See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/5314228

Methods for Protein Characterization by Mass Spectrometry,

Thermal Shift (ThermoFluor) Assay, and Multiangle or Static
Light Scattering

Article in Methods in molecular biology (Clifton, N.J.) · February 2008

DOI: 10.1007/978-1-60327-058-8_19 · Source: PubMed

CITATIONS READS

133 2,632

4 authors, including:

Joanne E Nettleship Matthew R Groves

Astra Health University of Groningen
122 PUBLICATIONS 4,056 CITATIONS 167 PUBLICATIONS 3,871 CITATIONS

SEE PROFILE SEE PROFILE

Arie Geerlof
Helmholtz Zentrum München Deutsches Forschungszentrum für Ge…
122 PUBLICATIONS 3,438 CITATIONS

SEE PROFILE

All content following this page was uploaded by Matthew R Groves on 03 June 2014.

The user has requested enhancement of the downloaded file.

 Chapter 19
Methods for Protein Characterization
by Mass Spectrometry, Thermal Shift
(ThermoFluor) Assay, and Multiangle
or Static Light Scattering
Joanne E. Nettleship, James Brown, Matthew R. Groves, and Arie Geerlof

Mass spectrometry (MS) is widely used within structural and functional

proteomics for a variety of tasks including protein quality assessment,
identification, and characterization. MS is used routinely for the determination
of the total mass of proteins, including N-glycosylated proteins, analysis of
selenomethionine incorporation, crystal content verification, and analysis of
N-glycosylation site occupancy. Protocols for sample preparation, data collection,
and analysis are given.
A recent development is the fluorescence-based thermal shift (ThermoFluor)
assay. It uses an environmentally sensitive dye, Sypro Orange, to monitor the
thermal stability of a protein and investigate factors (e.g., buffers, additives,
and ligands) affecting this stability. This chapter describes the application of
this method using a 96-condition in-house screen. The measurements are
performed on a commercially available real-time PCR machine.
Multiangle or static light scattering (SLS) is a very powerful technique to
determine the conformational state of proteins in solution, especially when
used in combination with size exclusion chromatography (SEC). In the
authors’ experimental set-up the SLS detector is connected in-line to a stand-
ard protein purification machine (e.g., the Äkta Purifier) equipped with an
analytical SEC column. The data collection and analysis are performed using
commercial software.

1. Introduction

Protein characterization plays an important role in two key aspects of

proteomics. The first is the quality assessment of the produced protein prepa-
rations. Obtaining protein of sufficient quality for structural and/or functional
studies is one of the major bottlenecks of most proteomics projects. This is
reflected in the overall statistics for the major structural genomics consortia
(for an overview, see http://sg.pdb.org/target_centers.html), which show that
so far the structures of less than 4% of the cloned genes could be determined.
Hence, it is essential to perform an extensive quality assessment of the protein
preparations prior to their application and use the results in the evaluation

From: Methods in Molecular Biology, Vol. 426: Structural Proteomics: High-throughput Methods
Edited by: B. Kobe, M. Guss and T. Huber © Humana Press, Totowa, NJ 299
300 Nettleship et al.

of the production process (1). The second is the determination of protein

properties such as domains, oligomeric state, posttranslational modifications
and protein–protein and protein–ligand interactions.
In most laboratories SDS-PAGE and dynamic light scattering (DLS) are
routinely used to determine the purity and monodispersity of a protein.
However, the application of these methods is well documented in the lit-
erature; therefore, it is not considered here. This chapter gives protocols
for the sample preparation and data collection and analysis of three other
protein characterization methods that are routinely used in the authors’
laboratories.

1.1. Protein Characterization by Mass Spectrometry

MS is a useful analytical technique and is widely used within structural pro-
teomics consortia (2–6). Analysis of both intact proteins and protein digests
allow for the verification of construct along with any modifications such as
selenomethionine incorporation or methylation as well as giving information
on posttranslational modifications.
The protocols in this section describe how to analyze intact proteins using
minimal sample preparation and automated procedures (Section 3.1.1.);
remove N-glycans prior to automated LC-MS (Section 3.1.2.); and prepare
proteins for LC-MS from crystals (Section 3.1.3.). A protocol for preparation
of peptides by digestion of proteins followed by automated LC-MS or LC-MS/
MS is given in Section 3.1.4., followed by methodology for the determination
of site occupancy of N-glycosylated proteins (Section 3.1.5.).

1.2. ThermoFluor Assay

Thermal shift (ThermoFluor) assays offer a rapid and simple technique for
assessing the thermal stability of proteins and to investigate factors affecting this
stability (1,7,8). An environmentally sensitive fluorescent dye is used to monitor
protein unfolding with respect to temperature. Melting curve analysis determines
the melting temperature, Tm (the midpoint of the unfolding transition); a shift
in Tm under different conditions indicates a change in stability. Commercially
available real-time PCR machines have sensitive thermal control and fluorescent
detection capabilities, allowing the assay to be routinely performed using low
amounts of protein. Thermal shift is amenable to drug screening and screening
of buffer conditions, additives, ligands, and cofactors to indicate promising crys-
tallization or storage conditions or to assign function (1,9,10).

1.3. Protein Characterization by Static Light Scattering

SLS is a noninvasive technique whereby an absolute molecular mass of a
protein sample in solution may be experimentally determined to an accuracy
of better than 5% through exposure to low-intensity laser light (690 nm). The
intensity of the scattered light is measured as a function of angle and may be
analyzed to yield the molar mass, root mean square radius, and second viral
coefficient (A2). The results of an SLS experiment can be used as a quality
control in protein preparation (e.g., for structural studies) in addition to the
determination of solution oligomeric state (monomer/dimer, etc.). SLS experi-
ments may be performed in either batch or chromatography modes. However,
as the measurement yields the volume-averaged molecular weight of the
 Chapter 19 Protein Characterization Methods 301

sample within the laser beam, it is more powerful to utilize the technique in
combination with protein purification. As the measurements are performed in
a flow cell, there is no loss of sample and the SLS detector can be integrated
easily into standard protein purification equipment. Due to the necessity of
obtaining good baselines in both 280 nm absorption measurements (UV)
and light scattering (LS) measurements, SEC represents a good choice of
separation media, due to the use of only a single buffer system for the entire
purification. Since the light scattering and concentration are measured for
each eluting fraction, the mass and size can be determined independently of
the elution position. This is particularly important for protein species with
nonglobular shapes, which may elute at positions distant from that predicted
by the calibration curve for the column.

