Analytical Biochemistry 301, 49–56 (2002)doi:10.1006/abio.2001.5478, available online at http://www.idealibrary.com on

High-Throughput Protein Characterization Using MassSpectrometric Immunoassay

Urban A. Kiernan,*,† Kemmons A. Tubbs,* Karl Gruber,* Dobrin Nedelkov,*Eric E. Niederkofler,*,† Peter Williams,† and Randall W. Nelson*,1

*Intrinsic Bioprobes, Inc., 625 S. Smith Road Suite, 22, Tempe, Arizona 85281; and †Department of Chemistryand Biochemistry, Arizona State University, Tempe, Arizona 85287-1604

Received May 29, 2001; published online December 10, 2001

A high-throughput mass spectrometric immunoas-say system for the analysis of proteins directly fromplasma is reported. A 96-well format robotic worksta-tion was used to prepare antibody-derivatized affinitypipette tips for subsequent use in the extraction ofspecific proteins from plasma and deposition onto 96-well format matrix-assisted laser desorption/ioniza-tion time-of-flight mass spectrometry (MALDI-TOFMS) targets. Samples from multiple individuals werescreened with regard to the plasma protein transthy-retin (TTR), followed by analysis of the same plasmasamples for the transthyretin-associated transportprotein, retinol-binding protein (RBP). Analyses wereable to detect the presence of posttranslationally mod-ified TTR and RBP, as well as a mutation present inthe TTR of one individual. Subsequent analyses ofwild-type and mutated TTR using enzymatically ac-tive MALDI-TOF MS targets were able to identify thesite and nature of the point mutation. The approachrepresents a rapid (;100 samples/2 h, reagent prepa-ration-to-data) and accurate means of characterizingspecific proteins present in large numbers of individ-uals for proteomic and clinical/diagnostic purposes.© 2001 Elsevier Science

Key Words: MALDI-TOF; mass spectrometry; high-throughput; affinity; immunoassay; protein.

Present day proteomics relies heavily on the massspectrometric identification of proteins separated us-ing two-dimensional polyacrylamide gel electrophore-sis (2D-PAGE)2 or tandem high performance liquid

1 To whom correspondence should be addressed. Fax: (480) 804-0778. E-mail: rnelson@intrinsicbio.com.

2

Abbreviations used: 2D-PAGE, two-dimensional polyacrylamidegel electrophoresis; MALDI-TOF MS, matrix-assisted laser desorp-

0003-2697/01 $35.00© 2001 Elsevier ScienceAll rights reserved.

chromatography (LC/LC) approaches (1–4). Such anal-yses are impressive in their ability to simultaneouslyinterrogate hundreds to thousands of proteins presentin complex biological media. When used in control ver-sus disease-state studies, the approaches are able toidentify proteins participating in, or serving as mark-ers for, disease states. The identification process servesas the starting point for subsequent studies gearedtoward, among other things, the more detailed struc-tural characterization of proteins. Due to the numer-ous genetic, transcriptional, and posttranslationalvariations possible for any given protein, it is essentialthat such follow-up studies investigate large numbersof samples and are able to characterize the structure ofa protein to the fullest degree possible.

Once a protein is identified and deemed relevant tofurther study, it is possible to replace 2D PAGE andLC/LC with affinity selection in preparation for directmass spectrometric characterization. In this manner,proteins of interest are extracted from the biologicalmedium to the exclusion of the rest of the complexmixture. Ideally, the extraction process is free of hin-drances from nonspecific binding of other proteins/com-pounds and exhibits an overall concentrating effecttoward the targeted protein. Indeed, numerous pastreports have demonstrated the potential of using affin-ity isolation in combination with mass spectrometry forthe analysis of proteins (5–18). Many of these reportsfocus on the use of matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS), which is used universally in the identifica-tion of affinity-isolated proteins (11, 12), the

tion/ionization time-of-flight mass spectrometry; MSIA, mass spec-trometric immunoassay; b2m, b-2-microglobulin; TTR, transthyre-tin; RBP, retinol binding protein; GA, glutaraldehyde; TFA,

trifluoroacetic acid; wt, wild type; MW, molecular weight.

