Westermeier R., Naven T., H?pker H.-R. Proteomics in Practice: A Guide to Successful Experimental Design

Подождите немного. Документ загружается.

2 Liquid Chromatography Techniques210

Tab. 2.11: Examples of loading capacities for

different columns diameters.

Column i.d. Practical loading capacity

6.4 mm 200–400 mg

4.6 mm 100–200 mg

2.1 mm 20–40 mg

1.0 mm 4–8 mg

Because – for protein RPC – the loading capacity under high-reso-

lution conditions is not a function of the column volume, it is neces-

sary to increase the column diameter, if more loading capacity is

required.

2.4.5

Fraction Size and Number of Fractions

Ideally a biological sample is pre-fractionated in such a way that each

fraction contains only one protein and that each protein is present in

one fraction only. Although practically not at all achievable, this

would lead to as many fractions as proteins, far too much to manage

for a complex biological sample with tens to hundreds of thousands

of proteins.

However, it is feasible to aim for the presence of any given protein

in not more than two fractions with a negligible overlap to the pre-

vious and subsequent fractions. An ideal chromatographic peak, as

shown in Figure 2.29, has a certain width at the base and at half

height that can be expressed in time or volume and can be calculated

and printed by state-of-the-art LC evaluation software packages.

For IEX this calculation results in some ten fractions: fraction 1 to

collect the flow-through of the unbound proteins, fraction 2 to 9 dur-

ing gradient elution and fraction 10 while washing down the remain-

ing proteins with high salt concentration.

In RPC, with its sharper peaks, some fifteen to twenty fractions

will be sufficient.

By applying the recommendations from the introduction in Section

2.2 – ideally 30 to 300 proteins per sample (=fraction) – a sample with

some 100,000 proteins would require approximately 300 to 400 frac-

tions, which is well in-line with the performance of the LC techniques

and the equipment used in protein pre-fractionation.

Naturally, less complex samples like yeast or E. coli with 4,000 to

6,000 proteins require a lower degree of protein pre-fractionation.

GraceVydac, The Handbook of

Analysis and Purification of

Peptides and Proteins by

Reversed-Phase HPLC, 3rd edn

(2002).

Decreasing the fraction size

further, results in too many

fractions, and an enormous

successive workload without

any tangible benefit.

2.5 Critical Review and Outlook 211

Fig. 2.29: In order to estimate the optimal fraction size for a protein

pre-fractionation experiment a fraction size of twice to three times the

peak width at half height can be recommended. If the apex of a peak is

in the middle of a fraction, this fraction contains almost the entire peak

volume. If a fraction change occurs at the peak maximum the peak is

split in equal parts in not more than two fractions.

2.5

Critical Review and Outlook

An internal comparison run with an E. coli sample (unpublished

data, work still in progress) has disclosed that identification rates and

sequence coverage raises significantly when using protein pre-fractio-

nation rather than applying shotgun approaches at peptide level only.

A preliminary interpolation claims that a complete analysis of all frac-

tions would give in the order of 2,000 identified proteins, a figure

matching the entire E. coli proteome.

Furthermore, protein pre-fractionation has additional benefits:

Subdivides complex protein mixtures into man-

ageable fractions;

Offers a wide dynamic range ;

Low m g to high mg initial sample load;

Very small to very large sample volumes;

Maintains the correlation between intact parent

proteins and their tryptic peptides;

2 Liquid Chromatography Techniques212

Facilitates convenient up-scaling of individual

fractions if more material is required, e.g. for

biomarker discovery;

Generates more and more confident protein

identification.

A sample that has undergone protein pre-fractionation, generally

does not require – after tryptic digestion – another multi-dimensional

separation at peptide level. Typically, a high-resolution one-dimen-

sional nanoRPC separation, followed by MS

analysis, will be the

method of choice. Figure 2.30 illustrates such a workflow.

Fig. 2.30: Complete LC-based workflow after the sample has

undergone a substantial de-complexification at protein level.