2. Materials
2.1. Mass Spectrometry
2.1.1. Preparation and Automated Mass Spectrometry
of Intact Protein Samples
1. Protein solution(s) to be analyzed.
2. 20 µM myoglobin solution (Sigma, Gillingham, UK). Dissolve protein in
water and store at −80°C in 15-µl aliquots.
3. Ultimate HPLC with autosampler (Dionex, Camberley, UK) coupled
to an electrospray ionization Q-Tof Micro mass spectrometer (Waters,
Manchester, UK) (see Note 1).
4. C4 PepMap 300 µ-precolumn cartridge (Dionex) (see Note 2).
5. Wash solvent: 97.3% water, 2% acetonitrile, 0.5% formic acid, 0.2% trif-
luoroacetic acid.
6. Solvent A: 95% water, 5% acetonitrile with 0.1% formic acid.
7. Solvent B: 20% water, 80% acetonitrile with 0.1% formic acid.
8. Skirted PCR plate (Abgene, Epsom, UK) and Pierceable Power Seals
(Greiner Bio-One, Stonehouse, UK).
2.1.2. Preparation of N-Glycosylated Samples
1. Peptide-N-glycosidase F (PNGase F, Sigma). Reconstitute the PNGase
F according to the manufacturer’s instructions, make 5-µl aliquots, and store
at −80°C.
2.1.3. Preparation of Samples from Crystals (11)
1. Extra fine long paper wicks (Hampton Research).
2. Acetonitrile.
3. Resuspension buffer: 20 mM Tris-HCl pH 7.5, 200 mM NaCl, 8 M urea.
4. Skirted PCR plate (Anachem, Luton, UK) and Pierceable Power Seals
(Greiner Bio-One).
2.1.4. Preparation and Automated Mass Spectrometry of Peptide Samples
1. Sequencing grade trypsin (Promega, Southampton, UK). Make 5-µl aliq-
uots and store at −80°C to cut down on degradation of the enzyme due to
freeze–thaw cycles.
2. Ultimate HPLC with autosampler (Dionex) coupled to an electrospray ioni-
zation Q-Tof Micro mass spectrometer (Waters) (see Note 1).
3. Ubiquitin (6 µM in water).
302 Nettleship et al.

4. Glufibrinopeptide B (3 µM in water).
5. Jupiter™ 4µ Proteo 90 Å column (Phenomenex, Macclesfield, UK).
6. Wash solvent: 97.3% water, 2% acetonitrile, 0.5% formic acid, 0.2%
trifluoroacetic acid.
7. Solvent A: 95% water, 5% acetonitrile with 0.1% formic acid.
8. Solvent B: 20% water, 80% acetonitrile with 0.1% formic acid.
9. Skirted PCR plate (Anachem) and Pierceable Power Seals (Greiner Bio-
One).
2.1.5. Preparation of Samples for N-Glycosylation Site Analysis
and Interpretation of Results
1. Sequencing grade trypsin (Promega). Make 5-µl aliquots and store at −80°C.
to cut down on degradation of the enzyme due to freeze–thaw cycles.
2. Peptide-N-glycosidase F (PNGase F). Reconstitute the PNGase F according
to the manufacturer’s instructions, make 5-µl aliquots and store at −80°C.
3. Zwitterion Chromatography-Hydrophilic Interaction Chromatography
(ZIC-HILIC) resin (HiChrom, Berkshire, UK).
4. Isopropanol.
5. Rainin Finepoint P20 filtered tips (Anachem, catalogue number: RT-20F).
6. Wash solvent: 20% water, 80% acetonitrile with 0.1% formic acid.
7. Resuspension solvent: 100% acetonitrile with 0.1% formic acid.
8. Elution solvent: 100% water with 0.1% formic acid.
9. 1 M Tris-HCl pH 7.5, 0.5 M NaCl.
10. Skirted PCR plate (Anachem), Pierceable Power Seal (Greiner Bio-One)
and Adhesive PCR seal (ABgene).
11. ProteoMass™ Guanidination Kit (Sigma) (optional depending on quality of
results).

2.2. ThermoFluor Assay

1. Solution of purified protein at 1 mg/ml (see Note 11).
2. Sypro Orange dye (Molecular Probes); 5,000× stock in DMSO diluted to
10× working stock in water and stored at 4°C (see Note 12).
3. 96-well thin-walled PCR plate (e.g., Thermo-Fast 96 skirted, Abgene) (see
Note 13).
4. Optically clear plate seals (e.g., Microseal “B” Film, Bio-Rad).
5. Plate sealer (not essential).
6. Buffer/additive/ligand screen. For example, buffers at 20–100 mM, salts
at 25–500 mM, ligands at suitable concentrations compared to protein and
other additives at concentrations comparable to those used in the Hampton
Research Additive Screen (see Note 14).

2.3. Static Light Scattering

1. Protein solution(s) to be analyzed (see Notes 22 and 23).
2. Äkta Purifier (GE Heathcare) applied with a size exclusion chromatography
column (see Notes 24 and 25).
3 Static light scattering detector (Wyatt Technology miniDAWN Tristar) (see
Note 26).
4. BSA solution (10 mg/ml in water).
 Chapter 19 Protein Characterization Methods 303

3. Methods

3.1. Mass Spectrometry

3.1.1. Preparation and Automated Mass Spectrometry
of Intact Protein Samples
1. Prepare 15 µl of a 20-µM solution for every protein to be analyzed by mass
spectrometry in buffer solution containing only ionic salts, for example,
20 mM Tris-HCl pH 7.5, 200 mM NaCl. Buffer components such as deter-
gents, polyethylene glycol (PEG) or high imidazole are to be avoided for best
results, since components of this nature are favorably ionized within the mass
spectrometer and therefore mask the protein signal (see Fig. 19.1).
2. Pipette 15 µl 20 µM myoglobin into the first well of the PCR plate. This is
used for the calibration of the instrument (see Note 3).
3. Aliquot 15 µl water into the second well to wash the precolumn, fol-
lowed by the sample(s) of interest, which are pipetted into the proceeding
wells. Preparation protocols for some specialized samples are given in the
following (see Sections 3.1.2. and 3.1.3.).
4. The PCR plate is sealed before automated running of the samples, which
can take place overnight.