49

50 KIERNAN ET AL.

investigation of posttranslational modifications (13),the differential display of proteins (14), and the quan-tification of proteins retrieved directly from biologicalmedia (8, 15–18).

Because of its directness in sample preparation, af-finity-isolation mass spectrometry represents an at-tractive means for “second phase” proteomics, wherethe same protein is repetitively under investigation.Such analyses have great potential in validation orprotein phenotyping studies where large populationsare under investigation and in more far-reaching ap-plication such as diagnostics. However, in order tomake full use of the speed of the approach, it is neces-sary to combine the affinity isolation/sample prepara-tion with robotics. In this manner, high-throughputplatforms (e.g., 96-well format titer plates) can be usedto address multiple samples in parallel. Such use ofrobotics in parallel processing requires the develop-ment of specialized affinity isolation devices and asso-ciated methods tailored specifically for the MALDI-TOF MS analysis. Moreover, to reach an additionallevel of analytical potential, it is often necessary toenzymatically digest small amounts of isolated pro-teins in order to gain a more detailed mass spectromet-ric characterization. Such processing requires yetanother set of specialized methods and devices forhigh-throughput application.

Recently, we have described prototype affinity pi-pette tips used in the mass spectrometric immunoas-say (MSIA) (17) of b-2-microglobulin (b2m) present invarious human biofluids (15). The affinity pipette tipswere found to perform exceptionally well in the quan-tification of b2m present in urine and overall repre-sented a vast improvement over rudimentary affinitypipette tips described previously (18). Additional im-provements upon the prototype design have yieldedaffinity pipette tips that performed equally well inother biofluids, in particular plasma (16). Parallel tothe development of affinity pipette tips has been thedevelopment of enzymatically active MALDI-TOF tar-gets for use in characterizing proteins postisolation(19–21). These activated targets serve as enzymaticmodifiers of proteins, as well as platforms for introduc-tion of samples into the mass spectrometer. Formattedto 96-well platforms, both the affinity pipette tips andthe enzymatically active targets serve as universal de-vices that can be coupled to robotics for the processingof numerous samples in parallel (see Fig. 1).

Presented here are high-throughput protein pheno-typing studies investigating the use of a parallel ro-botic system equipped with affinity pipette tips forpreparing transthyretin (TTR) and retinol binding pro-tein (RBP) directly from human plasma for MALDI-TOF MS characterization. Moreover, when needed, en-zymatically active targets were included into the

analyses for a more detailed structural analysis of the

targeted proteins. Of particular interest in the studywere: (1) the construction of affinity pipette tips exhib-iting low nonspecific binding properties when appliedto plasma proteins, (2) the development of high-sensi-tivity MSIA protocols that allowed minimally invasivesampling protocols, and (3) the identification of post-translational modifications and point mutationspresent in TTR and RBP.

MATERIALS AND METHODS

Affinity Pipette Tip Preparation

Affinity pipette tips were prepared following one ofthree methods. All three methods utilized porousamine frits described previously (15). In brief, the fritswere produced in bulk by loading soda lime glass beadsinto stainless steel annealing molds and baked to forma solid, yet porous frit. The frits were then removedand acid conditioned prior to a 12-h treatment with10% aminopropyl triethoxysilane. Once treated, thefrits are left with a functional amine surface that canbe used for coupling of ligands. In the first method,amine frits were activated by 19-h exposure to glutar-aldehyde (GA) (25% solution in 0.2 M phosphate buffer,pH 7.5) followed by rinsing (3 3 50 mL) with phosphatebuffer. The activated frits were then incubated withrabbit anti-TTR polyclonal antibody (Dako, Carpinte-ria, CA; Catalog No. A0002), diluted fivefold in HBSbuffer (0.01 Hepes, pH 7.4, 0.15 M NaCl, 0.005% (v/v)

