Expression, purification, and characterization of a recombinant flavin reductase from the luminescent marine bacterium Photobacterium leiognathi

A set of exercises for students


  • Thomas E. Crowley

    Corresponding author
    1. Department of Chemistry and Biochemistry, University of California, San Diego, La Jolla, California 92093
    • Mail code 0303,9500 Gilman Dr, La Jolla, California, United States 92093
    Search for more papers by this author
    • Tel.: 858-822-5653.


In Photobacterium, the flavin reductase encoded by luxG regenerates the reduced form of flavin mononucleotide (FMN). Reduced FMN is one of the substrates of the luciferase enzyme that catalyzes a light-emitting reaction. A set of experiments, that employs a luxG-expression plasmid construct (pGhis) and is suitable for an undergraduate laboratory course, is presented. Hexahistidine-tagged protein is expressed in E. coli from pGhis, with the T7 RNA polymerase/lac repressor induction system. Bacteria are lysed by sonication and the tag allows for purification by immobilized metal ion affinity chromatography. A gel filtration column is used to remove ions and the other small molecules. The Bradford assay, with multiwell plates and an automated plate reader, is used to identify protein concentration peaks from both columns. The concentration of purified enzyme is then calculated from its A280 using the predicted extinction coefficient. Yield and purity are further assayed with SDS-PAGE. Activity of purified enzyme is measured with riboflavin or FMN as substrate. Reaction rate is quantified by monitoring decrease in A340 as the redox partner, NADH, is oxidized.

Bioluminescence occurs in a variety of organisms; e.g., firefly, marine invertebrates and bacteria. Most of the luminescent bacteria exist in symbiosis with fish, proliferating within certain tissues and making the fish appear to be luminescent. The reaction that generates the light energy is catalyzed by the luciferase enzyme in all these organisms. However, the substrates of the reaction vary quite a bit. In the luminescent marine bacteria Photobacterium leiognathi and Photobacterium fischeri (previously known as Vibrio fischeri), one of the substrates is the fully-reduced, quinone form of flavin mononucleotide (FMN): FMNH [1].

equation image

As the light-generating reaction involves oxidation of FMNH to FMN, sustained luminescence necessitates the recycling of FMN back to the reduced form. Such a reaction (Fig. 1) is usually catalyzed by a flavin reductase enzyme [2] (the reduced form is often shown as FMNH2, however at pH > 6.3 it ionizes to FMNH [3]). NADH is oxidized as FMN is reduced.

equation image
Figure 1.

Reduction of FMN (and subsequent ionization) to generate FMNH.

In P. fischeri and P. leiognathi the luciferase enzyme is an αβ heterodimer [4]. The genes for these subunits, luxA and luxB, are encoded within the lux operon [5–8]. This operon also includes luxG that codes for a flavin reductase (protein is designated LuxG [9]). Thus, the translation of the lux operon-encoded mRNA should generate equimolar amounts of luciferase enzyme and flavin reductase.

Experiments with an E. coli-expressed, recombinant LuxG in Dr. P. Chaiyen's lab (Department of Biochemistry, Mahidol University, Thailand) have revealed that its characteristics are quite similar to that of the natural enzyme expressed in P. leiognathi [9]. The LuxG expression plasmid construct was provided by Dr. Chaiyen for development of student exercises.

LuxG from P. leiognathi is comprised of 237 amino acids and has a predicted molecular mass of ∼26.3 kDa (the recombinant protein, LuxG-His6, includes a hexahistidine on the carboxy terminus that increases it to ∼27.1 kDa). The 37% amino acid sequence identity between LuxG and the E. coli flavin reductase Fre suggests similar tertiary structures [2]. It has been predicted that the amino-terminal half of LuxG binds FMN/FMNH, while the carboxy-terminal half binds NADH/NAD+ [9]. The flavin-binding domain is likely to consist of one α helix and six β strands while the NADH/NAD+-binding domain probably contains four α helices and six β strands. LuxG appears to function as a homodimer.

The exercises described here demonstrate the power of recombinant DNA technology for high-yield expression and rapid purification of an enzyme. The first part of Exercise 1 requires that the students come back to the lab in the evening after inducing protein expression in the early afternoon, and Exercise 5 involves an overnight staining step. However, each of the remaining exercises may be completed in a 5-hour lab session.


Equipment for: Protein Expression and Purification, Concentration Assays, and Activity Assays

A shaking incubator set at 16 °C is necessary for growth of the bacterial culture. If a refrigerated incubator is not available, moving a heating incubator to a cold room is the most practical solution. A model S-450D sonifier (3 mm microtip) from Branson Ultrasonics was used for disruption of bacterial cell walls and membranes. A Beckman J2-HC refrigerated centrifuge and JA-14 rotor were used for pelleting bacteria and the post-lysis clearing spin. Spectrophotometers: Manual assays for protein concentration by the Bradford technique (A595), and enzyme activity assays (A340) were done with a Spectronics 20D+ (Thermo Instrument Systems) or an Amersham Ultrospec 2100 pro. Manual assays for protein concentration by direct measurement of A280 were done with the Ultrospec. Bradford and A280 assays were also performed with an Optima FLUOstar automated plate reader (BMG Labtech).

Equipment for Electrophoresis

Accupower 300 electrophoresis power supplies were from VWR Scientific. For gel casting, the JGC-2 kit from Thermo Scientific/Owl Separations Systems was used. The gel plates were 16 cm wide, 14 cm high and the gels were prepared with a thickness of 1.5 mm.