(A) 956
900
100
900 997
866 1067
834 1119
820 1147
791 1207
%

0 m/z
750 1000 1250 1500 1750 2000 2250
(B) 613.42
100
635.48

657.44

714.51
736.52
%

802.57

824.58
921.63

0 m/z
750 1000 1250 1500 1750 2000 2250

Fig. 19.1 (continued)

304 Nettleship et al.

(C) 625.44
100
625.48
669.51

713.51
713.58

%
757.54

801.60

802.59
0 m/z
750 1000 1250 1500 1750 2000 2250
(D) 792
100
791 836 880
669 902

946
1002
%

0 m/z
750 1000 1250 1500 1750 2000 2250

Fig. 19.1 Examples of raw data of good quality (A); with 10% PEG3350 contamination
(B); with 1% Triton X100 (detergent) contamination (C); and with 0.5 M imidazole
contamination (D).

5. Each sample is automatically loaded onto the C4 precolumn through the

autosampler and washed with wash solvent at 5 µl/min for 5 minutes.
6. The sample is then eluted at 5 µl/min in the reverse direction directly into
the mass spectrometer using a gradient of 5% to 80% solvent B over 1
minute. The concentration of mobile phase is held at 80% Solvent B for
10 minutes before re-equilibration of the column in 95% Solvent A, 5%
Solvent B.
7. Analysis of samples is by MS (as opposed to MS/MS). Data is processed
by combining the data under the peak on the chromatogram (Fig. 19.2A) to
give a raw data file showing % ions against mass/charge (see Fig. 19.2B).
This raw data is then deconvoluted to give a single mass peak using the
MaxEnt algorithm (see Fig. 19.2C). The MaxEnt algorithm is part of the
MassLynx software used to control the mass spectrometer.
3.1.2. Preparation of N-Glycosylated Samples
1. Boil 15 µl of a 20-µM protein sample in buffer solution containing only
ionic salts for 10 minutes to denature the protein. This can be done either in
a closed microcentrifuge tube or a sealed PCR plate.
2. After the solution has cooled to room temperature, add 1 µl of PNGase F
and incubate at 37°C for over 3 hours (see Note 4).
3. Perform LC-MS and data analysis as described in Section 3.1.1.
 Chapter 19 Protein Characterization Methods 305

(A) 11.94
100

0 Time
0.00 5.00 10.00 15.00 20.00
(B)
100 977 10071038
923 1072
898 1107
874 1146
852
1186
831
%

1230
810 1278
791 1384
1510 1581
773
1660 1845
739
0 m/z
750 1000 1250 1500 1750 2000 2250
(C) 33190
100
%

33215

0 mass
31000 32000 33000 34000 35000 36000

Fig. 19.2 Example data from a standard run showing the peak on the chromatogram
(A); the raw combined data (B); and the deconvoluted mass peak (C).

3.1.3. Preparation of Samples from Crystals (11)

1. Open the well containing the crystal(s) of interest and place the crystallization
plate under a microscope. This technique is limited by the concentration of
protein in the crystal(s) therefore larger crystal(s) will give better data.
2. Wick away the mother liquor from around the crystal(s) using an extra fine
long paper wick.
3. Carefully wash the crystal(s) in 15 µl of reservoir solution from the crystal-
lization experiment.
4. Remove all solution from around the crystal(s) using a wick.
306 Nettleship et al.

5. Carefully wash the crystal(s) twice with 15 µl acetonitrile, wicking away

any solution in between washes. The washes remove PEG and other
contaminants from the crystal(s). If any contaminants remain, this will
result in less satisfactory data.
6. Dissolve the crystal(s) in 15-µl resuspension buffer and transfer to a skirted
PCR plate and seal ready for mass spectrometry.
7. Analyze the sample using LC-MS as described in Section 3.1.1. (see Note 5).
3.1.4. Preparation and Automated Mass Spectrometry of Peptide Samples
1. To 15 µl of a 10-µM protein solution, add 1 µl trypsin solution and incubate
at room temperature for 30 minutes (see Note 4). For glycosylation site
analysis, follow the sample preparation protocol given in Section 3.1.5.
2. Aliquot calibrants into the first wells of a PCR plate. For this technique, ubiq-
uitin (6 µM) and glufibrinopeptide B (3 µM) are used as they can be loaded into
a well of the PCR plate and automatically injected onto the column as part of
the run. However, accurate calibration can also be performed by direct injection
using 10 µl of 0.05 µg/µl CsI in 50:50 isopropanol:water (see Note 3).
3. Pipette the sample(s) of interest into the proceeding wells and seal the plate.
4. Each sample is automatically loaded onto the Jupiter™ Proteo column
through the autosampler and washed with wash solvent at 8 µl/min for
5 minutes.
5. The sample is then eluted at 10 µl/min into the mass spectrometer using a
gradient of 95% solvent A:5% solvent B to 55% solvent A:45% solvent B
over 40 minutes. The concentration of Solvent B is then increased to 100%
over 1 minutes and held at 100% for 10 minutes before re-equilibration of
the column into 95% Solvent A:5% Solvent B.
6. Peptides can be detected using either MS or MS/MS modes as required (see
Note 6).
7. Analyze MS Data by combining the data under a peak of interest. MS/MS
data are analyzed using the ProteinLynx Global Server™ program, which
automatically processes the data and searches the online databases for pro-
teins containing the peptides in the sample of interest.

3.1.5. Preparation of Samples for N-Glycosylation Site Analysis

and Interpretation of Results
1. Pipette five samples of 20 µl of a 10-µM protein solution into five wells of
the skirted PCR plate.
2. Incubate the different samples as described in step 4 and in the summary
given in Fig. 19.3A. Timing of incubations can be combined as in Fig. 19.3A
to allow parallelization of the experiments in the PCR plate.
3. The following types of peptide are generated by these different samples:
• Sample 1: Nonglycosylated peptides
• Samples 2 and 3: Both glycosylated and nonglycosylated peptides
• Sample 4: Negative control for sample 5
• Sample 5: Glycosylated peptides only
4. Seal the plate using an adhesive PCR seal during incubations (see Notes 4
and 7).
• For sample 1: Add 1 µl of trypsin and incubate for 30 minutes at room
temperature.
 Chapter 19 Protein Characterization Methods 307

(A)
Sample 1

Sample 2 PNGase F
20µM Analysis
Sample 3 protein Trypsin PNGase F by LC-
solution ESI-MS
Sample 4
ZIC-HILIC
purification
Sample 5 PNGase F

Increasing time

Fig. 19.3 Scheme showing the workflow for glycosylation site analysis (A); and
construction of the ZIC-HILIC micro-column (B).