FIG. 1. High-throughput robotics platform for automated samplepreparation and MSIA analysis. Analytical samples are addressed inparallel using a 96-well robotic workstation fitted with affinity pi-pette tips (1). A six-stage work area for automated processing en-ables sample incubation through the tips (repetitive aspiration anddispensing, stage 2), rinsing (stages 3 and 4), and MALDI matrixaspiration (stage 5). The final step of the high-throughput MSIA isthe elution of the matrix/analyte mix from the affinity pipette tipsdirectly onto a 96-well format hydrophobic/hydrophilic contrastingMALDI-TOF MS target (stage 6).

polysorbate 20, 3 mM EDTA), for 19 h at 4°C. Follow-

51HIGH-THROUGHPUT PROTEIN ANALYSIS

ing the antibody linkage, the frits were rinsed vigor-ously with HBS and packed while still wet into wide-bore P-200 pipette tips and stored in HBS buffer. In thesecond method, frits were activated with GA as in thefirst method and incubated with 40-kDa polylysine (5.0mg/mL in phosphate buffer) for 19 h with gentle agi-tation. The frits were rinsed vigorously with phosphatebuffer and then activated with GA, derivatized withantibody, and packed into pipette tips using the proce-dures described for the first method. Using the thirdmethod, amine frits were packed into the pipette tip forsubsequent in-robotic (Beckman, Multimek 96, Fuller-ton, CA) amplification and antibody linkage. The tipswere first amplified by repetitively flowing GA (25%solution in HBS, pH 7.5) through the tips (100 mL, 120repetitions), followed by rinsing with HBS (100 mL, 20repetitions), exposure to polylysine (0.1 mg/mL in HBS,100 mL, 120 repetitions), and rinsing (HBS, 100 mL, 20repetitions). The amplified tips were then activatedwith GA (25% solution in HBS, pH 7.5, 100 mL, 120repetitions), exposed to anti-TTR antibody (0.05mg/mL in HBS, 20 mL, 120 repetitions), and rinsedwith HBS (100 mL, 20 repetitions). The same in-roboticprocedure was used for the preparation of anti-RBPtips using rabbit anti-RPB polyclonal IgG (Dako, Cat-alog No. A0054). The total time needed for in-roboticderivatization was approximately 50 min.

Sample Preparation

Whole blood samples were collected from six healthymale individuals. Samples (44.7 mL) were acquiredunder sterile conditions through a lancet-puncturedfinger using a heparinized microcolumn (DrummondScientific Co., Broomall, PA). The whole blood wasimmediately combined with 200 mL HBS and centri-fuged for 1 min (at 7000 RPM, 2500g) to pellet redblood cells. The supernatant (plasma/HBS) from eachsample was distributed in 25-mL aliquots randomlyamong the 96 wells of a titer plate, each well alreadycontaining 100 mL of HBS.

The plasma/HBS samples were addressed in parallelusing the 96-well format pipetting workstationequipped with either anti-TTR or anti-RBP affinitypipette tips. Sample incubation consisted of 20 cycles(aspiration and dispensing) of 100 mL of each samplethrough the affinity pipette tips. After incubation, tipswere rinsed using HBS (5 cycles, 100 mL), acetonitrile(3 cycles, 100 mL), and doubly distilled water (10 cycles,100 mL). Retained species were eluted by drawing 4 mLof MALDI matrix solution (saturated aqueous solutionof a-cyano-4-hydroxycinnamic acid, in 33% (v/v) aceto-nitrile, 0.2% (v/v) trifluoroacetic acid) into the tips anddepositing it directly onto a 96-well formatted MALDI-TOF target that had been treated as previously de-

scribed (16). Samples were allowed to air dry prior to

target insertion into the mass spectrometer. The totaltime required for preparation of the 96 samples wasapproximately 10 min.

TTR Mapping

TTR was isolated from plasma samples using theprotocols listed above. TTR was eluted from the tipsusing an acid solution (0.2% TFA, containing 1 mMn-octylglucoside) in place of the MALDI matrix solu-tion and applied directly to the trypsin-activatedMALDI-TOF MS target prepared as previously de-scribed (21–23). Digests were allowed to proceed for 15min at 40°C in high humidity enclave before termina-tion by the addition of 2 mL of matrix solution. Sampleswere air dried prior to insertion of the MALDI-TOFtarget into the mass spectrometer.