Conical-bottom, 15-mL plastic tubes were used for sonication (it was critical to use this size tube to keep the sonifier probe submerged while avoiding splashing). Quartz cuvettes used for UV spectroscopy had a narrow chamber that required only 0.5 mL of sample. Multiwells plates used for automated plate reader: 1) for Bradford assays, standard plastic 96-well plates (Nunc 269620); 2) for A280, UV-transparent, 96-well plates (Corning/Costar 3635). Plastic columns for affinity purification and desalting chromatography (1-cm inside diameter, 10-cm long) were obtained from Kimble-Chase Life Sciences, Kontes Glass Division.

LuxG-His6-expression Reagents

Isopropyl-α-D1-thiogalactopyranoside (IPTG) was obtained from Sigma Chemicals and dissolved at 100 mM in water. The solution was sterilized by filtration (0.2 μm pore). The pGhis plasmid construct (Fig. 2) that produces LuxG-His6 in response to IPTG induction has been described [9]. It was constructed by insertion of the LuxG coding sequence (CDS) in the pET24b vector which contains the gene for kanamycin resistance. The Tuner (DE3) strain of E. coli [FompT hsdSB(rBmB) gal dcm lacY1 (DE3)] was used as host because it allows for regulation of the level of expression from a pET24b construct by modulation of the IPTG concentration. Both vector and host originated at EMD Biosciences/Novagen. The host carrying the LuxG construct is referred to as Tuner/pGhis. Cultures of this strain were inoculated by the instructor and provided to the students to begin exercise 1. LB broth with 30 μg/mL kanamycin was used for growing bacteria.

Figure 2.

The pGhis plasmid construct used for T7 polymerase/lac repressor-regulated expression of hexahistidine-tagged LuxG in E. coli. The LuxG-His6 CDS can only be transcribed by the T7 polymerase. The lac operator (not indicated on map) is between the T7 promoter and the luxG CDS. The sequence encoding the histidines is at the 3′ end of the luxG CDS. Kan-res encodes an enzyme providing host cells with resistance to kanamycin, while lacI encodes the lac repressor protein. The map was drawn with the Vector-NTI program from Invitrogen.

LuxG-His6-purification Reagents

The proteinase inhibitor PMSF was dissolved at 10 mg/mL (60 mM) in ethanol and stored at −20 °C. Bacteria resuspension solution: 50 mM Tris pH 8, 100 μM PMSF, 10% glycerol (LuxG-His6 solubility increased by glycerol). His-Bind Resin from EMD Biosciences/Novagen: uncharged iminodiacetic acid (IDA) agarose resin. For each prep, 1 mL of settled resin is enough (binding capacity is ∼8 mg/mL bed volume). Resin-charging solution: 50 mM NiSO4. Solutions used to apply the cell lysate to the immobilized metal ion affinity chromatography (IMAC) column all contain: 50 mM Tris pH 8, 250 mM NaCl, 10% glycerol. Imidazole is included at: 20 mM for binding, 50 mM for washing and 250 mM for elution. Desalting column: Sephadex G-25 for molecular biology, DNA grade, Superfine (Sigma Chemicals). Desalting Column Buffer (also used for reductase activity assay): 50 mM Tris pH 8, 1 mM dithiothreitol (DTT), 10% glycerol.

Electrophoresis Reagents

Protein Sample Preparation

1) 10% β-mercaptoethanol (βME), 50 mM sodium phosphate pH 7.2; 2) 80% glycerol, 0.25% bromophenol blue (BPB); 3) 10% SDS.

Acrylamide Gel Preparation

1) 4× stacking gel buffer: 0.4% SDS, 0.5 M Tris pH 6.8; 2) 4× separating gel buffer: 0.4% SDS, 1.5 M Tris pH 8.8; 3) 30% acrylamide, 0.8% bisacrylamide; 4) polymerization catalysts:10% ammonium persulfate (prepared just before use) and N,N,N′,N′-tetramethylethylenediamine (TEMED).

10× Tris-glycine Electrophoresis Buffer

0.25 M Tris pH 8.8, 1.92 M glycine, 1% SDS.


Simply Blue (Invitrogen) was used to stain the proteins after electrophoresis. This solution contains Coomassie Brilliant Blue G-250 (rather than the R-250 version typically used to stain gels) in a nonorganic solvent. It is less hazardous than the methanol/acetic acid solution often used for gel staining. If Simply Blue cannot be obtained, substitution of ethanol for methanol makes this procedure somewhat less hazardous.

Electrophoresis Markers

1) Standard Polypeptide Mix (0.5 μg/μL, run 5 μg, six bands): BSA (66.6 kDa), bovine liver catalase (59.8 kDa), chicken ovalbumin (42.8 kDa), rabbit glyceraldehyde 3-phosphate dehydrogenase (GAPDH; 35.8 kDa), bovine carbonic anhydrase II (31.0 kDa), soybean trypsin inhibitor A (20.2 kDa); 2) Single-band marker: bovine heart lactate dehydrogenase (LDH, 36.6 kDa, 2 μg per lane); 3) Single-band marker: myoglobin (17.2 kDa, 5 μg per lane). On some gels a doublet was seen for carbonic anhydrase. It was assumed that the upper band was 31.0 kDa and the lower due to proteolytic degradation.

