6. Crystal staining dye (e.g. Hampton Research Izit Crystal, Jena Bioscience Black Light).

2.5. Data Collection and Structure Determination

2.5.1. Harvesting Crystals

1. Crystallization trays containing crystals.

2. Stereomicroscope (e.g. Zeiss SteREO Discovery.V8).

3. Cryo-protectant solution(s).

4. Magnetic wand (e.g. Hampton Research CrystalWand, Mitegen Magnetic Wand Straight).

5. Scalpel or razor blade.

6. Reverse (self-closing) tweezers (e.g. Dumont style N2a).

7. Petri dish.

8. Mounted crystallization pins/assemblies (loops/bases).

9. Uni-pucks (e.g. Mitegen Universal V1-Puck).

10. Puck tools and tongs.

11. Foam dewar (e.g. Spearlab Foam dewars, Mitegen Double Puck Loading Dewar).

12. Liquid nitrogen.

13. Puck log sheets.

2.5.2. Storing Crystals

1. Uni-pucks containing crystals.

2. Puck cane (e.g. Mitegen Shelved Puck Shipping Cane).

3. Liquid nitrogen.

4. Storage Dewar (e.g. Taylor-Wharton VHC35 Dewar or CX100 Cryo Dry Shipper).

2.5.3. Screening and Data Collection

1. X-ray source (synchrotron or home source including X-ray generator, optics, cryo system, and detector).

2. Frozen crystals in uni-pucks.

3. Puck tools and tongs.

4. Magnetic wand (e.g. Hampton Research CrystalWand, Mitegen Magnetic Wand Straight).

5. Cryo pin tongs (18 mm).

6. Foam dewar (e.g. Spearlab Foam dewars, Mitegen Double Puck Loading Dewar).

7. Workstation.

8. Data acquisition software (e.g. JBluIce, BluIce, SERGUI).

2.5.4. Diffraction Analysis and Structure Determination

1. Diffraction images.

2. Workstation.

3. Diffraction data processing and reduction software (e.g. HKL2000, MOSFLM, XDS, DIALS, Xia2).

4. Diffraction data analysis software (e.g. phenix.xtriage (PHENIX), Cell Content Analysis (CCP4)).

5. Structure determination software packages (e.g. PHENIX, CCP4, SHELX).

3. Methods

3.1. Cloning

3.1.1. Polymerase Chain Reaction (PCR)

1. Prepare 138 μL Phusion PCR master mix by combining 100.5 μL dH₂O, 30 μL 5X HF Phusion buffer, 4.5 μL 100% DMSO, and 3 μL dNTP mix (see Note 1).

2. For each PCR reaction (each construct you plan to clone), add 23 μL master mix, 1 μL template DNA, and 0.5 μL each forward and reverse primers to a thin-walled PCR tube. Lastly, add 0.5 μL HF Phusion polymerase to each reaction.

3. Place the tubes containing PCR reactions into the thermocycler and use according to manufacturer instructions. General PCR conditions include an initial denaturation step of 98°C for 10-30 s, 25-40 cycles of [98°C for 10 s, 55-60°C for 30 s, and 72°C for 15-30 s per kb of amplicon (for Phusion HF polymerase)], and a final extension of 72°C for 5-10 min before holding at 4°C until the reactions can be removed from the thermocycler (see Note 2).

4. Remove the PCR reaction from the thermocycler and prepare for gel purification as noted in the next section, or store at 4°C for short term or −20°C for long term.

3.1.2. Agarose Gel Electrophoresis and DNA Gel Extraction/PCR Cleanup

1. Prepare a 1% agarose gel by mixing 0.25 g agarose with 25 mL 1X TBE buffer into a conical tube or Erlenmeyer flask.

2. Microwave for 20-30 s or until fully dissolved.

3. Once dissolved, allow to cool until the container can be handled, then add 2.5 μL of ethidium bromide to the solution and mix gently by swirling the solution (see Note 3).

4. Carefully pour the agarose solution into the gel assembly, checking that the teeth of the comb are mostly covered for adequate well formation. Allow to cool for 15-20 min or until solidified (see Note 4).

