CA2228413A1 - Use of polyurethane membrane for maldi-tofms analysis of whole blood - Google Patents

Abstract

The use of non-porous membranes as sample supports for MALDI-TOFMS analysis of peptides and proteins as well as for the analysis of whole blood is herein described. Non-porous membranes have a uniform surface, allowing for greater sensitivity and accuracy. Specifically, the non-porosity favours crystal growth on the surface of the membrane only, thereby providing enhanced spectral quality over membranes with porous structures. Studies were performed using polyurethane (PU) membranes as an example of non-porous MALDI sample supports which showed that PU membranes yielded higher quality spectra compared with porous membrane sample supports such as polyether and polyvinyl difluoride) and the spectra obtained from PU membranes are of comparable quality as those obtained with metallic targets. However, the sample preparation for use with PU
membranes is much less arduous compared with the preparation necessary for metallic targets.

Description

NON-POROUS MEMBRANE FOR MALDI-TOFMS
The present invention relates to a non-porous membrane for use as a sample support in matrix-assisted laser desorption time-of-flight mass spectrometry.
BACKGROUND OF THE INVENTION
Within the last decade, with the advent of electrospray ionization (ESI) and matrix-assisted laser desorptioNionization (MALDI), mass spectrometry has been shown to be able to contribute to the rapid, sensitive and accurate characterization of biomolecules. These techniques have allowed for the development of mass spectromei.ry-based methods for investigation of biomolecular structure and function.
However, i n order to achieve the best analysis, these techniques must be performed quickly, accurately and with a minimum of sample loss.
For example, matrix-assisted laser desorptionlionization time-of-flight mass spectrometry (MALDI-TOFMS) (Karas et al., 1991, Mass Spectrom. Rev.
10:335, Hillenkamp and Karas, 1990, Meth. Enzym. 193:280, Zaluzec et al., 1995, Prot. Exp. Purifification 6:109) provides a rapid and convenient means for the characterization of proteins and peptides derived from biological samples.
This method has the advantage of being relatively tolerant of impurities, such as salts and buffers. As a result, molecular ions of peptides and proteins can still be produced by MALDI, even with salts or buffers at concentrations which would hamper other ionization processes such as electrospray ionization (ESI). Delayed extraction, combined with reflecting time-of-flight mass analysis, provides high resolution and allows accurate mass measurements for sample components in the parts-per-million range, even for fairly complex mixtures (Vestal et al.) 1995, Rapid Common.
Mass Spectrom. 9:1044). In spite of the relative tolerance to impurities mentioned above, biologically derived samples must still be isolated and purified prior to analysis to obtain the best results. Specifically, MALDI analysis of complex mixtures is often hindered by the suppression or quenching of the signals from some analytes. In addition, the presence of impurities will inhibit the formation of matrix-analyte crystals suitable for the MALDI ionization process. The impurities will also lead to the formation of adducts which will degrade the resolution and mass accuracy of the results. Typically, this is overcome by purification of the sample prior to analysis.
Several methods of sample purification prior to MALDI-TOFMS analysis have been developed including dialysis and chromatography. Both methods have limitations, such as sample loss and time-consuming sample preparation.
Alternatively, the desired analytes may be removed from the sample using affinity binding-baaed purification methods. Specifically, these methods selectively retain and concentrates the analytes of interest; however, these methods are time-consuming which may in turn result in degradation of the target analytes.
A different approach is to carry out the purification on the MALDI probe surface itself, which avoids many sources of sample loss. For example, a small amount of powdered chromatographic packing placed on the MALDI target allows for the selective removal of interfering components (Rouse and Vath, 1996, Anal.
Biochem. ;238:82). Surface modified agarose beads have also been used for this purpose (Hutchens and Yip) 1993, Rapid Common. Mass Spectrom. 7:576).
Furthermore, the use of films for sample supports for MALDI mass spectrometry of impure sarnples was patented by John S. Cottrell (US Patent 5,260,571 ), which discloses the application and use of thin films on the surface of the MALDI
probe as a method of preparing a sample for analysis. An alternative technique consists of chemically modifying the probe surface by the addition of coatings such as nitrocellulose (Liu et al., 1995, Anal. Chem. 6T:3482) and Nafion (Bai et al, 1994, Anal. Chem. 66:3423). As an extension of this approach, C-18 derivatized targets have been prepared (Brockman et al., 1997) Anal. Chem. 69:4716). Basically, if the analyte of interest is selectively adsorbed onto the modified probe, interfering substances can be washed off, while the analyte is retained. However, in the afore-mentioned examples, modification of the probe surface for sample binding is time-consuming and the probes are good for only a limited number of uses. In addition, samples must still be transported to the MALDI-TOFMS laboratory by conventional means, that is, in solution and on ice.
Similarly, MALDI-TOFMS has been used recently to analyze hemoglobin from whole blood (Houston and Reilly, 1997, Rapid Commun. Mass Spectrum. 11:1435). Rapid screening for hemoglobin abnormalities is of great importance as many health authorities now require the screening of new-burns' blood for hemoglobin-related.diseases, as it has been shown that early detection of sickle cell disease; significantly reduces infant mortality rates from this disease (Vichinsky et al., 1988, Pediatrics 81:749). In the Houston and Reilly protocol (supra)) whole blood samples ar~~ diluted and mixed with matrix solution. The resulting mixture is placed on a stainless steel MALDI probe and allowed to dry. While the MALDI spectra obtained with this method are of good quality, sample preparation requires a skilled mass spectrometrist familiar with MALDI matrix preparation methods. In addition, the analysis needs to be performed immediately or the sample processed and stored on ice.
The use of membranes as sample supports has recently been adopted as a means of both sample purification and sample delivery into the mass spectrometer (Vestling and Fenselau, 1995, Mass Spectrom. Rev. 14:169; Strupat et al. in Mass Spectrometry in the Biological Sciences, A.L Burlingame and S.A.
Carr editors, Humans Press: Totowa, New Jersey, 1996) p203). Several different membranes have been used for sample supports for MALDI mass spectrometry, two examples of which are: poly(vinylidene difluoride) (PVDF) (Vestling and Fenselau, 1994 Anal. Chem. 66:4371; Strupat et al., 1994, Anal. Chem. 66:464) and polyether (Blackledgea and Alexander, 1995, Anal. Chem. 67:843). Specifically, these membranes have been used to prepare samples for MALDI-TOFMS following sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and for collection of capillary electrophoresis samples. That is, purification of the analyte sample is performed prior to MALDI analysis. It is of note that deposition of aqueous protein solutions onto membrane supports has been shown to enhance MALDI signals for samples containing buffer components in higher concentrations than can generally be tolerated with traditional protocols. Subsequent purification by on-probe washing andlor enzymatic digestion results in low sample loss since proteins and peptides are bound fairly strongly to the membrane by ionic and hydrophobic interactions.
However, 'the above-described membranes are porous and therefore have a heterogeneous surface which reduces the accuracy of the analysis.
Specifically, the porosity permits distribution of the analyte and matrix within the membrane (Blackledge~ and Alexander, 1995, Anal. Chem. 67:843). This distribution of analyte within the ~>orous surface reduces the amount of analyte available near the top of the membrane for sampling with the laser to bring about a MALDI-MS spectrum. Thus, 5 increased Laser irradiance is required to penetrate deep into the pores.
This increased laser irradi~ance results ~ in charging which causes an increase in the flight time, which significantl~~ reduces spectral quality compared with metallic targets. The distribution also results in a non-uniform initial starting point for ions, again reducing the spectral quality by reducing the resolution.
Here we report the use of non-porous membranes as MALDI probes, which therE3fore have a uniform surface, allowing for greater sensitivity and accuracy.
Specifically) the non-porosity favours crystal growth on the surface of the membrane only, thererby providing enhanced spectral quality over membranes with porous structures. Studies were performed using polyurethane (PU) membranes as an example of non-porous MALDI supports. While PU membranes have been used previously for the separation and concentration of neutral metal complexes and organic dyes from aqueous solution (Oleschuk and Chow, 1996, Talanta 43:1545), they have not previously been used as probes for MALDI analysis. PU membranes possess a unique two-phase structure consisting of hydrophobic soft domains and relatively hydrophilic hard domains, and proteins and lipids have been shown to adsorb through hydrophobic interaction with the soft domains of the polymer (Sreenivasan et al., 1992, J. Appl. Polym. Sci. 45:2105). PVDF and PE, in addition to other membranes used for sample supports, typically bind through hydrophobic and s ionic interactions depending on the material. The PU is unique as the soft segments will bind tf~rough hydrophobic type interactions with a strength that is somewhat weaker and hence more reversible than PVDF which binds proteins quite strongly.
This binding property of the soft segments of the PU membrane allows more analyte to be removed from the surface of the membrane into the matrix solution which in turn results in more sample available for MALDI analysis. The PU membrane also swells upon the addition of methanol which increases the effective surface area available for protein binding. As a result, the PU membranes offer greater sensitivity for MALDI
analysis.
The use of PU membranes as sample supports for MALDI-TOFMS
analysis of peptides and proteins as well as for the analysis of whole blood is herein described. It is of note that non-porous membranes may be used, for example, as supports for analyzing blood plasma, cerebral fluid, spinal fluid, saliva, tears and other biofluids as well as for enviro-monitoring and the like.
SUMMARY' OF THE INVENTION
According to one aspect of the invention there is provided a non-porous membrane for use as an analyte sample support for matrix-assisted laser desorptionlionization time of flight mass spectrometry analysis of biomolecules.
Preferably, the non-porous membrane has a homogenous, uniform surface. The non-porosity favours crystal growth on the surface of the membrane only, thereby providing enhanced spectral quality over membranes with porous structures.

