386 MASS SPECTROMETRY / Matrix-Assisted Laser Desorption/Ionization 913
464
1985 Gal–GlcNAc - Gal–GlcNac Gal–GlcNacGal–GlcNac Fuc
Fuc
638
control the whole system, from separation and introduction devices to complex MS/MS experiments, the ESI-oa-ToF combination is capable of automated analyses on very small quantities of material, far exceeding the sensitivity of liquid SIMS. See also: Mass Spectrometry: Overview; Electrospray; Peptides and Proteins. Surface Analysis: Secondary Ion Mass Spectrometry of Polymers.
638
1087 1985
Gal–GlcNAc - Gal–GlcNac Gal–GlcNacGal–GlcNac 464 Figure 7 The fragmentation patterns induced after FAB ionization that distinguishes two isomeric sugars (1984; example provided by M-Scan Ltd.).
use of mass spectrometry as, arguably, the most important analytical tool for biochemistry, by itself this technique is almost obsolete. The development of atmospheric pressure electrospray (ESI), then microelectrospray sources, and of analyzers such as orthogonal axis time-of-flight (oa-ToF) has allowed analysis of biopolymers to become a routine in most biology research laboratories. Amenable to the application of sophisticated computing programs to
Further Reading Barber M, Bordoli RS, and Sedgwick RD (1980) Fast atom bombardment mass spectrometry. In: Morris HR (ed.) Soft Ionisation Biological Mass Spectrometry. London: Heyden. Clench MR (1992) A Comparison of Thermospray, Plasmaspray, Electrospray and Dynamic Fab. Manchester: VG Micromass. Gaskell SJ (ed.) (1986) Mass Spectrometry in Biomedical Research. Chichester: Wiley. Ito Y, Takeuchi T, Ishii D, and Goto M (1985) Direct coupling of micro high performance liquid chromatography with fast atom bombardment mass spectrometry. Journal of Chromatography 346: 161–166. McNeal CJ (ed.) (1986) Mass Spectrometry in the Analysis of Large Molecules. New York: Wiley. Morris HR, Panico M, and Haskins NJ (1983) Comparison of ionisation gases in FAB mass spectra. International Journal of Mass Spectrometry and Ion Physics 46: 363–366.
Matrix-Assisted Laser Desorption/Ionization D J Harvey, University of Oxford, Oxford, UK & 2005, Elsevier Ltd. All Rights Reserved.
Introduction Matrix-assisted laser desorption/ionization (MALDI) is a relatively new ionization technique that is capable of ionizing a large variety of compounds, particularly large proteins, for analysis by mass spectrometry. Its introduction, together with electrospray ionization in the late 1980s, extended the mass range of molecules that could be examined by mass spectrometry into the low megadalton range and provided a new tool that could be used for analysis of both biopolymers and synthetic polymers. The technique is comparatively simple to use; the sample is mixed with an excess of a suitable matrix, usually
a solid, and ionized with pulses from a laser (Figure 1). Ion detection is usually, but not exclusively, achieved by use of a time-of-flight (TOF) mass spectrometer. The purpose of the matrix is mainly to dilute the sample and dissipate the laser energy; however, the details of the ionization process are still poorly understood and subject to intensive research. The technique is capable of high-throughput spectral recording and typical target plates, such as those used for proteomics, can accommodate several hundred samples. Since its introduction, MALDI has become one of the major techniques for ionizing organic molecules and has largely replaced older, less sensitive, methods such as fast-atom bombardment. This article briefly reviews the technique and describes applications to several compound types. Two variations of MALDI mass spectrometry were invented almost simultaneously in the late 1980s. In
MASS SPECTROMETRY / Matrix-Assisted Laser Desorption/Ionization
Laser
Sample plate
Extraction lens
TOF mass spectrometer
387
provide collisional cooling of the ion beam have recently allowed hybrid quadrupole-TOF (Q-Tof) instruments to record spectra from MALDI ion sources and to produce high-quality fragmentation spectra by collisionally induced decomposition. The recently introduced TOF/TOF instruments have enabled high-energy fragmentation spectra to be obtained; such spectra provide additional information such as the structure of the amino acid side-chain of peptides and linkage position between the sugar rings of complex carbohydrates.