Miscellaneous Reagents

Bradford staining reagent [10]: 0.01% Coomassie Brilliant Blue G-250, 5% ethanol, 9% phosphoric acid. Standard protein solutions for Bradford assay: 1 mg/mL bovine γ-globulin (IgG fraction, Bio-Rad) and 1 mg/mL BSA (prepared from crystalline fraction V, EMD Biosciences). Reducing agent: 1 M DTT. Reductase activity assay substrates [9]: disodium-NADH: ϵ340 = 6.22 mM−1 cm−1, 20 mM stock; sodium-FMN: ϵ450 = 12.2 mM−1 cm−1, 4 mM stock; riboflavin: ϵ450 = 11.3 mM−1 cm−1, 400 μM stock.


Gloves should be worn when handling the following toxic chemicals: PMSF, NiSO4, acrylamide, TEMED, and Coomassie Brilliant Blue G-250 (a component of the Bradford reagent and the Simply Blue stain). Ear protection should be worn when operating the sonifier.


These experiments were all performed by students working in pairs. The schedule is summarized in Table I.

Table I. Schedule
11Exer 1, pt 1: Protein expression
 2Exer 1, pt 2 and 3: Lysis of cells and protein concentration by Bradford assay
23Exer 2: IMAC purification
 4Exer 3 & 4: Desalting. Protein concentration by A280
35Exer 5, pt 1: SDS-PAGE (stain gel till next session)
 6Exer 5, pt 2: Destain gel and data analysis
  Exer 6: Activity assays

Exercise 1: Expression of LuxG-His6 in E. coli and Release by Lysis

Part 1: Expression

The inducible expression system used here is a variation of that originally described by Studier [11, 12]. It makes use of the distinct promoter specificities of E. coli RNA polymerase and the RNA polymerase encoded by the T7 bacteriophage. Components of the well-characterized lac operon of E. coli [13, 14] are also used for this regulation. The lac repressor DNA-binding protein (encoded by lacI) and its target sequence, the lac operator, are used to prevent expression of the LuxG-His6 protein until the inducer, IPTG, is added to the medium. This expression control is necessary to prevent possible growth inhibition due to expression of a foreign protein. Although the early phase of culture propagation is at 37 °C (the optimal temperature for E. coli), after induction the culture is moved to a 16 °C incubator. This slows the cell cycle from ∼30 minutes to ∼7 hours and slows the rate of protein synthesis. The more gradual accumulation of the LuxG-His6 at the lower temperature is critical to keep the protein soluble [9].

A culture of 1 × 108 cells/mL has an A600 = 0.6. The culture will be grown at 37 °C until the A600 = 0.4–0.5 (induction density), at which time the inducer will be added. The incubation will then be shifted to 16 °C and allowed to continue for 7 hours. The A600 should double during this incubation. The ideal negative control is a Tuner/pGhis culture incubated for the same times at the same temperatures but not induced with IPTG. The instructor should prepare and lyse such a culture to provide a control sample for the SDS-PAGE in Exercise 5.

  • 1)At ∼8:00 AM: Inoculate a 200 mL LB/kanamycin culture with three colonies (1 mm diameter) of Tuner/pGhis and incubate at 37 °C with shaking. Inoculation with a small aliquot of a saturated overnight culture may give more consistent growth (but this has not been tested).
  • 2)Expect the culture to reach the induction density in ∼5 hours. Begin assaying absorbance after 4 hours of incubation.
  • 3)When induction density is attained: Cool to 16 °C, then add IPTG to 0.4 mM.
  • 4)Grow at 16 °C for 7 hours to allow expression.
  • 5)Assay A600 again to confirm growth during induction.

Part 2: Lysis of Tuner/pGhis Cells

Spin down culture using 250 mL centrifuge bottles: 6,000 rpm, 20 minutes at 4 °C. Resuspend (on ice) in 5 mL of Bacteria Resuspension Solution. Concentration should now be ∼6 × 106 cells/μL. Transfer to a 15 mL, screw-cap, conical bottom tissue culture tube (do not use a larger tube). Sample may be frozen in dry ice/isopropanol and stored at −70 °C until ready for sonication step.

Lyse cells with sonication (keep sample on ice, wear ear protection): This step is a bit tricky. If it does not work well it will be difficult to complete the remaining exercises. It is critical that: (a) ≥90% of the bacteria are lysed, (b) splashing is minimized so that the cells are consistently exposed to the full sonifier power and sample loss is avoided, and (c) the enzyme is not denatured by increase in temperature. Use a Branson 450 sonifier. Amplitude = 50% (50 Watts). Duty cycle: pulse on = 0.7 seconds, pulse off = 0.3 seconds. Three rounds, 30 seconds each. Despite being on ice during this procedure, the sample will warm up. After sonication, leave on ice for at least 10 minutes to bring temperature back down before proceeding. Transfer sample to a tube that will withstand 26,000 × g force and keep on ice. Centrifugation of sonicated lysate to remove membranes and membrane-associated proteins: 13,000 rpm (26,000 × g), 60 min, 4 °C. Transfer supernatant to a clear-plastic tube suitable for dry ice-freezing. The solution should be pale yellow. If it shows no color, the cells were not lysed. Concentration of total protein should be ∼14 mg/mL. This is the “initial supernatant”.

Examine pellet for inclusion bodies: Foreign proteins expressed in E. coli sometimes become sequestered in these structures if the tertiary structure has not formed properly. The membrane/cell wall portion of the pellet will be light brown whereas inclusion bodies are white. The pellet will be discarded because any LuxG-his6 in it is unlikely to have activity. Even if some of the enzyme is in the pellet, the majority should be in the supernatant.