The non-porous membrane may be composed of polyurethane.
Polyurethane membranes possess a unique two-phase structure consisting of hydrophobic soft domains and relatively hydrophilic hard domains, and proteins and lipids have been shown to adsorb through hydrophobic interaction with the soft domains of the polymer According to a second aspect of the invention, there is provided a method of preparing an analyte sample for matrix-assisted laser desorptionlionization time of flight mass spectrometry analysis, comprising:
(a) providing a non-porous membrane as a sample support;
(b) providing a matrix solution;
(c) applying the analyte sample directly to the non-porous membrane;
(d) allowing the analyte sample to dry;
(e) applying the matrix solution to the sample; and (f) allowing the analyte sample to dry.
Preferably, the non-porous membrane is composed of polyurethane. .
Polyurethane membranes possess a unique two-phase structure consisting of hydrophobic soft domains and relatively hydrophilic hard domains, and proteins and lipids have been shown to adsorb through hydrophobic interaction with the soft domains of the polymer.
Preferably, the method includes the additional steps of adding methanol to the anal~~te sample and allowing the analyte sample to dry following step (d). The addition of methanol to analyte samples deposited on the PU membrane causes swelling of the polyurethane which leads to enhanced protein sorption.
Specifically, , addition of methanol may cause disruption of the intermolecular forces holding the polymer chains together, thereby allowing an increase in the effective surface area available for protein sorption. Methanol also facilitates the partition of proteins and peptides from more polar components, for example salts.
The analyte sample may be whole blood. Specifically, the analyte sample may comprise a droplet of blood and said analyte sample may be applied directly to the non-porous membrane. That is, the droplet of blood is applied directly to the polyurethane membrane without any pre-treatment.
Preferably, the method includes the step of washing the analyte sample prior to step (e). In this manner, salts and the like may be removed from the analyte sample, thereby improving spectra quality. The washing solution may comprise for example, water, acidified water, a dilute trifluoroacetic acid solution, formic acid or acetic acid.
Preferably, the method includes the steps of adding a proteolytic enzyme in .a buffering solution to the analyte sample, allowing digestion of the analyte sample to occur and removing the buffering solution by washing the analyte sample prior to step (e).
Preferably, the method includes the step of adding additional analyte sample to the dried sample prior to step (e). In this manner, the analyte sample may be further concentrated.
The analyte solution may be extracted from a mixture. The analyte solution may be extracted from the mixture by placing the non-porous membrane in contact witi~ the mixture. In this manner, the analyte sample may be extracted from a mixture of samples andlor impurities and concentrated onto the non-porous membrane by placing the membrane in a solution containing the analyte sample.
The mixture may be a biological sample, for example, whole blood, blood plasma, cerebral fluid, spina!I fluid, saliva, or tears. Furthermore, the above-described process may be carried out in vivo.
One embodiment of the invention will now be described in conjunction with the ac~.~,ompanying drawings in which:
BRIEF DE;iCRIPTION OF THE DRAWINGS AND TABLES
Figure 1 is a side view of the MALDI-TOFMS probe.
Figure 2 is a schematic diagram of the structure of polyurethane (PU) in a membrane.
Figure 3 is the MALDI-TOF mass spectra of myoglobin using (a) a PU
membrane and (b) a metallic target.
Figure 4 is the MALDt-TOF mass spectra of 50 pmol bovine insulin using (a) a PU membrane, (b) a PVDF membrane and (c) a metallic target.
Figure 5 is the MALDI-TOF mass spectra of 5 pmol bovine insulin using (a) a PU membrane, (b) a PVDF membrane and (c) a metallic target Figure 6 is a comparison of measurements of the flight times of bovine insulin betvveen a PU membrane and a PVDF membrane.
Figure 7 is the MALDI-TOF mass spectra of myoglobin in the presence of NaCI using (a) a PU membrane and (b) a metallic target.
Figure 8 is scanning electron micrographs of myoglobin samples in the presence off NaCI on a PU membrane (a) untreated, (b) with matrix solution added, (c) following one wash and (d) following two washes and addition of matrix solution.
Figure 9 is the MALDI-TOF mass spectra of myoglobin in the presence of NaCI using a PU membrane (a) untreated, (b) after one wash, (c) after two washes, 5 and (d) after three washes.
Figure 10 is a MALDI-TOF mass spectra of tryptic digest products of citrate synthase (a) after a 2 minute digest performed directly on-membrane (PU) and (b) a 3 hour digest performed in an EppendorfT"" tube analysed on the PU
membrane with delayed extraction MALDI-MS.
10 Figure 11 is the MALDI-TOF mass spectra of tryptic digest products of citrate synthase after (a) a 2 minute, (b) a 5 minute , (c) a 10 minute, (d) a 30 minute and (e) a 60 minute digestion performed directly on the PU membrane.
Figure 12 is the MALDI-TOF mass spectra of (a) bovine serum albumin and (b) apotransferrin using PU membranes.
