**2.3 Restricted access material columns**

232 Biomarker

compared to off-line approaches and methodologies. Exposed surfaces are kept to a minimum that usually is the main cause of sample losses. Overall precision can also be controlled by having yields above 50 %. It is possible to handle yields that are lower; however, it generally is a real analytical challenge to obtain operational stability within such analytical processes. Direct injection of samples onto HPLC columns is substantially advantageous in the clinical laboratories in terms of its time- and labor-saving capabilities, in addition to other advantages given below. General direct injection methods have been devised which deal with the problem of many different proteins being present in the sample. The methods include the pre-column technique, restricted access materials, and chromatography in mobile phases containing surfactant. High performance affinity chromatography is also a direct injection technique will demonstrate its power in near feature (Figure 1 d). The characteristic and performance of each direct injection technique

The pre-column technique is the direct injection technique that is mostly reported. The precolumn technique utilizes two columns in series (pre-column and analytical) connected by a switching valve. The most common pre-column technique employs a reversed phase precolumn and a reversed-phase analytical column: the sample is injected into an aqueous mobile phase flowing through a pre-column (1-4 cm in length, 3 - 4.6 mm I.D.) which retains lipophilic compounds, passing non-retained hydrophilic compounds to waste. The switch in valve is then changed and components retained on the pre-column are eluted onto the analytical column by increasing the solvent strength of the mobile phase. This technique serves the dual function of concentration of analyte and removal of hydrophilic substances. There are many advantages of the pre-column injection technique in comparison to traditional sample preparation techniques: time saving in comparison to the labor-intensive liquid-liquid extraction and precipitation techniques, high reproducibility and high, also a superior detection limit capabilities due to its allowance for injection of large sample

Several types of column-switching designs have been applied. The back flush design is most often used because it reduces band broadening (Yamashita et al., 1992). However some prefer the forward-flush mode to protect the analytical column from possible impurities at the head of the column. Use of an on-line ~0.5µm filter is recommended, which needs periodic replacement. A design that incorporates a second pre-column parallel to the first has been employed which increases sample throughput, by alternating injection on one pre-

Most biofluids contain large amounts of well-known proteins such as albumin and IgGs, which overwhelm the separation system and make the detection of the low abundant proteins and peptides very difficult. It is thus advantageous to remove these proteins prior to digestion or direct separation. There are alternative ways of reducing the overall protein load by specific adsorption of albumin and IgG to affinity matrices (Nakamura et al., 2002, Wang et al., 2003, Govorukhina et al., 2003). While usually an affinity matrix is generally highly specific, in high content samples the affinity ligand is limited to exhibit its specificity. There are degrees of specificity between highly selective immunoaffinity matrices and less selective but more robust affinity supports using synthetic ligands. In an effort to reduce the amount of albumin from human serum, a number of affinity matrices has been evaluated based on antibodies or dye ligands. Antibody-mediated albumin removal was efficient and

column and back flushing of retained compounds on the other pre-column.

are comprehensively discussed below for the analysis of biological samples.

volumes.

The direct injection of biological samples onto the chromatographic column without any sample preparation is in most cases highly problematic and may lead to an irreversible contamination of the separation columns, which deteriorate selectivity and column performance. A powerful asset to circumvent all the named problems is the implementation of restricted access material for sample preparation. Special Solid Phase Extraction (SPE) supports possessing restricted access properties have been developed (Hagestam and Pinkerton, 1985, Yu et al., 1997, Boos and Rudolphi, 1997). In 1991, Desilets et al. (Desilets, et al., 1991) introduced the restricted access term. Silica based Restricted Access Materials have been developed for the clean-up in bioanalysis; first for low molecular weight compounds in biofluids (Rbeida, et al. 2005) and subsequently for biopolymers such as peptides (Wagner et al. 2001). Those supports were able to withstand several hundred plasma or serum injections (total volume of 5–7 ml) without losing performance. The concept and the methodology were successfully used for the sample clean-up of peptides and proteins out of biofluids by extending the range of available materials employing cation and anion-exchanger RAM (Machtejevas et al., 2004)). Specific non-silica based RAM were also developed for the investigation of the food (Bovanova & Brandsteterova 2000) and environmental matrices (Hogendoorn et al., 1999). Vijayalakshmi and co-workers (Pitiot et al., 2004) have presented a new RAM called a bi-dimensional chromatographic support operating on a size exclusion mode and an affinity or pseudoaffinty mode. A survey on the current state-of-art of RAMcolumns in sample pre-treatment is given Souverain et. al. (Souverain et al., 2004). A RAM support developed by Boos and Grimm (Boos & Grimm, 1999) is based on SCX-diol modification to improve performances in terms of efficiency, retention and reproducibility. Račaitytė et al. (Račaitytė et al. 2000) have shown that this type of RAMs is highly suitable for the on-line extraction and analysis of neuropeptides in plasma. Machtejevas et al. (Machtejevas et al. 2006) analyzed the pore structural parameters and size exclusion properties of LiChrospher strong cation-exchange and reverse phase restricted access materials. For peptide analysis out of the biofluids, the strong cation-exchange functionality seems to be particularly suitable mainly because of the high loadability of the strong cationexchange restricted access material (SCX-RAM) and the fact that one can work under nondenaturing conditions to perform effective chromatographic separations. The proper column operating conditions leads to the total effective working time of the RAM column to be equal to approximately 500 injections (depending on the type of sample).

