Anal Bioanal Chem (2004) 379: 383–390 DOI 10.1007/s00216-004-2633-y

R EV IE W

Sapna K. Deo

Exploring bioanalytical applications of assisted protein reassembly

Received: 14 January 2004 / Revised: 2 April 2004 / Accepted: 7 April 2004 / Published online: 4 May 2004
Springer-Verlag 2004

Abstract Reassembly of protein from its peptide fragments is a technique that can have many applications in the bioanalytical field. Typically, a reporter protein fragmented into its two peptides is employed as a label in this study. This fragments of peptide can reassemble yielding an active functional reporter. This reassembly of the protein can be assisted by non-covalently interacting peptides or proteins, which are attached to the fragmented reporter. This technique has been employed in several applications including study of protein–protein interactions, antibody screening, immunoassays, and high-throughput screening. This review focuses on different reporters employed in the study of reassembly of proteins and applications of this strategy in bioanalysis. Keywords Assisted protein reassembly Æ Protein-protein interactions Æ Immunoassasy Æ Reporter protein Æ Molecular rulers

Introduction Reassembly of proteins commonly known as ‘‘Proteinfragment complementation’’ is a technique in which protein is rationally dissected into two peptide fragments and is assembled back by using non-covalent interactions. There are two ways in which protein can reassemble, either by spontaneous reassembly or by an assisted reassembly (Fig. 1). Several proteins including ribonuclease, chymotrypsin inhibitor-2, t-RNA synthase, and inteins undergo spontaneous reassembly from their peptide fragments to yield active protein [1–4]. However, formation of aggregates of peptides is a major S. K. Deo Department of Chemistry, University of Kentucky, Lexington, KY 40506-0055, USA E-mail: [email protected] Tel.: +1-859-2579544

drawback of spontaneous reassembly of proteins (Fig. 1). To overcome this drawback, soluble oligomerization domains are attached to the peptide fragments, which assist in reassembly of proteins. This strategy has been successfully used in the reassembly of reporter proteins such as, dihydrofolate reductase, b-lactamase, ubiquitin, aminoglycoside kinase, and green fluorescent protein (GFP) [5–10]. The most critical issue in the success of the reassembly is the site of the cleavage in the protein. Availability of the 3-D structure and circular permutation studies can aid in the design of the protein fragments. Other considerations in the design of the successful protein reassembly include, small protein size, monomeric structure, and simplicity of the assay system. Another crucial factor in the design of the assay is at which terminus the oligomerization domains should be attached. Generally, proteins that have clear domains fusion at either termini work well. However, for proteins such as GFP, which has single domain or beta-lactamase, which do not have clear domains, selecting site of fusion is critical. A detail discussion about the design parameters and steps involved can be found in the review [11]. Protein reassembly technique has been shown to have several analytical and biochemical applications which will be discussed in this review. In addition, we will describe some examples of the reporter proteins that have been employed to date in this technique.

Reporters Reporters that can be fragmented and reassembled back need to fulfill certain selection criteria. Ideally, they should be small size proteins, which allow easy recombinant manipulation. They should be stable and active in monomeric form. The reporters should possess sufficient flexibility to allow fragmentation into peptides. In addition, the measurement of its activity should be easy and allow in vivo analysis. The reporters that have been employed to date in protein fragmentation assays include murine dihydrofolate reductase (mDHFR),

384 Fig. 1 Schematic representation of spontaneous and assisted reassembly where X and Y are peptide fragments of the reproter protein and A and B are the interacting domains

Spontaneous Reassembly

Assisted Reassembly

A

ubiquitin, b-lactamase, GFP, and luciferase. Each reporter has its advantages and disadvantages. For example, mDHFR is an enzyme and, therefore, needs addition of an external substrate, whereas GFP is an autofluorescent protein whose reassembly can be easily visualized without the need of an external substrate. mDHFR Murine dihydrofolate reductase (mDHFR) is a small enzyme that regenerates tetrahydrofolate from dihydrofolate in the presence of NADPH. It plays an essential role in the building of DNA and other processes. The three-dimensional structure of mDHFR shows that it has three structural fragments (F1, F2, F3) that form two domains: the adenine-binding domain (F2) and a discontinuous domain (F1, F3) [5]. The residues belonging to fragments F1 and F2 form the folateand NADPH-binding pockets. Only 4 of 29 residues that bind the substrate are part of the F3 fragment and these residues are not essential for catalysis. The circular permutation study has shown that the primary structure of mDHFR can be altered without affecting the enzyme activity. The amino acid residues 101–108, at the junction between F2 and F3, form a disordered loop, which can be disrupted without affecting the activity of the enzyme. Therefore, mDHFR was dissected by Michnick et al. [1] at position 107, which does not affect substrate and NADPH binding [5].