Mass Spectrometry

MS analysis was performed on a Bruker Biflex IIIMALDI-TOF mass spectrometer operating in lineardelayed-extraction mode with 19.00 kV full accelerat-ing potential. Draw-out pulses of 1.700 (300-ns delay)and 1.600 kV (300-ns delay) were used for TTR (singlycharged) and RBP (doubly charged), respectively. Massspectra were acquired automatically by summing ten10 laser shot spectra while gauging spectra qualityusing fuzzy logic routines (24). Mass maps (100 lasershots) were acquired in reflectron mode using an in-strument setting of full accelerating potential of 19.00kV, an ion-mirror voltage of 20.00 kV, and a draw-outpulse voltage of 1.65 kV (300-ns delay).

RESULTS AND DISCUSSION

Transthyretin is a small protein produced in theliver and found in serum and cerebral spinal fluid as ahomotetramer (25, 26). Functionally, TTR serves un-accompanied in the transport of thyroid hormones or incomplexes with other proteins in the transport of var-ious biologically active compounds. Structurally, wild-type (wt) TTR is composed of 127 amino acids and hasa molecular weight (MW) of 13,762.4. Over 80 pointmutations have been cataloged for TTR, with all but 10potentially leading to severe neurological complica-tions (26). The majority of mutation-related disordersare caused by amyloid plaques depositing on neuronsor tissue, eventually leading to dysfunctions includingcarpal tunnel syndrome and familial amyloid polyneu-ropathy (27–29). Retinol binding protein is composed of182 amino acids (MW(wt) 5 21,065.6) and is producedprimarily in the liver (30, 31). RBP, in a 1:1 complexwith TTR (one RBP molecule bound to one TTR tet-ramer) (32, 33), is a specific carrier of retinol (vitaminA) from the liver stores to the peripheral tissues (34).

Several variants/mutations have been found to disrupt

52 KIERNAN ET AL.

the TTR:RBP complex, which in turn leads to dimin-ished transport of vitamin A. These variants includepoint mutations in both TTR (Ile84Ser) (35) and RBP(Ile41Asn and Gly74Asp in different alleles in a singleindividual) (36) and posttranslationally truncated ver-sions of RPB (two variants: one missing the C-terminalleucine (RBP-C-Leu) and the second missing two C-terminal leucines (RBP-C-LeuLeu) (37).

Affinity Pipette Tip Optimization

Initial studies were performed to optimize the affin-ity pipette tips for the direct extraction of TTR andRBP from plasma. Based on past findings (16), amine-functionilized pipette tips were chosen as base chem-istries from which additional derivatization schemeswere applied and tested. Figure 2 shows a comparisonof methods for preparing the anti-TTR affinity pipettetips. Figure 2A is a mass spectrum of diluted humanplasma, which is given to illustrate the need to extractTTR and RBP from plasma in preparation for MALDI-TOF MS characterization. Signals from neither TTRnor RBP are observed in the spectrum. Figure 2B is aspectrum obtained using an affinity pipette tip pre-pared by glutaraldehyde activation of amine frits, fol-

FIG. 2. Affinity pipettes optimization. (A) Mass spectrum of dilutetip prepared via bulk derivatization by glutaraldehyde (GA) activatiand loading of the affinity frits into pipette tips. (C) Mass spectrumby GA activation of amine frits and linkage of 40-kDa polylysine, foland loading of the affinity frits into the pipette tips. (D) Mass sderivatization by GA activation of amine frits and linkage of 40-kDaanti-TTR antibody. (E) Mass spectrum obtained from an affinity pianti-RBP antibody.

lowed by exposure of the frits to anti-TTR antibody.