• For sample 2: Add 1 µl of PNGase F and incubate at 37°C for 3 hours. After
the reaction is complete, add 1 µl of trypsin and incubate for 30 minutes at
room temperature.
• For sample 3: Add 1 µl of trypsin and incubate for 30 minutes at room
temperature. After the reaction is complete, add 1 µl of PNGase F and
incubate at 37°C for 3 hours.
• For sample 4: Add 1 µl of trypsin and incubate for 30 minutes at room tem-
perature. Then perform the ZIC-HILIC purification (see the following).
• For sample 5: Add 1 µl of trypsin and incubate for 30 minutes at room
temperature, followed by ZIC-HILIC purification (see the following).
Afterward, add 4 µl of 1 M Tris-HCl pH 7.5, 1 M NaCl to increase the
pH before adding 1 µl of PNGase F. Incubate the sample for 3 hours at
37°C.
ZIC-HILIC purification:
• Resuspend the ZIC-HILIC resin in isopropanol to make a 50% (w/v) solution.
• Cut the end off of a filtered P20 tip about 5 mm after the filter and add 20 µl
of resin solution above the filter to form a micro-column (see Fig. 19.3B and
Note 8).
• Equilibrate the micro-column twice in 100 µl of wash solvent, using a P200
Pipetman to force the liquid through.
• Add 80 µl of resuspension solvent to the sample, which will give a final
acetonitrile concentration of 80% and load the sample through the tip.
• Wash the micro-column three times in 100 µl wash solvent.
• Ensure all solvent is removed from the end of the tip before elution of the
sample (see Note 9).
• Elute the sample in 20 µl of elution solvent.
308 Nettleship et al.

5. Remove the adhesive PCR seal and replace it with a Pierceable Power Seal
prior to analysis by LC-MS, which is run as in Section 3.1.4 using the MS
mode of operation.
6. For analysis, perform an in silico digest of your protein using the Peptide
Cutter tool (http://www.expasy.org/tools/peptidecutter/). Map the poten-
tial glycosylation sites onto the cleaved sequence. Glycosylation site
occupancy prediction is performed using NetNGlyc (http://www.cbs.dtu.
dk/services/NetNGlyc/) and NetOGlyc (http://www.cbs.dtu.dk/services/
NetOGlyc/) for N- and O-glycosylation, respectively.
7. Using the calculated monoisotopic masses of the peptides containing
potential glycosylation sites search for these in the MS data.
8. If the correct mass is found, this indicates the site is unoccupied. If the
mass is found to be 1Da larger than calculated, the site is occupied. This is
due to the PNGase F converting the asparagine to an aspartic acid during
the deglycosylation reaction.
9. As indicated, the following types of peptide should be detected in the samples
Sample 1: Nonglycosylated peptides.
Samples 2 and 3: Both glycosylated and nonglycosylated peptides.
Sample 4: Negative control for sample 5 (see Note 10).
Sample 5: Glycosylated peptides only (see Note 10).
10. A consensus of the results from the five samples allows each potential
site to be assigned as “glycosylated,” “not glycosylated,” or “partially
glycosylated.” The “partially glycosylated” category is for sites that appear
as both glycosylated and not glycosylated in the different experiments
(Fig. 19.4). It is these sites that may be mutated out to allow for secretion
of a homogeneous sample.
11. If results are unsatisfactory, the signal of peptides ending in lysine can be
enhanced using the ProteoMass™ Guanidination Kit (12). This converts the
lysine to homoarginine thus allowing the peptide to be more easily ionized.
Guanidination:
12. Add 10 µl of the Base Reagent into each well containing the digest sample
and mix well.
13. Add 10 µl of Guanidination Reagent to each well and mix. Incubate the
reactions at 65°C for 30 minutes.
14. Add 4 µl of 100% formic acid per well to stop the reaction instead of the
30–60 µl Stop Solution mentioned in Step 7 of the Sigma protocol. This
keeps the sample volume to a minimum.
15. Run the samples by LC-MS as described previously (Section 3.1.4.).

3.2. ThermoFluor Assay

1. The thermal shift assay can be performed in any commercially avail-
able real-time PCR machine. Although certain elements of the protocol
described here refer specifically to Opticon Monitor software version 3.1.32
(running a BioRad DNA Engine Opticon 2 real-time PCR machine), the
general principles are applicable to other systems.
2. The software must be programmed with the necessary details for running the
protocol, namely, which dye filters (see Note 12) and thermal parameters to
 Chapter 19 Protein Characterization Methods 309

(A) N-Glycosylation Site

N-Glycosylation Site N-Glycosylation Site
N-Glycosylation Site N-Glycosylation Site N-Glycosylation Site

OPPF3492
255 aa

(B) 691.88
100

692.36 [M+2H]2+
%

692.90

693.90

686.39 693.87 699.82

0 m/z
686 688 690 692 694 696 698 700
(C) 692.30
100
[M+2H]2+
692.81
%

693.32
689.30
689.82
690.30 693.74
687.85 695.69 697.77
0 m/z
686 688 690 692 694 696 698 700

Fig. 19.4 The assigned glycosylation sites mapped onto the amino acid sequence
with tryptic cleavage sites indicated by a perpendicular line, glycosylated sites by
a black triangle and partially glycosylated sites by a white triangle (A). Example
of ESI-LC-MS data for the peptide containing the third potential glycosylation site
(LSNLDPGNYSFR) showing the peak with no shift (B) and with a +1 Da shift (C).

use. For the Opticon Monitor software, this involves specifying a plate template
and running method. In the “Plate Setup” section, a new template should be
programmed, choosing the appropriate plate type (clear/white) and selecting
all wells to be read as samples using the SYBR Green filter (SBG1: see Note
12). In the “Protocol Setup” section, a new running method should be entered
as a melting curve, with heating from 20°C to 95°C and a 15-second hold
every 0.5°C, followed by a fluorescence reading. The assay volume should
be entered here to allow the software to calculate the actual temperature in the
reaction mixture. Also in this window, the authors set the temperature control
at “Sample Calculation” with “Lid Settings” set at “Constant 101°C (shut off
<20°C).” Save the templates with appropriate filenames.
310 Nettleship et al.