B

activity. A fluorescent b-lactamase substrate, CCF2/ AM, is most commonly employed in the study, which allows reproducibility and quantification of results in intact cells. By analyzing the b-lactamase structure, a site (between Gly196 and Leu198) located on a surface opposite to the active site has been proposed as an ideal site for cleavage of b-lactamase into two fragments [12]. Furthermore, this site contains no periodic secondary structure, and allows protein to fold easily. GFP Green fluorescent protein is an autofluorescent protein originally isolated from Aequorea Victoria [13]. It has a b-barrel structure with a central alpha helix that forms a chromophore. Initially it was thought that the structure of GFP is rigid and would not allow any mutations or peptide/protein insertion [14]. However, circular permutation and loop insertion studies have demonstrated that GFP structure is flexible to changes. Reagan et al. have dissected GFP into two fragments and showed its successful reassembly upon interaction of fused peptides (Fig. 2). GFP was dissected between residues 157 and 158 [15]. This position was selected because it has been shown to tolerate insertion of 20 amino acids. The advantages of using GFP as the reporter include high sensitivity, and capability of in vivo analysis. Luciferase

b-Lactamase The class A beta-lactamase TEM-1 is a key bacterial resistance enzyme against beta-lactam antibiotics. The TEM-1b-lactamase from E. coli meets the criteria for a protein reassembly candidate. It is well characterized structurally and functionally. It’s a monomeric protein and allows facile in vivo and in vitro assays. Furthermore, no orthologs of b-lactamase exist in eukaryotes, and therefore, it could be used in both eukaryotic and prokaryotic cells without any intrinsic background

The bioluminescent protein luciferase either isolated from firefly or from sea pansy Renilla reniformis has been utilized as a label in various biological and analytical applications [16]. Splitting of these proteins and directed reassembly has been demonstrated recently [17]. Renilla luciferase is a small monomeric 36-kDa bioluminescent protein that does not require posttranslational modification for its activity. This protein catalyzes oxidation of luciferin to produce light with an emission spectrum in the rage of 440–550 nm, in the case of Renilla luciferase, and in the range of 575–610 nm, for

385 Fig. 2 Dissection of GFP into two fragments and reassembly using leucine zipper peptides

firefly luciferase. The biggest advantage of using these bioluminescent proteins is high sensitivity of detection due to low background signal as molecules in biological fluids are not bioluminescent. Because of the difference in the emission wavelengths of the two luciferases they can also be used as a dual reporter.

Applications of protein reassembly Studying protein–protein interactions Protein–protein interactions are key in many biochemical processes, such as receptor activation, metabolic regulation, signal transduction etc. Therefore, understanding these interactions is essential in deciphering biochemical pathways. Several analytical and biochemical methods have been proposed to study protein–protein interactions. These include mass spectrometry, NMR, fluorescence resonance energy transfer (FRET), and the yeast two-hybrid system. Among these, the yeast two-hybrid system is commonly used to study protein– protein interactions in vivo. Assisted reassembly of proteins is an alternative method to existing approaches for studying protein–protein interactions. In this strategy, a reporter protein is dissected into two fragments. To these fragments two non-covalently interacting proteins are fused genetically. The interaction between the two test proteins allows reassembly of the reporter protein restoring its activity. The advantages of this method include in vivo analysis, homogeneous assay format, and moreover, detection of interaction leads to increase in the measured signal yielding high sensitivity. One disadvantage of this strategy is steric hindrance due to bulky proteins may prevent reassembly of the reporter protein. This problem, however, can be over come by designing flexible linker between test protein and the reporter fragment. Nevertheless, while designing this linker consideration should be given that this flexibility does not prevent reassembly of the reporter. Michnick et al. [11] have demonstrated that active mDHFR can be reassembled from its peptide fragments when fused to interacting peptides or proteins (Fig. 3)

