Mass Spectrometry of Self-Assembled Monolayers: a New Tool for Molecular Surface Science

REVIEW Mass Spectrometry of Self-Assembled Monolayers: A New Tool for Molecular Surface Science

Milan Mrksich*

Department of Chemistry and Howard Hughes Medical Institute, The University of Chicago, 929 East 57th Street, Chicago, Illinois 60521

he scientific and engineering disci- T plines have developed sophisti- ABSTRACT Most reactions can be performed in solution and on a surface, yet the challenges faced in applying cated toolboxes for making mat- known reactions or in developing entirely new reactions for modifying surfaces remain formidable. The products ter—including examples from organic of many reactions performed in solution can be characterized in minutes, and even products having complex chemistry to create small molecules with structures can be characterized in hours. When performed on surfaces, even the most basic reactions require a virtually unlimited complexity, molecular biology to prepare macromolecules with a substantial effort—requiring several weeks—to characterize the yields and structures of the products. This precise arrangement of amino acids and a contrast stems from the lack of convenient analytical tools that provide rapid information on the structures of defined tertiary structure, and lithographic molecules attached to a surface. This review describes recent work that has established mass spectrometry as a methods to fashion inorganic and metallic powerful method for developing and characterizing a broad range of chemical reactions of molecules attached to materials into integrated circuits and de- self-assembled monolayers of alkanethiolates on gold. The SAMDI-TOF mass spectrometry technique will enable a vices. Each of these approaches is based on next generation of applications of molecularly defined surfaces to problems in chemistry and biology. distinct methods for assembling matter, and each relies on analytical methods that KEYWORDS: biochip · interfacial reactions · label-free · self-assembled can characterize the structures that they monolayer · SAMDI-TOF mass spectrometry generate. In the bottom-up approaches common to synthetic chemistry, a range of methods including NMR and X-ray diffraction are used to delineate the connectivity layer membranes.3 This review deals with of atoms within molecules. The biopolymers the analogous challenge of characterizing created using enzyme-mediated conver- the molecular structures of organic surfaces. sions are characterized using electro- The development of self-assembled mono- phoretic methods and high-resolution mass layers (SAMs) has, in principle, made it spectrometry, while integrated circuits are straightforward to functionalize surfaces characterized with electron microscopy, with a broad range of molecular composi- Downloaded via NORTHWESTERN UNIV on June 24, 2020 at 21:00:34 (UTC). scanning probe tools, and sensitive spec- tions to enable applications in biology and troscopies. The tools remain fundamental chemistry.4 Yet, the common methods used to each of these areas, enabling further ad-

See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles. in surface analysis do not have the resolu- vances in the preparation of their respective tion and sensitivity required to analyze targets. complex molecular compositions and, in The pursuit of structures with at least turn, have limited efforts to pursue the de- one dimension sized between 1 and 100 sign and application of surfaces to these ar- nm—which underlies the nanoscience and eas. In this review, the recent development technology discipline that has grown over of mass spectrometric methods that are en- the past decade—is likewise dependent on abling applications of surface science in analytical methods, and in several cases it chemistry and biology is described. The re- *Address correspondence to has been hampered by a lack of methods view begins with an overview of the current [email protected]. 1 having appropriate performance. For ex- methods that are broadly available, a de- Received for review December 10, 2007 ample, it remains difﬁcult to characterize scription of recent work that expands the and accepted December 17, 2007. the products resulting from treatment of scope of surfaces that can be analyzed, and Published online January 22, 2008. proteins or carbon nanotubes with organic provides examples that illustrate the im- 10.1021/nn7004156 CCC: $40.75 reagents,2 and it is still difﬁcult to character- pact that these methods are having in the ize the dynamic structures of lipid rafts in bi- biological and chemical sciences. © 2008 American Chemical Society

www.acsnano.org VOL. 2 ▪ NO. 1 ▪ 7–18 ▪ 2008 7 ANALYTICAL METHODS IN SURFACE SCIENCE molecules in a monolayer.9 Yet, the skills and facili- Atomic Composition—The UHV Toolbox. Several surface ties required to perform these measurements limit analytical techniques have been developed and now their use to a relatively small number of aﬁcionados. provide a remarkable set of tools for the analysis of me- Atomic and lateral force microscopies that use a tallic and inorganic surfaces; these were recently high- scanning probe to image the forces with a substrate lighted with the award of the 2007 Nobel Prize in Chem- are now widely available and provide straightfor- istry to Gerhard Ertl. These methods offer exceptional ward characterization of the shapes, sizes, and sensitivity in characterizing the atomic composition and chemical properties of patterned nanostructures.10 structure of clean substrates and in turn have been criti- The extension of these methods with scanning