The principle of the restricted access support is based upon the presence of two chemically different surface properties of porous silica particles (Figure 3). The outer surface of the particles (25 – 40 mm O.D.) is highly biocompatible: it possesses diol modification and hence is hydrophilic, while the pore surface chemistry is tailored as a hydrophobic dispersion phase with C-18 functionality or as a strong cation exchanger with SO3 functionality. An advantage of these adsorbents relies on the simultaneous occurrence of two chromatographic separation mechanisms: selective reversed-phase interaction or ion exchange chromatography of lower molecular mass analytes and size exclusion

Profiling of Endogenous Peptides by Multidimensional Liquid Chromatography 235

with buffer solution containing a high salt concentration, e.g. 0.1 M, at pH 5 to 7. The high salt concentration is needed to suppress electrostatic interactions between the solute and the charged surface. In the sample clean up of the RAM-SCX column the concentration of salt is much lower e.g smaller than 20 mM and the pH is kept at approximately 3. Under these conditions, electrostatic attraction forces are dominant between the positively charged peptides and proteins whereas the negatively charged species are excluded from the pores

**b**

**0 5 10 15 20 25 30 35 40 45 min**

Fig. 4. Typical SCX-RAM column separation profile: peaks (a) represent physical exclusion by pore size. Trapped retained bio-molecules are separated by a gradient in the second step (b). Conditions: column - LiChrospher 60 XDS (SO3/Diol), 25 x 4 mm I.D., flow rate - 0.5 ml/min, gradient from 0 to 1 M NaCl in 20 mM KH2PO4 pH 2.5, containing 5 % ACN in 30

After loading the RAM-SCX column the washing step elutes at isocratic conditions all the excluded compounds between the start and 15 minutes. After 15 minutes switching occurs and the trapped analytes are eluted from the RAM-SCX column with a strong eluent under gradient condition in the period between 15 and 45 minutes (Figure 4 b). Thus it is a charge and charge distribution selective process combined with SEC. Use of RAM-SCX allows one the direct application of biofluids onto the column. Small peptides are selectively trapped in the pores by cationic functional groups while large molecular weight biopolymers are directed to waste. This strategy performs the sample clean up and selective peptide

Mass loadability of SPE and RAM columns play a key role in executing the sample clean-up. It is advisable to work below the overload regime of the column. Otherwise, displacement effects and other phenomena such as secondary interaction by adsorbed species might take place, which will lead to non-reproducible results (Wilemsen et al., 2004). Last statement is particularly important when the task is to monitor medium to low abundant proteins, therefore, usually large sample volumes in the millilitre range are applied. As the column lifetime is known to be limited a control measure has to be applied to check the condition of the RAM-SCX column and, if necessary, replace it by a new one. In our experience the

Although compatibility of these stationary phases with direct biological sample injection is high, one still has to keep in mind that samples have to be filtrated or centrifuged prior to

min. Sample: 100 μl Human Hemofiltrate (3.7 mg/ml), UV detection at 214 nm.

of the RAM-SCX column through electrostatic repulsion forces (Figure 4 a).

**mAU**

**<sup>300</sup> a**

enrichment in one simultaneous step (Figure 1 c).

column endured about 200 injections of urine.

chromatography for the macromolecular sample constituents. By regulating the pore size of the particles, a physical restriction barrier is adjusted to regulate the interval of molecules that may penetrate and, in the case of penetration, may be trapped in the functionalized pore structure. A pore size of 6 nm allows access to the pores only for analytes with a molecular mass below 15 kDa. The proteins (>15 kDa) can thus be eluted with the void volume directly into the waste. Smaller analytes, however, such as drugs and metabolites from body fluids, pesticides or hormone residues from milk or animal tissue samples, may enter the pores and interact with the n-alkyl chains or ion-exchange groups bound to the inside of the pores. When dealing with complex samples i.e. human bio-fluids, the sample clean-up and fractionation of the sample into matrix and target analytes can be achieved. Depending on the ligands present in the pores, small molecules with hydrophobic or ionic properties are selectively enriched.