[7]. In this study, GCN4 leucine zipper peptides that interact noncovalently were fused to the dissected mDHFR. The enzyme mDHFR was dissected at the amino acid residue 107 and the leucine zipper peptides were genetically fused to the dissected enzyme fragments. Reassembly of mDHFR was monitored by three different types of assays, namely, colorimetric, fluorometric and E. coli survival assay. In vitro mDHFR activity can be monitored by adding a substrate that yields either a colored or a fluorescent product. For example, mDHFR activity can be monitored by following the appearance of the fluorescent product tetrahydrofolate upon the addition of the substrate dihydrofolate. In vivo reassembly of mDHFR can be monitored by transfecting the constructs into E. coli. Bacterial DHFR is inhibited by methotrexate whereas, mammalian DHFR has a 120,000-fold less affinity for methotrexate. Therefore, only the E. coli cells that have reassembled mDHFR due to interaction of fused peptides survive in medium containing methotrexate. Using this strategy, detection of two known protein–protein interactions was demonstrated in vivo. Specifically, the two pairs were p21 ras GTPase that interacts with rasbinding domain of the ser/Thr kinase and raf (rapamycin-mediated interaction of the immunophillin FKBP) that binds to Saccharomyces cerevisiae target of rapamycin (TOR2). Furthermore, E. coli survival assay was used to screen two libraries of complementary designed leucine-zipper forming peptides. For that, the leucine zipper libraries randomized at the position adjacent to the hydrophobic core, were genetically fused to either one of two mDHFR fragments. Three levels of stringency were applied in the selection process. For the lowest stringency level, two libraries were screened in a single-step. For the second level of stringency, fragments of a mutant of mDHFR was employed which needs better interactions between leucine zippers for forming the active enzyme. In the third level of stringency, the peptides selected from the second step were allowed to compete with each other. This strategy of selection ensures that the sequence of interacting peptides selected have high stability, ability to form heterodimers that have less charge repulsion and advantageous in vivo

386 Fig. 3 Schematic diagram of protein–protein interaction study using dissected mDHFR. Interaction between proteins inside cell results in functional reassembly of mDHFR enzyme, which can be monitored by formation of colored product

Protein B

Protein A

Protein-Protein Interaction

mDHFR Fragment

mDHFR Fragment

Protein A Protein B

Cell

SUBSTRATE

properties such as solubility and protease resistance. The simplicity of this approach and the specificity of the information obtained will allow protein–protein interaction studies. This approach would be also useful in protein design and directed evolution. Michnik et al. have also demonstrated the use of split ubiquitin for studying protein–protein interactions [18]. Ubiquitin-specific proteases (UBPs) can recognize and cleave only the folded ubiquitin. The reporter protein (hemaglutinin (HA)-tagged mDHFR) was fused to the N-terminal half (Nub) and was coexpressed with the C-terminal half (Cub) of ubiquitin in the yeast cells. The two halves retain their affinity for each other and spontaneously reassemble to form the so-called splitubiquitin from which the reporter protein can be cleaved off. A point mutation was performed in the N-terminal half of ubiquitin (NubG) that completely abolishes the affinity of the two halves for each other. Since the separate NubG and Cub parts are not recognized by ubiquitin-specific proteases (UBPs), no cleavage of the reporter protein takes place. The two interacting proteins, leucine zippers, were then fused to the two halves along with the reporter protein. Interaction of the two proteins brings the NubG and Cub domains close enough to reconstitute split-ubiquitin, resulting in the release of the reporter protein by the action of the UBPs. This method allows study of kinetic and equilibrium aspects of protein–protein interactions. Furthermore, this strategy can be used for screening cDNA libraries for studying interactions of target protein with unknown proteins.

