ASSAY and Drug Development TechnologiesVolume 1, Number 2, 2003© Mary Ann Liebert, Inc.

Literature Search and Review

J. Fraser Glickman

In each issue of ASSAY and Drug Development Technologies, our Literature Editor, J. Fraser Glickman,Ph.D., selects several significant papers covering timely and pertinent topics that will keep our readers up-to-date on the ever-changing field of assay and drug development technologies. Dr. Glickman will provide rel-evant commentary on each of the cited abstracts.

High Throughput Mice

385

Peters LL, Cheever EM, Ellis HR, Magnani PA, Sven-son KL, Von Smith R, Bogue MA: Large-scale, high-throughput screening for coagulation and hemato-logic phenotypes in mice. Physiol Genomics 2002;11:185–193.

Abstract: The Mouse Phenome Project is an internationaleffort to systematically gather phenotypic data for a de-fined set of inbred mouse strains. For such large-scale proj-ects the development of high-throughput screening proto-cols that allow multiple tests to be performed on a singlemouse is essential. Here we report hematologic and coag-ulation data for more than 30 inbred strains. Completeblood counts were performed using an Advia 120 analyzer.For coagulation testing, we successfully adapted the DadeBehring BCS automated coagulation analyzer for use inmice by lowering sample and reagent volume require-ments. Seven automated assay procedures were developed.Small sample volume requirements make it possible to per-form multiple tests on a single animal without euthanasia,while reductions in reagent volume requirements reducecosts. The data show that considerable variation in manybasic hematological and coagulation parameters exists

among the inbred strains. These data, freely available onthe World Wide Web, allow investigators to knowledge-ably select the most appropriate strain(s) to meet their in-dividual study designs and goals.

Commentary: Inbred mouse strains have become im-portant models for the understanding of the genetic ba-sis of complex diseases. For example, screening of ran-domly mutagenized inbred strains can reveal phenotypesthat match human disease states. However, in order tofully utilize sequence information from the mouse genomeproject, it is necessary to precisely define what is the nor-mal phenotype and rapidly identify what is an alteredphenotype among a large population of individuals. Pe-ters et al. present high-throughput methods to analyzeblood chemistry and biology in mice. Seven different as-say procedures are described that rapidly measure var-ious coagulation factors and complete blood cell counts,for example. The methods require low sample volume, sothat euthanasia is not necessary. The 42 most commoninbred strains were tested in these parameters. The sys-tem described would be useful in determining those mu-tations that have effects on hematological parameters.

� �

Kinase Substrates

Brinkworth RI, Breinl RA, Kobe B: Structural basisand prediction of substrate specificity in protein ser-ine/threonine kinases. Proc Natl Acad Sci U S A2003;100:74–79.

Abstract: The large number of protein kinases makesit impractical to determine their specificities and sub-strates experimentally. Using the available crystal struc-tures, molecular modeling, and sequence analyses of ki-

nases and substrates, we developed a set of rules gov-erning the binding of a heptapeptide substrate motif(surrounding the phosphorylation site) to the kinase andimplemented these rules in a web-interfaced programfor automated prediction of optimal substrate peptides,taking only the amino acid sequence of a protein kinaseas input. We show the utility of the method by analyz-ing yeast cell cycle control and DNA damage check-point pathways. Our method is the only available pre-

dictive method generally applicable for identifying pos-sible substrate proteins for protein serine/threonine ki-nases and helps in silico construction of signaling path-ways. The accuracy of prediction is comparable to theaccuracy of data from systematic large-scale experimen-tal approaches.

Commentary: The identification of increasing numbersof protein kinases that are important in human pathol-ogy has resulted in a demand for methods that rapidlyidentify their cognate substrates. This area of researchis also important to those seeking to map out signaltranduction pathways and to develop physiologicallyrelevant assays. Brinkworth et al. have identified a com-putational algorithm for predicting oligopeptides thatwill act as substrates for serine/threonine kinases. Con-ventional approaches to identifying kinase substrates

utilize phage display techniques, which can be very timeconsuming.

The computational method relies upon input of the com-plete amino acid sequence of the kinase, and uses a modelbased on the known three-dimensional structures of thecatalytic domain of several classes of serine/threonine ki-nases bound to cognate substrates. The known kinaseshave certain specificity determinants that can be used astemplates for the unknown kinase. The system was able topredict with high probability a good peptide substrate forthe enzymes, and also to probe the human genome forprobable physiological protein substrates. The system wastested against a known phage-display methodology andwas found to agree well with this technique. Such meth-ods would be useful to those interested in developing ser-ine/threonine kinase assays, and needing a tool for rapidlydesigning peptide substrates for synthesis.