In the middle of the nineties of the last century there was hype for

high-throughput approaches in proteome analysis. Now, some ten

years later the success stories are still quite rare. Obviously, the chal-

lenges and requirements for meaningful proteomics have been

underestimated. It seems as if there are no quick achievements in

sight. However, time was not wasted as scientists have characterized

2.5 Critical Review and Outlook 213

and cataloged thousands of proteins in data bases, strong and undis-

putable foundations for the future work in proteomics.

The extra workload associated with running multiple replicates of

protein and peptide separations – electrophoresis or chromatography

– has not always been appreciated. However, recently several papers

can be found, which are requesting and proposing pre-fractionation

strategies and techniques in order to cope with the enormous com-

plexity of human samples for the deciphering of significant biomar-

kers; for instance the publications by Righetti et al., Guerrier et al.,

Wang and Hanash (all from 2005).

The strategies and application examples as outlined in the chapter

above, if performed properly, are somewhat more time consuming

compared to traditional concepts. It has proven that “quick and dirty”

approaches can only give access to the “low hanging fruit”. From a

rational and logical standpoint it appears more likely that more intel-

ligent sample preparation methods as e.g. protein pre-fractionation

will fractionate and de-complexify samples to such an extent that con-

secutive separation- and detection-devices will be enabled to analyze

the sample to the desired level. (This will – of course – require a cer-

tain amount of time and work effort.)

Righetti PG, Castagna A, Anto-

nioli P, Boschetti E. Electro-

phoresis 26 (2005) 297–319.

Guerrier L, Lomas L, Boschetti

E. J Chromatogr. 1073 (2005)

25–33.

Wang H, Hanash S. Mass Spec

Reviews 24 (2005) 413– 426.

215

Mass Spectrometry

This section discusses the use of mass spectrometry (MS) for proteo-

mics applications; protein identification, characterization and quanti-

fication. The section addresses key parameters for protein identifica-

tion and the systems available to perform the listed applications.

As explained in previous chapters, large format two-dimensional

gel electrophoresis enables the resolution of several thousand pro-

teins in a reproducible fashion in a relatively short period of time.

The instrumental developments in two-dimensional gel electrophor-

esis offered the momentum for proteomics. Similarly, technology

advancement for protein identification was essential if proteomics

was going to expand into the dominating field we know today. Edman

sequencing was the principal method of protein identification and

though it was used with considerable success for routine protein

identification, the method was relatively slow (one or two peptides

per day) and relatively insensitive (upper fmol to low pmol amounts;

Pappin et al. 1995). Developments in the field of mass spectrometry

addressed speed and sensitivity issues, which enabled biomolecular

analysis, particularly the analysis of peptides and proteins to be per-

formed routinely with confidence. Together with the development of

database search engines and the population of sequence databases,

researchers have employed MS with great success in proteomics. The

evolution of MALDI and ESI ionization sources has impacted sensi-

tivity significantly and mass analyzer development continues to

extend the boundaries of mass accuracy, resolution and functionality.

Mass spectrometry is an analytical technique that measures the

mass-to-charge ratio (m/z) of ions based upon their motion in an elec-

tric or magnetic field. Sample molecules are converted into ions in

the gas phase and separated according to their m/z ratio; positively

and negatively charged ions can be formed. This technique is per-

formed by a mass spectrometer, which typically consists of three com-

ponents (see Figure 3.1) and these components dictate the level of

performance and the type of analysis a mass spectrometer can per-

form.

Proteomics in Practice. A Guide to Successful Experimental Design 2

Ed.

Reiner Westermeier, Tom Naven, and Hans-Rudolf Hçpker

ISBN: 978-3-527-31941-1

Pappin DJC, Rahman D,

Hansen HF, Bartlet-Jones M,

Jeffery W, Bleasby AJ. Mass

Spectrometry in the biological

sciences. (1996) 135–150

3 Mass Spectrometry216

Fig. 3.1: A schematic indicating the three regions of a mass

spectrometer: Ion source: the component which ionizes the

sample in question, i.e. ESI and MALDI. Mass analyzer: the component

that separate the ions according to their m/z ratio. Detector: the

final component which records the signal produced by an ion.