PRODUCT

A bioluminescent reporter protein, renilla luciferase, has been employed by Gambhir et al. to demonstrate its applicability in studying protein–protein interactions by splitting it into two parts and by following its reassembly upon interactions between the fused protein pair [17]. Since the signal generated from the active luciferase is due to a bioluminescent reaction, high sensitivity of detection can be achieved due to low background. Due to the unavailability of a 3-D X-ray crystal structure of this protein, six different cleavage sites were analyzed to yield the optimal reassembly. Reassembly of the protein was demonstrated using the interacting protein pair (MyoD-Id), which belongs to the family of loop-helixloop nuclear proteins. The authors have also demonstrated that firefly luciferase can be employed in the fragment-complementation assays to detect protein– protein interactions. They further demonstrated noninvasive imaging of protein–protein interactions in living mice using a cooled charge-coupled device camera and split firefly luciferase [19]. This approach will allow continuous monitoring of protein–protein interactions in a particular biological pathway. Furthermore, this approach can be extended to detecting drugs that modulate protein–protein interactions in vivo. High throughput screening High throughput screening (HTS) has become an integrated component of today’s drug discovery process. Several HTS platforms have been constructed and are

387

commercially available. The ideal platform should be fast, automated, miniaturized, easy to use, compatible with small and large molecules, tolerant to various solvents, and should yield a lower number of false positive. Several available HTS assays possess these characteristics, but there is room for improvement. Therefore, there is ongoing research to develop new and improved HTS assays. Palmer et al. have developed a novel cell-based functional assay to monitor G protein-coupled receptor (GPCR) activation in a high-throughput format using the strategy of protein reassembly [20]. The strategy employs a pair of inactive b-galactosidase (b-gal) deletion mutants as fusion partners to the interacting proteins, b-arrestin and GPCR that can be activated by a ligand. To monitor GPCR activation, stable cell lines expressing both GPCR- and b-arrestin-b-gal fusion proteins are generated. Upon activation by ligand, b-arrestin binds to the activated GPCR, and this interaction allows functional reassembly of b-gal from its mutant fragments. The activation of GPCR is measured directly by quantitating restored b-gal activity. The validation of this strategy was demonstrated using two GPCRs: the b2-adrenergic amine receptor and the CXCR2 chemokine-binding receptor. Using a 96-well microtiter plate format the Library of Pharmacologically-Active Compounds (LOPAC, RBI-Sigma, Natick, MA, USA) was screened against the b2-adrenergic receptor cell line to identify both agonists and antagonists. This assay system has also been performed in a high-density (384-well) microtiter plate format. The assay system developed is a specific, sensitive, and robust method for studying and screening GPCR-mediated signaling pathways.

Applied Biosystems has marketed a high throughput screening platform based on enzyme complementation for the measurement of any analyte of interest, called HitHunter enzyme fragment complementation (EFC) (Fig. 4). The EFC assay is based on an engineered b-galactosidase enzyme that consists of two fragments, the enzyme acceptor (EA) and the enzyme donor (ED). These two fragments self associate to form active b-galactosidase. The EFC assay utilizes an ED-analyte conjugate in which the analyte may be recognized by a binding protein, receptors or an antibody. The ED-analyte conjugate can complement with EA to form active b-galactosidase, producing a luminescence signal. However, the binding of analyte to its respective binding protein prevents complementation of ED-analyte with EA, and therefore no luminescence is generated. Free analyte can compete with the ED-analyte conjugate for binding to the specific binding protein, thereby, releasing ED-analyte conjugate which can complement with EA, producing a signal dependent upon the amount of free analyte present in the sample. The EFC assay is a rapid, robust and homogeneous methodology for analyte detection. Furthermore, it is a highly sensitive assay and can be miniaturized and automated. This method can be used for the screening of library of compounds that are either activators or inhibitors of a receptor or an enzyme in drug discovery process. A disadvantage of this assay is that since it works on the principle of self-assembly, it may suffer form protein aggregation problems. Adenosine 3¢, 5¢-cyclic monophosphate (cAMP) is a tightly regulated signaling molecule involved in G protein coupled receptor (GPCR) activation. Therefore, measurement of intracellular cAMP concentrations is an