Remove appropriate aliquots of the initial supernatant: (a) Determination of protein concentration (40 μL). (b) Analytical SDS-PAGE (60 μL). (c) Activity assays (540 μL). (1 μL should contain lysate from ∼6 × 106 cells. Usually run lysate from ∼15 × 106 cells on gel. For activity assay, usually use ∼30 μL per mL of reaction mix.)

Remainder should be frozen as one aliquot in dry ice/isopropanol and stored at −70 °C. This aliquot will be put through affinity purification.

If the three assays mentioned earlier will not be done immediately, also freeze the small aliquots. Storage: −20 °C for the concentration and gel samples, −70 °C for the activity sample.

Part 3: Protein Concentration in the Initial Supernatant with the Bradford Assay

Use the following procedure for assays of multiwell-plate samples in an automated plate reader (reaction volume ∼0.2 mL). For standard curve: use aliquots of bovine γ-globulin solution containing: 1, 2, 3, 4, 5, and 6 μg. The highest concentration will be 30 μg/mL. Mix each protein aliquot (<0.02 mL) with 0.2 mL of Bradford reagent. Allow 20 minutes at room temperature before quantifying the A595. Expect an absorbance of 0.4–0.5 for the 30 μg/mL reaction. To assay absorbances manually with a Spectronics D20+, reaction volume must be at least 3 mL. In this case, the standard masses must be increased.

Exercise 2: Affinity Purification of LuxG-His6 by IMAC

IMAC purification of His-tagged recombinant proteins is a very commonly-used technique [15]. In most cases, the majority of the tagged protein can be recovered from a complex mixture with very little contamination of undesired proteins. The imidazole structure in the histidines linked to the LuxG sequence will bind the Ni2+ ions on the column when applied in the low-imidazole buffer. All the host cell (E. coli) proteins flow through in these conditions. After washing the column-bound protein with the intermediate-imidazole buffer, the tagged LuxG is eluted with the high-imidazole buffer. The procedure is performed in the cold room to maintain enzyme activity. Although LuxG activity is stabilized by the presence of a reducing agent such as DTT [9], this is not added until after the nickel column. Otherwise the DTT will reduce the Ni2+ ions and inhibit binding of the hexahistidine tag. Expect ∼4% of Tuner/pGhis lysate protein mass to be LuxG-His6. For ∼50 mg of total protein, 1 mL of packed resin (the “bed volume”) is sufficient. Once the lysate is loaded, it takes less than an hour to collect the necessary fractions. Therefore an automated fraction collector is not needed. Fractions may be collected manually in 1.5 mL snap-cap tubes.

  • 1)Set up column in cold room. Fill half-way with water, then pipette in appropriate amount of resin.
  • 2)After resin settles, estimate the bed volume and add more if necessary.
  • 3)Charge resin with the NiSO4 solution followed by ten bed volumes of binding solution. Resin will turn green when Ni2+ ions bind.
  • 4)Apply initial supernatant to column, then follow with six bed volumes of binding solution to push most of the non-tagged protein through.
  • 5)Wash with five bed volumes of wash solution to ensure that all nontagged protein has been removed.
  • 6)Elute LuxG-His6 with five bed volumes of elution solution, collecting 1 mL fractions.
  • 7)If the bed volume is ∼1 mL, expect tagged protein to be distributed in three fractions.
  • 8)Close stopcock so that the (reusable) resin does not dehydrate.
  • 9)Use Bradford assay as described in Exercise 1, Part 3 to identify the fractions containing the majority of the LuxG-His6. Assay 10 μL of each column fraction with 0.2 mL of Bradford reagent.
  • 10)Pool fractions containing the majority of the protein and add DTT to 1 mM final.
  • 11)Use the data from Step 9 to calculate the concentration of protein in the pool.
  • 12)Remove aliquots of pool for subsequent assays: (a) SDS-PAGE: an aliquot containing 24 μg protein; (b) Activity assay: 360 μL (need 20 μL per mL of reaction mix).
  • 13)If not ready for the desalting step, freeze remainder of the affinity-purified protein in dry ice/isopropanol and store at −70 °C. If the gel and activity assay will not be done immediately, also freeze these samples and store at: −20 °C for the gel aliquot and −70 °C for the activity aliquot.

Exercise 3: Desalting of LuxG-His6 by Column Chromatography

A desalting column can be used to remove ions and small molecules from a solution containing protein or nucleic acid [15]. In this exercise, Sephadex G-25 resin will be used. The tagged protein will be excluded from the beads and migrate in the void volume; i.e. it will not be slowed down by the resin. This is because the molecular mass of the protein is greater than the exclusion limit (5 kDa). The protein concentration should not change significantly during this procedure. The smaller molecules (imidazole) and ions (Na+ and Cl) will be slowed down because they weave through the interior of the beads. Thus, the LuxG-His6 protein is recovered in the desalting column buffer. This is an appropriate solution for the activity assay. To measure the concentration of desalted protein by the UV-absorbance method (Exercise 4), it is critical that all of the imidazole (which absorbs at 280 nm) be removed. As with the IMAC column, the procedure is performed in the cold room to avoid loss of enzyme activity and fractions are collected manually in 1.5 mL snap-cap tubes.

