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

Подождите немного. Документ загружается.

2 Liquid Chromatography Techniques190

The first step in protein pre-fractionation can still be regarded as sam-

ple preparation for the purpose of eliminating irrelevant proteins

such as albumin and immuno-globulins and other high-abundancy

proteins in plasma or serum where affinity chromatography is the

method of choice. The flow-through fraction is then subject to intelli-

gent multidimensional protein pre-fractionation and sample de-com-

plexification, while the remaining proteins on the column will be

stored for other investigations or simply discarded during column

equilibration and cleaning.

And that is going to require development of new technologies as

well as the refinement of already existing tools, like e.g. affinity chro-

matography. It can be efficiently used to simplify protein samples by

enriching subsets of the proteome such as phosphorylated proteins

by IMAC, glycoproteins on lectin columns, just to mention the most

frequently used applications. In this case the retained proteins are of

main interest and the flow-through fraction is less important but

might be stored for continuative investigations.

Samples that do not contain any major contaminating proteins can

be processed directly. For reasons of simplicity and clarity the follow-

ing explanations are covering two-dimensional separations only,

which in practice will be relevant in most cases.

& Note: Too many steps can result in an

unacceptable loss of proteins. In this case less

can be more. As much separation as necessary,

but as little as possible. Sticking to the “keep it

simple” rule can be a good advice and a

consideration for success.

See also Figure 6 on page 9.

Even with the most sophisticated strategy, and the best analysis

techniques, it will – for the foreseeable future – not be possible to iso-

late all proteins in a complex proteomics sample.

In order to identify and characterize proteins present at very low

concentrations, it is inevitably to start with large sample amounts

and/or volumes in order to assure their presence in sufficient concen-

tration. One of the most exiting samples is probably blood. Dealing

just with complex samples is already difficult enough; however, if the

composition of a sample is similar to plasma, the challenges are

increasing exponentially. About 90% of the amount of plasma pro-

teins is albumin and IgG and just 25 high-abundant proteins repre-

sent 99% of the total amount. The remaining thousands (or mil-

lions?) of proteins are distributed over a huge concentration range

(see Figure 2.16).

Post-translational modifications

are perhaps the largest frontier

for proteomics.

Sample complexity and

dynamic range

2.3 Liquid Chromatography Techniques and Applications in Proteome Analysis 191

& Plasma is the largest and deepest version of the

human proteome: Largest, i.e. most proteins

(estimates range from >30,000 to >10,000,000).

Deepest, i.e. widest dynamic range (>10

). See

paper by Anderson and Anderson (2002).

Fig. 2.16: The analogy with an iceberg illustrates the challenges in

proteomics in terms of sensitivity and dynamic range. In order to

detect a biomarker present at a concentration of 1 fmol/L, theoretically

a sample volume of 1L is required, if the LOD of the MS equipment

is 1fmol.

Currently there is research ongoing for developing strategies on how

to tackle this demanding dynamic range and sensitivity challenge.

Key to this task is to get rid of as many of the low-interest high-

abundant proteins as possible (tip of the iceberg), without losing too

many of the free low-abundant proteins and those bound to carrier

proteins like e.g. albumin. High-abundant protein affinity depletion

followed by a 2DLC (IEX, RPC) strategy on protein level as described

further down seems to be the most promising approach. As much as

possible a sample load combined with high-recovery/low-loss separa-

tion steps and a high-sensitivity mass spectrometer are essential pre-

requisites for the successful deciphering of at least parts of the

human plasma proteome.

A simple calculation may help to highlight the current limits and

limitations. Let us assume that a state-of-the-art mass spectrometer

can routinely operate at the low femtomole level. If you start with

1 milliliter of sample you theoretically can come to the 10

–12

mmol/L

Anderson NL, Anderson NG.

Mol Cell Proteomics 1 (2002)

845–867.

2 Liquid Chromatography Techniques192

range in the iceberg picture, providing a 100% recovery, which is very

unlikely, since complex separations cannot be performed quantita-

tively. Although this is a question of simple mathematics, it is not

always appreciated. A possible solution might be to apply more sam-

ple, if possible or available, or improve techniques, methods, and

skills in order to attack the atto- or even zeptomolar range. Such

results are published these days and achievable by the top research

labs of the world.