Samples Figure 1 MALDI ion source.
the version invented by Koichi Tanaka in Japan, a matrix of glycerol and cobalt powder was mixed with the sample and ionized with a laser. In the other version, invented by Michael Karas and Franz Hillenkamp in Germany, and which has enjoyed much more popularity, the sample was co-crystallized with a ultraviolet absorbing matrix, nicotinic acid, and ionized with a frequency-quadrupled Nd-YAG laser. The main products were singly charged ions such as [M þ H] þ or [M þ Na] þ and the technique was shown to be capable of ionizing proteins with masses in excess of 100 kDa.
Instrumentation Commercial instruments now almost invariably use nitrogen lasers emitting at 337 nm although several other wavelengths including those in the infrared (IR) region of the spectrum have been used experimentally. As MALDI naturally produces a pulsed ion beam, it is ideally interfaced to TOF mass spectrometers that are ideally suited to record the massive, largely singly charged ions produced in most analyses. The mass range of the technique is usually limited, not by its ability to ionize molecules but by the drop in signal strength with increasing mass imposed by most common mass spectrometric detectors. In reality, the method is useful up to B300 kDa on commercial instruments although ions can sometimes be seen at much higher mass. Other types of mass spectrometer have been used to record MALDI spectra; magnetic sector instruments equipped with an array detector initially offered the advantage of higher resolution than was available with the early linear TOF instruments but these spectrometers have now largely been replaced with reflectron-TOF instruments using delayed extraction ion sources. Mass spectrometers operated with higher than normal ion source pressures to
Matrices Most types of molecule can be ionized given a suitable matrix and a considerable amount of work has been performed on the determination of the best matrices for various compounds. Many of the most popular matrices are listed in Table 1. Some of these matrices, such as 2,5-dihydroxybenzoic acid (DHB), are suitable for a broad range of compounds whereas others, such as sinapinic acid, tend to work with only a few compound types, such as proteins in the case of sinapinic acid. As the technique has matured, most analyses for a particular compound type will be conducted with only a limited number of matrices. In order to function as a satisfactory matrix, a compound should possess the following properties: 1. The ability to form crystals incorporating the analyte into the crystal lattice or the ability to dissolve the analyte if the matrix is a liquid. 2. The ability to absorb the laser energy. 3. Possess sufficiently poor volatility that sample spots remain on the target in the vacuum system of the mass spectrometer for a reasonable time but, at the same time, be sufficiently volatile that they can be vaporized with the laser. 4. Possess suitable chemistry such that they are able to ionize the sample, usually by proton transfer, but, at the same time, they should not modify the analyte chemically. Most of the early matrices were small organic acids capable of protonating the proteins and amino acids that were the main early targets of the technique, but many neutral and even basic compounds are now employed for different compound types. The most common method for preparing a MALDI target, known as the dried droplet technique, is to mix the sample and matrix, usually in a ratio of about 1:5000 in B1 ml of solvent, and allow the mixture to dry on the target. Mixing can occur either before addition to the target or, more commonly, solutions of the sample and matrix are added to the target independently and allowed to mix. Most
Table 1 Common matrices for MALDI mass spectrometry Matrix