REVIEW cal to developing materials and probes that are functionalized with chemical or bio- processes in semiconductor fab- logical molecules are enabling unprecedented stud- VOCABULARY: mass spectrometry –an rication—in part for their high ies of the properties of individual proteins with new analytical technique that measures the mass- sensitivity in measuring the insights into the roles of force in biology.11,12 Clearly, to-charge ratio of ions; a means of detecting densities of impurities and these methods have provided powerful tools for molecules according to their molecular weight dopants—and to understand- characterizing the shapes, mechanics, and proper- • self-assembled monolayer – a nanoscale ing the mechanisms of hetero- ties of nanoscale structures, but they do not provide two-dimensional structure consisting of a geneous catalytic processes.5,6 general routes for the molecular characterization of single layer of molecules on a substrate; An excellent survey of these complex interfaces. important for modifying the physical methods is provided in the text Molecular Structure—The Chemistry Toolbox. The introduc- 7 properties of a surface • biochip –anarrayof by Adamson and Gast. In- tion of SAMs—including alkylsiloxanes on oxide sub- peptides, proteins, or other molecules that can cluded is X-ray photoelectron strates and alkanethiolates on the coinage metals— be treated with samples to identify spectroscopy (XPS), where a made it possible to display organic molecules and carry protein–protein interactions and enzyme monoenergetic X-ray source is out reactions of those molecules at surfaces.13,14 These activities • SAMDI – the combination of self- used to eject inner-shell elec- advances gave way to a substantial body of work that assembled monolayers and matrix-assisted trons that are then analyzed to emphasized the need for surface analytical methods laser desorption/ionization mass spectrometry, quantitatively identify the com- that could provide information on the molecular struc- which permits rapid characterization of the position of elements comprising ture—in contrast to information on elemental composi- masses of alkanethiolates in a monolayer; a surface. Electron diffraction tion described above—of monolayers. Several meth- useful for characterizing chemical and and ion scattering methods are ods have been important to identifying molecular biochemical reactions at surfaces powerful methods for charac- fragments, or functional groups, including grazing terizing the structures of or- angle infrared spectroscopy and secondary-ion mass dered surfaces. A variety of optical methods—includ- spectroscopy (SIMS). The latter identiﬁes masses of mo- ing ellipsometry and surface plasmon resonance (SPR) lecular fragments that derive from the high-energy spectroscopy—measure changes in refractive index to photodissociation of molecules at the surface. The data provide information on the amount of molecules local- can be analyzed to derive important information on ized at the interface. the molecular surface structure, but they are compli- It is clear that these methods were and remain criti- cated, and the technique is not routinely applied by the cal to the physics and fabrication of electronic struc- non-expert.15 tures and to surfaces used in catalysis—which have Needs in Molecular Surface Science. The preceding discus- relatively simple elemental structure and oftentimes sion points to a clear disconnection between current rely on dopants present at low concentrations. Yet, they surface analysis tools and applications in the chemical are less valuable for analyzing surfaces that are func- and biological sciences. Whereas the bulk characteriza- tionalized with complex chemical or biopolymeric tion of organic and biological macromolecules is now structures. In these cases, it is the connectivity of the at- routine, these methods cannot be applied to the small oms into molecules, and changes in this connectivity amounts of molecules that are present on surfaces (at following chemical or biological reactions, that are im- densities of approximately 5 ϫ 1014 molecules/cm2). At portant, more so than the elemental composition. the same time, the methods described above that have Morphology, Shape, and Forces—The Scanning Probe Toolbox. ultrasensitivity and can reveal elements present at low The characterization of the small feature sizes and densities at surfaces do not offer information on the out-of-plane structure of nanoscale assemblies has molecular structures of surfaces. Those methods that beneﬁted enormously from the family of scanning do provide molecular information (for example, SIMS) probe techniques. Scanning tunneling microscopy are not easily accessible or usable by the non-expert has exquisite sensitivity and spatial resolution in and therefore do not provide a routine tool, ultimately characterizing the electronic properties of a sur- limiting their utility. face.8 Among the most impressive applications to The consequences of this lack of tools are strik- chemistry was reported by Allara, Tour, and Weiss, ing. For example, there are far greater than 100,000 who imaged the switching of individual conjugated reports of homogeneous phase reactions,16 yet only

8 VOL. 2 ▪ NO. 1 ▪ MRKSICH www.acsnano.org REVIEW about 100 reports of reactions at SAMs.17 Applica- substrates are conductive, SAMs can be engineered to tions in biology routinely use biochips, where arrays be redox-active and thereby allow applied potentials to of peptides, proteins, or other molecules are treated be used to modulate the structures and properties of with samples to identify protein–protein interac- the monolayer.28 Finally, SAMs are compatible with sev- tions and enzyme activities.18 These applications eral analytical methods that are important in biology. now rely on fluorescent or radioisotopic labels to SAMs supported on thin gold films (approximately 10 identify activities, with the limitation that many bio- nm) are still optically transparent and permit optical chemical activities are not easily labeled. Of greater and fluorescence microscopy. Monolayers on thicker concern, the need for labeling strategies requires gold films (40 nm) can be analyzed with SPR that the activity to be measured is already defined, spectroscopy.29 and therefore these techniques are not open to the My group’s work in biosurfaces has revealed the discovery of unanticipated activities. many challenges associated with creating effective

HO SURFACE CHEMISTRIES FOR BIOLOGY A A host of interesting biological processes operate OH at surfaces and provide opportunities to develop and OH OH OH OH apply biologically active monolayers.19 For example, molecular recognition and lateral organization of proteins at lipid bilayer membranes are important in signaling pathways and have prompted the development of model surfaces for studies of immune synapse func- tion.20 Most cells adhere to an insoluble protein ma- S S S S S trix, and studies of the ligand–receptor interactions that mediate adhesion have benefited from the develop- B ment of monolayers that present peptide and carbohy- HO O O 220 mV drate ligands.21 Immobilized molecules are now fre- OH O O quently used to perform high-throughput assays of -150 mV biochemical activities and benefit from the use of well- C defined surface chemistries.22 Self-assembled monolay- 1st Scan ers of alkanethiolates on gold have been the most important route to these biosurfaces and are the focus of this review. Other important routes to biosurfaces have been reviewed elsewhere.23 2 µA SAMs combine several properties that make them the best available chemistries for modeling biological surfaces. First, monolayers that present short oligomers of the ethylene glycol group are highly effective at pre- -400 -200 0 200 400 600 venting non-specific adsorption of proteins.24 As a gen- Potential (mV) versus Ag/AgCl eral rule, most proteins will adsorb non-specifically to most man-made surfaces. This unwanted adsorption D 1.0 can block the interaction of the immobilized ligands 0.9 with soluble partners and can give rise to non-specific 0.8 0.7 interactions. Hence, it is vital to have a non-interacting It 0.6 Io background for the presentation of ligands that can 0.5 then interact with components of a proximal fluid. Sec- 0.4 ond, these surfaces permit wide flexibility in tailoring 0.3 0.2 the structure and properties, either by preparing mono- 0243 6 9 12 15 18 21 layers from terminally substituted alkanethiols or by Time (min) carrying out reactions to elaborate the structure of a Figure 1. Interfacial reactions can be characterized with monolayer after it has assembled. This flexibility is im- electrochemical methods. In this example, a hydroquinone portant to implementing immobilization chemistries group attached to a monolayer (A) undergoes oxidation to that give good control over the density and orienta- give the corresponding benzoquinone group, which selectively reacts with cyclopentadiene to produce the cycload- tion—and therefore the activity—of attached biomole- duct (B). Cyclic voltammetry reveals waves for the oxidation cules. Third, several methods are available that can pat- and reduction reactions of the hydrquinone and shows a de- tern monolayers, to give control over the shapes of creasing intensity of those waves as the reaction proceeds (C). Integration of the area under the waves provides kinetic attached cells and to create arrays of distinct biomolec- information and shows that the reaction is first order in ben- ular content.25–27 Fourth, because the underlying gold zoquinone (D). www.acsnano.org VOL. 2 ▪ NO. 1 ▪ 7–18 ▪ 2008 9 models of biological surfaces for studies of cell adhesion and in applications of biochips. These challenges do not stem from a shortage of reactions that can be employed to elaborate monolayers, but rather from the characterization of the products of chemical and biochemical reactions. In an early example, we demonstrated the use of a Diels–Alder reaction to immobilize biomolecules to a monolayer. The strategy relied on monolayers that presented a benzoquinone mol-