5. Place the gel into a DNA gel electrophoresis apparatus. Add 1X TBE running buffer to the “max” line of the of gel apparatus or until the buffer is just above the gel surface.

6. Add 5 μL DNA ladder to the first well.

7. Mix 5 μL 6X DNA loading dye with each 25 μL PCR reaction.

8. Carefully pipette each into separate wells of the gel and record.

9. Run the gel at 100 volts for 25-40 min or until the dye front is at the bottom of the gel.

10. Remove the gel from the apparatus and, with proper protection, use a UV transilluminator to visualize the bands in the gel.

11. Use a scalpel to carefully extract the slice of the gel containing your PCR insert, ideally removing a 300-800 mg slice of agarose, being sure to exclude any excess gel not containing your sample.

12. Following manufacturer instructions, use a DNA gel extraction kit to dissolve the gel slice and purify the DNA, eluting with 50-100 μL of ultrapure microbiology grade water (see Note 5).

3.1.3. Restriction Enzyme Digestion

1. Measure the concentration of the purified PCR insert using standard methods for DNA, preferably using a nanoliter volume instrument such as a NanoDrop or Epoch2 (see Note 6).

2. Combine the gel purified PCR insert(s) (100 μL) with 1 μL XhoI, 1 μL NcoI, and 11.3 μL of 10X Cutsmart buffer in a 1.5 mL microcentrifuge tube.

3. Repeat steps 1-2 with the vector DNA, adjusting the volumes of the components according to the concentration and volume being used.

4. Incubate for at least 1 h at 37°C for adequate digestion, however, longer incubation times even up to overnight may be used to ensure complete digestion (see Note 7).

5. Once the digest incubation period is complete, perform a DNA cleanup of the PCR insert using a DNA extraction/cleanup kit, eluting with 50-75 μL of ultrapure microbiology grade water.

6. For the vector DNA digest reaction(s), first run on a 1% agarose gel as instructed in section 3.1.2. Then excise the digested vector band and use a DNA gel extraction kit to dissolve the gel slice and purify the DNA, eluting with 50-75 μL of ultrapure microbiology grade water.

3.1.4. Ligation of Insert and Vector

1. Measure the concentration of the digested PCR insert and vector as indicated in section 3.1.3 – step 1.

2. Mix the digested PCR insert and vector at a molar ratio of 5-10:1 (insert:vector); then add T4 DNA ligase (1 μL per 10 μL of ligation reaction volume), and 10X T4 ligase buffer to a final 1X concentration.

3. Incubate at room temperature for at least 1 h, however, longer incubation times up to overnight will ensure maximal ligation yields (see Note 8).

3.1.5. Transformation of Ligation Reaction into Chemically Competent DH5α

1. Thaw 50 μL of chemically competent DH5α cells on ice. Once thawed, add 1-5 μL of the ligation reaction to the cells, mix by gently tapping the tube, and incubate on ice for 20 min (see Note 9).

2. During the incubation, place an LB-agar plate containing the appropriate antibiotic in a 37°C incubator to pre-equilibrate.

3. Heat shock the cells in a 42°C water bath for 45 s and immediately place back on ice for 2 min.

4. Add 200 μL of SOC outgrowth medium and shake at 37°C for 30-60 min at ~1000 RPM using a thermomixer.

5. Using a flame and sterile cell spreader, plate all the cells onto the pre-equilibrated agar plate containing the appropriate antibiotic(s) (e.g. For a pHIS2-parallel vector, use an LB-agar/carbenicillin or LB-agar/ampicillin plate).

6. Incubate the agar plate(s) overnight (~20 h) at 37°C inverted.

7. Check the plate(s) and count the number of colonies present for the ligation reaction and compare to the control plate with vector only (if this was also made). While the colonies are best when fresh, the plates may be sealed with parafilm and stored at 4°C for up to two weeks.

3.1.6. Overnight Cultures and DNA Plasmid Minipreps

1. Select 2-5 colonies to prepare 5 mL cultures of LB with an appropriate antibiotic for each in separate 14-mL culture tubes.

2. Use a flame and a sterile loop each time to carefully transfer each colony into the pre-made, pre-labeled culture tube, ensuring the cap is loose to allow proper aeration.