Abbreviation
Molecular weight
Solvent
Analyte
3-Amino-4-hydroxybenzoic acid
–
153.1
MeCN/H2O
Sugars
Structure
COOH
NH2 OH 2,5-Dihydroxybenzoic acid
DHB
154.1
MeCN, acetone
General, particularly sugars
COOH OH
HO DHB/2-hydroxy-5-methoxybenzoic acid
Super-DHB
154.1/168.1
MeCN, acetone
Peptides, sugars
COOH
COOH
OH
OH + CH3O
HO DHB/1-hydroxy-iso-quinoline
DHB/HIQ
154.1/145.2
MeCN/H2O
Sugars
COOH OH +
N
HO
2-(40 -Hydroxy-phenylazo)benzoic acid
HABA
242.2
Acetone
Polymers, sugars
OH COOH N
N
OH Continued
Cinnamic acid
–
148.1
MeCN/H2O
General
Nicotinic acid
–
123.1
H2O
Proteins
COOH
COOH
N Sinapinic acid (3,5-dimethoxy-4hydroxycinnamic acid)
–
224.2
MeCN/H2O
Proteins, polymers
CH3O
COOH
HO OCH3 a-Cyano-4-hydroxycinnamic acid
4-HCCA
189.2
MeCN/H2O, acetone
Peptides, lipids
COOH CN HO
2,4,6-Trihydroxyacetophenone
THAP
167.2
Ethanol
Sugars. Particularly those containing sialic acid, oligonucleotides
O HO
OH
OH Arabinosazone
–
328.4
EtOH, MeOH
Sugars
N
H
N
NH
N HC
OH
HC
OH
CH2OH
Continued
Table 1 Continued Matrix
Abbreviation
Molecular weight
Solvent
Analyte
6-Aza-2-thiothymine
ATT
143.2
MeCN/H2O
Gangliosides
Structure
OH CH3 N N HS
6,7-Dihydroxycoumarin (Esculetin)
–
178.2
H2O
Sugars
N
HO
HO
Caffeic acid
–
180.2
MeCN/H2O
Glycosaminoglycans
O
O
HO
COOH
HO
Ferulic acid
–
194.2
MeCN/H2O
Glycoproteins
CH3O
COOH
HO
Harmane derivatives
–
182.2
MeCN/H2O
Cyclodextrins. sugars
N N H CH3
3-Hydroxypicolinic acid
HPA
139.1
Ethanol
Oligonucleotides
OH
N
COOH
CH3
CH3
CH3 – All-trans-retinoic acid
300.4
Tetrahydrofuran (THF)
Synthetic polymers
CH3
OH Synthetic polymers Tetrahydrofuran (THF) – 1,8,9-Trihydroxyanthracene (dithranol)
226.2
– 3b-Indoleacrylic acid
187.2
Acetone
Synthetic polymers
OH
N H
OH
COOH
CH3
COOH
MASS SPECTROMETRY / Matrix-Assisted Laser Desorption/Ionization
391
samples are allowed to dry under ambient conditions but drying is sometimes accelerated by heat, currents of air or nitrogen, or by placing the target in a desiccator. Although this method works well for the majority of compounds, many variations have been developed to accommodate different analyte properties or simply to improve signal strength or quality. For example, in the layering technique, a film of matrix is deposited on the target in an organic solvent and allowed to dry quickly. The sample solution, with or without additional matrix, is then placed on top and allowed to dry as normal. Sometimes an additional matrix is added at this stage and again allowed to dry. The small matrix crystals from the original layer are thought to catalyze the formation of small crystals in the final target spot, thus producing a more homogeneous target. Some matrices, such as DHB, produce only a few large crystals, usually originating at the periphery of the target and pointing inwards. Consequently, with instruments that are not fitted with a camera that is able to image the target, it can be difficult to find a suitable spot from which to record a signal. Thus, in addition to the layering technique, several other procedures have been introduced to overcome the problem. One such method, used with carbohydrates, is to recrystallize the sample spot, once dried, with ethanol. This procedure not only reduces the crystal size but is also thought to produce better uptake of the sample by the matrix. Another method is to use a mixed matrix where the added substance, usually in a ratio of about 10:1 sample to additive, forms small seed crystals for the main matrix. Examples are the addition of 2-hydroxy-5-methoxybenzoic acid or 1-hydroxy-iso-quinoline (HIQ) to DHB, the former mixture being known as super-DHB. Additives may also be used to form ion pairs with the sample or impurities in the target solution. Ion pairing of sulfated carbohydrates with amino acids (see below) is an example of the first additive whereas the use of citric