Fig. 3. Artistic representation of a SCX-RAM silica particle (LiChrospher 60 XDS (SO3/Diol), Merck KGaA, Germany). The external surface is coated with hydrophilic, electroneutral diol-groups for the exclusion of high molecular weight components (>15 kDa); the internal surface is functionalized with ion-exchange groups accessible for low molecular weight components which may be trapped by electrostatic interaction.

The diffusion barrier can be accomplished in two ways: (i) the porous adsorbent particles have a topochemically different surface functionalization between the outer particle surface and the internal surface. The diffusion barrier is then determined by an entropy controlled size exclusion mechanism of the particle depending on the pore size of adsorbent (Pinkerton 1991); (ii) the diffusion barrier is accomplished by a dense hydrophilic polymer layer with a given network size over the essentially functionalized surface. In other words, the diffusion barrier is moved as a layer to the interfacial layer inside the adsorbent particles, the exclusion properties are controlled by the size of the polymeric network protecting the internal surface and is no longer dependant on the average pore diameter of the adsorbent (Mazsaroff & Regnier, 1988).

The SEC process is entropically driven; i.e proteins with decreasing shape and size penetrate an increasing volume of the porous particles. The SEC of proteins is commonly carried out

chromatography for the macromolecular sample constituents. By regulating the pore size of the particles, a physical restriction barrier is adjusted to regulate the interval of molecules that may penetrate and, in the case of penetration, may be trapped in the functionalized pore structure. A pore size of 6 nm allows access to the pores only for analytes with a molecular mass below 15 kDa. The proteins (>15 kDa) can thus be eluted with the void volume directly into the waste. Smaller analytes, however, such as drugs and metabolites from body fluids, pesticides or hormone residues from milk or animal tissue samples, may enter the pores and interact with the n-alkyl chains or ion-exchange groups bound to the inside of the pores. When dealing with complex samples i.e. human bio-fluids, the sample clean-up and fractionation of the sample into matrix and target analytes can be achieved. Depending on the ligands present in the pores, small molecules with hydrophobic or ionic

Fig. 3. Artistic representation of a SCX-RAM silica particle (LiChrospher 60 XDS (SO3/Diol), Merck KGaA, Germany). The external surface is coated with hydrophilic, electroneutral diol-groups for the exclusion of high molecular weight components (>15 kDa); the internal surface is functionalized with ion-exchange groups accessible for low molecular weight

The diffusion barrier can be accomplished in two ways: (i) the porous adsorbent particles have a topochemically different surface functionalization between the outer particle surface and the internal surface. The diffusion barrier is then determined by an entropy controlled size exclusion mechanism of the particle depending on the pore size of adsorbent (Pinkerton 1991); (ii) the diffusion barrier is accomplished by a dense hydrophilic polymer layer with a given network size over the essentially functionalized surface. In other words, the diffusion barrier is moved as a layer to the interfacial layer inside the adsorbent particles, the exclusion properties are controlled by the size of the polymeric network protecting the internal surface and is no longer dependant on the average pore diameter of the adsorbent

The SEC process is entropically driven; i.e proteins with decreasing shape and size penetrate an increasing volume of the porous particles. The SEC of proteins is commonly carried out

components which may be trapped by electrostatic interaction.

properties are selectively enriched.

(Mazsaroff & Regnier, 1988).

with buffer solution containing a high salt concentration, e.g. 0.1 M, at pH 5 to 7. The high salt concentration is needed to suppress electrostatic interactions between the solute and the charged surface. In the sample clean up of the RAM-SCX column the concentration of salt is much lower e.g smaller than 20 mM and the pH is kept at approximately 3. Under these conditions, electrostatic attraction forces are dominant between the positively charged peptides and proteins whereas the negatively charged species are excluded from the pores of the RAM-SCX column through electrostatic repulsion forces (Figure 4 a).

Fig. 4. Typical SCX-RAM column separation profile: peaks (a) represent physical exclusion by pore size. Trapped retained bio-molecules are separated by a gradient in the second step (b). Conditions: column - LiChrospher 60 XDS (SO3/Diol), 25 x 4 mm I.D., flow rate - 0.5 ml/min, gradient from 0 to 1 M NaCl in 20 mM KH2PO4 pH 2.5, containing 5 % ACN in 30 min. Sample: 100 μl Human Hemofiltrate (3.7 mg/ml), UV detection at 214 nm.