Both the activation and the antibody linkage stepswere performed in bulk prior to loading of the affinityfrits into pipette tips. TTR signals are observed in thespectrum; however, in the presence of other signalsattributed to the nonspecific binding of apolipoproteinsC-I, C-I9 and C-III. Figure 2C shows a spectrum ob-tained using an affinity pipette tip prepared by GAactivation of amine frits and linkage of 40-kDa polyly-sine, followed by GA activation of the polylysine andlinkage of anti-TTR antibody. All activation and cou-pling steps were performed in bulk prior to loading ofthe affinity frits into the pipette tips. The spectrumdemonstrates a reduction of nonspecific binding and ahigh-quality signal profile for the TTR. Figure 2Dshows a spectrum obtained from an affinity pipette tipto which polylysine-intermediate and antibody werecoupled using a robotic protocol. In this method, aminefrits were prepared in bulk and loaded into pipette tipsand the fully assembled tips were then loaded into a96-well format robotic workstation. Polylysine inter-mediate was coupled to the tips (via GA activation)followed by antibody linkage (through GA activationand exposure to anti-TTR antibody). The methodshows the lowest nonspecific binding while still main-

uman plasma. (B) Mass spectrum obtained using an affinity pipettef amine frits, followed by exposure of the frits to anti-TTR antibodyained using an affinity pipette tip prepared via bulk derivatizationed by GA activation of the polylysine, linkage of anti-TTR antibodytrum obtained using an affinity pipette tip prepared via roboticlylysine, followed by GA activation of the polylysine and linkage ofe tip prepared in the same manner as in (D), except for the use of

d hon oobtlowpecpo

pett

taining high-quality TTR signals. Improved results

two

53HIGH-THROUGHPUT PROTEIN ANALYSIS

were most likely due to increased functionalizationwithin the active microflow channels of the frits. Bulkpreparation has a possibility of having dead space vol-ume, leaving large areas unblocked or without Ab cou-pling, which may lead to higher nonspecifically boundcompounds. Figure 2E shows a spectrum obtained froman affinity pipette tip prepared in the same manner,

FIG. 3. High-throughput MSIA analysis of transthyretin and retirandomly arranged in a 96-well titer plate. (A) Mass spectra resultiShown is the region of the singly charged TTR signals. Highlightedother. (B) Mass spectra resulting from the MSIA analysis of the sameof the doubly charged RBP signals. The highlighted cells represent

except for the use of anti-RBP antibody. As in Figure

2D, the target protein was sufficiently extracted fromthe plasma sample with minimal binding of nonspecificcompounds. As previously mentioned, RBP forms com-plex with TTR in vivo and it would be expected thatsome of the RBP affinity-retrieved in the MSIA tipswould be in the form of an RBP–TTR complex. How-ever, it is presumed that the acetonitrile wash per-

binding protein using human plasma samples from six individualsfrom the MSIA analysis utilizing anti-TTR derivatized pipette tips.s show spectra with resolvable parent ion differences between eachples utilizing anti-RBP derivatized pipette tips. Shown is the regionspectra with resolvable parent ion differences.

nolngcellsam

formed following the affinity retrieval disrupts the

54 KIERNAN ET AL.

complex, as a result of which TTR signals are notobserved in the resulting mass spectrum. In all, thein-robotic tips derivatization method was chosen forthe preparation of affinity pipette tips for use in sub-sequent TTR and RBP screening experiments.

Screening

Plasma/HBS samples from six individuals were ali-quotted into wells of a 96-well titer plate, screened inparallel using anti-TTR affinity pipette tips, and datawere acquired using automated MALDI-TOF MS. Fig-ure 3A shows the results of the screening. The spectrashow a generally high reproducibility, with wild-typeTTR detected in all samples. Mass accuracy for wt-TTRusing external calibration (96 samples: MW(range) 513,745–13,787 (;0.11% variation)) was reasonable,considering the dimensions of the titer plate format ofthe MALDI target. The same plasma/HBS sampleswere next screened using anti-RBP affinity pipettetips. Due to resolution limitations at higher mass, dou-bly charged RBP was used for the data analyses. Fig-ure 3B shows the results of the screening. The auto-mated preparation routine was sufficient in producingsamples from which adequate ion signals were ob-tained with mass accuracies (externally calibrated)over the surface of the 96-well format target similar tothose observed for TTR (MW(range) 5 21,040–21,094,calculated from doubly charged (;0.14% variation)).