Figure 13 is the MALDI-TOF mass spectra of (a) the alpha chain of human hemoglobin and (b) the beta chains of human hemoglobin using PU
membrane.
Figure 14 is the MALDI-TOF mass spectra of the alpha and beta chains of human hemoglobin digested on-membrane (PU) with trypsin for (a) 2 minutes, (b) 5 minutes, (c) 10 minutes and (d) 60 minutes.
Figure 15 is the MALDI-TOF mass spectra of canine plasma on PU
membrane.
Figure 16 is the MALDI-TOF mass spectra of human plasma standards on PU membrane.
Table 1 is the mapping of selected tryptic fragments of citrate synthase from Figure 10.
Table 2 is the mapping of selected tryptic fragments of human hemoglobin from Figure 14C.
DETAILED DESCRIPTION
Referring to Figure 1, the non-porous probe 1 comprises a probe body and a non-porous membrane 14. The details of the probe body 10 are not shown 10 as the probe body 10 is a target stage of a mass spectrometer and is therefore known in the art. In this embodiment) the probe body 10 comprises a probe cap 12 and a support disk 16. The probe cap 12 is arranged to be fitted onto the probe body 10 so as to secure the non-porous membrane 14 to the probe body 10 as described below.
In this embodiment) the non-porous membrane 14 comprises non-porous ether-type PU membrane 18, 50 ~,m in thickness (catalog number XPR625-FS, from Stevens Elastomerics, Northampton, MA, USA) and the support disk 16 comprises a silver disk 20.
For MALDI-TOFMS analysis, the non-porous probe 1 is assembled as follows. The PU membrane 18 is washed with water and methanol prior to use.
Sample preparation is based on the dried-drop method (Karas and Hillenkamp, 1988, Anal. Chern. 60:2299). 2 ~I of an analyte-containing solution is placed on the PU
membrane 18 and allowed to dry slowly. 2 ~I methanol is added to the dried analyte and also dried. The addition of methanol to samples deposited on the PU
membrane 18 causes swelling of the polyurethane which leads to enhanced protein sorption.
Specifically, addition of methanol may cause disruption of the intermolecular forces holding then polymer chains together, thereby allowing an increase in the effective surface arena available for protein sorption. Methanol also facilitates the partition of proteins and peptides from more polar components, for example salts. At this stage, the dried analyte may be washed by applying 20 ~I aliquots of deionized, filtered water obtained from a Barnstead Nano-PureT"" water filtration system to the dried analyte an~j then removing the water repeatedly at intervals of one minute. 2 ~I of matrix solution (sinapinic acid saturated in 70:30 water-acetonitrile) is then added and allowed to crystallize slowly. At this time, the silver disk 20 is coated with a thin layer of adhesive. After the matrix had dried, the PU membrane 18 is placed on the silver disk 20) excess PU membrane 18 is trimmed away from the silver disk 20 and the silver disk 20 is placed onto the probe body 10 for MALDI analysis as described below. Alternatively, because the analyte-containing solution is dried onto the PU
membrane 18, the PU membrane 18 can be stored andlor sent to the MALDI
laboratory for analysis. The physical properties of the membrane facilitates its transfer through regular mail service allowing ease of transport from distant outside locations to the MAl_DI laboratory. Furthermore, samples can be applied directly to the PU
membrane 18 for analysis without dilution or pre-treatment, thereby reducing sample loss and degradation and improving accuracy and sensitivity of the MALDI
analysis.
As shown in Figure 2, PU membranes 18 possess a unique two-phase structure consisting of hydrophobic soft domains 22 and relatively hydrophilic hard domains 24.
The hard domains 24 are microcrystalline regions on the surface of the polymer, where the isocyanate portions of the polymer chains are aggregated by hydrogen bonding between the carbamate groups on adjacent polymer chains. These hard domains 24 are relatively polar in comparison to the soft segment domains 22.
The soft domains 22 consist of long chain polyethers which are relatively amorphous in character compared with the hard domains 24. The two-phase structure of the polyurethane elastomer provides two different regions of possible membrane-protein interaction:c, differing in polarity and in ability to form hydrogen bonds.
Hydrogen bonding wii:h the hard domains 24 and hydrophobic interactions with the soft domains 22 are believed to take place between the protein and the PU membrane 18, resulting in relatively strong binding (Sreenivasan et al., 1992). Furthermore, as noted above, the PU membranes 18 have a uniform surface, allowing for greater sensitivity and accuracy.
Alternatively, the probe body may comprise a commercially available MALDI probe having a flat, metallic surface to which the non-porous membrane is affixed using an adhesive. The non-porous membrane is then trimmed to fit the surface of the probe body prior to sample application.
Alternatively, the non-porous membrane may be affixed to a commercial probe body using an adhesive.
Alternatively, other washing solutions known in the art, for example, acidified water) a dilute trifluoroacetic acid solution, formic acid or acetic acid, may be used instead of water.
It is of note that any of the matrix solutions known in the art may be used in place of sinapininc acid.