acid to bind interfering cations in hydroxybenzophenone matrices is an example of the second. L-Fucose has also been used as a matrix adduct to improve spectral quality; it has been proposed that this decomposes to carbon dioxide and water within the laser plume to enhance the number of ion– molecule reactions. It appears to be important for the sample to enter the crystal lattice of the matrix for satisfactory ion production to occur and proteins tagged with a dye have been used to demonstrate that inclusion does indeed take place. Under ideal conditions high sensitivity can be achieved; some samples can be analyzed from as little as a few hundred attomoles applied
392 MASS SPECTROMETRY / Matrix-Assisted Laser Desorption/Ionization
to the target and often, after spectral acquisition, the remaining sample can be recovered. The technique is reasonably sensitive to the presence of contaminants such as salts and buffers but is more tolerant to the presence of these compounds than techniques such as electrospray. A slight drawback to the technique results from the inhomogeniety of the sample surface that causes the signal strength to vary across the target. Consequently, it is sometimes necessary to hunt for so-called sweet spots when acquiring a spectrum. MALDI is a relatively quantitative technique providing that sufficient laser shots are acquired per spectrum to even out the variations in intensity caused by the inhomogeneous target. However, although accuracy can be relatively high, precision is usually only in the order of 710%. Except for some of the larger proteins, MALDI produces almost exclusively singly charged ions. The reason is thought not to be associated with the initial ionization process, but with subsequent reactions in the plume of material desorbed by the laser. It is thought that, as in electrospray ionization, multiply charged ions are initially formed but that these undergo collisions in the plume, predominantly with matrix molecules, allowing neutralization to occur. It has been proposed that such processes are mainly under thermodynamic control, allowing predictions to be made as to the nature of the resulting spectra. Such ion–molecule reactions could explain, for example, why ionization from matrices such as DHB doped with divalent metals such as copper, produce only singly charged ions. Because the second ionization potential of copper (20.29 eV) is above that for DHB (8.05), collisions between ionized copper and DHB result in charge transfer to the matrix. Similar processes are thought to account for phenomena such as matrix suppression and specific analyte ion suppression that are seen under some conditions. Sample purity is an important factor in obtaining a good MALDI spectrum as contaminants such as salts and buffers tend to inhibit efficient crystallization of sample targets. Among micropurification methods in common use is drop-dialysis in which droplets (about 1 ml) of the sample solution are spotted onto a dialysis membrane floating on the surface of water. A variant, used in carbohydrate work, is to use a Nafion-117 membrane that additionally adsorbs hydrophobic compounds. Various ion-exchange resins are in common use; for MALDI work, these are frequently used as microcolumns packed into small pipette tips. Thus, peptides can be efficiently desalted by adsorption onto a C18 solid phase often packed into a pipette tip and sold under the name of ZipTip. Various other such tips including ion-exchange resins have recently become available. Another desalting
technique is to add a few beads of an ion-exchange resin, such as AG-50, to the MALDI target. The beads can either be removed mechanically before introduction of the target into the mass spectrometer, or they can be left on the target with little apparent effect on spectral quality.