REVIEW ecule—which reacts with cyclopentadiene to give the expected 4ϩ2 cycloaddition product—at a low density against a background of tri(ethylene glycol) groups.30 We used the benzoquinone in part because it is redox active and can be reduced to the corresponding hydroquinone by applying an electrical potential to the underlying gold (Figure 1). We could carry out cyclic voltammetry in the presence of a diene and quantitate the amount of benzoquinone in each scan and, therefore, the rate of the Diels–Alder reaction. To gain evidence that the loss of redox activity of the quinone was indeed due to the anticipated reaction, we verified that a range of dienes reacted with the benzoquinone and that analogous compounds that did not contain a conjugated diene unit had no affect on the quinone redox couple. Further, we could quantitatively derive second- Figure 2. Biochemical interactions can be assayed using biochips and fluorescence detection. In this example, an ar- order rate constants for the reaction and found that ray of 10 carbohydrates is prepared and treated with a galac- the relative rate constants for a series of dienes agreed tosyltransferase enzyme (A). Identical arrays (B) are treated with the expected trend based on electronic effects. To with carbohydrate-binding proteins before and after the enzyme reaction to reveal changes in the structures of the im- gain structural evidence for the product, we used graz- mobilized sugars and, therefore, the sugars that are sub- ing angle infrared spectroscopy to observe the carbonyl strates for the enzyme (C). functional group of the quinone and cycladdition prod- changes to the mass of molecules at the surface, includ- uct. We found that many of the established tech- ing those that result from enzymatic modification of niques—including XPS and ellipsometry—were less the immobilized ligand. For example, the enzyme galac- useful in identifying the products and yields of the reac- tosyltransferase can append a galactose residue to ter- tions. The point of this example is that no single minal N-acetylglucosamine groups. To characterize the method—and certainly no method that is as conve- products that result from treatment of the array with nient as NMR in synthetic chemistry—could provide the enzyme, we used a panel of lectins having known strong evidence for the product and yield of the reac- specificities to infer the pattern of reactivity of GalTase. tion; therefore, a combination of techniques had to be By probing the array with fluorescently labeled lectins exploited to make the case for the product. A related example comes from our development of before and after treatment with GalTase, we could de- a carbohydrate array to profile the binding specificities termine those carbohydrates that were modified by the of proteins and the preferred substrates for enzymes.31 enzyme and could therefore identify preferred sub- We prepared a series of carbohydrateϪdiene conju- strates for the enzyme. gates and spotted these reagents onto a monolayer This need for a direct labeling of the reaction prod- presenting the benzoquinone group to prepare an ar- uct is common in all assays of biological activities and ray of 10 monosaccharides (Figure 2). We could easily can be a challenge in implementing high-throughput characterize the binding specificities of lectins—pro- assays. This requirement is particularly limiting in ex- teins that bind to carbohydrates—by imaging mono- periments that seek to identify unanticipated biochemi- layers that were treated with fluorescently labeled pro- cal activities, where a lack of knowledge of the modifi- teins. Alternatively, SPR spectroscopy could be used to cation prevents implementation of the appropriate monitor the interactions of non-labeled proteins with labeling strategy. Outside of biology, there exists a simi- immobilized carbohydrates, and recent work by Corn lar need for analytical methods that can rapidly assess enables the application of this technique in an imag- the products and yields of reactions performed on suring format,32 as do alternate configurations that are faces. We have found that matrix-assisted laser desorp- based on plasmonic devices.33 Yet, SPR does not have tion/ionization time-of-flight mass spectrometry the sensitivity to monitor events that result in small (MALDI-TOF MS) is well-suited to the characterization

10 VOL. 2 ▪ NO. 1 ▪ MRKSICH www.acsnano.org REVIEW of SAMs, and in the following section present the background to this technique, along with several ex- Detector amples of its application to chemical and biological monolayers. Alkanethiolate LASER SAMDI-TOF MASS MS SPECTROMETRY Alkyldisulfide Mass spectrometry methods, like the optical methods described above, are label-free in that they de- tect molecules according to their m/z molecular weight and do not require modified forms of the analyte. But Figure 3. Combination of self-assembled monolayers and matrix-assisted laser desorption/ionization mass spectrometry—in a technique termed SAMDI MS—permits rapid characterization they carry significant advantages in of the masses of alkanethiolates in the monolayer. Both alkanethiolate and dialkyl disulfide providing molecular information that forms of the molecules are observed. This method enables applications of model surfaces to a can discriminate the analytes. Be- broad range of problems in chemistry and biology. cause MS methods can identify species by their masses, they can be applied to the analy- straightforward manner. Indeed, in a first example, we sis of mixtures of activities, provided that each activity characterized the products resulting from reactions in- can be identified by a peak at a unique m/z ratio. MS volving cycloaddition to an immobilized maleimide methods have the further advantage that they are not group and condensation of a carboxylic acid on the intrinsically limited in detecting small changes in a sub- monolayer. These first examples revealed the informa- strate. As described further below, whereas SPR meth- tion that can be collected on a monolayer in a rapid ods are not able to identify phosphorylation of a pep- analysis and easily interpreted. Indeed, by comparing tide substrate (a mass change of 80 Da), MS can identify SAMDI spectra of a monolayer before and after a chemi- an exchange of hydrogen for deuterium in a molecule cal treatment, it is possible to rapidly assess the num- attached to a monolayer. The following section gives ber and approximate yields of distinct products and, examples of the use of SAMDI for characterizing bio- from consideration of the mass of the new adducts, the chemical activities and application of this in chemistry nature of the product. We note earlier work that used and biology. These examples demonstrate the utility of SAMs as substrates for MALDI-TOF MS, including the SAMDI MS methods as a general tool for characteriz- pioneering work of Nelson in the development of im- ing chemical and biochemical reactions at surfaces. munoassays and of the surface-enhanced laser desorp- Wilkins and Hanley have applied laser desorption tion/ionization (SELDI) methods that have been com- mass spectrometry to the characterization of SAMs mercialized for identification of possible and observed adducts corresponding to intact al- biomarkers.37,38 These examples were not concerned kanethiols, including disulfides and their complexes with characterizing the alkanethiolates of the mono- with gold atoms.34,35 The nature of the adducts and layer, as is SAMDI, but rather with characterizing pro- the degree of fragmentation of the parent molecular teins that interacted with the monolayer. ions were dependent on the laser fluence. This work Several examples of the use of SAMDI to character- used home-built mass spectrometers, and it did not in- ize chemical and biochemical reactions at surfaces fol- vestigate monolayers derived from alkanethiols that low, and these serve to demonstrate the merits of this were substituted with functional groups or molecular method and the scope of problems for which it is fragments and did not characterize the products result- suitable. ing from interfacial reactions. On the basis of these im- Electrochemical Interfacial Reactions. Electrochemical re- portant examples, we reasoned that MALDI-TOF MS actions of molecules attached to a monolayer are well- with a commercial instrument might be applicable to known for several redox-active groups and most exten- a far broader set of applications in surface chemistry sively studied with the ferrocene redox couple.39 The (Figure 3). Indeed, we found this to be the case. Treat- ability to perform oxidation or reduction reactions by ment of monolayers with the common energy- applying electrical potentials to the gold film underly- adsorbing matrices used in MALDI MS resulted in spec- ing the monolayer permits opportunities to create “dy- tra having major peaks corresponding to the masses namic substrates” whose structures and properties can of the alkanethiolates (and the respective disulfides).36 be switched. We have developed a family of strategies Hence, this technique, which we have termed SAMDI- to electrochemically modulate the activities of biologi- TOF MS owing to its combination of self-assembled cal ligands attached to a monolayer and have shown monolayers and desorption/ionization mass spectrom- that these approaches could be employed to control etry, provides information that can be interpreted in a the adhesion, migration, and organization of mamma- www.acsnano.org VOL. 2 ▪ NO. 1 ▪ 7–18 ▪ 2008 11 EtO C validating the preparation of these “biosurfaces”. In 2 OH one example, a monolayer presenting a maleimide group at low density against a background of oligo- O O (ethylene glycol) groups is treated with a thiol-tagged 1.0 ligand.43 The maleimide group selectively reacts with OMe O OMe 943.5 the thiol functionality of the ligand, and the glycol . O O O t n