' 14 Alternatively, additional sample may be applied to the membrane to further con~.~,entrate the sample prior to the addition of matrix solution.
In other embodiments, the analyte may be extracted from a mixture of samples andlor impurities and concentrated onto the membrane by placing the membrane in a solution containing the analyte. The solution may be a biological solution, for example, whole blood, blood plasma, cerebral fluid, spinal fluid, saliva, or tears. Furthermore) the above-described process may be carried out in vivo.
In other embodiments, the membrane is dissolved in a solvent and then applied directly to the probe body, thereby forming a film on the probe body.
The sample is then applied to the film as described above.
The use of PU membranes as sample supports for MALDI-TOFMS
analysis of peptides and proteins as well as for the analysis of whole blood is described in the Examples listed below. It is of note that other compounds if prepared so as to Ibe non-porous may be suitable for use as non-porous membranes.
Furthermore) it is of note that non-porous membranes may be used, for example, as supports for MALDI analysis of blood plasma, cerebral fluid, spinal fluid, saliva, tears and other biofluids as well as for enviro-monitoring and the like.
EXAMPLE I - PREPARATION OF SAMPLES FOR PEPTIDE AND PROTEIN MASS
SPECTROMETRY
Solutions of horse heart myoglobin (16,951 Da), bovine insulin (5,733 Da) and bovine serum albumin (66,430 Da) obtained from Sigma Chemicals (St.
Louis, MO, USA) and bovine apotransferrin (78,030 Da) obtained from Calbiochem (LaJolla, C~4, USA) were made up in deonizied, filtered water (10'i-10'i M) and used without further purification. The samples were prepared and applied to the PU
membrane 18 as described above.