The resolution of the data (m/dm ;800) was suffi-cient to resolve a second TTR signal in all samples (dueto cysteinylation of Cys10; Dm 5 119 Da; (38, 39)) anda second set of satellite signals at ;30 Da higher inmolecular weight than wt-TTR and cys-TTR in 14 rep-licate analyses of the plasma sample taken from oneindividual (Fig. 4A). Such splitting of signals is a gen-eral indicator of a point mutation present within theprotein and suggest that one individual in the studywas of heterozygous phenotype, with a genetic poly-morphism resulting in a TTR variant shifted in massby ;130 Da. Regarding RBP, the resolution and qual-ity of the data were sufficient to reproducibly recognizemoderate levels of RBP-Leu (the truncated RBP form)in all but one of the individuals (individual 4; 15 rep-licate analyses, Fig. 4B). Currently, there is no cleardefinition of the RBP-to-RBP-Leu ratio that is consid-ered normal in humans; however, recent results sug-gest that an overabundance of RBP-Leu (exceeding theconcentration of RBP) is associated with chronic renalfailure (37). Quantitative assays for both TTR and RBPare currently under development.

TTR Mapping

Other groups have previously demonstrated the use

of enzymatic mass mapping in combination with

MALDI-TOF MS for the analysis of point mutationspresent in TTR (40, 41). Of particular interest is thework of Theberge et al., who used enzymatically activeMALDI-TOF MS targets in mapping TTR for the pres-ence of mutations (40). In their work, trypsin-activatedgold targets were manufactured as previously de-scribed (21–23) and applied in the rapid mass mappingof TTR immunoprecipitated from serum/plasma sam-ples. The approach was shown able to detect the pres-ence of a Val30Met mutation, which is known to be thecause of familial amyloidic polyneuropathy, the mostcommon form of amyloidosis (29). Moreover, and fittingwith the work presented here, the authors commentedon the suitability of the enzymatic-target approach tohigh-throughput point mutation analyses. Followingsuit, TTR from the six individuals participating in thisstudy were mapped in parallel using a trypsin-active

FIG. 4. Phenotypic differences in RBP and TTR within the studygroup. (A) Mass spectra resulting from the MSIA analysis of twosamples showing the presence of wild-type and cysteinylated formsof TTR (wt-TTR and cys-TTR) in both samples and the existence of apoint mutation in one of the individuals (bottom trace), indicated bythe splitting of the two TTR signals. (B) Mass spectra resulting fromthe MSIA analysis of two samples showing the presence of onlynative RBP in one individual (upper trace) and both native RBP andRBP-Leu in a second individual (bottom trace). Shown is the regionof the doubly charged RBP signals.

MALDI-TOF MS target.

55HIGH-THROUGHPUT PROTEIN ANALYSIS

Trypsin digests of TTR from the six individuals wereperformed in parallel by eluting the TTR with a low-pHsolution directly from the affinity pipette tips ontotrypsin-activated sites on a bioreactive probe. Follow-ing a previously defined protocol (22), the trypsin-ac-tive sites were precoated with Tris–HCl, which actedas a compensating buffer, raising the pH of the eluent(pH ;2.5) to that of the working range of trypsin (pH;7–8). The digests were allowed to proceed for 15 minin a humidified environment (maintained at 37°C) be-fore termination by the addition of an acidified MALDImatrix. Figures 5A and 5B (upper traces) are spectrarepresentative of the mass maps obtained for five of thesix individuals. The map yielded ;80% TTR sequencecoverage (Fig. 5A, inset), which is sufficient to analyzeall but 4 of the 80 known mutations (26). Moreover, themapping data are consistent with the presence of onlywt-TTR. Mass maps of the presumably mutated sam-ple demonstrated the same sequence coverage as ob-served for wt-TTR. However, an additional signal as-

FIG. 5. Tryptic mapping of MSIA-isolated TTR via utilization ofbioreactive probes. (A) Mass spectra showing two tryptic maps ob-tained after MSIA analysis of the two samples shown in Fig. 4A. Theinset illustrates a coverage map of the TTR sequence with the ob-served peptide fragments. (B) Expanded regions of the two massspectra showing the exact mass difference between the wt and mu-tant peptide fragments.

sociated with the T12 fragment (at Dm 5 129.988 Da)

was present in the mass spectrum (Figs. 5A and 5B,lower traces). There are two possible (known) pointmutations in this tryptic fragment that would result ina nominal 130-Da mass shift. Of these, Thr119Met(Dm 5 29.992 Da) agrees most closely with the ob-served mass difference (versus Dm 5 30.011 Da forthe other possible mutation; Ala109Thr). TheThr119Met variant has been characterized previouslyand is referred to as TTR “Chicago” variant, which isfound to be nonamyloidogenic (42, 43).