Theoretically, applying 1 L of serum or a zeptomolar sensitivity

mass spec, or better both, would provide the desired performance of a

12 to 15 orders of magnitude dynamic range.

LC-MS manufactures claim that the combination of nanoLC and

nanoESI at peptide level covers a dynamic range of 10

(see Figure

2.17). A typical peak volume at this scale is ideally in the range of

100 nL. Closing the gap with the desired dynamic range is only possi-

ble if more sample can be acquired at the very beginning and concen-

trated in an intelligently designed workflow with high recovery. The

more we (can) load, the more we see.

Fig. 2.17: Expanding the overall dynamic range by increasing

the initial sample load.

The HUPO organization currently promotes a series of different

initiatives to work on the proteomes of plasma (discussed above),

brain, liver, and more are expected to follow soon. For all these com-

plex samples in principle – with specific and dedicated sample pre-

paration – very similar strategies can be applied. A schematic general

workflow is shown in Figure 2.18.

2.3 Liquid Chromatography Techniques and Applications in Proteome Analysis 193

Fig. 2.18: Schematic flow diagram showing the separation

steps on protein and peptide level.

Peak capacities and separation performance in protein

pre-fractionation

& Today, even with the most sophisticated tools and

intelligent separation strategies it is impossible to

achieve high purity for each protein in a

proteomics sample.

Human plasma consists of at least some 100,000 proteins, post trans-

lational modifications, splice variants and isoforms not yet included.

A pre-fractionation into 1,000 fractions still results in more than 100

proteins per fraction in average, but each single fraction probably

simple enough to be identified and characterized after digestion by

peptide 2DLC-MS. Literature reports of capacities for just 1D RPC of

200 to 500 peaks; however, in practice under normal lab conditions

2 Liquid Chromatography Techniques194

one can be happy if 10% of the theoretical values will be achieved.

Table 2.9 shows expected peak capacities for real samples in practice.

Tab. 2.9: Expected peak capacities on intact protein level under

practical laboratory conditions.

Affinity chromatography >2

Ion exchange chromatography ~20

Gel filtration ~10

Reversed phase chromatography ~50

2.4

Practical Considerations and Application of LC-based Protein

Pre-fractionation

In contrast to the well established electrophoresis techniques, the

application of liquid chromatography in proteomics is a – surpris-

ingly – rather young discipline. Fortunately, especially for the widely

used separation of tryptic peptides, both for one- and two-dimen-

sional separations, most researchers worldwide rely more or less on

the same recipe of running conditions with a very few minor – indivi-

dual – modifications only.

In protein pre-fractionation – a very promising approach still in its

infancy – we have not yet achieved any kind of standardization or

standard operation procedures (SOP). Based on the well founded and

long lasting history with chromatography in general and protein pur-

ification in particular, it was not too difficult to come up with a range

of suitable approaches and technical solutions as described by Apffel

(2003) and Dixon et al. (2006).

Although the purpose of protein purification is somewhat different

compared to protein pre-fractionation (PPF), as shown in Table 2.10,

by far most of the well established rules and basic considerations are

applicable for this new strategy in proteomics, too. Although the

same equipment, columns and expertise is used, the objective of each

approach is different.

Apffel A. In Simpson RJ, Ed.

Purifying proteins for proteo-

mics: A laboratory manual.

Cold Spring Harbor Laboratory

Press, Cold Spring Harbor,

New York (2003) pp. 75–100.

Dixon SP, Pitfield ID, Perrett

D. Biomed Chromatogr 207

(2006) 508–529.

2.4 Practical Considerations and Application of LC-based Protein Pre-fractionation 195

Tab. 2.10: The differences between protein purification and pre-

fractionation.

Protein purification „ protein pre-fractionation

The aim of protein purification is to

obtain a single fraction with a single

pure protein for continuative studies.

The protein pre-fractionation procedure

in proteomics results in very many

fractions with each still containing tens

to hundreds of proteins

During the purification process the

number of target protein containing

fractions decreases.

During the pre-fractionation process the

number of fractions increases exponen-

tially.

The fractions containing contaminant

proteins are discarded.

All fractions are of the same high value.

Knowing that within the next future a lot of new and probably even

more successful approaches will enter the scene, this section will

only cover one protein pre-fractionation workflow as a typical exam-

ple. This universal two-dimensional workflow combines ion-exchange

with reversed phase chromatography (see Figure 2.19). With some

individual modifications only, it can be used for a variety of samples.