MALDI-TOFMS was performed in the linear mode on Manitoba II, an instrument constructed in our laboratory, using an accelerating potential of 25 kV. In order to avoid saturation of the detector by low mass matrix ions, the detector was pulsed on ~~ 19,000 ns after each laser shot. Delayed extraction experiments were 10 performed on the same instrument with a delay time of 700 ~s with a pulse height of 3 kV and an accelerating potential of 20 kV. Spectra were obtained using a laser fluence (337 nm) adjusted slightly above threshold. Each spectrum presented here results from the sum of either 50 or 100 consecutive shots. Furthermore, external and internal calibration modes were used. For comparison purposes samples were also 15 applied to a metallic target and a PVDF membrane. In all cases, external calibrations for measurements using the various MALDI supports were performed with standards prepared on similar targets.
EXAMPLE III - COMPARISON BETWEEN PU AND METAL
Figure 3 shows MALDI-TOF mass spectra of 50 pmol of myoglobin using a PU membrane (Figure 3a) and a metallic target (Figure 3b). As can be seen, the spectra are essentially identical. It is of note that while equivalent resolution and mass accuracy were obtained, preparation of impure samples for MALDI analysis using PU membranes is much less arduous than preparation of the same samples on metallic probes. In addition, as noted above, samples bound to PU membranes do not need to be shipped on ice and may be stored at ambient temperatures.
EXAMPLE IV - COMPARISON BETWEEN PU, METAL AND PVDF
Figure 4 shows MALDI-TOF mass spectra of 50 pmol of bovine insulin obtained using a PU membrane (Figure 4a), a PVDF membrane (Figure 4b) and a metallic target (Figure 4c). As can be seen, the PU membrane and the metallic target yield equiv<alent resolution and mass accuracy for bovine insulin. However, the mass accuracy olbserved with the PVDF membrane was not as satisfactory. A
comparison was also made with 5 pmol of bovine insulin (Figure 5). In this case, PU
(Figure 5a) and the metallic target (Figure 5c) yielded comparable spectra to that observed with 50 pmol of sample while the PVDF (Figure 5b) membrane again produced poor quality spectra. In the case with PVDF, the laser intensity required to obtain ionization threshold vvas higher than that required for PU and the metal target. This was attributed to the porosity of PVDF) which permits distribution of the analyte and matrix within the membrane (Blackledge and Alexander, 1995, Anal. Chem. 6T:843). This necessitates an increase in the laser irradiance in order to sample within the pores which results in charging. The charging in addition to a non-uniform initial starting point for ions reduces the spectral quality by reducing the resolution. In comparison, the non-porous nature of the PU membrane, as with a metallic target, favours crystal growth on the surface of the membrane only, thereby providing enhanced spectral quality over membranes with porous structures. Thus, the non-porous membranes provide improved spectra compared to porous membranes. In fact, the quality is similar to that obtained with the metallic targets.
EXAMPLE 'V - COMPARISON OF FLIGHT TIMES FOR PU AND PVDF
Figure 6 shows the distribution in the flight times for bovine insulin ions desorbed from PU and PVDF membranes. It is of note that the distribution for the PU
membrane shows a similar variance to what is typically obtained on a metal target and a smalller variance than observed for ions desorbed from PVDF membranes.
In general, peak shapes were better on PU membranes compared to metal targets, making centroid assignment more systematic. This was possibly due to partitioning of the bound protein molecules from interfering adducts, for example, salts, which can affect the position of the peak centroid. PU-deposited samples were also tolerant of a large range of laser intensities, without observation of peak broadening due to charging or adduct formation. Conversely, a larger variance in flight times was observed with PVDF due to the spatial distribution of sample within the pores and the larger laser intensity required to generate spectra.
EXAMPLE VI - COMPARISON OF PU AND METAL USING NaCI-DOPED
SOLUTIONS
Figure 7 shows mass spectra of 200 pmol of myoglobin with 200 nmol of NaCI obtained using a PU membrane (Figure 7a) and a metallic target (Figure 7b). As can be seen, use of PU membranes brought a substantial improvement to the quality of the data obtained compared with metal targets for the analysis of NaCI-doped solutions. li: is apparent that selective partitioning of the protein molecules, NaCI, and matrix components between the aqueous phase and the surface of the PU membrane likely occurs, as some areas of the target produced good quality spectra even in the presence of excess NaCI. This was not the case with the metal target. This clearly shows that not only does the PU membrane provide comparable spectra quality to the metallic targets, in some instances, such as samples with high salt concentrations, provide spe:ctra of higher quality than metallic targets.
EXAMPLE VII - WASHING PROTOCOL
Scanning electron micrographs of 200 pmol myoglobin samples in the presence of 200 nmol NaCI on a PU membrane are shown in Figure 8. Application of NaCI-containing solutions to PU membranes resulted in a marked difference in the crystallization patterns of the protein and NaCI mixture before washing, after washing and with addition of matrix solution, as shown in Figure 8. Prior to washing (Figure 8a), NaCI is visible on the surface. After addition of matrix (Figure 8b), disrupted crystallization is observed, due to the presence of the NaCI. Following one washing step (Figure 8c), the visible amount of NaCI is removed and only a small amount of sample containing salt remains on the membrane. After two or more washing steps, protein and salt are not visible by scanning electron microscopy on the membrane surface. However, a sufficient amount of protein remains bound to the membrane as MALDI still produces strong signals. After washing, the addition of matrix (Figure 8d) results in the formation of analyte-matrix crystals, typical of a clean sample which will produce a c,~ood MALDI spectra.