Variants of the Basic MALDI Technique Atmospheric Pressure MALDI
In atmospheric pressure MALDI (ap-MALDI), the sample is ionized outside the vacuum system and ions are captured through a small orifice, usually into an ion-trap mass spectrometer. Its advantage appears to be a considerable amount of rapid collisional cooling of the ions by the high atmospheric pressure, leading to stabilization of sensitive compounds. On the other hand, some compounds, such as carbohydrates, appear to suffer increased fragmentation under these conditions. Surface Enhanced Laser Desorption/Ionization
Surface enhanced laser desorption/ionization (SELDI) is a variant of MALDI in which the MALDI probe is derivatized with various substances that have affinity for the analyte. The probes are then used to extract the analyte directly from mixtures thus avoiding sample loss through more complicated procedures such as column chromatography. Contaminants can be washed from the probe with appropriate buffers or solvents leaving the purified analyte ready for analysis. Many adsorbents have been used; typical examples are hydrophobic or ionic compounds, enzymes, various receptors, antibodies, and nucleic acids. Although most applications have been reported with proteins, the technique is potentially applicable to any type of compound for which a specific adsorbent can be attached to the probe. Three types of SELDI are recognized, surfaceenhanced affinity capture (SEAC), surface-enhanced neat desorption (SEND), and surface-enhanced photolabile attachment and release (SEPAR). With SEND, the energy-absorbing molecules are bound to the probe surface, often covalently and no additional matrix is necessary. In recent years, porous silica has been found to be an effective SEND surface. Although large molecules can be desorbed, the technique has largely been restricted to small-molecule analysis. Additional binding between the analyte and the energy-absorbing compound is involved with SEPAR and analytes are released photochemically during the ionization process. SEAC is the most widely used SELDI technique; the probe essentially
MASS SPECTROMETRY / Matrix-Assisted Laser Desorption/Ionization
acts as the purification device via affinity capture and then spectra are recorded, as in classical MALDI, following the addition of an appropriate MALDI matrix. Various affinity probes are now commercially available and the technique offers great potential for examination of trace material in complex biological matrices.
Applications of MALDI to Specific Compound Types Proteins
Proteins produce mainly singly charged ions (Figure 2), usually by protonation and the larger ones are usually ionized from sinapinic acid. However, DHB, ferulic acid, and nicotinic acid can also be effective. A small amount (usually 0.1%) of organic acid, such as trifluoroacetic acid, is often added to aid protonation. In addition to the main singly charged monomeric ions, the resulting spectra usually contain weak doubly charged ions and oligomers, usually dimers and occasionally trimers. For many proteins, these additional oligomeric ion peaks are artifacts of the MALDI process and are more abundant from samples containing high concentrations of protein. However, their observation has prompted research into whether MALDI can be used to study protein– protein interactions. The answer appears to be yes in certain circumstances although, in general, milder techniques such as electrospray, which maintain more realistic physiological conditions, are preferred. Nevertheless, MALDI has been used for measurements of physiologically relevant protein–protein interactions under appropriate conditions. These conditions include proper choice of matrix and pH, the use of chemical cross-linking, or by recording spectra from only one laser shot aimed at any particular region of the target in order to minimize laserinduced dissociation of the complex. 12 359.2 [M + H]+
Relative abundance (%)
100
50 [M + 2H]2+ 6183.3 7909.4
[2M + H]+ 24 720.0
0 6000
10 000 14 000 18 000 22 000 26 000 m/z
Figure 2 Typical positive ion MALDI mass spectrum of a protein.
393
The accuracy of protein mass measurement depends on a number of factors because many proteins are sufficiently massive that resolution is not achieved from additional species such as salts or matrix adducts. Under these conditions, such species cause peak broadening and result in a small shift in measured mass to higher values. Post-translational modifications such as glycosylation can often be resolved for the smaller proteins such as ribonuclease B (Figure 3) with a single glycosylation site, but larger compounds with multiple modifications often produce only broad peaks. Peptides
Peptides are usually ionized as [M þ H] þ ions from a-cyano-4-hydroxycinnamic acid (4-HCCA) or DHB and produce very different molar responses depending on the proton affinity of the constituent amino acids. Arginine-containing peptides, for example, usually produce very strong signals and can be detected at the attomole level. Glycopeptides, on the other hand, frequently give very weak signals in the presence of peptides due to ion suppression phenomena and can ionize as [M þ H] þ and/or [M þ Na] þ species depending on the size of the carbohydrate and the proton affinity of the peptide. A major application of MALDI to peptide analysis is in the area of proteomics where MALDI can provide a very rapid throughput of the resulting tryptic peptides. Unlike electrospray, however, under MALDI conditions, tryptic peptides will produce singly rather than doubly or triply charged ions. Carbohydrates
Many matrices have been used for carbohydrate analysis but 2,5-DHB has proved to be the most versatile and widely used. Esculetin is more specific, however, and ionizes sugars in preference to peptides and glycopeptides in mixtures of these compounds (Figure 4). Carbohydrates form [M þ Na] þ ions with most matrices but other adducts can be produced if the matrix is doped with the appropriate salt. Anionic carbohydrates additionally form [M H] ions. Sensitivity in both modes is lower than with peptides as the latter can be efficiently protonated because of their high proton affinity. However, sensitivity can be increased by permethylation or derivatization at the reducing terminus with a compound that can be protonated, such as a tertiary amine or one that possesses a constitutive charge. Alkyltrimethylammonium compounds have been used in the latter context. Mixtures of polymeric carbohydrates, such as dextrans, tend to show a drop in sensitivity for the
394 MASS SPECTROMETRY / Matrix-Assisted Laser Desorption/Ionization Ribonuclease B
14 900
100
= Mannose
Relative abundance (%)
= GlcNAc 15 062
Ribonuclease A 13 684 50
15 386 15 225 15 549
0 13 500
14 000
14 500
15 000
15 500
16 000
m/z Figure 3 Positive ion MALDI mass spectrum of ribonuclease B, a glycoprotein with a single N-linked glycosylation site.