3. Shake the culture tubes at ~200 RPM overnight at 37°C or until stationary phase.

4. Centrifuge the cultures at 4,500 RPM using a benchtop centrifuge for 15 min at 4°C.

5. Pour off the medium, being sure to remove as much of the liquid as possible.

6. The plasmid DNA may be purified using a commercially available DNA plasmid isolation/miniprep kit according to the manufacturer instructions.

7. Measure the concentration of each of the miniprep DNA samples using standard methods for DNA as described in section 3.1.3 – step 1.

3.1.7. DNA Sequencing Analysis

1. For each miniprep DNA sample, adjust the concentration and volume required by the sequencing facility being used. Do the same for each sequencing primer being used.

2. Submit samples for sequencing analysis.

3. Use a vector analysis program or a pairwise BLAST program to compare the sequencing results to the known sequence to determine if the cloning was successful.

3.2. Expression

3.2.1. Transformation into BL21 (DE3) Cells.

1. Place an LB-carbenicillin agar plate into a 37°C incubator to pre-equilibrate.

2. Thaw a 50 mL tube of chemically competent BL21(DE3) cells on ice (see Note 10).

3. Pipette 1 μL of your sequence verified plasmid into the cells, close the cap, and mix by gently tapping the end of the tube.

4. Place back on ice for 20 min.

5. Heat shock the cells in a 42°C water bath for 45 s and immediately place back on ice for 2 min.

6. Add 200 mL of SOC medium to the tube.

7. Place into a thermomixer and shake at 1000 RPM at 37°C for at least 30 min (see Note 11).

8. Pipette 30 μL of the transformation reaction onto one side of the pre-warmed LB-carb agar plate and use an inoculation loop to spread the cells in a streaking fashion across the plate, rotating by 90° and repeating until back at the starting position.

9. Allow the plate to dry completely and then incubate at 37°C inverted overnight (16-24 h).

3.2.2. Small-scale Expression Test

1. Inspect your plate for well grown colonies. Select one or more colonies, circle, and label them on the bottom of the plate (see Note 12).

2. Prepare a 14 mL culture tube with 3-5 mL of LB medium supplemented with ampicillin (50 μg/mL).

3. Use a sterile loop to gently pick up your colony and transfer into your medium with rapid stirring for 2-5 s.

4. Repeat steps 1-3 for each colony you have chosen to test for expression.

5. Shake (~200 RPM) the cells at 37°C until they have reached an OD₆₀₀ between 0.6 and 1.0.

6. For each culture in step 5, transfer 1 mL of the cells into a 1.5 mL microcentrifuge tube and label as pre-induction control (Pre).

7. Prepare the pre-induction controls (Pre) for SDS-PAGE analysis as described in section 3.2.3

8. Transfer 1 mL of the cells from step 5 into a fresh 14 mL culture tube and label as post-induction sample (Pre).

9. Store the remaining cultures from step 5 at 4°C.

10. Induce the post-induction cultures from step 8 by adding IPTG to a final concentration of 1 mM (5 μL of stock).

11. Allow the cells to continue to grow an additional 1-2 h.

12. Prepare the post-induction samples (Post) for SDS-PAGE analysis as described in section 3.2.3.

3.2.3. Preparing Cells for SDS-PAGE Analysis

1. Adjust the OD₆₀₀ of the cells to 0.6 in a final volume of 1 mL in a new 1.5 mL microcentrifuge tube.

2. Spin down the cells using a microcentrifuge at 16,000 x g for 1 min.

3. Remove and discard of the medium supernatant.

4. Add 200 μL of 2x SDS-Loading Buffer and gently resuspend the cells by pipetting.

5. Heat the samples for 5 min at 95°C.

6. Remove and centrifuge the samples for 5 min at full speed using a microcentrifuge (see Note 13).

7. Remove the samples and be careful not to disturb or shake them.

8. Analyze the pre- and post-induction samples, including a suitable protein marker, using SDS-PAGE using manufacturer instructions and stain/destain the gel to visualize the results.

9. Compare the pre vs post-induction samples to determine if an increase in the levels of protein expression is observed in the presence of IPTG. Make a note if there were differences between the different colonies, if more than one was analyzed (see Note 14).