Figure 9 shows the MALDI-TOF mass spectra of 200 pmol of myoglobin in the presnece of 200 nmol of NaCI on a PU membrane untreated (Figure 9a), and washed once (Figure 9b), twice (figure 9c) and three (Figure 9d) times. As can be seen from Figure 9, the relatively strong interactions of the PU membrane with proteins and peptides enables the introduction of a washing step.
Specifically, samples of myoglobin were prepared in a 1000-fold excess of NaCI and applied to the membrane. In this case MALDI spectra were obtained using a wide laser beam to ensure sampling of the entire surface of the target including areas which contained NaCI, myoc,~lobin and matrix. An overall improvement in peak shape and resolution was observed with increasing numbers of washes, as shown in Figures 9a-d. Some peaks in the spectrum of the untreated sample (Figure 9a) correspond to Na adducts.
These adducts cause peak broadening and make accurate mass assignment difficult.
After successive washing steps) the peaks become narrower as the abundance of Na adducts decreases with the removal of NaCI. The resulting increase in resolution enables coiTect mass assignment.
Thus, as shown in Figures 8 and 9, the PU membranes bind the analytes tightly enough to allow washing of the membrane, which in tum results in improved spectra quality. UVhile porous membranes also allow the use of washing steps, each time the analyte is solubilized in buffer, wash solution or matrix solution, the analyte is further distributed into the membrane. This distribution of analyte within the porous surface reduces the amount of analyte available near the top of the membrane for sampling with the laser to bring about a MALDI-MS spectrum. Thus, increased Laser irradiance is required to penetrate deep into the pores. This increased laser irradiance results in charging which causes an increase in the flight time, which significantly reduces spectral quality compared with metallic targets and the PU
membrane. The distribution also results in a non-uniform initial starting point for ions, again reducing the spectral quality by reducing the resolution.

EXAMPLE VIII - ON-MEMBRANE PROTEOLYTIC DIGESTION OF CITRATE
SYNTHASE=
Figure 10 shows the MALDI-TOF mass spectra of the tryptic digest products of 40 pmol of citrate synthase following a 2 minute digest performed directly 10 on-membrane (Figure 10a) and a 3 hour digest performed in an EppendorfT""
tube followed by analysis using delayed extraction (Figure 10b). Figure 11 shows the MALDI-TOI= mass spectra of tryptic digest products of citrate synthase after 2 minutes (Figure 10a), 5 minutes (Figure 10b), 10 minutes (Figure 10c), 30 minutes (Figure 10d) and 60 minutes (Figure 10e) of digestion performed directly on a PU
membrane.
15 In this manner, application of our membrane methodology to real samples was carried out by performing tryptic digests of citrate synthase directly on the PU
membrane and comparing 'the results to those obtained from samples digested in EppendorfT""
tubes (Figure 10). It is of note that similar spectra were obtained for samples digested in EppindorfT"' tubes and on the PU membrane (Figure 10), indicating that digestion of 20 citrate synthase was not hindered as a result of citrate synthase binding to the PU
membrane. Furthermore, most segments of the protein were mapped against calculated fragments, shown in Table 1, thereby providing further evidence that tryptic digestion w,as not hindered.