  • 1)Prepare the Sephadex column (in water initially) in a manner similar to that used for the IMAC column. If the volume of the pooled fractions from the affinity column is 3–4 mL, a 5 mL bed volume will be sufficient.
  • 2)Only one buffer (recipe in Equipment, Supplies and Reagents) is used for a desalting column. After the resin settles in water, equilibrate it in desalting buffer.
  • 3)Remove excess buffer from above the surface. Apply the affinity-purified protein solution and allow to flow into the resin.
  • 4)When ∼50% of the sample has entered the resin, begin collecting 1 mL fractions.
  • 5)After all of the sample has entered the resin, fill the upper part of the column with desalting column buffer and continue until 12 fractions have been collected.
  • 6)After collection is finished, close stopcock so that the (reusable) resin will not dehydrate.
  • 7)Identify fractions containing the majority of the protein: Assay a 10 μL aliquot of each by the Bradford method as in Exercise 1, Part 3. Use 0.2 mL of Bradford reagent for each reaction.
  • 8)Pool the fractions containing the majority of the protein and use the data from step 7 to calculate the concentration in the pool.
  • 9)Remove aliquots of the pool for subsequent assays: (a) Direct spectrophotometry, 500 μL; (b) SDS-PAGE, an aliquot containing 24 μg protein; (c) Activity assay, 360 μL (need 20 μL per mL of reaction mix).
  • 10)Freeze remainder of desalted LuxG-His6 in dry ice/isopropanol and store at −70 °C. If the next three assays will not be done immediately, also freeze the small aliquots and store at: −20 °C for the spectrophotometry and gel aliquots, and −70 °C for the activity aliquot.

Exercise 4: LuxG-His6 Concentration by Direct Spectrophotometry

The concentration of a pure protein may be determined by quantifying absorbance at 280 nm [16–18]. Most of the absorbance of proteins at this wavelength is due to the sidechains of the aromatic amino acids tryptophan and tyrosine. A small amount may also be contributed by cystines; i.e., cysteine sidechains in a disulfide linkage (each cysteine in such a disulfide structure is a half cystine). Thus if the sequence of the protein is known and the possible presence of cystines predicted, an extinction coefficient (ϵ280) may be calculated (in M−1 cm−1). Protein ϵ280 values are available in the Expert Protein Systems Analysis (ExPASy) database, within the Proteomics and Sequence Analysis Tools/Protein Parameters section (website below). This assay should only be done on the desalted sample because the imidazole present after the affinity purification absorbs at 280 nm.

  • 1)Search the UniProt Knowledgebase ( for P. leiognathi LuxG entries.
  • 2)Using one of the obtained accession numbers, calculate the parameters for this protein using the Protein Parameters tool at:
  • 3)There should be two ϵ280 values given: one value assumes the maximum number of cystines the other assumes no cystines. LuxG has seven cysteines and therefore the native protein may have three cystines. However the reducing agent DTT that has been added to the purified LuxG-His6 may have disrupted some or all of the cystines. Therefore it is appropriate to use the second ε280 value mentioned earlier.
  • 4)Measure the A280 of the desalted protein: (a) Sample aliquot: If using a conventional spectrophotometer and a quartz cuvette it is best to use a cuvette with a narrow chamber so that only. 0.5 mL of sample is sacrificed. If using a UV-transparent, 96-well plate and a plate reader, 0.2 mL should be enough; (b) Set the baseline: As the 10% glycerol in the desalting column buffer gives an A280 of ∼0.1, use this as the blank; (c) Measure the A280 of the protein sample. This type of spectrophotometry is accurate in the range: 0.1–1.0.
  • 5)Use the ϵ280 of LuxG to calculate the concentration of the desalted protein in μM and mg/mL. Use this concentration for preparing aliquots for the remaining experiments.

Exercise 5: SDS-PAGE of Tuner/pGhis Lysate and Purified LuxG-His6

The standard method of denaturing proteins and separation by electrophoresis in acrylamide is used [19–22]. Although the LuxG expressed in P. leiognathi has been shown to function as a homodimer, it migrates as a monomer on a denaturing gel. Samples to be examined: 1) uninduced Tuner/pGhis negative control lysate (provided by instructor), 2) initial supernatant of induced Tuner/pGhis from Exercise 1, 3) the affinity-purified sample from Exercise 2 and 4) the desalted sample from Exercise 3.

Part 1: Preparation of Gel and Samples, Electrophoresis, and Staining

  • 1)Prepare the 15% acrylamide separating gel: 15 mL 30% acrylamide, 0.8% bisacrylamide; 7.5 mL 4× separating gel buffer, 7.5 mL water, 400 μL of 10% ammonium persulfate, 15 μL TEMED.
  • 2)Prepare samples for electrophoresis: (a) The amount of each sample to prepare for electrophoresis is given in Table II; (b) The amounts of denaturing reagents and tracking dye to add to each protein sample are given in Table III.
  • 3)Prepare the 4.5% acrylamide stacking gel: 2.5 mL 4× stacking gel buffer, 1.5 mL 30% acrylamide, 0.8% bisacrylamide, 6.0 mL water, 100 μL 10% ammonium persulfate, 20 μL TEMED.
  • 4)Just before loading on gel: Heat samples in boiling water bath for 60 seconds, then spin to bring all liquid to bottom of tube.
  • 5)Electrophoresis: (a) Dilute 10× Tris-glycine running buffer to 1×; (b) Run at 125 V, 55 mA until dye enters sep gel (∼30 minutes); (c) Increase to 250 V, 110 mA. Run until dye is at bottom (∼60 minutes).
  • 6)Staining: (a) Mix 135 mL of Simply Blue with 15 mL of 20% NaCl; (b) Rinse the gel with water three times; (c) Submerge gel in stain solution until next lab session.
Table II. Amounts of protein samples to analyze by SDS-PAGE
SampleNumber of cellsMass of protein (μg)
Standard polypeptidesn.a.5
Uninduced lysate15 × 10624
Initial supernatant15 × 10624
30 × 10648
Affinity-purified LuxG-His6n.a.2
Desalted LuxG-His6n.a.2
Table III. Amounts of denaturing reagents for SDS-PAGE samples
ComponentVolumeFinal concentration
  • a

    Concentrations of the components of these stocks are in Equipment, Supplies and Reagents.