3. Prepare the assay plate using a multichannel and/or repeater pipette to

reduce pipetting errors. Total reaction volume is 50 µl (see Note 15). To each
well, add 5 µl protein solution, 5 µl 10× Sypro Orange solution and 40 µl of
whichever screen is under investigation. The order of addition is unimportant.
Up to 96 additives or buffers can be tested on each plate; replicates can be
used for more reliable evaluation of fewer conditions.
4. Seal the plate carefully (see Note 16) and centrifuge for 1 minute at 500g to
mix components.
5. Place the plate into the PCR machine, close the lid and set the program
running. For the Opticon Monitor software, select the appropriate plate and
method, click “Run” and give a filename when prompted. After about 1
hour the assay is complete.
6. Assay progress can usually be followed in real-time. In the Opticon Monitor
software, this is done via the “Status” window where the “Optical Read Status”
tab allows the user to highlight wells and follow the melting curve(s).
7. When the assay is complete, the melting curves can be analyzed and Tms
calculated. Some software packages perform these calculations automati-
cally while others require that the data be exported into statistical analysis
software where curve fitting can be performed to extract Tms. The “Melting
Curve” window of the Opticon Monitor software calculates the Tm from
the maximum value of the first derivative curve of the melting curve (Fig.
19.5). In the “Graphs” tab, selecting a well displays the data for that well,
with radio buttons allowing display of the intensity curve, the first deriva-
tive curve or both. The “Calculations” tab at the bottom of the window
displays the Tms. An ideal melting curve would be sigmoidal but this is not
always the case (see Notes 17–19) and adjustment of the “Peak Location
Boundaries” in the “Graphs” tab might be necessary to define the correct
region for the calculation. Care must be taken with the interpretation and
it is advisable to visually inspect the melting curves before accepting the
calculated Tms. A “smooth” function which removes noise from the curves
can also affect calculated Tms and should be used with care (see Note 20).
8. Upon completion of the protocol, all melting curves can be compared
and the results correlated with the various screens (see Note 21). Running
multiple samples allows comparison of data from well-to-well; while large
thermal shifts are obvious, such direct comparison can aid in the detection
of smaller shifts. If necessary, raw data can usually be exported for curve
fitting and further analysis.
9. Ideally, follow-up experiments should be performed using titrations of
ligand or additive and assaying in replicate.

3.3. Static Light Scattering

3.3.1. Experimental Set-up and Calibration
1. For SLS measurements in the chromatography mode the miniDAWN Tristar
(SLS detector) should be placed in the flow path of the Äkta Purifier. The
instrument is typically equipped with a Superdex 200 10/300 GL SEC col-
umn. The SLS detector uses the UV signal from the Äkta Purifier to measure
protein concentrations and care needs to be taken to ensure that a suitable
delay volume is programmed into the ASTRA software (Wyatt Technology,
Santa Barbara, CA) to correct for the difference in flow path between the UV
 Chapter 19 Protein Characterization Methods 311

Fig. 19.5 Example thermal shift data showing the raw data curve alongside the first
derivative curve. The temperature at the peak of the first derivative curve is the melting
temperature, Tm (dotted line).

and SLS measurements. This can be calculated from the difference in elution
volume between the UV and SLS signals. A simple way to ensure the correct
determination of the delay volume is available through the analysis software
(see “Alignment” in the “View” menu). An overlaid display is given of both
the SLS and UV signals and a right mouse click and drag between the two
respective peaks allows the user to manually overlay the two signals. High
precision is achievable by zooming in on the peaks (Control-left mouse to
drag a zoom area). Once this delay volume is defined it should only need to
be redefined if tubing length between the UV and SLS detectors is altered.
2. A 0.22-µm prefilter should be placed immediately upstream of the SLS
detector in order to remove large particles (e.g., produced by the pumps),
which will disturb the measurements (see Note 27).
3. The system is equilibrated with 2–3 column volumes of an appropriate
buffer (see Note 28).
4. Once the system has been correctly set up and equilibrated a standard protein
is used for calibration purposes. Bovine serum albumin (BSA) represents a
good choice of calibration sample, as it forms a number of known oligomeric
states in solution (monomer, dimer, and tetramer of 66-kDa subunits).
The calibration is performed as follows:
1. Inject 100 µl BSA solution and start data collection (see Section 3.3.2.).
2. Adjust the value of the “AUX1 calibration constant” in the “system set-up”
window of the “Collect” menu during data analysis until the correct mass
for BSA is obtained (see Section 3.3.3.). This value needs to be determined
periodically to correct for changes in the intensity of the UV light.
However, any protein of “known” molecular weight and specific absorp-
tion coefficient may be used and after initial installation and calibration it
is recommended that the user confirms the calibration using, e.g., lysozyme
(molecular weight of 14 kDa).
312 Nettleship et al.

3.3.2. Data Collection

1. Start a new experiment by selecting “New” from the “File” menu of the
ASTRA software. A window appears with two screens monitoring the UV
and LS signals.
2. Set-up the experimental parameters in the “Collect” menu. Enter in the
“System set-up” window “solvent” (typically water), “flowrate” and “AUX1
calibration constant.” Correct setting of the flow rate is essential as the
device uses timers to define data collection windows over the required vol-
ume ranges. Hence incorrect flow rates will result in miscollection of the
data. The new “AUX1 calibration constant” is calculated by correcting the
value determined during the BSA calibration (see Section 3.3.1.) for the dif-
ference in the absorption coefficients of the sample protein and BSA:

0.1% 0.1%
(AUX1)new = (AUX1)BSA calibration X A 280, BSA X A 280, sample

Where (AUX1)BSA calibration is the AUX1 calibration constant determined

during BSA calibration; A0.1% 280, sample
is the absorption coefficient at 280 nm
for a 1 mg/ml sample solution (see Note 29); A0.1% 280, BSA
is the absorption coeffi-
cient at 280 nm for a 1 mg/ml BSA solution (0.68 cm−1).
Enter in the “Collection set-up” window the following parameters:
“operator,” “sample ID,” “injection-to-collect-delay volume,” “collection
duration,” and “collection interval” (see Note 30). Save the experiment.
3. Fill the injection loop with 100 µl of the sample protein solution (see Note 22).
4. Start the chromatography method on the Äkta Purifier (see Note 31).
5. The data collection on the SLS detector has to be started manually: choose
“single injection” in the “Collect” menu. Press “OK” to start data collection
at the same time the sample is injected on the SEC column (indicated by the
shift of the injection valve).
3.3.3. Data Analysis
The data obtained from the SLS detector is analyzed using the ASTRA
software.
1. Choose “Baselines” in the “View” menu. Draw the baselines in the UV and
the three LS signals by clicking on the traces in the linear parts to either
side of the peak(s). It is worthwhile optimizing the baselines since they are
crucial for the outcome of the data analysis. An option is available to define
a baseline only in the second LS window, through “Auto baseline” in the
“Options” menu, but the user is strongly recommended to visually check
the first and third windows manually.
2. Select “Peaks” in the “View” menu. To check if the peaks in the UV and
LS signals overlay choose two traces (AUX1 and one LS) by clicking on
the “Data” button in the window. After a second click on “Data” both
traces appear in the window. This can be repeated for the other LS signals.
When the peaks overlay well select the area of the peak(s) to be analyzed
by left mouse clicking and dragging over the area to be analyzed. The
selected area is temporarily visualized by a gray bar and can be adjusted if
needed.
3. Select “Report” in the “View” menu and choose “Summary.” This window
summarizes the results of the data analysis. A graphical display of the
 Chapter 19 Protein Characterization Methods 313