After loading the RAM-SCX column the washing step elutes at isocratic conditions all the excluded compounds between the start and 15 minutes. After 15 minutes switching occurs and the trapped analytes are eluted from the RAM-SCX column with a strong eluent under gradient condition in the period between 15 and 45 minutes (Figure 4 b). Thus it is a charge and charge distribution selective process combined with SEC. Use of RAM-SCX allows one the direct application of biofluids onto the column. Small peptides are selectively trapped in the pores by cationic functional groups while large molecular weight biopolymers are directed to waste. This strategy performs the sample clean up and selective peptide enrichment in one simultaneous step (Figure 1 c).

Mass loadability of SPE and RAM columns play a key role in executing the sample clean-up. It is advisable to work below the overload regime of the column. Otherwise, displacement effects and other phenomena such as secondary interaction by adsorbed species might take place, which will lead to non-reproducible results (Wilemsen et al., 2004). Last statement is particularly important when the task is to monitor medium to low abundant proteins, therefore, usually large sample volumes in the millilitre range are applied. As the column lifetime is known to be limited a control measure has to be applied to check the condition of the RAM-SCX column and, if necessary, replace it by a new one. In our experience the column endured about 200 injections of urine.

Although compatibility of these stationary phases with direct biological sample injection is high, one still has to keep in mind that samples have to be filtrated or centrifuged prior to

Profiling of Endogenous Peptides by Multidimensional Liquid Chromatography 237

One of the major of those special features is low column backpressure. Low backpressure is not only nice to have, but a must, as setting up multidimensional separation platform for proteomics it allows one to select a desired flow-rate from a broader range (see Figure 6).

012345

0.00

0.02

0.04

0.06

Mass loadability per column (mg)

0.08

0.10

0.12

Column I.D. (mm)

Fig. 6. Estimation of monolithic silica columns flow rates and mass loadability per column. Columns 2, 3 and 4.6 mm I.D. are 10 cm long; columns 50 µm, 100 µm and 200 µm are 15 cm

Prof. Regnier group, demonstrated the advantage of flow rate variation possibility for 4.6 mm I.D. Chromolith Performance column. It was concluded that silica monolith reversedphase chromatography columns show little loss in the resolution of peptides ranging up to several thousand in molecular weight as mobile phase velocity is elevated from the conventional 2.5–25 mm/s (Xiong et al., 2004). Moreover, at 25 mm/s with a 100 mm length column, operating pressure did not exceed 150 bar. This is well within the pressure limit of most commercial LC instruments. The separation of a tryptic digest of cytochrome *C* in 6 and 60 min seemed almost identical. Resolution at 25 mm/s linear velocity was 77% of that at 2.5 mm/s. It was conluded that the fact that peptide separations could be achieved 10 times faster than with a conventional packed column with moderate loss in resolution could

Combination of different sizes fulfils the injection volume requirement for various samples. The possibility of being able to vary the flow-rate over a large area up to very high linear flow velocities combined with the robustness of the monoliths also reduce considerably the

Important to notice, that comparing a particulate and a silica monolithic guard column showed that the particulate column was clogging much faster than the monolithic column (Machtejevas et al., 2007). 120 injections of plasma (50 µl each injection) led to an increase of approximately 6 bar at the particulate column back pressure, while at the monolithic column the back pressure rise was only approximately 1 bar. The life time of the short silica monolithic columns used as a trap column or as a guard column heavily depends on a type

"down times" during washing and re-equilibration of the column (Rieuxet al., 2005).

0.1µl

1µl

10µl

0.1ml

Flow rate per min

long.

1ml

10ml

Flow rate

have a major impact on analytical throughput in proteomics.

Mass loadability per column

injection to remove the solid contaminants and precipitations. Even so, some components tend to agglomerate/precipitate with the time while samples queue up in autosampler. Therefore, an additional in-line filter is highly recommended. One should keep in mind that the operational flow rate has enormous impact on the molecular size distribution when employing a RAM column. Higher flow rates can shift the molecular range of the trapped molecules to lower values as smaller molecules need less time to penetrate the pores. Also higher flow rates could alter the hydrodynamic volume of the biomolecules. Higher molecular mass molecules will be enriched operating at lower flow rates. Column temperature affects the viscosity of the mobile phase and, consequently, the diffusion ratio and influences the speed of mass transfer. Carefully performed optimizations of the chromatographic parameters ensure the success of the analysis.