I groups are effective at preventing unwanted protein 0.5 . l

e 721.6 adsorption in subsequent assays. Characterization of O O O R the monolayer with mass spectrometry shows clear 0 REVIEW O O O peaks corresponding to the maleimide-terminated al- 5957358751015kanethiolate prior to the reaction and the immobilized m/z adducts following reaction. A more complicated example is encountered in ex- 900 mV periments that require the selective immobilization of H proteins. For routes that rely on the selective reaction of O a functional group on the protein with a capture group 1.0 on the surface—for example, the reaction of a protein OMe O OMe 721.6 having a single cysteine residue with a maleimide . O O O t n I

group on the surface—it is necessary to ﬁrst purify the

. 0.5 l e 749.6 recombinant protein to prevent competing reactions of O O O R other components in the sample.44 Because the puriﬁ- O O O 0 cations are often not complete, it can be important to

595 735 875 1015 verify that the protein of interest, and only this protein, m/z was efficiently immobilized to the substrate. This prob- Figure 4. SAMDI MS can be applied to characterize the lem is extremely challenging because the number of products of electrochemical reactions of monolayers. In this possible proteins in the sample that can compete for example, a monolayer is functionalized with an acetal derived from hydroquinone. Application of a 900 mV poten- the surface is vast—making it impossible to explicitly tial to the underlying gold film results in oxidative hydroly- check every other protein—and the structures common sis of the acetal to reveal the aldehyde group. SAMDI spectra to all proteins make it difficult to apply spectroscopic before and after the reaction show clear peaks for the disulfide containing the functionalized monolayer (944 and 750 methods to this problem. Here, too, the mass-resolving Da, respectively) and for the disulfide derived from the back- ability of mass spectrometry methods makes it straight- ground alkanethiolates (721 Da). forward to verify the integrity of the immobilization lian cells.40,41 Again, the development and optimiza- process. We developed an active-site-directed method tion of these strategies requires a characterization of whereby an irreversible inhibitor of an enzyme is pre- the products and yields of the electrochemical reac- sented on the monolayer surface and is used to co- valently immobilize a fusion protein containing the tar- tions. The SAMDI method has proven critical toward get enzyme and the protein of interest to be displayed this activity. Figure 4 shows an example of an electro- at the surface.45 One example characterized the immo- chemical protecting group for the aldehyde functional- bilization of a 40 kDa protein having cutinase linked to ity. The protecting group is based on the redox-active the cell adhesion domain from fibronectin to a mono- hydroquinone unit and employs a pendant hydroxym- layer presenting a phosphonate capture ligand.46 ethyl group to form an acetal from the aldehyde carbo- Kinase Activity Assays. The SAMDI assay is well-suited nyl group. This protected structure is stable under neu- for performing a broad range of enzyme activity as- tral conditions but can be efficiently deprotected says. In the “solid-phase” format, the substrate is first at- through oxidation of the hydroquinone ring. The tached to the monolayer and then treated with the en- SAMDI spectra show clearly that the deprotection pro- zyme to give the corresponding product, usually with ceeds in high yield and provides the aldehyde product, a change in mass. Our first examples of this assay were 42 which can then be used to immobilize ligands. directed toward measuring the activities of kinases, Immobilization of Ligands. A common step in the prepa- which represent a family of approximately 500 human ration of biologically active surfaces is the immobiliza- enzymes that are responsible for the phosphorylation tion of molecules to a surface having appropriate chem- of proteins and play a role in essentially all cellular pro- istries, including activated esters, maleimides, epoxides, cesses. Functional assays of these enzymes are impor- and nucleophilic groups. It can be difficult to ensure tant in biochemical research, drug discovery, and diag- that the immobilization reaction has proceeded to nostics. Current solid-phase assays use surfaces that completion, or that other molecules in the sample have present peptide substrates and monitor the phosphory- not been co-immobilized with the desired ligands. lation of the peptide using either antibodies that recog- SAMDI provides a convenient and effective method for nize the modified peptide or adenosine triphosphate

12 VOL. 2 ▪ NO. 1 ▪ MRKSICH www.acsnano.org REVIEW

A B label-based assays, a lack of a signal is interpreted as a Ac-RKRSRAEC lack of activity, but it could be due to a failed immobili- S zation of the peptide at the surface. Because SAMDI O monitors all species on the monolayer, it can discrimi- N nate between true lack of activity and the false nega- O tives that arise from defect substrates. Assays of Multiple Activities. This feature of MS methods PKG 1200 1600 2000 2400 ATP m/z to identify multiple species on the surface can be ex- P ploited for performing multi-analyte assays. Current ap- PKG 80 proaches to assaying several distinct kinase activities Ac-RKRSRAEC ATP S in a cell lysate, for example, would either separate the O sample into several aliquots and perform individual as- N says on each or would pattern a set of peptides—where O each was phosphorylated by only one kinase in the mix- 1200 1600 2000 2400 m/z ture—in an immobilized array. With mass spectrom- C etry, however, the peptides could be immobilized as a + [M2+Na] mixture to a single monolayer, and, provided that the 100 + + [M3+H] [M4+H] + 2018 %

[M1+H] mass of each peptide is resolved from the others in the , 2293 2515 y

t 1839 i

s set, the phosphorylation of each peptide could be fol-

n 50 e t lowed independently. Figure 5 demonstrates this strat- n I egy using a monolayer to which four peptide substrates 0 were immobilized. A SAMDI spectrum shows clear 1000 1500 2000 2500 3000 m/z peaks for each of the peptideϪalkanethiol conjugates. Treatment of the surface with a single kinase—here, casein kinase 1—results in a spectrum in which the