Ions produced in the source are separated in the analyzer according

to their mass-to-charge ratio (m/z) and the analyzer, to a large degree,

determines the MS performance and functionality. Resolution, and

mass accuracy are standard MS instrument specifications and tan-

dem mass spectrometry is an essential functionality. All three are key

for successful protein identification (ID) and characterization.

For instance, the higher the mass accuracy a mass spectrometer

can afford, then the greater the tolerance limits can be set for data-

base searches and ultimately the greater the confidence in the subse-

quent results. Resultantly, for protein identification and characteriza-

tion it is essential to assign the mass of a given peptide as accurately

as possible and as such a differentiation between the average and

monoisotopic mass needs to be made.

Average and Monoisotopic Mass Assignment Amino acids consist of

the elements carbon, hydrogen, nitrogen and oxygen, and to a lesser

extent sulfur. Each of these elements exists naturally as a mixture of

isotopes. For instance, carbon exists as a mixture of the

C isotope

(98.9%) and the

C isotope (1.1%). As such the abundance of the two

isotopes in a given compound will be reflected in a mass spectrum by

the isotope envelope of the compound. There are two types of mass

measurement for a given compound:

Average mass: the measurement which reflects

the contribution of the isotopes in an isotopic

envelope. The average mass is taken at the cen-

troid of the isotopic envelope.

Monoisotopic mass: is the sum of the masses of

the atoms in a molecule using the principle iso-

tope mass of each atom instead of the isotope

averaged atomic mass. In practice, this is the

217

mass of the first peak in a peptide isotopic envel-

ope. With respect to carbon, this is termed the

C peak.

Figure 3.2 illustrates the predicted isotope distribution of the peptide

HLKTEAEMK at resolution 5000 (FWHM).

Fig. 3.2: Predicted isotope distribution of the peptide

HLKTEAEMK at resolution 5000 (FWHM).

The spectrum demonstrates the resolution of the

C and

C iso-

topes. As a result the monoisotopic mass,

C peak, can be clearly

assigned.

The mass difference between the average and monoisotopic masses

is 0.73 Da.

Resolution and mass accuracy specifications are determined in a

large part by the analyzer technology. Improvements in analyzer per-

formance have enabled the mono-istopic mass of a peptide, rather

than the average mass of a peptide, to be almost exclusively used for

protein identification, characterization and quantification. Mass mea-

surement accuracies of less than 2 ppm have been reported in the

routine analyzes of complex peptide mixtures and intact proteins. It

is possible to increase this further with internal mass calibration, as

described by Matthias Mann and co-workers, with mass accuracies of

less than 1 ppm (r.m.s.) reported.

3 Mass Spectrometry

The second peak in the

envelope represents the peptide

structure where one of the

carbon atoms in the peptide is

now a

C atom.

This is a significant difference

which can be capitalized on for

protein identification by redu-

cing the error tolerances for

database searches.

Yates JR, Cociorva D, Liao J,

Zabrouskov V. Anal Chem 78

(2006) 493–500.

Macek B, Waanders, L, Olsen

JV, Mann M. Mol Cell Proteo-

mics 5 (2006) 949–958.

Olsen JV, de Godoy LMF, Li G,

Macek B., et al.; Mol Cell

Proteomics 4 (2005) 2010–

2021.

3 Mass Spectrometry218

3.1

Ionization

The ion source is the region of the mass spectrometer where the gas

phase ions are produced from sample molecules. The method of pro-

ducing the ions is termed the ionization technique. Several ionization

techniques have been developed; the earliest incarnations include

electron impact and chemical ionization, which are useful for ioniz-

ing small molecular weight molecules, but less applicable for larger

(bio)molecules. The first example of ionizing larger biomolecules was

reported by MacFarlane and Torgerson in 1976 using the technique

plasma desorption. The introduction of fast atom bombardment in

1981 (Barber et al. 1981) enabled the ionization and detection of a

range of intact biomolecules with relatively good sensitivity. The FAB

source was coupled with a magnetic sector analyzer, enabling biomo-

lecule tandem mass spectrometry analysis to be performed with rela-

tively good sensitivity, high resolution for the first time.