βME, sodium phosphatea5 μL1% βME, 5 mM phos.
Glycerol, BPBa10 μL16% glyc, 0.05% BPB
10% SDS1 μL0.2%
 50 μL total 

Part 2: Destaining and Data Analysis

Rinse gel with water. Background should not be as high as on a gel stained with Coomassie/methanol/acetic acid. Bands may be visible immediately. If not, allow several hours for destaining.

Data Analysis: (a) Prepare a semilog plot of standard protein migration: log10(molecular mass) on the y axis and distance migrated (in mm) on the x axis; (b) Compare the pattern of bands in the initial supernatant lanes to that in the uninduced negative control lane. Is there evidence that the induction worked; (c) Determine the molecular mass of the LuxG-His6 monomer by interpolation on the semilog plot; (d) Comment on the effectiveness of the affinity purification and desalting procedures.

Exercise 6: Flavin Reductase Activity Assays

NADH absorbs at 340 nm much more strongly than does NAD+. Flavin reductase activity is quantified by measuring the rate of decrease in the A340. The assay is done on aliquots of: the initial supernatant, affinity purified LuxG-His6 and desalted LuxG-His6.

  • 1)For each of the three samples to be assayed, calculate the [LuxG-His6] in μM. The initial supernatant contains many proteins. However the affinity purification provides the total mass of LuxG-His6 that was in the initial supernatant. Just divide this mass by the volume of the supernatant, then convert from mg/mL to μM. Calculate the volume of each of the three samples to add to a 3-mL reaction to give 0.4 μM enzyme final (50–100 μL if the yield of LuxG-His6 was close to the predicted value given in Exercise 2).
  • 2)Components of the various reactions are shown in Table IV. Reactions 1–3 should be done for each of the three protein samples (reaction 3 is a “no flavin” negative control). As reactions 4 and 5 are “no protein” negative controls with each of the two flavin substrates, only one of each of these needs to be done. The last component to be added is NADH and this should be done just before placing the tube in the spectrophotometer (exposure to atmosphere may prematurely oxidize NADH).
  • 3)Record the A340 of the reaction immediately after mixing and at 30 seconds timepoints for at least 3 minutes. The A340 should initially be 1.0–1.2 and the rate of change should be 0.1–0.3 min−1. Any change observed in the negative controls should be subtracted from the experimental values.
  • 4)Use the ϵ340 value for NADH to convert the absorbance changes to enzyme activity values in μmol/min/mL. Do this calculation for each of the protein samples with each of the two flavin substrates.
  • 5)Calculate the specific activity of the affinity-purified and desalted samples in μmol/min/mg. This calculation is not done for the initial supernatant because the Fre flavin reductase expressed by the E. coli host cells is likely to contribute to the activity [2].
  • 6)Calculate the turnover number (in s−1) for the affinity-purified and desalted samples. This is the number of substrate molecules (or moles) converted to product by each enzyme active site (or mole of active sites) each second [23, 24]. LuxG-His6 functions as a homodimer. However there is one active site in each subunit, so use the monomer molecular mass for this calculation. Compare the value obtained to those previously reported for LuxG and other flavin reductases [9, 25]. Suggest explanations for significant differences.
Table IV. Flavin reductase activity assay of the initial supernatant or purified LuxG-His6
ComponentReac 1Reac 2Reac 3Reac 4Reac 5Final conc.
  • a

    The protein volume is usually <0.1 mL so it does not increase the total reaction volume significantly. If the protein volume is greater than this, decrease the volume of buffer accordingly.

  • b

    Use the desalting column buffer (Equipment, Supplies and Reagents).

ProteinVariableaVariableaVariableaNoneNone0.4 μM LuxG
NADH0.03 mL0.03 mL0.03 mL0.03 mL0.03 mL200 μM
FMN0.03 mL  0.03 mL 40 μM
Riboflavin 0.3 mL  0.3 mL40 μM
Bufferb2.9 mL2.6 mL2.9 mL2.9 mL2.6 mL 
Total∼3 mL∼3 mL∼3 mL∼3 mL∼3 mL 


Protein Concentration Assays

Concentrations reported in all figures and Tables were initially calculated from a BSA standard curve. However, after it was learned that BSA binds Coomassie Blue more efficiently than most other proteins, these values were corrected. Bovine γ-globulin binds Coomassie to an extent similar to that of most other proteins so it is the better standard [26]. Comparison of the staining-signal slopes for a range of BSA and γ-globulin concentrations (Fig. 3) showed a BSA:γ-globulin signal ratio of 1.63:1 and 1.87:1 in duplicate trials (average = 1.75:1). Although the range used in these trials (0–20 μg/mL in the staining reaction) was the same as that used for all assays reported below, a higher range is recommended in Experimental Procedures for future experiments. As all concentrations determined by reference to BSA were underestimates, they have been multiplied by 1.75 to give more accurate values.