Fig. 19.6 Analysis of the solution oligomeric state of a 7.0-kDa protein by static light
scattering in combination with size exclusion chromatography. The separation of the
oligomers was performed on a HiLoad 16/60 Superdex 75 column (GE Healthcare).
The light scattering signal (dots) is shown as the mass distribution in a slice in each of
the two peaks in the elution profile monitored by the absorbance at 280 nm (solid line).
The molecular masses were calculated to be 7.3 and 13.7 kDa, respectively, which
correlates well with the monomeric and dimeric protein forms.

masses present in each data slice is available through “MM vs. Volume”
in the “Distribution” menu (Fig. 19.6). In order to measure only regions
of the chromatogram in which a single species is eluting it may be neces-
sary to adjust the analysis region (see Step 2) such that a horizontal line is
obtained. The presence of nonhorizontal sections within the analysis area
indicates the presence of two species, whose relative concentrations are not
constant—leading to a change in the mass averaged molecular weights of
these sections.

4. Notes

1. All protocols can be adapted for use with LC-MS systems from other manu-
facturers.
2. C4 precolumns can be reused, however some samples “stick” to the column
and re-elute in all subsequent samples. On average each precolumn lasts for
around 2 months.
3. Calibration can be performed either before running the samples or retro-
spectively and then applied to relevant data during analysis.
4. As PNGase F reactions and tryptic digests need to go to completion and pro-
tein denaturation is not an issue, longer incubation times—up to overnight—
can be used. This will ensure a complete reaction without jeopardizing the
sample.
314 Nettleship et al.

5. When running samples prepared from crystals, it is essential to either use

a new precolumn on the LC-MS or to make sure the precolumn is clean as
the concentration of protein is likely to be below ideal.
6. MS analysis gives the accurate mass of a peptide, whereas MS/MS analysis
gives both the accurate mass and fragmentation data. The fragmentation
data can be used to give information on the sequence of the peptide. MS/
MS data for a digested protein provide a mass fingerprint, which can be
used to identify a protein by BLASTing the fingerprint against databases
such as SwissProt.
7. This protocol can be used with digestion enzymes other than trypsin,
which may lead to more informative results depending on the positioning
of the cleavage sites in relation to putative glycosylation sites in the pro-
tein of interest. Other enzymes used routinely are chymotrypsin, glutamyl
endopeptidase, thermolysin, and lysyl endopeptidase, although any diges-
tion enzyme is appropriate for this method.
8. The ZIC-HILIC resin will migrate into the filter of tips from some manufactur-
ers; however, Rainin Finepoint tips have proved successful for this method.
9. Removal of wash solvent or sample from the end of the tip can be performed
using a P20 pipetman.
10. The ZIC-HILIC resin binds glycosylated peptides; however, some non-
glycosylated peptides do bind such as a peptide containing a His6 tag.
Purification of the glycosylated peptides away from non-glycosylated ones
increases the chance of detection
11. The authors’ standard protocol uses protein at 1 mg/ml, with each well
containing 5 µg protein. This amount was selected to ensure that most pro-
teins give an adequate signal when first assayed. As little as 1.5 µg have
been used per well with reproducible results.
12. The dye generally used is Sypro Orange, which fluoresces strongly when
bound in the hydrophobic regions exposed in unfolded proteins. Its fluo-
rescent properties are similar to those of SYBR Green, commonly used in
real-time PCR applications, and hence the SYBR Green filters on real-time
PCR machines can be used for excitation and detection.
13. Various types of plate are available depending on which PCR machine
is being used (white or clear plates, skirted or non-skirted, and so on).
Although the authors prefer opaque white plates, they have used clear
plates with success.
14. The authors use a 96-conditions in-house screen (Table 19.1) for standard
thermal shift assays which is stored as a master block at −20°C. As well as
testing protein stability in several buffer types at different pHs and varying
salt concentrations, the screen also assesses an assortment of metal salts,
nucleotide cofactors, sugars, reducing agents, and other chemicals. When
required, the block is defrosted and 40 µl are taken from each well for assay-
ing. Each block contains enough material for five thermal shift assays.
15. The basic experimental conditions are widely applicable to different
proteins. Assay volume and protein concentration can be lower (as low
as 15 µl with as little as 1.5 µg protein/well), but this may not give an
adequate signal for all proteins.
16. Care should be taken not to touch the seals with ungloved hands. Wear
gloves to avoid smearing across the surface, which may affect optical
properties.
Table 19.1 The 96 conditions of the in-house screen, which is routinely used for the thermal shift assay
1 2 3 4 5 6 7 8 9 10 11 12
A 50 mM 50 mM Sodium 50 mM Citric 50 mM Citric 50 mM Sodium 50 mM Sodium 50 mM 50 mM HEPES 50 mM 50 mM 50 mM 50 mM CAPSO
Sodium acetate pH acid pH 5.0 acid pH 5.4 cacodylate cacodylate HEPES pH 7.4 Tris-HCl Tris-HCl CAPSO pH 9.4
acetate 4.4 pH 6.0 pH 6.4 pH 7.0 pH 8.0 pH 8.4 pH 9.0
pH 4.0
B 50 mM 50mM Sodium 50mM Citric 50 mM Citric 50 mM Sodium 50 mM Sodium 50 mM 50 mM HEPES 50 mM 50 mM 50 mM 50 mM CAPSO
Sodium acetate acid pH5.0 acid pH 5.4 cacodylate cacodylate HEPES pH 7.4 Tris-HCl Tris-HCl CAPSO pH 9.4 100 mM
acetate pH4.4 100mM 100 mM pH 6.0 pH 6.4 pH 7.0 100 mM pH 8.0 pH 8.4 pH 9.0 NaCl
pH4.0 100mM NaCl NaCl 100 mM 100 mM 100 mM NaCl 100 mM 100 mM 100 mM
100mM NaCl NaCl NaCl NaCl NaCl NaCl NaCl
NaCl
C 50 mM 50 mM Sodium 50 mM Citric 50 mM Citric 50 mM Sodium 50 mM Sodium 50 mM 50 mM HEPES 50 mM 50 mM 50 mM 50 mM CAPSO
Sodium acetate acid pH5.0 acid pH5.4 cacodylate cacodylate HEPES pH7.4 Tris-HCl Tris-HCl CAPSO pH 9.4 200 mM
acetate pH4.4 200 mM 200 mM pH 6.0 pH 6.4 pH 7.0 200 mM pH 8.0 pH8.4 pH 9.0 NaCl
pH4.0 200 mM NaCl NaCl 200 mM 200 mM 200 mM NaCl 200 mM 200mM 200 mM
200 mM NaCl NaCl NaCl NaCl NaCl NaCl NaCl
NaCl
D 50mM 50mM Sodium 50mM Citric 50 mM Citric 50 mM Sodium 50 mM Sodium 50 mM 50 mM HEPES 50 mM 50 mM 50 mM 50 mM CAPSO
Sodium acetate acid pH5.0 acid pH 5.4 cacodylate cacodylate HEPES pH 7.4 Tris-HCl Tris-HCl CAPSO pH 9.4 500 mM
acetate pH4.4 500 mM 500 mM pH 6.0 pH 6.4 pH7.0 500 mM pH 8.0 pH 8.4 pH 9.0 NaCl
pH4.0 500 mM NaCl NaCl 500 mM 500 mM 500 mM NaCl 500 mM 50 mM 500 mM
500 mM NaCl NaCl NaCl NaCl NaCl NaCl NaCl
NaCl
E 2 mM ATP 2 mM ADP 2mM AMP 2 mM 2 mM AMPcPP 2 mM 2 mM 2 mM GTP 2 mM GDP 2 mM GTP- 2 mM 2 mM FAD
AMPCCP AMPPcP UTP γS TMP
F 2 mM β- 2 mM β−γ 2mM dCMP 2 mM dGMP ssDNA 7mer ssDNA 9mer 1% glyc- 5% glycerol 10% 20% glycerol 1 mM DTT5 mM DTT
NAD MethylGTP erol glycerol
G 10 mM 10 mM MgCl2 10mM MnCl2 10 mM ZnCl2 10 mM FeCl3 100 mM KCl 100 mM 200 mM 10 mM 10 mM 3% DMSO10 mM NiCl2
CaCl2 LiCl NaThyioc- L-Proline Phenol
yanate
H 100 mM 10 mM 10mM Urea 5% PEG 400 3% D(+)- 3% D- 10 mM L- 10 mM L- 10 mM 10 mM 0.5% n- 0.5% n-Dodecyl
Glycine Spermidine Glucose Galactose Alanine Methionine L-Serine L-Arginine Octyl-β−D- b-D-maltoside
glucoside
The screen has been composed by Erika Mancini and Christian Siebold (Oxford, UK).
316 Nettleship et al.