100 peak for the selective peptide substrate has shifted by

% + 80 Da and the peaks corresponding to substrates for

, [pM4+H] y

t 47 i 2595

s other kinases remain unchanged. Further, this con-

n 50 e

t cept can be applied to the analysis of unrelated activi- n I ties—for example, those that report on protease, ki- 0 nase, and glycosylation events—because the assay 1000 1500 2000 2500 3000 m/z does not require processing steps that are often mutu- ally exclusive for different classes of activity. With the Figure 5. SAMDI MS can be applied to assays of enzyme ac- sensitivities of modern MALDI-TOF instruments, we ex- tivities. In this example, a peptide substrate for protein kinase G was immobilized to a monolayer presenting a male- pect that more than a dozen assays can be easily ac- imide group (A). SAMDI spectra of the monolayer before (B) commodated in a common substrate. and after (C) the enzyme reaction reveal that the mass of the Assays of Endogenous Cellular Activities. Characterizing peptide-terminated alkanethiol increased by 80 Da, corresponding to the expected phosphorylation. A multi-analyte the results of an immobilized format assay becomes assay was performed by immobilizing a mixture of four pep- more difficult with complex samples, including cell ly- tide substrates. Treatment of the monolayer with casein kinase 1 resulted in phosphorylation of only a single peptide, sates, blood, and other humoral fluids, since there are a which was easily detected in the SAMDI spectra. nearly unlimited number of competing analytes that can interact with the surface or give cross-reactivities. (ATP) that is labeled at the terminal phosphate group One common example comes from the analysis of the with the 32P isotope. The SAMDI format avoids the need family of caspase enzymes that underlie the apoptosis for radioisotopic or antiobody reagents and can be ap- programs that cause cells to die. The proteases have plied to multiple assays in the same mixture. In one ex- distinct sequence specificities for the substrates that ample, a monolayer presenting the RKRSRAE peptide they cleave, but current assays based on fluorescence 47 was treated with protein kinase G (PKG) and ATP. are not effective at resolving these activities. This limita- Spectra before and after treatment with PKG revealed tion derives from the use of tetrapeptide substrates that the mass of the peptide had increased by 80 Da, as that have a fluorescent molecule conjugated to the expected for a single phosphorylation reaction (Figure amide targeted by the protease. The presence of this 5). Further, the spectrum shows that all of the peptide non-natural residue, together with the deletion of pep- was active toward the PKG. This ability of SAMDI to tide sequence that can be important for enzyme–sub- monitor both the substrate and the product of a bio- strate discrimination, leads to substantial cross- chemical reaction is an important benefit because it al- reactivity of the peptides for the family of caspases. lows the yield to be estimated. It is also important be- Label-free assays permit the use of a native peptide cause it serves as a quality control in the assay. With that more closely resembles the endogeneous sub- www.acsnano.org VOL. 2 ▪ NO. 1 ▪ 7–18 ▪ 2008 13 100 100 + [M] (His)6 AB(His)6 ) 50 ) 50 % 2+ % ( (

O [M] I I (His)6

O 0 0 10 20 30 40 50 60 70 80 90 10 20 30 40 50 60 70 80 90 m/z (kDa) m/z (kDa) N Ni + HN 2 REVIEW N C 100 100 (His)6 (His)6 + )

[M] ) 50 50 % % ( (

(His) I

I 6

0 0 10 20 30 40 50 60 70 80 90 10 20 30 40 50 60 70 80 90 m/z (kDa) m/z (kDa)

Figure 6. Protein interactions can be monitored with the SAMDI method. A monolayer presenting a chelate of the Ni(II) ion was used to immobilize His-tagged proteins (A). SAMDI MS shows clear peaks for a 70 kDa protein (B). The monolayer was treated with a mixture of this protein and a 14 kDa protein known to interact with the ﬁrst and then analyzed by SAMDI to reveal the presence of both proteins. The same interaction could be identiﬁed in the opposite orientation, by using a His- tagged version of the smaller protein (C).

strates for the enzymes. We designed several peptides mass spectrum of this surface revealed a sharp peak at with this motivation in mind and performed assays of m/z of 70 kDa for the protein (Figure 6). When the sur- cell lysates on monolayers presenting the selective pep- face was treated with a mixture of the His-tagged pro- tide substrates.48 By analyzing the activities of lysates tein and a second protein that is known to interact with that were prepared at several times after induction of the first, we observed separate peaks for each protein. apoptosis, we found that the SAMDI assay provided a We did not observe the second protein when it was ap- clearer resolution of specific caspase activities, with sen- plied alone to the monolayer, demonstrating the effec- sitivity comparable to that of the fluorescence assays. tiveness of the tri(ethylene glycol)-terminated monolay- Again, multiple caspase activities could be assayed with ers for preventing non-specific adsorption of proteins. a single monolayer by relying on the resolving ability Further, we found that the same interaction could be of the mass spectra. We have also reported examples observed in the other orientation, when the second of the measurement of kinase activities in cellular ly- protein was prepared in its His-tagged form. Because sates49 and of protein antigens in cerebral spinal fluid.50 this assay uses microliter quantities of protein and does Protein–Protein Interactions. Many applications in bio- not require the protein to be rigorously purified, it could interfacial science rely on measuring the binding of probe well-suited to the global mapping of protein–proteins in a sample to ligands on a biochip surface. The tein interactions. current approaches rely almost exclusively on fluores- High-Throughput Screening. Analytical techniques that cence measurement, either by directly labeling the tar- can rapidly and quantitatively report on enzyme activi- get protein or by using labeled antibodies that are ties are valuable in screening programs that evaluate brought to the surface by way of the target protein. many thousands of reactions. In drug discovery, it is Here, too, the expense associated with preparing high- common to perform enzyme assays in the presence of affinity and selective antibodies for analytes has moti- small molecules from a library, and to then identify vated the investigation of label-free methods for mea- those molecules that inhibit (or activate) the enzyme suring protein binding. Surface plasmon resonance activity. We demonstrated that SAMDI is an effective ap- spectroscopy and analogous methods that measure proach to these screening applications. In one ex- changes in the refractive index of the medium near an ample, we developed an assay for the anthrax lethal fac- interface are the most important tools, but they are lim- tor toxin, which has proteolytic activity toward peptide ited in that specific and non-specific binding are not re- substrates, and we used this assay to identify inhibitors solved. We found that SAMDI is well-suited to analyz- from a library containing 10,000 small molecules (Fig- ing proteins bound to the monolayer substrates and ure 7).54 Droplets containing the enzyme and eight can identify proteins with masses up to 100,000 Da.51,52 compounds were applied to a target plate having im- In one approach to characterizing protein–protein in- mobilized peptide and incubated for 1 h. The plate was teractions with SAMDI, we immobilized His-tagged pro- then rinsed and analyzed by mass spectrometry to teins to a monolayer presenting a Ni(II) chelate.53 The identify those droplets that had no activity. Using this