Figure 3.

Comparison of Bradford Staining Intensity of γ-globulin and BSA. The A595 of the staining reactions are plotted as a function of the mass of protein in each (0.2 mL) reaction. (a) First trial. (b) Second trial.

The correspondence between A280 and the Bradford assay (γ-globulin standard) for quantification of LuxG-His6 concentration was also evaluated. A sample of affinity-purifed and desalted LuxG-His6 (distinct from thosereported below) was assayed by the instructor with both techniques. The A280 assay gave a concentration of 0.14 mg/mL (5.3 μM) while Bradford gave 0.21 mg/mL (7.8 μM).

Instructor's Expression of LuxG-His6 in Tuner/pGhis Cells, Lysis by Sonication and Quantification of Total Protein

A Tuner/pGhis culture was incubated and induced with IPTG as described in Experimental Procedures. The majority (98%) of the cells were lysed by sonication (followed by a clearing spin) for the purpose of purifying native LuxG-His6. Concentration of total soluble protein in the “initial supernatant” was determined with the Bradford assay to be 14 mg mL−1. The remainder of the culture was pelleted, then boiled in SDS to provide a denatured lysate. Aliquots of the initial supernatant and the denatured lysate were then examined by electrophoresis (Fig. 4). A band of the expected molecular mass (27.1 kDa) is very prominent in initial supernatant samples corresponding to 18 or 72 × 106 cells. Although the band at this position in the denatured and 6 × 106 cell-initial supernatant lanes is very faint, this result is consistent with efficient induction of LuxG-His6 expression in Tuner/pGhis cells.

Figure 4.

Induction of LuxG-His6 expression in Tuner/pGhis bacteria by IPTG. SDS-PAGE (15% acrylamide) and Simply Blue staining were used. Lanes 1–4: (1) LDH, (2) myoglobin, (3) Standard Polypeptides (see Equipment, Supplies, and Reagents), (4) SDS-lysate of 15 × 106 cells. Lanes 5–7: aliquots of sonicated lysate (membranes removed by centrifugation). Number of cells: lane 5 = 6 × 106, lane 6 = 18 × 106, lane 7 = 72 × 106. Molecular masses are given in kDa. Arrow on right marks LuxG-His6 at 27.1 kDa.

Instructor's Purification of LuxG-His6 from the Tuner/pGhis Initial Supernatant: IMAC Followed by Desalting Chromatography

The His-bind resin was charged with Ni2+ ions and used to purify the tagged protein. Aliquots of the flow-through during the binding and wash steps, and the elution fractions, were examined by SDS-PAGE (Fig. 5). LuxG-His6, migrating at 27.1 kDa, was efficiently selected by the His-bind resin. The majority of this protein was in the first four elution fractions, while very little was detected in the flow-through fractions. There was no sign of degradation and only slight contamination (∼36–80 kDa) was detected. The remainder of fractions 1–4 was pooled and protein concentration in the pool determined to be 0.88 mg/mL by Bradford assay.

Figure 5.

Instructor's evaluation of fractions from the IMAC purification of LuxG-His6. SDS-PAGE (15% acrylamide) and Simply Blue staining were used to detect the proteins in each fraction. Markers in the first three lanes: (1) LDH, (2) myoglobin and (3) Standard Polypeptide mix. Lanes 4–10, aliquots of column fractions: (4) 20 μL of the flow-through during binding, (5) 50 μL of the flow-through during washing, (6–10) 33 μL aliquots of each 1 mL elution fraction.

The LuxG-His6 eluted from the IMAC column was further purified by desalting chromatography. Protein concentration in the 1 mL fractions from this second chromatography procedure was determined with the Bradford assay. A pool of the peak fractions had a concentration of 0.75 mg/mL. Aliquots with 3.5, 18, or 35 μg were examined by SDS-PAGE (Fig. 6). The LuxG-His6 migrating at 27.1 kDa showed no sign of degradation. Only a very slight contamination (∼70–80 kDa) was detected.

Figure 6.

Instructor's evaluation of purity and integrity of LuxG-His6 after desalting chromatography. Samples were examined by SDS-PAGE (15% acrylamide) and Simply Blue staining. Markers in the first three lanes: (1) LDH, (2) myoglobin, and (3) Standard Polypeptide mix. Lanes 4–6 contain: 3.5, 18, or 35 μg of desalted LuxG-his6, respectively.

Students' Electrophoresis of a Tuner/pGhis Initial Supernatant and Purified LuxG-His6

LuxG-His6 was expressed and purified as described in Experimental Procedures. As a negative control, the instructor prepared a Tuner/pGhis lysate from cells grown under conditions identical to the experimental culture but not induced with IPTG. Aliquots from these samples were analyzed by SDS-PAGE as described in Experimental Procedures (except that the masses of protein loaded were greater). A protein of the molecular mass of LuxG-His6 (27.1 kDa) was induced by IPTG in Tuner/pGhis cells and the IMAC procedure selected the induced protein cleanly from the lysate (Fig. 7). As expected, the band at 27.1 kDa was of similar intensity for the 3.5 μg affinity-purified and desalted samples, and also similar for the 18 μg aliquots of each of these samples. There was no evidence of protein degradation.

Figure 7.