Fig. 2.19: Schematics of the protein part of the LC-based

proteome analysis workflow.

2 Liquid Chromatography Techniques196

2.4.1

Sample Extraction and Preparation

In principle exactly the same methods for sample preparation as

described in Section 1.5.1 can be applied. As liquid chromatography

deals with UV detection the purity requirements for the reagents can

be significantly higher compared to electrophoresis.

& Note: Very hydrophobic proteins, e.g. membrane

proteins, are not solubilized with this type of

buffers and are thus not accessible with the

methods described below.

2.4.1.1 Plasma, Serum, CSF, and other Body Fluids

A special sample preparation step for the removal of the highest-

abundant proteins is required in most of the cases. Such a depletion

step can be performed by applying immuno-affinity techniques.

Despite its unique selectivity immuno-affinity chromatography does

not meet the requirements for loading capacity or economy. Just a

few microliters of plasma can be loaded and the running costs are

high. In addition, caused by unavoidable leaking effects, small

amounts of antibodies can be released into the elution buffer that

might result in ambiguous protein identification further downstream

in the workflow.

Recently, a hypothesis is under discussion that claims that the low-

molecular-weight region of the blood proteome, which is a mixture of

small intact proteins plus fragments of the large proteins, represents

all classes of proteins. As stated by Liotta et al. (2003) and Tirumalai

et al. (2003) this fraction is expected to be the ideal source for biomar-

ker discovery.

Such an approach, described by Tanaka et al. (2006), suggests to

cut-off the high-molecular-weight and concentrate the low-molecular-

weight proteins by dedicated membrane or hollow-fiber devices and

to continue the analysis with the low-molecular-weight fraction only.

Another remarkable paper by Wagner et al. (2002) describes the

application of protein mapping of biological samples of human

hemofiltrate for the analysis of proteins and peptides with a molecu-

lar weight below 20 kDa.

Once the sample has been extracted and prepared in a suitable way,

a two-dimensional protein pre-fractionation – ion-exchange chroma-

tography in combination with reversed phase chromatography – will

follow.

Liotta LA, Ferrari M, Petricoin

E. Clinical proteomics: written

in blood. Nature 425 (2003)

905.

Tirumalai RS, Chan KC, Prieto

DA, Issaq HJ, Conrads TP,

Veenstra TD. Mol Cell Proteo-

mics 2 (2003) 1096–1103.

Tanaka Y, Akiyama H, Kuroda1

T, Jung G, Tanahashi K, Sugaya

H, Utsumi J, Kawasaki H,

Hirano H. Proteomics 6

(2006) 4845–4855.

Wagner K, Miliotis T, Marko-

Varga G, Bischoff R, Unger KK.

Anal Chem 74 (2002) 809–

820.

2.4 Practical Considerations and Application of LC-based Protein Pre-fractionation 197

& Before starting with the most valuable and

difficult sample it is strongly recommended to

learn and establish the new method in the

laboratory by using a well known simple standard

sample first. This period can take or save several

weeks and months but is a basic prerequisite for

the application of each new technique.

2.4.2

Experimental Setup

2.4.2.1 Scale of Operation

As mentioned earlier in Section 2.2 it is advantageous to start with a

higher initial amount of protein in order to extend the overall

dynamic range, enable access to lower-abundant proteins and last but

not least to compensate for the inevitable losses in each of the indivi-

dual steps in the entire workflow until the sample finally enters the

mass spectrometer.

Protein pre-fractionation as described here permits the separation

of initial sample loads between » 100 mg and » 100 mg of protein.

Protein pre-fractionation at an even higher, industrial scale has been

reported.

2.4.2.2 Type of Instrumentation

In principle any kind of biocompatible HPLC system equipped with a

fraction collector can be used to perform protein pre-fractionation.

However, for the efficient separation of proteins and peptides it is

very often necessary to use buffers with high salt concentrations at

very acidic or basic pH that are extremely corrosive to stainless steel

components which are commonly used in conventional HPLC equip-

ment. Besides contaminating the sample with corrosion products

that are released from the metal surfaces it may also result in a mod-

ifications of the proteins and finally in increased maintenance and

replacement costs for the instrumentation.