17. Not all proteins give good melting curves with a defined transition between
folded and unfolded. This does not necessarily reflect instability, since the
authors have seen poor melting curves from otherwise well-behaved and
crystallizable proteins.
18. Following the complete unfolding of the protein (where fluorescence
peaks), it is common to see a drop off in fluorescence, possibly due to
exclusion of the dye during aggregation of denatured protein. Also, some
proteins display a slight initial drop in fluorescence before the true transi-
tion begins, perhaps due to dye binding weakly in hydrophobic surface
pockets and dissociating at low temperatures.
19. Sometimes, other interactions can be observed in the thermal shift profile,
e.g., a plateau partway along the melting curve indicating melting of oli-
gomers or complexes before the transition corresponding to unfolding of
the individual proteins.
20. Thermal shifts are frequently significant enough to see clearly and there is
often no need for accurate Tm determination. Following visual inspection,
the authors generally take Tm values as calculated by the Opticon Monitor
software. If necessary, data can be exported to other software for poten-
tially more accurate Tm determination via curve fitting.
21. We have observed thermal shifts in excess of 12°C, but it is possible to see
much smaller shifts of less than 1°C, especially when melting curves are
compared side by side. For very low shifts, although they might turn out
to be significant, experimental error must be considered as a factor.
If necessary, this could be investigated by further testing in replicate.
22. The nature of the protein affects the upper and lower amounts that can
be accurately analyzed. Proteins with a relatively high absorption coef-
ficient ( > 0.7 cm−1) will produce significantly more deflection in the
UV measurements than those with low absorption coefficients ( <
0.5 cm−1) for identical sample loading. It is possible to overload the UV
signal used by the SLS detector, so care must be taken that sample con-
centrations do not become excessive. Conversely, insufficient sample con-
centration in the elution fractions will result in weak LS and UV signals.
Such weak signals then result in large errors in the analysis. As a general
rule of thumb, 100 µg to 1 mg of protein sample applied to a Superdex 200
10/300 GL column (GE Healthcare, Piscataway, NJ) result in reasonable
LS and UV signals. When a preparative SEC column is used [e.g., the
HiLoad 16/60 Superdex 200, GE Healthcare] 5 to 10 mg of sample should
be applied.
23. The results of the SLS experiment are volume averaged. Hence, the best
results are obtained when highly pure samples are analyzed by SEC since
it is relatively easy to separate different oligomeric species of one protein
from one another. However, when impure samples are applied increased
errors appear due to insufficient separation of different protein species.
24. Other protein purification equipment can be used provided the UV signal
of the instrument can be used by the SLS detector.
25. Both analytical and preparative SEC columns can be used for the analysis
(see Note 22).
26. This protocol is based on the authors’ experience with the Wyatt
Technology mini-DAWN Tristar detector connected in-line with an Äkta
Purifier. Other equipment combinations are also possible in order to
 Chapter 19 Protein Characterization Methods 317