However, FAB was supplanted by two ionization techniques devel-

oped in the late 1980s. Karas and Hillenkamp introduced matrix

assisted laser desorption ionization (MALDI) in 1988 as a technique

that could readily ionize (large) biomolecules in a very sensitive man-

ner. MALDI is a pulsed ionization technique which utilizes the

energy from a laser to desorb and ionize the analyte molecules in the

presence of a light absorbing matrix. In another breakthrough, Fenn

and co-workers demonstrated that electrospray ionization could also

ionize large biomolecules with high sensitivity.

3.1.1

Matrix Assisted Laser Desorption Ionization

Matrix Assisted Laser Desorption Ionization (MALDI) ions are cre-

ated by mixing the analyte with a small, organic molecule which

absorbs light at the wavelength of the laser, the matrix. The analyte

becomes incorporated into the crystal lattice of the matrix and is then

irradiated with a laser. The laser causes desorption and ionization of

the matrix and analyte, either by protonation or cationation (positively

charged ions) or by deprotonation (negatively charged ions; see Fig-

ure 3.3). The ions are then accelerated into the MS analyzer (see Sec-

tion 3.2)

& MALDI produces predominantly singly charged

ions.

McFarlane R, Torgerson DF.

Science 191 (1976) 920–925.

Barber M, Bordoli RS, Sedg-

wick RD, Tyler AN. Nature 293

(1981) 270–271.

These two ionization techniques

have become the standard for

ionization of protein and

peptide samples.

Fenn JB, Mann M, Meng CK,

Wong SK, Whitehouse C.

Science 246 (1989) 64–71.

Tanaka K, Ido Y, Akita S,

Yoshida Y, Yoshida T. 35-kai

Shitsuryo Bunseki Rengo

Toronkai, Yoshishu (1987)

22–23.

Karas M Hillenkamp F. Anal

Chem 60 (1988) 2299–2301.

3.1 Ionization 219

Fig. 3.3: Schematic of the MALDI process.

For a tryptic digest, a-cyano-4-hydroxy cinnamic acid is typically the

standard matrix of choice (Beavis and Chait 1989). The matrix affords

high sensitivity for the detection of peptides (and proteins) and exhi-

bits negligible matrix adduction. A second matrix, 2,5-dihydoxyben-

zoic acid (DHB) initially described by Strupat et al. 1991; enables pep-

tide analysis, detection of high molecular weight proteins and the

analysis of oligosaccharides released from glycoproteins (Mock et al.

1991; Stahl et al. 1991; Harvey 1993).

Tab. 3.1: Common MALDI matrices used in biological applications.

Matrix Matrix Structure Application

a-Cyano-4-hydroxy-

cinnamic acid

COOH

UV laser: peptide analy-

sis and protein digests.

Analytes <10 kDa

Sinapinic acid

(4-hydroxy-3,5-di-

methoxycinnamic acid)

OMe

COOH

Analysis of large poly-

peptides and proteins

>10 kDa

2,5-Dihydroxybenzoic

acid (2,5 DHB)

COOH

UV laser: protein di-

gests and proteins, oli-

gosaccharides released

from glycoproteins

Beavis, RC, Chait BT. Rapid

Commun Mass Spectrom

3(1989) 432–435.

Strupat K, Karas M, Hillen-

kamp F. Int J Mass Spectrom

Ion Proc 111 (1991) 89–102.

Mock KK, Davy M, Cottrell JS.

Biochem Biophys Res Commun

177 (1991) 644–651.

Stahl B, Steup M, Karas M,

Hillenkamp F. Anal Chem 63

(1991) 1463–1466.

Harvey DJ. Rapid Commun

Mass Spectrom 7 (1993) 614–

619.

Beavis RC, Chaudhary T, Chait

BT. Org Mass Spectrom 27

(1992) 156–158

Beavis RC, Chait BT. Rapid

Commun Mass Spectrom 3

(1989) 432–435.

Strupat K, Karas M, Hillen-

kamp F. Int J Mass Spectrom

Ion Proc 111 (1991) 89–102.