3.2.4. Prepare a Glycerol Stock of the Transformed Cells

1. Collect and label a 1.5 mL microcentrifuge tube for each sample in section 3.2.3 – step 9 with confirmed expression (see Note 15).

2. Pipette 250 μL of your overnight culture from section 3.2.2 – step 9.

3. Pipette 250 μL of 60% glycerol/LB solution and mix by pipetting.

4. Place your glycerol stock into a freezer at −80°C until future use; flash freezing in liquid nitrogen is preferred when possible.

3.2.5. Prepare a 5 mL Overnight Starter Culture for Large-scale Expression

1. Prepare a 14 mL culture tube with 6 mL of LB medium supplemented with ampicillin (50 μg/mL) (see Note 16).

2. Use a sterile loop to inoculate the culture using the glycerol stock prepared in section 3.2.4. Do this by scraping across the top of the frozen glycerol stock with the sterile loop and transferring to the culture medium with rapid stirring for 2-5 s. Alternatively, fresh colonies may be used if no difference in expression was observed across multiple colonies in section 3.2.3 – step 9.

3. Shake (~200 RPM) the culture at 37°C until the OD₆₀₀ is ~0.5, do not overgrow (see Note 17).

3.2.6. Large-scale Expression

1. Prepare five 2 L baffled flasks with 1 L of TB medium in each, autoclave using manufacturer instructions (see Note 18).

2. Let the flasks cool and add carbenicillin to a final concentration of 100 μg/mL.

3. Collect the overnight culture from section 3.2.5 – step 3.

4. Pipette 1 mL of the starter culture into each of the 1 L TB/carb baffled flasks.

5. Shake (~200 RPM) the flasks for 72 h at 20°C.

6. Harvest the cells by centrifugation at 8,000 x g for 15 mins at 4°C.

3.3. Purification

3.3.1. Preparation of Lysate

1. Weigh the cell pellet (either fresh or frozen) from section 3.2.6 – step 6 and add 10 mL of 1xPBS per gram of cells.

2. Stir the cells either on ice or in a cold room rapidly until fully resuspended (see Note 19).

3. Add DNase to a final concentration 10 μg/mL, mix well, and incubate for 10 min.

4. Add PMSF to a final concentration of 200 μM and immediately lyse the cells by three passages through an Emulsiflex C3 high-pressure homogenizer (Avestin) using manufacturer instructions maintaining a pressure of ~15,000 PSI (see Note 20).

5. The lysate is now ready for further purification (see Note 21).

3.3.2. Solubilization from Whole Cell Lysate

1. Centrifuge the lysate at 12,000 x g for 10 min to remove large insoluble cell debris.

2. Pour the supernatant into a fresh beaker and add DDM to a final concentration of 0.5% and stir overnight at 4°C.

3. Centrifuge the solution at 40,000 x g to remove large insoluble debris.

4. Collect the supernatant and filter (0.45 μm); the solubilized protein is in the supernatant.

5. Proceed with further purification as described in section 3.3.5 (see Note 21).

3.3.3. Solubilization from Isolated Membranes

1. Centrifuge the lysate at 12,000 x g for 10 min to remove large insoluble cell debris.

2. Pour the supernatant into round-bottom ultracentrifuge tubes (e.g. for a Type 45 Ti rotor which has a maximum volume of 70 mL) (see Note 22).

3. Centrifuge at 235,000 x g for 90 min using an ultracentrifuge. The resulting pellet contains the cellular membranes.

4. Pour off the supernatant and transfer the membrane pellet into a 40 mL dounce homogenizer.

5. Homogenize the membranes in ~60 mL of 1xPBS.

6. Bring the volume to 80 mL with 1xPBS, then add Elugent to a final concentration of 5% with rapid stirring, and stir overnight at 4°C.

7. Centrifuge the solubilized sample at 235,000 x g for 60 min.

8. Collect the supernatant and filter (0.45 μm); the solubilized protein is in the supernatant.

9. Proceed with further purification as described in section 3.3.5 (see Note 21).

3.3.4. Isolation of Inclusion Bodies and Refolding

1. After cell lysis, harvest the inclusion bodies by centrifuging at 12,000 x g for 10 min.

2. Transfer the inclusion bodies into a dounce homogenizer and wash by homogenizing in 40 mL of 1xPBS+2% Triton X-100.

3. Centrifuge at 12,000 x g for 10 min to isolate the inclusion bodies, discarding the supernatant.

4. Repeat the washes in steps 2-3 two more times. The inclusion body pellet should ideally become whiter with each wash (see Note 23).