1257.4
100
Glycopeptides
2,5-DHB
1419.5
Relative abundance (%)
50
1581.5 1299.6
1095.4
1461.5
1853.6 1682.5 1743.6 1824.6
1950.6
2015.6
0 100
(A)
1257.4
Esculetin
50
1419.5
1095.4
1581.5
1156.4
1663.5
1743.6
1853.6
0 1100 (B)
1200
1300
1400
1500
1600
1700
1800
1900
2000
m/z
Figure 4 Positive ion MALDI mass spectrum of N-linked glycans released from ribonuclease B and recorded from (A) 2,5-DHB and (B) esculetin.
larger molecules such as those with masses B10 kDa (Figure 5) but signals can be increased by removal of the shorter polymers by size-exclusion chromatography. For smaller compounds such as
N-linked glycans, ionization appears reasonably consistent over the mass range of 2–3 kDa that most of these compounds fall into but smaller O-linked glycans suffer a drop in sensitivity and interference
MASS SPECTROMETRY / Matrix-Assisted Laser Desorption/Ionization
100
395
1175.4 1499.5 1013.3 1337.4
Relative abundance (%)
1661.5 1823.5 1985.6 2147.7
851.2
2471.8 2309.7 2633.8
50
2795.9 2958.0 3120.0 3282.1 3444.1 3606.2 3768.2 3903.3 0 1000
1400
1800
2200
2600
3000
3400
3800
m /z Figure 5 Positive ion MALDI mass spectrum of glucose oligomers recorded from 2,5-DHB.
by ions from the matrix. The consistent ionization shown by glycans of similar mass is related to the ionization being produced by sodium adduction rather than protonation, a mechanism that is similar for all compounds. Sialylated glycans are unstable under MALDI conditions and eliminate substantial amounts of sialic acid with the result that broad metastable ions appear in spectra recorded with reflectron-TOF instruments. Methylation or salt formation can stabilize the sialic acids by removing the labile acidic proton that is involved in the sialic acid loss. Sulfated glycans, such as glycosaminoglycans, eliminate sulfate but can be examined by ion pairing with a suitable amino acid such as (RG)15. Nucleic Acids
Small oligonucleotides with four to six base pairs were first analyzed by MALDI mass spectrometry in 1990 but it was soon clear that oligonucleotide analysis by this technique was beset by two problems: salt formation and fragmentation. Salts were the result of the affinity of the phosphate groups for alkali metals whereas fragmentation appeared to be related to protonation of the base. Salt formation can be reduced by use of ammonium salts, either added to the matrix or exchanged into the sample before target preparation. The ammonium ion transfers a proton to the phosphate to create the free acid by dissociation of the ammonium phosphate ion pairs following desorption. Fragmentation can be reduced by chemical modification. Thus, 7-diazaanalogs of purine groups confer stability whereas adenine and guanine bases can be stabilized by replacement of the 7-nitrogen group with carbon.