Digests were also performed for periods of time varying from 2-60 minutes directly on the PU membrane (Figure 11 ). Good quality MALDI spectra were observed following removal of the buffer components with the washing protocol described above. It is of note that over the duration of the digest, the initially abundant high mass ions were replaced with lower mass ions (Figure 11 a-e).
Furthermore, it is of note that the protein underwent significant digestion after only two minutes. This may indicate that the protein denatures upon sorption and drying on the PU
membrane, thus facilitating rapid digestion. It is of note that samples prepared on metallic targets did not produce spectra at all.
The 3 hour proteolytic digestion was used to investigate the advantages of using delayed extraction with samples deposited on the PU membrane. The results presented iin Figure 10b indicate a peak profile similar to the earlier digest profiles.
The use of delayed extraction resulted in a substantial increase in resolution as shown in the inset where the oxidation product of the compound, producing a peak at m/z 5759, may be observed. This enabled more accurate mass assignments as shown in Table 1. Furthermore, as noted above, this procedure is simply not possible with metallic targets without the use of extensive sample pre-treatment.
EXAMPLE IX - APPLICATION TO HIGH MASS PROTEINS
Figure 12 shows MALDI-TOF mass spectra of 20 pmol of high molecular weight proteins bovine serum albumin (Figure 12a) and apotransferrin (Figure 12b) on PU membranes. Specifically, high molecular weight proteins may be defined as those with a mass greater than 20,000 Da. It is of note that the results obtained are comparable: to that obtained using the metallic target. While a slight increase in mass was observed for the samples deposited on the PU membrane, this is likely due to charging and to use of external calibration. This phenomenon was observed only for higher m/z values and may be corrected with calibration in similar experimental conditions. Thus, PU membranes may be used for MALDI-TOFMS analysis of high molecular vveight proteins and peptides.
EXAMPLE X - APPLICATION TO WHOLE BLOOD ANALYSIS
Figure 13 shows the MALDI-TOF mass spectra of the alpha and beta chains of h~.~man hemoglobin. Approximately 0.5 ~.I of whole blood, that is, a droplet of blood, was collected directly onto the PU membrane and allowed to dry. The sample was then vvashed as described above. Spectra were acquired in linear mode with external calibration. Peaks correspond to alpha and beta chains of hemoglobin, in +1 and +2 charge states. 'Figure 14 shows the MALDI-TOF mass spectra of the alpha and beta chains of human hemoglobin following trypsin digestion of 2 minutes (Figure 14a), 5 minutes (Figure 14b), 10 minutes (Figure 14c) and 60 minutes (Figure 14d).
The sample was digested directly on the membrane with trypsin. Spectra were acquired in linear mode with calibration performed internally on known peaks.
The peak assignments corresponding to those observed for the 10 minute digest (Figure 14c) are given in table 2. Clearly, the spectra obtained are of good quality, indicating that the PUI membranes may be used for MALDI-TOFMS analysis of whole blood. In addition, hemoglobin variants may be rapidly and easily characterized.

DISCUSSION
As can be seen from the above Examples, the use of PU membranes as sample supports for MALDI-TOFMS analysis of proteins and peptides yields equivalent accuracy and resolution to values obtained with metallic targets and, produces superior resolution when, for example, analysing impure samples.
Specifically, it is the non-porous nature of the membrane that facilitates crystal growth on the surface of the membrane only and thus provides for enhanced spectral quality over porous membranes. Furthermore, the relatively strong interactions of the PU
membranes with bound proteins and peptides enables the introduction of a washing step in order to remove salt and buffer components, which may interfere with MALDI
analysis. As a result, the quality of the spectra obtained may be improved by repeated washes. Finally, tryptic digestion of citrate synthase performed on the membrane surface yielded characteristic fragments, allowing for successful peptide mapping.
It is of note that the above-described experiments were carried out using standard analytes under controlled conditions. Clearly, this does not in any way predict success for use of MALDI-TOFMS analysis of more complex solutions, such as biological fluids. However, we have now applied our PU membrane technology for the analysis of whole blood by MALDI-TOFMS. A sample is acquired in an outside location by a health care practitioner as follows: a lancet is used to prick a person's finger and the droplet of blood is collected directly onto the pre-washed PU
membrane simply by touching the finger to the membrane. The sample is then allowed to dry under ambient conditions. A chemical modifier may be added, for example, methanol, acetic acid or the like to disrupt coagulation or digestion may be pertormed at this time. The sample is then sent to a distant laboratory for MALDI-TOF analysis.
Sample preparation in the mass spectrometry lab is simple and may be pertormed by an inexperienced person according to the above-described protocol. That is) 2 ~,I
of methanol i:; applied to the dried sample on the membrane and the membrane is allowed to dry. The sample is then washed one to three times with cold water.
Matrix solution is added and allowed to dry. The sample is then affixed to a target and placed in the mass spectrometer for MALDI-TOF analysis. If desired) proteolytic digestion of the sample may be performed directly on the membrane prior to the washing step in order to obtain more information about the sample. As noted above, this may be done at the time of sample acquisition or at any time thereafter. Digestion proceeds according to the protocol described in Example IX. That is, between 2 and 10 ~,I of a 1 mglml solution of a proteolytic enzyme, for example, trypsin, in buffer is added to the blood :sample directly on the membrane and mixed with the sample.
Digestion may be performed for varying periods of time from 2 minutes to greater than 60 minutes, dependant upon the extent of digestion required.
As cyan be seen in Figure 13, good quality MALDI mass spectra of whole blood can be obtained which are comparable to purified samples. Furthermore, the procedure is easy to follow and may be performed by an inexperienced user.
Sample manipulation is facilitated by using the membrane as liquid samples are not required.
Analysis is rapid with minimal time required for hands on preparation. Samples may be collected in any setting with a lancet and a small piece of PU membrane.
The analysis is sensitive because there is minimal loss of sample during the digestion process and washing steps as the entire process is performed on the membrane.