5. Repeat the washes in steps 2-3 three more times with 1xPBS only to remove as much Triton X-100 as possible.

6. After the final wash step, dissolve the inclusion bodies in 40 mL of freshly prepared 8 M urea.

7. Add DDM to a final concentration of 0.1%.

8. Allow the urea-solubilized sample to rock at room temperature for 30 min.

9. Centrifuge the sample at 12,000 x g for 10 min and collect the supernatant.

10. Transfer the supernatant into pre-equilibrated dialysis tubing and dialyze against 1 L of 1xPBS overnight at 4°C (see Note 24).

11. Collect the dialyzed sample and centrifuge at 40,000 x g for 30 min to remove insoluble debris.

12. Collect the supernatant and filter (0.45 mm); the refolded protein is in the supernatant.

13. Proceed with further purification as described in section 3.3.5 (see Note 21).

3.3.5. Purification of Solubilized Membrane Proteins

1. Attach the 5 mL HisPur Ni-NTA column to the FPLC system according to manufacturer instructions.

2. Write a standard method using a method wizard (or manually) to perform affinity chromatography as follows monitoring at 280 nm: (i) equilibrate the column with 5 column volumes (CV) of Buffer A, (ii) apply the solubilized protein sample using a sample pump connected to an air sensor, (iii) wash out unbound sample using 5 CV of Buffer A, (iv) perform elution with fractionation (2 mL per fraction) using a linear gradient from 0 mM to 500 mM imidazole over the course of 25 CV; or by using a step gradient with 25, 50, 100, 250, and 500 mM imidazole steps with 5 CV each, and lastly (v) re-equilibrate the column with 5 CV of Buffer A.

3. Use SDS-PAGE analysis of the peak fractions to determine which contain the target protein.

4. Combine the fractions containing the protein of interest (see Note 25).

5. Concentrate the sample to 5 mL or less and run over a Sephacryl S-300 HR16/60 Hi-Prep size-exclusion column (GE Healthcare) in SEC buffer using the FPLC system with a flow rate of 0.5 mL/min with fraction volumes of 1 mL each (see Note 26).

6. Use SDS-PAGE analysis of the peak fractions to determine which contain the target protein.

7. Combine the purest fractions containing the protein of interest.

8. Concentrate the sample to 10 mg/mL (see Note 27).

3.4. Crystallization

3.4.1. Aliquoting Commercial Crystallization Broad-matrix Screens (Hanging Drop)

1. Centrifuge the commercial screen deep well block at 250 x g for 2 min.

2. Label the needed number of round-bottom plates (see Note 28).

3. Carefully remove the seal of the deep well block and place into the pipetting robot according to manufacturer instructions.

4. Add the labeled round-bottom plates into the available positions on the robot, according to manufacturer instructions.

5. Pipette 50 μL of solution from the deep well block into each of the round-bottom plates. If the instrument has more than a single dispensing position, a single aspiration may allow multiple dispenses; follow manufacturer instructions.

6. Replace the filled plates in the dispensing positions with empty ones and repeat step 5.

7. Repeat steps 5-6 until all the solution in the deep well block has been dispensed (see Note 29).

8. Seal the round-bottom plates using an automated microplate heat sealer or manually using single plate seals (see Note 30).

9. Repeat steps 1-8 for each commercial broad-matrix screen that will be used. Alternatively, steps 1-8 can be performed in parallel for all the screens to improve efficiency.

10. Store the aliquoted plates between 4-10°C or freeze at −20°C.

3.4.2. Preparing 96-well Optimization Screens

1. Gather the information for the lead crystallization condition that will be optimized.

2. Prepare sufficient quantities of the necessary stock solutions for each of the components (see Note 31).

3. Using the screen design software of the liquid handling robot, design an appropriate screen around the lead condition that will pipette 50 – 100 μL into each well. Ideally, precipitants (± 10%) are varied along the X-axis and salts are varied along the Y-axis (± 25-100 mM, depending on the starting concentration), while buffers are kept constant across each single plate, yet, can be varied across multiple plates. Alternatively, buffer pH can also be varied along one of the axes if desired (see Note 32).