CONCLUSION

The use of affinity isolation in combination withMALDI-TOF MS holds much promise for the detectionof protein variants stemming from genetic, transcrip-tional, or posttranslational affects. However, in orderto reach its fullest potential, the approach must bescaled for the analysis of significantly large numbers ofsamples, thereby allowing correlations that are able tostatistically differentiate between nondetrimental pro-tein variations and those linked to disease/ailment. Wehave reported here progress in the development of ahigh-throughput system, composed of parallel roboticsfitted with specialized affinity extraction and sampleprocessing devices and automated MALDI-TOF MS,that is capable of assaying specific proteins present incomplex biological media. By addressing multiple sam-ples robotically and in parallel, the time required forreagent and sample preparation is on the order of ;100samples/h. Data acquisition using automated MALDI-TOF MS is equally rapid and when performed on na-tive proteins is able to provide a “first glance” into thenature of protein variants found in individuals. Whenwarranted, more detailed structural analyses are pos-sible by digesting proteins enzymatically and analyz-ing the results using high-performance MALDI-TOFMS. Given the ability to rapidly perform such analyses,the high-throughput approach demonstrated here islikely to find much use in proteomics/clinical studieswhere the exceptional structural characterization ofspecified proteins present in large populations is re-quired.

ACKNOWLEDGMENTS

This publication was supported in part by Grants 2 R44 GM56603-01, 2 R44 GM56580-0, and R01 GM55872-02 from the NationalInstitutes of Health. Its contents are solely the responsibility of theauthors and do not necessarily represent the official views of theNational Institutes of Health.

REFERENCES

1. Jensen, O. N., Larsen, M. R., and Roepstorff, P. (1998) ProteinsSuppl., 74–89.

2. Dunn, M. J. (1997) Biochem. Soc. Trans. 25, 248–254.

56 KIERNAN ET AL.

3. Patterson, S. D., and Aebersold, R. (1995) Electrophoresis 16,1791–1814.

4. Yates, J. R., 3rd (1998) J. Mass. Spectrom. 33, 1–19.5. Hutchens, T. W., and Yip, T. T. (1993) Rapid Commun. Mass

Spectrom. 7, 576–580.6. Papac, D. I., Hoyes, J., and Tomer, K. B. (1994) Anal. Chem. 66,

2609–2613.7. Brockman, A. H., and Orlando, R. (1995) Anal. Chem. 67, 4581–

4585.8. Wang, R., Sweeney, D., Gandy, S. E., and Sisodia, S. S. (1996)

J. Biol. Chem. 271, 31894–31902.9. Krone, J. R., Nelson, R. W., Dogruel, D., Williams, P., and Gran-

zow, R. (1997) Anal. Biochem. 244, 124–132.10. Rudiger, A. H., Rudiger, M., Carl, U. D., Chakraborty, T., Roep-

storff, P., and Wehland, J. (1999) Anal. Biochem. 275, 162–170.11. Posewitz, M. C., and Tempst, P. (1999) Anal. Chem. 71, 2883–

2892.12. Nelson, R. W., Jarvik, J. W., Taillon, B. E., and Tubbs, K. A.

(1999) Anal. Chem. 71, 2858–2865.13. Nedelkov, D., and Nelson, R. W. (2000) J. Mol. Recognit. 13,

140–145.14. Merchant, M., and Weinberger, S. R. (2000) Electrophoresis 21,

1164–1177.15. Tubbs, K. A., Nedelkov, D., and Nelson, R. W. (2000) Anal.