Furthermone, using PU membrane with MALDI adds an insignificant cost to the analysis. Finally, on-membrane digestion followed data base mapping will facilitate the identification of abnormal amino acid sites in the sequence of the alpha and beta hemoglobin chains as spectra of intact chains may yield insufficient information.
5 Furthermore, we have results of MALDI-TOFMS analysis of human plasma standards (Figure 16;) and fresh plasma from canine samples (Figure 15). Several plasma proteins were observed to bind to the PU membrane and were characterized, as shown in Fiigures 15 and 16. This application to plasma demonstrates that PU
may be used as model biomaterial and this procedure may be used to model protein 10 sorption in-vivo.
It is of note that in other embodiments, membranes composed of other polymers may be used provided the membranes are manufactured so as to be non-porous. Furthermore, in other embodiments, any of the appropriate proteolytic enzymes known in the art may be used for on-membrane digestion of the sample.
15 Finally, it is of note that non-porous membranes may be used, for example, as supports for analyzing blood plasma, cerebral fluid, spinal fluid, saliva, tears and other biofluids as well as for enviro-monitoring and the like.
Since various modifications can be made in my invention as herein above des~~ribed, and many apparently widely different embodiments of same made 20 within the spirit and scope of the Gaims without department from such spirit and scope, it is intended that all matter contained in the accompanying specification shall be interpreted as illustrative only and not in a limiting sense.

Claims (16)

1. A non-porous membrane for use as an analyte sample support for matrix-assisted laser desorption/ionization time of flight mass spectrometry analysis of biomolecules.

2. The membrane according to claim 1 wherein the non-porous membrane has a homogenous, uniform surface.

3. The membrane according to claim 1 or 2 wherein the non-porous membrane is composed of polyurethane.

4. The membrane according to claim 3 wherein the non-porous membrane comprises a thin film of polyurethane applied to a probe body.

5. A method of preparing an analyte sample for matrix-assisted laser desorption/ionization time of flight mass spectrometry analysis, comprising:
(a) providing a non-porous membrane as a sample support;
(b) providing a matrix solution;
(c) applying the analyte sample directly to the non-porous membrane;
(d) allowing the analyte sample to dry;
(e) applying the matrix solution to the analyte sample; and (f) allowing the analyte sample to dry.

6. The method according to claim 5 wherein the non-porous membrane is composed of polyurethane.

7. The method according to claim 5 or 6 including the steps of adding methanol to the analyte sample and allowing the analyte sample to dry after step (d).

8. The method according to any one of claims 5 to 7 wherein the analyte sample is whole blood.

9. The method according to claim 8 wherein in step (c) the analyte sample comprises a droplet of blood and said analyte sample is applied directly to the non-porous membrane.

10. The method according to any one of claims 5 to 9 including the step of washing the analyte sample prior to step (e).

11. The method according to any one of claims 5 to 10 including the steps of adding a proteolytic enzyme in a buffering solution to the analyte sample, allowing digestion of the analyte sample to occur and removing the buffering solution by washing the analyte sample prior to step (e).

12. The method according to any one of claims 5 to 11 including the step of applying additional analyte sample to the dried analyte sample on the non-porous membrane prior to step (e).

13. The method according to any one of claims 5 to 12 including extracting the analyte sample from a mixture comprising the analyte sample and impurities.

14. The method according to claim 13 wherein the analyte solution is extracted from the mixture by placing the non-porous membrane into the mixture.

15. The method according to claim 14 wherein the mixture is a biological solution.

16. The method according to Claim 14 performed in vivo.