14 VOL. 2 ▪ NO. 1 ▪ MRKSICH www.acsnano.org REVIEW approach, we could evaluate several thousand compounds in a day, and we identified a small molecule that inhibited the lethal factor protease with a micro- molar dissociation constant and that was active in cell culture models. Mass spectrometry had not previously been applied to screening applications, primarily because the preparation of samples for analysis—which includes removing salts and enriching the desired analyte in the sample—is too slow and expensive. With the SAMDI method, a straightforward rinse of the surface accomplishes both objectives and is therefore compatible with large numbers of assays. The use of a label-free detection method also reduces the time to develop the assay and will be particularly important for assaying those activities that are difficult to label. Chemical Reaction Discovery. These benefits of SAMDI also allow its application to the identification of new synthetic transformations in chemistry. The development of new reactions still employs a traditional route, which often begins with an unexpected observation and is followed with a linear sequence of optimizations. Recent work has employed parallel screening to identify reagents that effect known reactions in high yield55 or with high enantioselectivity.56 In both cases, the need for labeling strategies makes these strategies ill- suited for the discovery of unexpected, and potentially new, reactions. The SAMDI method, because it can identify any product that results from a reaction (provided that the mass has changed), can be applied to this latter goal. Indeed, SAMDI is well-suited to identifying the products that result from treatment of immobilized Figure 7. SAMDI was used to perform a screen of 10,000 molecules with reagents. Figure 8 shows examples of small molecules to identify inhibitors of the anthrax lethal the replacement of the hydrogen of a terminal alkyne factor protease. (A) A peptide substrate for lethal factor was immobilized to a monolayer presenting maleimide groups. with deuterium and the palladium-catalyzed coupling (B) Treatment of the monolayer with recombinant protease 57 of the alkyne with iodobenzene. In both cases, SAMDI resulted in cleavage of the peptide, which could be analyzed shows clear peaks corresponding to the mass of the ex- by SAMDI mass spectrometry. (C) Chemical screens were performed by arraying 100 droplets that contained the pro- pected product and has a mass resolution better than tease and eight compounds from the library, followed by 1 Da. We recently reported 15 reactions that were char- analysis of the spots with mass spectrometry, which clearly acterized in this way and extended that work to a identified those spots having an inhibitor in the reaction mixture (D, pink spectrum). screen to identify new reactions. In one example, we found an unexpected reaction of primary aldehydes used an Applied Biosystems Voyager DE-PRO instru- with an immobilized amine to give N-alkylpyridinium ment, and we have found that current MALDI-TOF mass products. spectrometers from other manufacturers are also effective for performing SAMDI. We note that the cost of PERSPECTIVES the instruments is substantially lower than that of many This last example illustrates the unique capabilities UHV instruments used in surface science. This combina- that the SAMDI MS method brings to molecular surface science. No other technique could be used to rap- tion of cost, speed, and commercial instrumentation idly survey hundreds of reaction zones on a monolayer makes SAMDI-TOF MS a powerful tool—and for many to identify those combinations of reactants and re- programs, the enabling tool—for researchers that use agents that give products in high yield. The SAMDI SAMs in studies of chemistry and biology. method can do so at a rate exceeding 200 spots per What are the current limitations of the SAMDI hour and has a mass resolution that allows the identifi- method? The primary variable in this method concerns cation of products that differ in mass from the substrate the choice and application of matrix. We have used the by a single dalton. A further benefit of this method is standard matrices developed for MALDI-TOF MS and that it can be performed using commercially available find that the optimal matrix can vary with the struc- instruments. The examples described in this review tures of monolayer and analytes that are being charac- www.acsnano.org VOL. 2 ▪ NO. 1 ▪ 7–18 ▪ 2008 15 throughput of the method. Finally, the monolayers do A 100 779.6779.6 not have the stability required in some applications. y t i

H s

n Temperatures in excess of 70 °C, ultraviolet light in the e t 50 n I 965.6965.6

e presence of oxygen, and chemical reagents that are v i t a l

e strong oxidants, bases, or acids cause damage to the R 0 600 800 1000 1200 monolayer. For biological assays, which require moder- 600 800 m/z m/z 1000 1200 ate temperatures and neutral aqueous environments, NaH the stability of the monolayers is not limiting. D2O 100 780.6 Self-assembled monolayers are now a basic compo- y t i REVIEW s nent of the toolbox for nanoscience. At the most basic n e t n

D I 50

e 967.5 level, these nanoscale two-dimensional structures are v

i 967.5 t a l

e important for modifying the physical properties of a R

0 surface. In current applications, the monolayers pro- 600 800 1000 1200 600 800 m/z 1000 1200 vide a route to preparing surfaces that are decorated m/z with a vast array of molecular content and that, in turn, 583 583 B enable applications in chemistry and biology. These ap- H plications often require a sequence of reactions to as- 769769 semble the ﬁnal monolayer and to analyze it after its application. The SAMDI method brings unprecedented