Fig. 4 Schematic of a Hit–Hunter Enzyme complementation assay

Analyte-ED

EA

Antibody

Active Enzyme

Inactive Enzyme

Active Enzyme Free Analyte

388

important assay in drug discovery processes. The EFC assay from Applied Biosystems was evaluated for measuring cAMP in cells overexpressing a Gas-coupled receptor Golla by and Seethala [21]. For that, the bgalactosidase (b-gal) donor fragment was conjugated to cAMP (ED-cAMP), which can complement with b-gal enzyme acceptor (EA) fragment to form an active b-gal enzyme. Binding of ED-cAMP conjugate to the anticAMP antibody prevented its complementation with the EA fragment to form an active enzyme. Addition of samples containing free cAMP allowed complementation of ED-cAMP with EA depending upon the levels of free analyte. The assay was then employed to monitor the activation of adenylate cyclase. When glucagon-like peptide (GLP)-1 binds to GLP-1 receptor (with a Kd of 0.2 nmol/l), it activates adenylate cyclase, which results in an increase in intracellular camp level (EC of 0.3 nmol/l). This increase in levels of cAMP allowed restoration of b-gal activity. The assay was further validated with forskolin, exendin, and several active GLP-1 peptide analogues. The adenylate cyclase inhibitors, MDL-12330A and SQ-22536, effectively inhibited the stimulation of cAMP by GLP-1 and forskolin. The assay was validated with several low molecular weight nonpeptide compounds and peptide agonists with different binding constants. The cAMP assay developed has a larger dynamic range of detection compared to other existing methods for cAMP detection. The assay is nonradioactive, sensitive, robust, has minimal interference from DMSO and colored compounds, and is amenable for automation. Nevertheless, self-complementation methodology suffers from the drawback of yielding false positives, which is a disadvantage of this type of assay. This is because, if the interactions between fragmented beta-galactosidase is stronger than the interaction between ligand and antibody complementation will occur even in the absence of free ligand. In addition this will yield very high background signal. Immunoassays Sandwich enzyme-linked immunosorbent assay (ELISA) is a routinely used technique in diagnostics laboratories due to high sensitivity, selectivity, low background and wide dynamic range. The limitation of this type of assays is the requirement of two epitopes on antigen. As an alternative to this technique, an open sandwich ELISA was developed by Ueda et al. [22]. In this assay, the antigen allows reassociation of the antibody with weak variable regions VH and VL by bridging between them. This method was further improved by attaching peptide fragments of the enzyme b-galactosidase to the VH and VL fragments. Binding of the antigen allows reassociation of the antibody variable region and also it leads to the complementation of fragments of the b-galactosidase yielding an active enzyme. For the purpose of assay, an antibody heavy chain variable region fragment was genetically fused to the N-terminal deletion mutant of

b-galactosidase and the light chain variable region was fused to the C-terminal deletion mutant of b-galactosidase. Different length linkers were engineered between the antibody chain and the b-gal fragment. Using antihen egg lysozyme antibody, an antigen dependent enhancement in enzyme activity was monitored. An optimum assay provided 1000-fold improvement in sensitivity compared to the traditional immunoassay method. The possible reason for improved sensitivity is the requirement of lower reagent concentration, which reduces the interchain interaction of the fusion proteins, which in turn enhances the effect of antigen at its lower concentration. The assay developed is a noncompetitive homogeneous type assay and, hence, is easy and quick to perform. One of the drawbacks of this assay is the requirement of a suitable antibody with a weakly interacting VH/VL region. This drawback can be overcome by using a phage display technique to select an antibody that is compatible with an open sandwich assay. In addition, mutations in the antibody sequence can be performed to yield weakly interacting VH/VL domains. Residues that can be mutated to alter interactions in variable region of antibody have also been identified. Antibody screening Commonly used methods of antibody screening include phage display, ribosome display, surface display on yeast and bacteria, and the yeast two-hybrid system. One of the disadvantages of these methods is the requirement of a purified antigen. To overcome this limitation and to develop a fast as well as less error prone method, Pluckthun et al. have suggested enzyme complementation assay as an alternate strategy [23]. In this assay, the antibody and the antigen are fused to two fragments of murine DHFR. Binding of the two interacting partners allows reassembly of the enzyme, thereby, allowing growth on minimal media containing methotrexate. Anti-GCN4 scFv antibody, which recognizes random-coil conformation of the yeast transcriptional activator, GCN4, containing leucine zipper peptide motif was fused to either of the two-mDHFR domains. GCN4 leucine zipper was also fused to either of the two-mDHFR domains. Amino acid linkers were engineered between mDHFR fragments and antibody/ antigen. The antibody containing plasmid was cotransformed with a plasmid-encoding antigen and plated under selective conditions. Several colonies (about 1·108) were obtained upon expression of the fusion protein between fragment of mDHFR and antigen/ antibody due to reassembly of mDHFR. As a negative control same experiment was performed without inducing expression of fusion proteins or without the specific antigens. As expected very few colonies (about 20) were obtained in this experiment (Table 1). Method describe here was also utilized to screen for other antigen-antibody pairs to demonstrate universal application of this technique. The method is simple requiring only