4. Insert/load the stock solutions onto the automated liquid handing robot and prime the system according to the manufacturer instructions.

5. Load the round-bottom 96-well plate onto the instrument and run the program to dispense the optimization screen. If desired, larger volumes can be designed and dispensed into 0.5 or 1 mL deep well blocks.

6. Mix the optimization plate using a microplate mixer or vortexer fitted with a microplate adapter. If making a deep well block, seal and then mix by inverting continuously until homogeneous. Then centrifuge at 250 x g for 2 min prior to using.

7. Repeat steps 1-6 for each lead crystallization condition that will be optimized.

8. Additionally, an additive optimization screen can be prepared by preparing 5 mL of each lead crystallization condition and dispensing 50 μL into each well of a 96-well round-bottom plate. Then, 5.6 μL of each additive from a 96-well additive screen is pipetted into each well of the round-bottom plate using a 12-channel multipipetter, and follow step 6.

9. Store the optimization plates between 4-10°C or freeze at −20°C.

3.4.3. Performing Broad-matrix and Optimization Crystallization Screening Using an Automated Crystallization Robot (Hanging and Sandwich Drop)

1. Design a hanging drop method that dispenses 200 nL of protein solution and 200 nL of well solution onto the hanging drop crystallization sheet using an automated crystallization robot, according to manufacturer instructions. Where possible, utilize a humidification chamber or alternatively, a humidity flow from a standard humidifier may directed over the plate deck (see Note 33).

2. Load a V-bottom 96 well polypropylene plate (for a SPT Labtech mosquito LCP, but may vary for other crystallization robots) onto the plate deck according to the designed method and pipette a sufficient volume of the protein sample for setting 1-5 plates into a full column.

3. Centrifuge the broad-matrix or optimization 96-well crystallization screen at 250 x g for 2 min and load onto the robot plate deck according to the designed method.

4. Add a hanging drop crystallization sheet onto the plate deck, according to robot manufacturer instructions (see Note 34).

5. Run the designed method to dispense the hanging drop crystallization tray.

6. When finished dispensing, immediately invert the crystallization sheet from step 4 and carefully affix to the top of the crystallization plate from step 3 such that the drops are facing towards the well solution, being sure that the crystallization drops are properly aligned to the corresponding wells. Use a plate roller to ensure proper sealing across the entire plate.

7. For crystallization in bicelles, mix 40 μL of protein solution with 10 μL of bicelles and incubate on ice for at least 30 min. Then repeat steps 2-6 for each crystallization screen being used.

8. For crystallization using LCP, first prepare the LCP/protein mix according to published protocols. Then attach the syringe containing the LCP/protein mix according to instrument manufacturer instructions. Next, prepare a new method for LCP crystallization with sandwich plates consisting of 50-100 nL of LCP/protein mix and 1 μL of well solution. Run the method according to manufacturer instructions for each crystallization screen being used (see Note 35).

9. Transfer the crystallization trays to an automated crystallization imaging robot with an attached hotel according to manufacturer instructions at an appropriate temperature, if available. Alternatively, crystallization trays may be stored in standalone temperature-controlled incubators.

3.4.4. Performing Optimization Screening Manually

1. Gather the information for the lead crystallization condition that will be optimized.

2. Prepare sufficient quantities of the necessary stock solutions for each of the components (see Note 31).

3. Design the optimization screen for either a 15 or 24-well large format optimization tray using an online tool such as Hampton Research’s Make Tray Tool (https://hamptonresearch.com/make-tray.php) (see Note 36). Print the summary which lists the pipetting volumes of each component for each well of the tray.

4. Pipette the components into each well of the tray as indicated and mix the tray thoroughly using a plate mixer or vortexer with a plate adapter.