& It is recommended to use a dedicated

biocompatible, high-performance bio molecule

separation system. Such systems have a

completely bio-inert flow path and can withstand

the even most aggressive buffers, solvents and

high salt concentrations (like 8 mol/L urea).

Rose K, et al. Proteomics 2004,

4, 2125–2150.

Conventional HPLC systems

are intended for small molecule

analytical rather than bio mole-

cule preparative use.

2 Liquid Chromatography Techniques198

For convenience and ease-of-use the fraction collector should be com-

patible with all different kind of vessels including microtiter plates.

KTA and Ettan LC have proven to be suitable for protein pre-

fractionation and can be recommended without any restriction or

limitation.

2.4.2.3 Choice of Chromatography Column

The number of HPLC columns on the market is huge and unma-

nageable. It is recommended to use columns that have proven to per-

form well for protein separations. Except for RPC it is essential that

the columns are made of biocompatible material such as glass,

PEEK



or other suitable plastic material.

& Packing material that is based on silica is not

recommended as it is unstable at pH above 7.5. In

order to avoid cross-contamination from one

sample to another the column must survive a

cleaning-in-place (CIP) process, i.e. the column

must tolerate wash steps with 1 mol/L NaOH or

HCl. A column packed with porous polymer based

material is recommended.

Non-porous (NPS) columns show superior resolution but offer only

low loading capacity. Besides a good separation performance protein

recovery is a key feature.

2.4.2.4 Monitoring Each Separation Step, Monitoring Each Dimension

In general, but especially during the method development process it

is beneficial to monitor progress and performance of each protein

pre-fractionation step by analyzing each, or representative fractions

by SDS electrophoresis, and – if appropriate or necessary – optimize

the procedure, as outlined in the poster of K. Marcus et al. (2005).

2.4.3

Ion Exchange Chromatography and Protein Pre-fractionation

An ideal technique to deal with samples of crude protein extracts –

small to large amounts or volumes, native, denaturing or reducing

conditions – is ion exchange chromatography. This technique can be

operated at any scale, from micro to macro scale. Due to its adsorptive

mode diluted protein solutions can be concentrated and separated

with high resolution in less than an hour. Its ease of use and its logi-

Good batch-to-batch reproduci-

bility and long-term availability

is best with big and well estab-

lished manufactures.

K. Marcus, et al. “A novel

approach towards differential

proteomics with multidimen-

sional intact protein prefractio-

nation of brain samples” Poster

at the HUPO 4th Annual

World Congress Munich

(2005).

2.4 Practical Considerations and Application of LC-based Protein Pre-fractionation 199

cal, directly adaptable theory made IEX the most widely used techni-

que for protein separations.

In order to benefit from the concentration effect all proteins should

bind to the column. In theory this can be achieved by loading the

sample at very low pH (e.g. glycine/HCl buffer) to a cation, or at very

high pH (e.g. piperidine/HCl buffer) to an anion exchange column.

However, these extreme pH conditions are not compatible with pro-

tein stability. Physiological pH or slightly basic conditions (Tris/HCl

buffer) are more favorable conditions for high protein recovery.

In cases, where AIEX does not give satisfying results, or in the tan-

dem IEX approach, see below in Figure 2.20, the application of CIEX

should be considered.

Recently, another technique based on charge, or more precisely, on

isoelectric point (IEP, or pI) had its renaissance in the area of protein

pre-fractionation: chromatofocusing (CF). CF can be regarded as the

LC analogue to isoelectric focusing in electrophoresis, however, with-

out achieving its separation performance. Although the theory is very

convincing, in practice this technique is suffering from a very

restricted pH range, protein precipitation at the isoelectric point,

high running costs and the necessity to remove the pH gradient

creating ampholytic reagents (Polybuffer) after the separation. It is

quite unlikely that CF will have a major breakthrough in the near

future.

Fig. 2.20: Potential operating area for charge- and pI-based protein

separation (theoretical spot map modified from Link et al. 1997).

Experience from several

decades indicates that anion

exchange chromatography is

the first choice for the separa-

tion of intact proteins.

In general anion-exchange-

chromatography is the first

choice for the separation of

intact proteins.

Link AJ, Robison K, Church

GM. Electrophoresis 18 (1997)

1259–1313.