389 Table 1 Number of colonies obtained after cotransformation of the plasmid encoding antibody-DHFR and the plasmid encoding antigen-DHFR sequences Antibody-DHFR fusion

Specific antigenDHFR fusion

Non-specific antigenDHFR fusion

Anti-GCN4 Anti-FkpA Anti-HAG

3.3·108 5.6·107 1.3·108

20±13 20±7 13±13

three steps, namely, transformation of plasmid, expression of fusion protein, and analysis of colonies. Therefore, this method will be useful for library screening in a high throughput format. The major drawback of this system is that antibodies should be stable under reducing conditions since they are expressed in cytoplasm. In addition, the assay is performed in E. coli, therefore, preventing the analysis of post-translationally modified proteins. Molecular rulers The success of protein fragment reassembly strategy is dependant on proper folding of the enzyme. Therefore, by knowing the 3-D structure of the reporter, it is possible to predict the distance required between the two fragments for proper folding of the reporter. This can alternatively be used to study the conformational changes and allosteric transitions between interacting Protein–protein/protein-ligand pairs. This hypothesis was tested by Michnick et al. [11] to study a structural allosteric model for the ligand-induced activation of erythropoietin receptor (EpoR) [24]. Dimers of unliganded erythropoietin receptor exist in a conformation that prevents activation of JAK2. Upon ligand-induced conformational change, JAK2 gets activated. A fluorescence assay based on dimerization-induced complementation of fragments of the mDHFR was developed. Complementation of mDHFR upon interaction between fused pairs allows binding to the high-affinity fluorescein-conjugated inhibitor methotrexate (fMTX) in a 1:1 complex. The fMTX is retained in cells by this complex, whereas unbound fMTX is transported out of the cells, which can be monitored by fluorescence microscopy, fluorescence-activated cell sorting (FACS), or spectroscopy. EpoR receptor dimer transmembrane domains are separated by the distance observed in the crystal structure of unliganded EpoR (73 A˚). For proper folding of mDHFR fragments, it is essential that the NH2-termini of the DHFR fragments be 8 A˚ apart. Several flexible linker peptides of 5, 10, and 30 amino acids corresponding to lengths of 20, 40, and 120 A˚, respectively (4 A˚ per peptide bond) were inserted between the transmembrane domain and DHFR fragments. It was expected that with 5-amino acid (5-aa) linker, complementation of the DHFR fragments would only occur in the presence of ligands (Fig. 1b), with 10-amino acid (10-aa) linker, some complementation could occur in the

absence of ligand, but with 30-amino acid (30-aa) linker, it would be independent of ligand. Fluorescence flow cytometric analysis of the cotransfectants cells treated with the ligand erythropoietin and incubated with fMTX showed an eightfold increase in fluorescence over the background signal. In the absence of ligand, cells transfected with EpoR-DHFR fragment fused through 5-aa linkers showed no fluorescence, cells with a 10-aa linker showed a weak fluorescence, and cells with a 30-aa linker showed the same level of fluorescence in the presence or absence of ligand. These results demonstrated that ligand binding and induced conformational change is essential for EpoR activation. Furthermore, the protein reassembly strategy can be employed to probe the distance changes upon conformational changes in protein and can serve as the molecular ruler.

Biochemical pathways Profiling of the protein–protein interactions allow us to map biochemical pathways. There can be three steps involved in this process, screening, pharmacological profiling, and determining cellular location. All of these three steps were performed by using the survivalscreening fluorescence protein reassembly assay employing DHFR as the reporter [25]. Initial screening was performed on the basis that only the cells that have reassembled DHFR will survive in media depleted of nucleotides. Pharmacological profiling that involves studying effect of agonist/antagonist on protein–protein interaction was monitored by measuring changes in the fluorescence. In addition, fluorescence imaging allows one to find cellular location of these interactions. Analysis of the results allowed representation of biochemical pathway in time, space, and in response to specific stimuli. Using this method signal transduction pathway mediated by receptor tyrosine kinases was studied. The map generated was consistent with the known interactions and it also included additional novel interactions. Conclusions and future perspectives Methods developed based on protein reassembly have several advantages such as rapidity, homogeneous format, and ability to perform in vivo screening. In addition when the reassembly is assisted by interacting pairs of proteins used such as leucine zippers it reduces the possibility of false positives compared to the strategy where self-complementation is used for the reassembly of proteins as in the case of beta-galactosidase. Limitations of this method include, high background signal, possible interference in reassembly by the matrix components. Nevertheless, background signal is minimized in case of assisted reassembly since reporter does not self-assemble even at high concentration. The biggest challenge of this method is designing fragments of