Like other compounds, successful analysis of oligonucleotides depends on selection of a suitable matrix. An early system used frozen aqueous solutions on a copper target with irradiation with visible light at 581 nm. DNA with more than 600 bases could be desorbed intact. However, molecules with over 2000 nucleotides have been detected using IRMALDI. Possibly the most useful matrix for oligonucleotides is 3-hydroxypicolinic acid (HPA). It produces both positive and negative ions although better results have been reported for the negative ion spectra. The addition of a polyamine comatrix such as spermine has been reported to improve spectral quality. Double-stranded DNA becomes denatured under MALDI conditions with only ions corresponding to the monomer chains appearing in the spectra. Sensitivity appears to be a function of base composition. Thus, homopolymers containing thymine yield stronger signals than homopolymers containing the other three bases. The reason is thought to be related to fragmentation rather than absorption into the MALDI crystal lattice as thymine is more difficult to protonate than the other bases. Reduced fragmentation has been proposed as the reason why RNA produces stronger signals than DNA; the presence of the 20 -hydroxy group prevents the occurrence of the proposed 10 ,20 -trans-elimination involved in glycosidic cleavage of the ribose moiety from the base. Lipids
Analysis of lipids by MALDI mass spectrometry has received much less attention than analysis of proteins and peptides but lipids can, nevertheless, be induced to produce strong signals with suitable matrices.
396 MASS SPECTROMETRY / Matrix-Assisted Laser Desorption/Ionization 1375.9 1331.9 1419.9 1464.0 1287.9 1508.0
100
Relative abundance (%)
785.6
637.6
741.5
829.6 873.6
697.5
1552.0
1243.8 1199.8
1596.1
917.7 1155.8
961.7
569.5
1640.1
50 1005.7 1111.8
1684.2
1049.8 1728.2 1773.2 1817.2
0 600
800
1000
1200 m /z
1400
1600
1800
Figure 6 Positive ion MALDI mass spectrum of polyethylene glycols recorded from 2,5-DHB.
Phospholipids can be ionized from the popular matrices such as DHB or 4-HCCA but positive ion signal strengths vary considerably with the different types of lipid; phosphatidylcholine, with its quaternary nitrogen atom, produces very strong positive ion signals whereas phospholipids such as phosphatidylinositol, which do not possess a positive charge or amine group, are more difficult to ionize and produce weaker signals. Glycosphingolipids generally behave as derivatized carbohydrates and ionize well from DHB with the production of [M þ Na] þ ions. However, deacylation as with lyso-sphingolipids, converts the amide group into an amine with the result that both [M þ H] þ and [M þ Na] þ ions are formed. MALDI is used extensively in the analysis of bacterial glycolipids such as lipid A, usually in negative ion mode, because of the presence of phosphate or other acidic groups. Synthetic Polymers
A considerable amount of work has been reported on the use of MALDI mass spectrometry to examine synthetic polymers. The most popular matrices are DHB, HABA, 1,8,9-trihydroxyanthracene (dithranol), and all-trans-retinoic acid, doped with a cationization reagent such as copper or silver. The oligomer distribution, the composition of the end groups, the repeat unit, and the nature of any chemical modifications can all be determined if oligomer
resolution can be attained but one of the major applications is in measurement of the mean molecular weight of polydisperse fractions (Figure 6). The main problem with analyses of this type is mass discrimination against the larger compounds, caused both by instrumental and matrix factors. Instrumental factors are usually the result of detector saturation as most detectors do not have the dynamic range necessary to accommodate the concentration differences between the various constituents of the polymer mixture. Solvents also have a marked effect on the accuracy of the results. As most synthetic polymers are insoluble in aqueous solvents, the presence of water can have an adverse effect on crystallization. In mixtures containing water, if the organic solvent is more volatile it will evaporate first, causing precipitation of the polymer and inefficient incorporation into the matrix. Nevertheless, for mixtures covering a relatively narrow mass range, MALDI can give an accurate value for the average molecular weight. See also: Laser-Based Techniques. Liquid Chromatography: Size-Exclusion. Mass Spectrometry: Time-of-Flight. Proteomics.