5. For detergent and bicelle crystallization, pipette 1-2 μL of sample onto an 18 mm round coverslip, followed by an equal volume of well solution. Then carefully invert the coverslip and place over the corresponding well and gently press slightly into the grease until a seal is formed; it is not necessary to press fully down to the plate rim. Repeat for each well of the optimization plate.

6. For LCP crystallization, prepare the LCP/protein mix in a 10 μL Hamilton syringe according to instrument manufacturer instructions. Install the syringe containing the LCP/protein mix into a repeating dispenser. Pipette 200 nL (a single dispense) of LCP/protein mix and 1 μL of well solution into each well of a 2 × 2 section of a 96-well sandwich plate. Seal with an 18 mm square coverslip. Repeat for each of the remaining 2 × 2 sections of the optimization plate.

7. Store the optimization trays at an appropriate temperature ranging 4-21°C in standalone temperature-controlled incubators (see Note 35).

3.4.5. Visualizing and assessing crystallization leads

1. If using an automated crystallization imaging system, image the trays on a routine schedule at least once per week for up to 3 months. Images are typically visualized remotely and interactively scored and annotated (see Note 37).

2. If using a stereomicroscope to manually visualize the crystallization trays, monitor the trays at least one time per week for up to 3 months. Keep a log of the most promising conditions and those conditions producing crystals.

3. For lead conditions that produce crystals, image the drop using UV imaging either on the automated imaging robot or manually using a standalone UV imaging microscope (see Note 38).

4. If UV imaging is inconclusive, use a stereomicroscope and a cryoloop or microtool to crush one of the smaller crystals in the drop. Protein crystals are fragile and should crush relatively easily, while salt crystals would require significant force and would typically break off in chunks. Alternatively, small molecule dyes may be used to stain protein crystals (see Note 39).

3.5. Data Collection and Structure Determination

3.5.1. Harvesting and Storing Crystals

1. Use the puck tools to open a uni-puck, placing the base into an empty foam dewar.

2. Add liquid nitrogen to the dewar until it just covers the uni-puck base. Allow to cool for a few minutes until the liquid nitrogen is no longer boiling.

3. Fill the dewar to the rim and place the cover/cap on to allow the dewar and uni-puck to equilibrate. Ensure that the dewar stays topped off the entire time you are harvesting crystals and avoid frost buildup and ice in the liquid nitrogen.

4. To harvest crystals from 96-well plates, use a razor blade to carefully cut around the drop of interest by scoring a box, being sure to cut completely through the crystallization sheet.

5. Gently lift the edge of the excised square, cut the underlying glue, and quickly invert the drop.

6. Use reverse action tweezers to hold the drop or place it onto a petri dish to allow easy movement on the microscope. Align the drop to the focus region of the microscope (see Note 40).

7. Place a crystal loop/base onto a magnetic crystal wand and quickly, but carefully, bring the loop into proximity of the drop and gently submerge the loop in the drop, capture a crystal, and then rapidly freeze the crystal in the liquid nitrogen, removing the dewar lid immediately before plunging the loop (see Note 41).

8. Once frozen, carefully place the loop in the appropriate position within the puck, being sure not to re-expose it to air.

9. Repeat steps 7-8 for each crystal in the drop that you wish to harvest. Record the necessary crystal information in the puck log sheet for each position.

10. Repeat steps 4-9 for each drop in the tray you wish to harvest crystals from.

11. If cryoprotectants are being used, they can be pipetted in close proximity to the drop with the crystals being directly transferred and incubated for the desired time period prior to looping and freezing as outlined in steps 7-8. Alternatively, the cryoprotectant can be directly added to the drop. If longer incubation periods are desired, the drop can be inverted and placed back over the well solution and taped to reduce evaporation.

12. Once a puck is filled, use the puck tools to re-install the puck top and place the puck into a pre-cooled puck cane according to manufacturer instructions and place into a storage dewar until X-ray diffraction screening and data collection. Repeat for other pucks that are filled while harvesting crystals.

3.5.2. Screening and Data Collection

1. If screening at a home source, transfer the puck to a pre-cooled dewar filled with liquid nitrogen.

2. Use the puck tools to open the top of the puck to allow access to the crystal pins.

3. Using a magnetic wand and cryo pin tongs, transfer the crystal pin of interest from the puck to the goniometer as instructed by the facility scientist.