600 800 1000 1200 abilities to identify the alkanethiolates and their reac- 600 800 1000 1200 tion products, to identify proteins, polymers, or other I m/z Pd(PPh3)4 CuI molecules that are associated with the monolayer, and TBAF, THF 659 659 to verify the integrity of the monolayer in experiments. Several researchers have recently adopted this method, and we expect that its acceptance as a basic compo- nent of the characterization toolbox for nanoscience 921921 will continue to grow. 600 800 1000 1200 600 800 1000 1200 m/z Acknowledgment. My laboratory’s development of the SAMDI method has been supported by the National Science Founda- Figure 8. Chemical reactions of molecules attached to a tion and the National Institutes of Health. I am grateful to Zack monolayer are rapidly characterized with SAMDI MS. Treat- Gurard-Levin and Susan Reed-Strowhorn for assistance with ment of a monolayer presenting a terminal alkyne group preparation of the figures and references. with sodium hydride and deuterated water results in the exchange of the terminal hydrogen for a deuterium atom (A). SAMDI MS identifies the product with a mass change of 1 Da REFERENCES AND NOTES (at m/z 780.6) and the associated disulfide product (at m/z 1. Whitesides, G. M. Nanoscience. Nanotechnology and 967.5). A Sonagashira coupling of iodobenzene to the alkyne Chemistry. Small 2005, 1, 172–179. also proceeds in high yield to produce the phenylalkyne ad- 2. Antos, J. M.; Francis, M. B. Transition Metal Catalyzed duct (B). Methods for Site-Selective Protein Modification. Curr. Opin. Chem. Biol. 2006, 10, 253–262. terized. We also find that the deposition of matrix can 3. Jacobson, K.; Mouritsen, O. G.; Anderson, G. W. Lipid Rafts: result in patches of the surface giving good or poor sig- At a Crossroad Between Cell Biology and Physics. Nat. Cell Biol. 2007, 9, 7–14. nal. Current efforts to develop strategies to uniformly 4. Love, J. C.; Estroff, L. A.; Kriebel, J. K.; Nuzzo, R. G.; apply the matrix may improve this step. MALDI can be Whitesides, G. M. Self-Assembled Monolayers of Thiolates on Metals as a Form of Nanotechnology. Chem. Rev. 2005, used to provide a spatial mapping of the chemistry of 105, 1103–1169. the monolayer. Modern instruments have a 1 ␮m scan- 5. Vogel, E. Technology and Metrology of New Electronic ning capability and sufficient resolution to provide im- Materials and Devices. Nat. Nanotechnol. 2007, 2, 25–32. 6. Bell, A. T. The Impact of Nanoscience on Heterogeneous ages of products with better than 10 ␮m sensitivity. We Catalysis. Science 2003, 299, 1688–1691. also note that SAMDI does not have the quantitative 7. Adamson A. W.; Gast A. P. Physical Chemistry of Surfaces, 6th ed.; John Wiley & Sons: New York, 1997; Table 8. character that other surface analytical methods do. For 8. Binnig, G.; Rohrer, H. Scanning Tunneling Microscopy: a given analyte on the monolayer, the intensity of the From Birth to Adolescence. Rev. Mod. Phys. 1987, 59, 615– m/z peak is generally related to the density of that spe- 625. 9. Donhauser, Z. J.; Mantooth, B. A.; Kelly, K. F.; Bumm, L. A.; cies, but different molecules ionize with varying effi- Monnell, J. D.; Stapleton, J. J.; Price, D. W.; Rawlett, A. M.; ciencies, making it difficult to directly compare the rela- Allara, D. L.; Tour, J. M.; Weiss, P. S. Conductance Switching tive amounts of distinct molecules. Peaks in Single Molecules Through Conformational Changes. Science 2001, 292, 2303–2307. corresponding to the unfunctionalized alkanethiolate 10. Giessibl, F. J. Advances in Atomic Force Microscopy. Rev. in the monolayer can be used to calibrate the intensi- Mod. Phys. 2003, 75, 949–983. 11. Bustamante, C.; Chemla, Y. R.; Forde, N. R.; Izhaky, D. ties of peaks corresponding to different molecules, but Mechanical Processes in Biochemistry. Annu. Rev. Biochem. doing so requires authentic standards and reduces the 2004, 73, 705–748.