Further Reading Ashcroft AE (1997) Ionization Methods in Organic Mass Spectrometry. Cambridge: Royal Society of Chemistry.
MASS SPECTROMETRY / Mass Separation 397 Dong X, Gusev A, and Hercules DM (1998) Characterization of polysiloxanes with different functional groups by time-of-flight secondary ion mass spectrometry. Journal of the American Society for Mass Spectrometry 9: 292–298. Farmer TB and Caprioli RM (1998) Determination of protein–protein interactions by matrix-assisted laser desorption/ionization mass spectrometry. Journal of Mass Spectrometry 33: 697–704. Festag R, Alexandratos SD, Joy DC, Wunderlich B, Annis B, and Cook KD (1998) Effects of molecular entanglements during electrospray of high molecular weight polymers. Journal of the American Society for Mass Spectrometry 9: 299–304. Harvey DJ (1999) Matrix-assisted laser desorption/ionization mass spectrometry of carbohydrates. Mass Spectrometry Reviews 18: 349–451. Jackson AT, Yates HT, Scrivens JH, Green MR, and Bateman RH (1998) Matrix-assisted laser desorption/ionization-collision induced dissociation of poly(styrene). Journal of the American Society for Mass Spectrometry 9: 269–274.
Merchant M and Weinberger SR (2000) Recent advancements in surface-enhanced laser desorption/ionizationtime of flight-mass spectrometry. Electrophoresis 21: 1164–1167. Montaudo G and Lattimer RP (eds.) (2002) Mass Spectrometry of Polymers. Boca Raton, FL: CRC Press. Parees DM, Hanton SD, Cornelio Clark PA, and Willcox DA (1998) Comparison of mass spectrometric techniques for generating molecular weight information on a class of ethoxylated oligomers. Journal of the American Society for Mass Spectrometry 9: 282–291. Tang N, Tornatore P, and Weinberger SR (2004) Current developments in SELDI affinity technology. Mass Spectrometry Reviews 23: 24–44. Zhu H, Yalcin T, and Liang L (1998) Analysis of the accuracy of determining average molecular weights of narrow polydispersity polymers by matrix-assisted laser desorption ionization time-of-flight mass sepctrometry. Journal of the American Society for Mass Spectrometry 9: 275–281.
Mass Separation K J Welham, University of Hull, Hull, UK & 2005, Elsevier Ltd. All Rights Reserved.
Introduction A mass spectrometer can be broken down into four basic components: sample introduction, ionization, mass separation, and detection/recording. This article will concentrate on mass separation, or ‘analysis’, although the components should not be treated in complete isolation as there are interactions between them. The modes of mass analysis that will be considered here can be classified as follows: *
*
*
Continuous mode: * magnetic sectors, possibly with electric sectors * quadrupole mass filters Pulsed mode: * time-of-flight (TOF) analyzers Ion trapping devices: * quadrupole traps * Fourier-transform ion cyclotron resonance traps.
These various methods differ widely in cost, size, and complexity, and also in terms of performance (i.e., mass/charge range and resolving power). The
preferred method of mass analysis depends on the type of problem being investigated.
Types of Mass Analyzer Magnetic Sectors
Magnetic sector instruments are relatively expensive but they generally give high performance when combined with electric sectors, especially in terms of resolution, mass measurement accuracy, sensitivity, and m/z range. They employ high accelerating potentials (several kilovolts) in comparison with quadrupole instruments. An ion of mass m and charge ze accelerated out of an ion source through potential V will acquire a velocity v, and under the influence of a magnetic field of intensity B will follow a circular path of radius r. The following equations apply: Ion kinetic energy: mv2 =2 ¼ zeV
½1
Deflecting force: BzeV ¼ centrifugal force ¼ mv2 =r Combining these: m=z ¼ B2 r2 e=2V
½2 ½3
The number of charges, z, is usually equal to one, so eqn [3] shows that a spectrum of masses can be obtained by changing one of the three variables, B, r,