4. Align the crystal to the center position as instructed, rotating the crystal 90° and re-aligning to the center position. Repeat until the crystal is properly aligned along the axis of rotation and always in the beam.

5. Move the detector to where the edge would produce ~3 Å resolution diffraction.

6. Set the parameters for collecting initial images at 0 and 90°. For example, for a Rigaku system equipped with a MicroMax-007 HF X-ray generator and a CCD detector, one would screen with an initial exposure of 30 - 300 s over 0.5 - 1° oscillation. The goal here is to determine if the crystal diffracts and what an approximate resolution may be. All final data collection should be performed at a synchrotron as described below.

7. If screening at a synchrotron, coordinate with the beamline staff to load the pucks into the robotic dewar (see Note 42).

8. Using the puck log sheet, use the data acquisition software to mount the desired crystal by selecting the puck and position to be loaded onto the goniometer.

9. Align the crystal to the center position as instructed by the beamline staff, rotate the crystal 90° and re-align to the center position. Repeat until the crystal is properly aligned along the axis of rotation and always in the beam.

10. Move the detector to where the edge would produce ~3 Å resolution diffraction.

11. Setup for initial screening by attenuating the X-ray beam for only a short exposure. For example, at the GM/CA ID-B beamline with a 20 μm beam size and an Eiger 16 M detector, we routinely screen with 80% attenuation for 0.25 s over 0.25 – 0.5° oscillation, collecting at 0, 45, and 90°.

12. Once the set of initial images have been collected, inspect them to determine if there are diffraction spots, the shape and intensities of the diffraction spots, the resolution, and whether there is ice present or not. From these initial observations, decide if the crystal is suitable to continue with data collection or if you should move on to screen other crystals. Keep in mind that now, it is often faster to just collect a dataset rather than spending time discussing it with a colleague or carefully inspecting the diffraction images.

13. To set up for data collection, refer to the strategy program within the data acquisition software for parameters to load and adjust as necessary, including detector distance. When in doubt, just collect a full 180 or 360° dataset if using molecular replacement to solve the structure. For experimental phasing using an anomalous signal, coordinate with the beamline staff to perform an edge scan and collect at 1-3 wavelengths (peak, edge, remote) being sure to collect as highly redundant data as possible, yet avoiding radiation damage (see Note 43).

3.5.3. Diffraction Analysis and Structure Determination

1. Most synchrotron beamlines now employ automated data processing and analysis tasks once a dataset is collected. These are convenient for quickly analyzing the data to assist in deciding how to proceed further.

2. Other programs such as phenix.xtriage (PHENIX) and Cell Content Analysis (CCP4) may also be used to analyze the data to determine other important information including the number of molecules per asymmetric unit, space group options, strength of anomalous signal, overall data quality, if twinning is present, and if pseudo translational symmetry is present.

3. If solving the structure using molecular replacement, use the information gathered in step 2 to set up a molecular replacement job within PHENIX or CCP4 using standard protocols. For example, if using PHASER, select and load your sequence file, your processed data file, and your cleaned search model. Complete the cell composition details and set up for a search procedure, being sure to turn on the option to try ‘All possible in same pointgroup’. Once a correct solution has been found, proceed to iterative cycles of refinement and model building until the model matches the data as best as possible with suitable R/R_free values.

4. If solving the structure using experimental phasing, use the information gathered in step 2 to set up an experimental phasing job within PHENIX, CCP4, SHELX or other suitable program using standard protocols. For example, if using AutoSol (PHENIX), define the phasing method and select and load your sequence file, your processed data file(s) and corresponding values for wavelength, f’, f’’, the anomalous scattering atom, and the estimated number of anomalous atoms present. Adjust the parameters for the substructure search as desired, including whether or not a model will be built and the number of processors to use. Once a correct substructure solution has been found, a density-modified map will be calculated which you can open and view using COOT or PyMOL. A structure is deemed ‘solved’ if secondary structure elements and side chains are observed in the electron density. From here, perform an autobuilding job that should build much of the structure (with good density) and then proceed to iterative cycles of refinement and model building until the model matches the data as best as possible with suitable R/Rfree values.