Biochem. 289, 26–35.16. Niederkofler, E. E., Tubbs, K. A., Gruber, K., Nedelkov, D.,

Kiernan, U. A., Williams, P., and Nelson, R. W. (2001) Anal.Chem. 73, 3294–3299.

17. Nelson, R. W., Krone, J. R., Bieber, A. L., and Williams, P. (1995)Anal. Chem. 67, 1153–1158.

18. Krone, J. R., Nelson, R. W., and Williams, P. (1996) SPIE 2680,415–421.

19. Dogruel, D., Williams, P., and Nelson, R. W. (1995) Anal. Chem.67, 4343–4348.

20. Nelson, R. W., Dogruel, D., Krone, J. R., and Williams, P. (1995)Rapid Commun. Mass Spectrom. 9, 1380–1385.

21. Nelson, R. W. (1997) Mass Spectrom. Rev. 16, 353–376.22. Nelson, R. W. (2000) in Polymorphism Detection and Analysis

(J. D. Burczak and E. Mardis, Eds.), pp. 393–396, Eaton Pub-

lishing, Natick.

23. Lewis, J. K., Krone, J. R., and Nelson, R. W. (1998) Biotechniques24, 102, 104, 106, 108 passim.

24. Jensen, O. N., Mortensen, P., Vorm, O., and Mann, M. (1997)Anal. Chem. 69, 1706–1714.

25. Ingenbleek, Y., and Young, V. (1994) Annu. Rev. Nutr. 14, 495–533.

26. Schreiber, G., and Richardson, S. J. (1997) Comp. Biochem.Physiol. B Biochem. Mol. Biol. 116, 137–160.

27. Damas, A. M., and Saraiva, M. J. (2000) J. Struct. Biol. 130,290–299.

28. Plante-Bordeneuve, V., and Said, G. (2000) Curr. Opin. Neurol.13, 569–573.

29. Benson, M. D., and Uemichi, T. (1996) Amyloid-Int. J. Exp. Clin.Invest. 3, 44–56.

30. Goodman, D. S. (1980) Ann. N. Y. Acad. Sci. 348, 378–390.31. Blaner, W. S. (1989) Endocr. Rev. 10, 308–316.32. Monaco, H. L. (2000) Biochim. Biophys. Acta 1482, 65–72.33. Naylor, H. M., and Newcomer, M. E. (1999) Biochemistry 38,

2647–2653.34. Kanai, M., Raz, A., and Goodman, D. S. (1968) J. Clin. Invest. 47,

2025–2044.35. Waits, R. P., Yamada, T., Uemichi, T., and Benson, M. D. (1995)

Clin. Chem. 41, 1288–1291.36. Biesalski, H. K., Frank, J., Beck, S. C., Heinrich, F., Illek, B.,

Reifen, R., Gollnick, H., Seeliger, M. W., Wissinger, B., andZrenner, E. (1999) Am. J. Clin. Nutr. 69, 931–936.

37. Jaconi, S., Rose, K., Hughes, G. J., Saurat, J. H., and Siegent-haler, G. (1995) J. Lipid Res. 36, 1247–1253.

38. Terazaki, H., Ando, Y., Suhr, O., Ohlsson, P. I., Obayashi, K.,Yamashita, T., Yoshimatsu, S., Suga, M., Uchino, M., and Ando,M. (1998) Biochem. Biophys. Res. Commun. 249, 26–30.

39. Theberge, R., Connors, L., Skinner, M., Skare, J., and Costello,C. E. (1999) Anal. Chem. 71, 452–459.

40. Theberge, R., Connors, L. H., Skinner, M., and Costello, C. E.(2000) J. Am. Soc. Mass Spectrom. 11, 172–175.

41. Bergquist, J., Andersen, O., and Westman, A. (2000) Clin. Chem.46, 1293–1300.

42. Harrison, H. H., Gordon, E. D., Nichols, W. C., and Benson, M. D.(1991) Am. J. Med. Genet. 39, 442–452.

43. Scrimshaw, B. J., Fellowes, A. P., Palmer, B. N., Croxson, M. S.,

Stockigt, J. R., and George, P. M. (1992) Thyroid 2, 21–26.