390

reporter without affecting the 3-D structure of the protein drastically. Several unique reporters have been used in this technique, nevertheless, there is a need to find other reporter proteins with exquisite properties that can be fragmented easily and can be reassembled. Future applications of this technique can be imagined to include solving of more sophisticated problems such as protein design and folding. The method will also prove useful in mapping of biochemical pathways. In addition, application of the protein reassembly strategy in studying enzymes and in designing novel enzyme catalyst can be envisioned. The assays developed using protein reassembly show great promise for applications in the biochemical and the analytical fields. Nevertheless, this technique is still in its infancy and needs to be explored. Acknowledgements The author would like to thank Prof. Sylvia Daunert from the University of Kentucky for her guidance. This work was supported by grant number P42 ES 07380 from the National Institute of Environmental Health Sciences, National Institutes of Health, and National Science Foundation.

References 1. Richards FM, Vithayathil PJ (1959) J Biol Chem 234:1459– 1465 2. Gay GP, Fersht AR (1994) Biochemistry 33:7957–7963 3. Shiba K, Schimmel P (1992) Proc Natl Acad Sci U S A 89:1880–1884 4. Southworth MW, Adam A, Panne D, Byer R, Kautz R, Perler FB (1998) EMBO J 17:918–926

5. Pelletier JN, Campbell-Valois FX, Michnick SW (1998) Proc Natl Acad Sci USA 95:12141–12146 6. Remy I, Michnik SW (1999) Proc Natl Acad Sci U S A 96:5394–5399 7. Pelletier JN, Arndt KM, Pluckthun A, Michnick SW (1999) Nat Biotechnol 17:683–690 8. Karimova G, Pidoux J, Ullman A, Ladant D (1998) Proc Natl Acad Sci U S A 95:5752–5756 9. Johnsson N, Varshavsky A (1994) Proc Natl Acad Sci U S A 91:10340–10344 10. Rossi F, Charlton CA, Blau HM (1997) 94:8405–8410 11. Michnick SW, Remy I, Valois CFX, Belisle A, Pelletier JN (2000) Methods Enzymol 328:208–230 12. Galarneau A, Primeau M, Trudeau L, Michnick SW (2002) Nat Biotechnol 20:619–622 13. Deo SK, Daunert S (2001) Anal Biochem 289:52–59 14. Baird GS, Zacharias DA, Tsien RY (1999) Proc Natl Acad Sci U S A 96:11241–11246 15. Ghosh I, Hamilton AD, Regan L (2000) J Am Chem Soc 122:5658–5659 16. Chiu NHL, Christopoulos TK (1999) Clin Chem 45:1954–1959 17. Paulmurugan R, Gambhir SS (2003) Anal Chem 75:1584–1589 18. Pelletier JN, Remy I, Michnick SW (1998) J Biomol Technol 10:32–39 19. Paulmurugan R, Umezawa Y, Gambhir SS (2002) Proc Natl Acad Sci U S A 99:15608–15613 20. Yan Y, Boldt-Houle DM, Tillotson BP, Gee MA, D’Eon BJ, Chang X, Olesen CEM, Palmer MAJ (2003) J Biomol Screen 7:451–459 21. Golla R, Seethala R (2003) J Biomol Screen 7:515–525 22. Ueda H, Yokozeki T, Arai R, Tsumoto K, Kumagai I, Nagamune T (2003) J Immunol Methods 279:209–218 23. Mossner E, Koch H, Pluckthun A (2001) J Mol Biol 308:115– 122 24. Remy I, Wilson IA, Michnick SW (1999) Science 283:990–993 25. Remy I, Michnick SW (2001) PNAS 98:7678–7683