16 VOL. 2 ▪ NO. 1 ▪ MRKSICH www.acsnano.org REVIEW

12. Lee, H.; Scherer, N. F.; Messersmith, P. B. Single-Molecule 34. Trevor, J. L.; Lykke, K. R.; Pellin, M. J.; Hanley, L. Two-Laser Mechanics of Mussel Adhesion. Proc. Natl. Acad. Sci. U.S.A. Mass Spectrometry of Thiolate, Disulfide, and Sulfide Self- 2006, 103, 12999–13003. Assembled Monolayers. Langmuir 1998, 14, 1664–1673. 13. Nuzzo, R. G.; Allara, D. L. Adsorption of Bifunctional 35. Gong, W.; Elitzin, V. I.; Janardhanam, S.; Wilkins, C. L.; Organic Disulfides on Gold Surfaces. J. Am. Chem. Soc. Fritsch, I. Effect of Laser Fluence on Laser Desorption Mass 1983, 105, 4481–4483. Spectra of Organothiol Self-Assembled Monolayers on 14. Wasserman, S. R.; Tao, Y. T.; Whitesides, G. M. Structure Gold. J. Am. Chem. Soc. 2001, 123, 769–770. and Reactivity of Alkylsiloxane Monolayers Formed by 36. Su, J.; Mrksich, M. Using MALDI-TOF Mass Spectrometry to Reaction of Alkyltrichlorosilanes on Silicon Substrates. Characterize Interfacial Reactions on Self-Assembled Langmuir 1989, 5, 1074–1087. Monolayers. Langmuir 2003, 19, 4867–4870. 15. Benninghoven, A. Chemical Analysis of Inorganic and 37. Nedelkov, D.; Nelson, R. W. Surface Plasmon Resonance Organic Surfaces and Thin Films by Static Time-of-Flight Mass Spectrometry: Recent Progress and Outlooks. Trends Secondary Ion Mass Spectrometry. Angew. Chem., Int. Ed. Biotechnol. 2003, 21, 301–305. Engl. 1994, 33, 1023–1043. 38. Engwegen, J. Y.; Gast, M. C.; Schellens, J. H.; Beijnen, J. H. 16. Bishop, K. J. M.; Klajn, R.; Grzybowski, B. A. The Core and Clinical Proteomics: Searching For Better Tumour Markers Most Useful Molecules in Organic Chemistry. Angew. with SELDI-TOF Mass Spectrometry. Trends Pharmacol. Sci. 2006, 27, 251–259. Chem., Int. Ed. 2006, 45, 5348. 39. Gooding, J. J.; Mearns, F.; Yang, W; Liu, J. Self-Assembled 17. Chechik, V.; Crooks, R. M.; Stirling, C. J. M. Reactions and Monolayers into the 21st Century: Recent Advances and Reactivity in Self-Assembled Monolayers. Adv. Mater. 2000, Applications. Electroanalysis 2003, 15, 81–96. 12, 1161–1171. 40. Hodneland, C. D.; Mrksich, M. Biomolecular Surfaces that 18. Ng, J. H.; Ilag, L. L. Biochips Beyond DNA: Technologies Release Ligands under Electrochemical Control. J. Am. and Applications. Biotechnol. Annu. Rev. 2003, 9, 1–149. Chem. Soc. 2000, 122, 4235–4236. 19. Mrksich, M. What Can Surface Chemistry Do for Cell 41. Yousaf, M. N.; Houseman, B. T.; Mrksich, M. Turning on Cell Biology. Curr. Opin. Chem. Biol. 2002, 6, 794–797. Migration with Electroactive Substrates. Angew. Chem., Int. 20. Mossman, K. D.; Campi, G.; Groves, J. T.; Dustin, M. L. Ed. 2001, 40, 1093–1096. Altered TCR Signaling from Geometrically Repatterned 42. Yeo, W. S.; Mrksich, M. Electroactive Substrates that Reveal Immunological Synapses. Science 2005, 310, 1191–1193. Aldehyde Groups for Bio-Immobilization. Adv. Mater. 2004, 21. Mrksich, M. Model Organic Surfaces for Mechanistic 16, 1352–1356. Studies of Cell Adhesion. Chem. Soc. Rev. 2000, 29, 43. Houseman, B. T.; Gawalt, E. S.; Mrksich, M. Maleimide- 267–273. Functionalized Self-Assembled Monolayers for the 22. Houseman, B. T.; Huh, J. H.; Kron, S. J.; Mrksich, M. Peptide Preparation of Peptide and Carbohydrate Biochips. Chips for the Evaluation of Protein Kinase Activity. Nat. Langmuir 2003, 19, 1522–1531. Biotechnol. 2002, 20, 270–274. 44. Hong, H.; Jiang, M.; Sligar, S. G.; Bohn, P. W. Cysteine- 23. Kodadek, T. Development of Protein-Detecting Specific Surface Tethering of Genetically Engineered Microarrays and Related Devices. Trends Biochem. Sci. Cytochromes for Fabrication of Metalloprotein 2002, 27, 295–300. Nanostructures. Langmuir 1994, 10, 153–158. 24. Mrksich, M.; Whitesides, G. M. Using Self-Assembled 45. Hodneland, C. D.; Lee, Y. S.; Min, D. H.; Mrksich, M. Monolayers That Present Oligo (ethylene glycol) Groups Selective Immobilization of Protein to Self-Assembled To Control the Interactions of Proteins with Surfaces. In Monolayers Presenting Active Site Directed Capture Poly(ethylene glycol): Chemistry and Biological Applications; Ligands. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 5048–5052. American Chemical Society Symposium Series 680; 46. Murphy, W. L.; Mercurius, K. O.; Koide, S.; Mrksich, M. American Chemical Society: Washington, DC, 1997; pp Substrates for Cell Adhesion Prepared via Active Site- 361Ϫ373. Directed Immobilization of a Protein Domain. Langmuir 25. Xia, Y.; Rogers, J. A.; Paul, K. E.; Whitesides, G. M. 2004, 20, 1026–1030. Unconventional Methods for Fabricating and Patterning 47. Min, D. H.; Su, J.; Mrksich, M. Profiling Kinase Activities by Nanostructures. Chem. Soc. Rev. 1999, 99, 1823–1848. Using a Peptide Chip and Mass Spectrometry. Angew. 26. Mirkin, C. A. The Power of the Pen: Development of Chem., Int. Ed. 2004, 43, 5973–5977. Massively Parallel Dip-Pen Nanolithography. ACS Nano 48. Su, J.; Rajapaksha, T. W.; Peter, M. E.; Mrksich, M. Assays of 2007, 1, 79–83. Endogenous Caspase Activities: A Comparison of SAMDI 27. Dillmore, W. S.; Yousaf, M. N.; Mrksich, M. A Photochemical Mass Spectrometry and Fluorescence Formats. Anal. Chem. Method for Patterning the Immobilization of Ligands and 2006, 78, 4945–4951. Cells to Self-Assembled Monolayers. Langmuir 2004, 20, 49. Su, J.; Bringer, M. R.; Ismagilov, R. F.; Mrksich, M. 7223–7231. Combining Microfluidic Networks and Peptide Arrays for Multi-Enzyme Assays. J. Am. Chem. Soc. 2005, 127, 28. Mrksich, M. Dynamic Substrates for Cell Biology. MRS Bull. 7280–7281. 2005, 30, 180–184. 50. Patrie, S. M.; Mrksich, M. SAMDI-TOF Mass Spectrometry 29. Mrksich, M.; Sigal, G. S.; Whitesides, G. M. Surface Plasmon for Diagnostic Immunoassays of Human Protein Antigens. Resonance Permits In Situ Measurement of Protein Anal. Chem. 2007, 79, 5878–5887. Adsorption on Self-Assembled Monolayers of 51. Yeo, W. S.; Min, D. H.; Hsieh, R. W.; Greene, G. L.; Mrksich, Alkanethiolates on Gold. Langmuir 1995, 11, 4383–4385. M. Label-Free Detection of Protein-Protein Interactions on Ϫ 30. Yousaf, M. N.; Mrksich, M. Diels Alder Reaction for the Biochips. Angew. Chem., Int. Ed. 2005, 44, 5480–5483. Selective Immobilization of Protein to Electroactive Self- 52. Marin, V. L.; Bayburt, T. H.; Sligar, S. G.; Mrksich, M. Assembled Monolayers. J. Am. Chem. Soc. 1999, 121, 4286– Functional Assays of Membrane-Bound Proteins by 4287. SAMDI-TOF Mass Spectrometry. Angew Chem., Int. Ed. 31. Houseman, B. T.; Mrksich, M. Carbohydrate Arrays for the 2007, 46, 8952–8954. Evaluation of Protein Binding and Enzyme Activity. Chem. 53. Sigal, G. B.; Bamdad, C.; Barberis, A.; Strominger, J.; Biol. 2002, 9, 443–454. Whitesides, G. M. A Self-Assembled Monolayer for the 32. Lee, H. J.; Nedelkov, D.; Corn, R. M. SPR Imaging Binding and Study of Histidine-Tagged Proteins by Surface Measurements of Antibody Arrays for the Multiplexed Plasmon Resonance. Anal. Chem. 1996, 68, 490–497. Detection of Low Molecular Weight Protein Biomarkers. 54. Min, D. H.; Tang, W. J.; Mrksich, M. Chemical Screening by Anal. Chem. 2006, 78, 6504–6510. Mass Spectrometry to Identify Inhibitors of Anthrax Lethal 33. Malyarchuk, V.; Stewart, M. E.; Nuzzo, R. G.; Rogers, J. A. Factor. Nat. Biotechnol. 2004, 22, 717–720. Spatially Resolved Biosensing with a Molded Plasmonic 55. Evans, C. A.; Miller, S. J. Proton-Activated Fluorescenc as a Crystal. Appl. Phys. Lett. 2007, 90, 203113. Tool for Simultaneous Screening of Combinatorial www.acsnano.org VOL. 2 ▪ NO. 1 ▪ 7–18 ▪ 2008 17 Chemical Reactions. Curr. Opin. Chem. Biol. 2002, 6, 333–338. 56. Korbel, G.; Lalic, G.; Shair, M. D. Reaction Microarrays: A Method for Determining the Enantiomeric Excess of Thousands of Samples. J. Am. Chem. Soc. 2001, 123, 361– 362. 57. Li, J.; Thiara, P. S.; Mrksich, M. Rapid Evaluation and Screening of Interfacial Reactions on Self-Assembled Monolayers. Langmuir 2007, 23, 11826–11835. REVIEW

18 VOL. 2 ▪ NO. 1 ▪ MRKSICH www.acsnano.org