Student's demonstration of induction of LuxG-His6 in Tuner/pGhis cells and purification by IMAC. SDS-PAGE (15% acrylamide) and Simply Blue staining were used to separate and detect the proteins. Lanes contained: (1) Standard Polypeptides (smallest protein, soybean trypsin inhibitor A, was not detected), (2) 42 μg of the negative control, uninduced initial supernatant, (3 and 4) 42 and 84 μg of the induced initial supernatant, (5 and 6) 3.5 and 18 μg of the affinity purified LuxG-His6, (7 and 8) 3.5 and 18 μg of the desalted LuxG-His6.

Instructor's Assay for Flavin Reductase Activity in the Tuner/pGhis Lysate and Purified LuxG-His6

Reactions were prepared as described in Experimental Procedures. Aliquots of the initial supernatant, the affinity-purified protein and desalted protein (prepared by the instructor as described earlier) were tested along with no-protein controls. Quantification of enzyme activity, and calculation of activity concentration, specific activity and turnover number were as described in Experimental Procedures. The data and calculations (Table V) revealed a specific activity of 2.8 μmol/min/mg and a turnover number of 1.3 s−1 for the most purified (desalted) enzyme sample.

Table V. Instructor's data for flavin reductase activity in Tuner/pGhis lysate and purified LuxG-His6a
  • a

    Control reactions with no protein and either the FMN or riboflavin substrate showed no activity.

  • b

    Although reaction volume is given as 3 mL in Experimental Procedures, these reactions were scaled down to 2 mL.

  • c

    Specific activity was not calculated for initial supernatant because the assay may detect host cell (E. coli) reductase activity in this fraction.

SampleInit supInit supAffipurAffipurDesaltedDesalted
Conc (mg/mL)14140.880.880.750.75
Aliquot assayedb (mL)0.0300.0300.0200.0200.0250.025
dA/dT (min−1)0.300.
dC/dT (μM min−1)484519162626
Aliquot activ (μmol min−1)0.0960.0900.0380.0320.0520.052
Activ conc (μmol min−1 mL−1)
Spec activc (μmol min−1 mg−1)n.a.n.a.
Turnover (s−1)n.a.n.a.1.00.811.31.3

Students' Assay for Flavin Reductase Activity in Purified LuxG-His6

LuxG-His6 was expressed and purified, and aliquots assayed for reductase activity, as described in Experimental Procedures. The students' data and calculations for the affinity–purified enzyme (Table VI) revealed an activity concentration 79–83% of that determined by the instructor for the affinity-purified sample. The student's values for specific activity (0.73–0.90 μmol/min/mg) and turnover number (0.33–0.41 s−1) were 41% of the values for the instructor's affinity-purified enzyme.

Table VI. Students' data for flavin reductase activity in purified LuxG-His6a
  • a

    Control reactions with no protein and either the FMN or riboflavin substrate showed no activity.

  • b

    Reaction volume was 3 mL as described in Experimental Procedures.

Conc (mg/mL)1.751.75
Aliquot assayedb (mL)0.0320.032
dA/dT (min−1)0.1050.084
dC/dT (μM min−1)16.913.5
Aliquot activ (μmol min−1)0.05070.0405
Activ conc (μmol min−1 mL−1)1.581.27
Spec activ (μmol min−1 mg−1)0.900.73
Turnover (s−1)0.410.33


Although the A280 quantification of LuxG-His6 concentration was 33% lower than the Bradford result, it is still worthwhile to include both techniques. It is important for students to learn multiple options for determining concentrations. Although the student's enzyme activity results were not as good as the instructor's, the values were close enough to indicate that students taking their first biochemistry lab course can perform these techniques successfully. The greatest technical hurdle in these exercises was in obtaining the appropriate concentration of protein in the Tuner/pGhis lysate generated by sonication. The challenge is to expose the cells to sufficient high-frequency pulses to disrupt the majority of the cells without allowing the temperature to increase to the point where LuxG-His6 and other proteins become denatured. In addition to loss of the flavin reductase activity, denaturation may generate aggregates of LuxG-His6 with E. coli proteins. Such aggregates will then pellet with the membranes in the subsequent centrifugation step. Keeping the sample on ice during sonication is certainly recommended. However, even under these circumstances, the solution can warm up to 75 °C.

Although the techniques for purifying proteins directly from cells or tissues should always be included in an undergraduate protein chemistry lab course, it is becoming increasingly important to also teach the recombinant DNA-expression methods. The recombinant DNA techniques provide a practical way to purify mg amounts of many proteins that would be very difficult to purify directly from their natural source. It also allows for expression/purification of a truncated protein or a mutant version that has not been observed in nature. Recombinant proteins such as this are excellent tools for structure-function studies. However, for some investigations, purification of a protein from its natural source is the only means of answering questions regarding: posttranslational modifications, regulation of expression and stability, protein–protein interactions and the effect of naturally occurring mutations. For this reason, protein chemistry courses should also include purification and characterization of endogenous enzymes from tissues.

The LuxG enzyme is well suited to a course for students just beginning to learn biochemistry lab skills because: 1) it is a soluble protein, 2) the subunit molecular mass is in a range easily measured by SDS-PAGE, 3) the only specialized equipment necessary for the activity assay is a Spectronics 20D+ spectrophotometer (common in teaching labs) and 4) in nature it plays a key role in the intriguing phenomenon of bioluminescence. Although a luminescence assay is not included in these exercises, inclusion of this topic in lectures accompanying the lab increases students' interest in the course.


The author thanks Dr. Pimchai Chaiyen at Mahidol University, Thailand for providing the the pGhis plasmid construct. The technical assistance provided by Peter Wotruba in the UCSD Chemistry and Biochemistry department is also greatly appreciated.