increase the accuracy of the method (e.g., the installation of a refractive

index concentration measurement device). Such additional equipment will
improve the accuracy of the technique, with an attendant increase in the
cost of the installation. There are currently at least two companies which
provide SLS detectors: Wyatt Technology (http://www.wyatt.com/) and
Viscotek GmbH (http://www.viscotek.com). The reader is strongly recom-
mended to browse the company Web pages for further information.
27. The addition of a prefilter immediately upstream of the SLS detector is a
sensible step to prevent particles entering the device and affecting meas-
urements. A standard 0.22-µm paper filter supported by a metal frit is suf-
ficient for this purpose, although the backpressure of the system needs to
be monitored. Increases in the backpressure may indicate that the filter has
become clogged and needs replacing. Care also needs to be taken that the
column itself suffers no damage due to increases in backpressure at stand-
ard operating flow rates. In the authors’ experience it is permissible to add
the measured backpressure of the filter and SLS detector to the maximum
recommended backpressure rating of the column(s) used. This value can
be determined by selecting a column bypass and noting current pump pres-
sure (typically 0.3 MPa). Thus the Superdex 200 10/300 GL column (rated
for a maximum backpressure of 1.5 MPa) can now be run at 1.8 MPa.
28. The SLS measurement is highly sensitive to baseline errors in both the
UV and LS signals and as a result a thorough equilibration is needed for
precise SLS measurements. Typically 2–3 column volumes of buffer rep-
resents a suitable equilibration volume for size exclusion columns. The
components of the buffer should also not present too strong a background
in either absorption of the UV signal or the LS signal. High concentrations
of reducing agents (e.g., dithiothreitol) or glycerol should also be avoided
if possible. If essential for the experiment these buffers should be prepared
immediately prior to the experiment. As is true for all size exclusion
experiments the buffer should be well filtered and degassed prior to use.
29. Inaccurate absorption coefficients represent the major source of errors in
the data analysis. While absorption coefficients generated from the linear
sequence are, in the main, sufficiently accurate it is recommended that
UV absorption spectra are recorded of the sample in both the size exclu-
sion buffer and a denaturing buffer in order to experimentally establish
the extinction coefficients of the folded proteins. The presence of cofac-
tors (e.g., nucleotides), which have significant absorption at 280 nm,
also results in inaccurate concentration estimates from the UV signal.
Additionally, samples that show significant absorption at the wavelength
used for the SLS measurements (690 nm) result in errors in the LS meas-
urements.
30. The maximal number of data points that can be collected per experiment is
14,400. Hence the minimal collection interval that can be chosen depends
on the collection duration. For an experiment performed with a Superdex
200 10/300 GL column a typical collection duration is 30 minutes, which
allows a collection interval of 0.125 seconds.
31. To allow simultaneous sample injection and start of the data collection it
is advisable to program a short column equilibration step (e.g., 0.1 column
volumes for the Superdex 200 10/300 GL column) in the chromatography
method.
318 Nettleship et al.

Acknowledgments

The work described here was supported by the MRC and the European
Commission as SPINE, contract-no. QLG2-CT-2002-00988.

References
1. Geerlof, A., Brown, J., Coutard, B., Egloff, M.-P., Enguita, F. J., Fogg, M. J.,
Gilbert, R. J. C., Groves, M. R., Haouz, A., Nettleship, J. E., Owens, R. J., Ruff, M.,
Sainsbury, S., Svergun, D. I., and Wilmanns, M. (2006) The impact of protein
characterization in structural proteomics. Acta Crystallogr. D Biol. Crystallogr. 62,
1125–1136.
2. Yokoyama, S., Hirota, H., Kigawa, T., Yabuki, T., Shirouzu, M., Terada, T., Ito, Y.,
Matsuo, Y., Kuroda, Y., Nishimura, Y., Kyogoku, Y., Miki, K., Masui, R., and Kuramitsu,
S. (2000) Structural genomics projects in Japan. Nat. Struct. Biol. 7, 943–945.
3. Jeon, W. B., Aceti, D. J., Bingman, C. A., Vojtik, F. C., Olson, A. C., Ellefson, J. M.,
McCombs, J. E., Sreenath, H. K., Blommel, P. G., Seder, K. D., Burns, B. T., Geetha,
H. V., Harms, A. C., Sabat, G., Sussman, M. R., Fox, B. G., and Phillips, G. N., Jr.
(2005) High-throughput purification and quality assurance of Arabidopsis thaliana
proteins for eukaryotic structural genomics. J. Struct. Funct. Genomics 6, 143–147.
4. Pantazatos, D., Kim, J. S., Klock, H. E., Stevens, R. C., Wilson, I. A., Lesley, S. A.,
and Woods, V. L., Jr. (2004) Rapid refinement of crystallographic protein construct
definition employing enhanced hydrogen/deuterium exchange MS. Proc. Natl.
Acad. Sci. USA 101, 751–756.
5. Jawhari, A., Boussert, S., Lamour, V., Atkinson, R. A., Kieffer, B., Poch, O.,
Potier, N., van Dorsselaer, A., Moras, D., and Poterszman, A. (2004) Domain
architecture of the p62 subunit from the human transcription/repair factor TFIIH
deduced by limited proteolysis and mass spectrometry analysis. Biochemistry 43,
14420–14430.
6. Potier, N., Billas, I. M., Steinmetz, A., Schaeffer, C., van Dorsselaer, A., Moras, D.,
and Renaud, J. P. (2003) Using nondenaturing mass spectrometry to detect fortui-
tous ligands in orphan nuclear receptors. Protein Sci. 12, 725–733.
7. Pantoliano, M. W., Petrella, E. C., Kwasnoski, J. D., Lobanov, V. S., Myslik, J.,
Graf, E., Carver, T., Asel, E., Springer, B. A., Lane, P., and Salemme, F. R. (2001)
High-density miniaturized thermal shift assays as a general strategy for drug dis-
covery. J. Biomol. Screen. 6, 429–440.
8. Lo, M. C., Aulabaugh, A., Jin, G., Cowling, R., Bard, J., Malamas, M., and
Ellestad, G. (2004) Evaluation of fluorescence-based thermal shift assays for hit
identification in drug discovery. Anal. Biochem. 332, 153–159.
9. Carver, T. E., Bordeau, B., Cummings, M. D., Petrella, E. C., Pucci, M. J.,
Zawadzke, L. E., Dougherty, B. A., Tredup, J. A., Bryson, J. W., Yanchunas, J., Jr.,
Doyle, M. L., Witmer, M. R., Nelen, M. I., DesJarlais, R. L., Jaeger, E. P., Devine,
H., Asel, E. D., Springer, B. A., Bone, R., Salemme, F. R., and Todd, M. J. (2005)
Decrypting the biochemical function of an essential gene from Streptococcus pneu-
moniae using ThermoFluor technology. J. Biol. Chem. 280, 11704–11712.
10. Matulis, D., Kranz, J. K., Salemme, F. R., and Todd, M. J. (2005) Thermodynamic
stability of carbonic anhydrase: measurements of binding affinity and stoichiometry
using ThermoFluor. Biochemistry 44, 5258–5266.
11. Nettleship, J. E., Walter, T. S., Aplin, R., Stammers, D. K., and Owens, R. J. (2005)
Sample preparation and mass-spectrometric characterization of crystal-derived
protein samples. Acta Crystallogr. D Biol. Crystallogr. 61, 643–645.
12. Brancia, F. L., Oliver, S. G., and Gaskell, S. J. (2000) Improved matrix-assisted
laser desorption/ionization mass spectrometric analysis of tryptic hydrolysates of
proteins following guanidination of lysine-containing peptides. Rapid Commun.
Mass Spectrom. 14, 2070–2073.

View publication stats