Literature Search and Review386

Zero-Mode Waveguides

Levene MJ, Korlach J, Turner SW, Foquet M, Craig-head HG, Webb WW: Zero-mode waveguides for sin-gle-molecule analysis at high concentrations. Science2003;299:682–686.

Abstract: Optical approaches for observing the dynamicsof single molecules have required pico- to nanomolar con-centrations of fluorophore in order to isolate individual mol-ecules. However, many biologically relevant processes oc-cur at micromolar ligand concentrations, necessitating areduction in the conventional observation volume by threeorders of magnitude. We show that arrays of zero-modewaveguides consisting of subwavelength holes in a metalfilm provide a simple and highly parallel means for study-ing single-molecule dynamics at micromolar concentra-tions with microsecond temporal resolution. We present ob-servations of DNA polymerase activity as an example ofthe effectiveness of zero-mode waveguides for performingsingle-molecule experiments at high concentrations.

Commentary: Levene et al. have developed methods thatare alternatives to confocal fluorescence correlation spec-troscopic techniques for observing single-molecule bind-ing and enzymatic processes. The events are observedthrough zero-mode waveguides and are more suited tomeasuring the activity of enzymes and ligands with mil-limolar affinity, because the analysis is performed at vol-umes much lower than those of confocal detectors. Themethod is based on the assays being performed in small(subwavelenth) holes in a metal film deposited in a mi-croscope coverslip, upon which millions of such holes canbe made. The fluorescence correlation approach for mea-suring ligand binding depends upon the measurement ofaltered diffusion times in a very small volume of mea-surement—the difficulty in its high throughput being the

long times for data acquisition and the susceptibility to ar-tifacts. By using zero-mode waveguides, the observationvolumes are reduced to such a small volume that the timeit takes for diffusion out of the chamber is extremely re-duced, and thus the temporal resolution is improved, al-lowing for faster acquisition of fluorescence correlationdata. In proof-of-concept experiments, fluorescence cor-relation was carried out on DNA polymerase activity us-ing fluorescently coumarin-labeled dCTP as a probe. Thesystem seems to extend the ability to measure enzyme ac-tivities that require high concentrations of substrate. Thetechnology is still in its infancy and would require somedevelopment in order to bring it to the level of robustnessneeded for routine assay development.

Illumination

Dichroic filter

Collected fluorescence

Fluorescent ligand

Enzyme

Metal film

Fused silica

Reprinted from Levene MJ, Korlach J, Turner SW, Foquet M, Craig-head HG, Webb WW: Zero-mode waveguides for single-molecule anal-ysis at high concentrations. Science 299:682–686. © 2003.

Jaiswal JK, Mattoussi H, Mauro JM, Simon SM:Long-term multiple color imaging of live cells usingquantum dot bioconjugates. Nat Biotechnol 2003;21:47–51.

Abstract: Luminescent quantum dots (QDs)—semicon-ductor nanocrystals—are a promisingalternative to or-ganic dyes for fluorescence-based applications. We havedeveloped procedures for using QDs to label live cellsand have demonstrated their use for long-term multicolorimaging of live cells. The two approaches presented are(i) endocytic uptake of QDs and (ii) selective labeling ofcell surface proteins with QDs conjugated to antibodies.Live cells labeled usingthese approaches were used forlong-term multicolor imaging. The cells remained stablylabeled for over a week as they grew and developed.These approaches should permit the simultaneous studyof multiple cells over long periods of time as they pro-ceed through growth and development.

Wu X, Liu H, Liu J, Haley KN, Treadway JA, Lar-son JP, Ge N, Peale F, Bruchez MP: Immunofluores-cent labeling of cancer marker Her2 and other cellu-lar targets with semiconductor quantum dots. NatBiotechnol 2003;21:41–46.

Abstract: Semiconductor quantum dots (QDs) are amongthe most promising emerging fluorescent labels for cellu-lar imaging. However, it is unclear whether QDs, whichare nanoparticles rather than small molecules, can specifi-cally and effectively label molecular targets at a subcellu-lar level. Here we have used QDs linked to immunoglob-ulin G (IgG) and streptavidin to label the breast cancermarker Her2 on the surface of fixed and live cancer cells,to stain actin and microtubule fibers in the cytoplasm, andto detect nuclear antigens inside the nucleus. All labelingsignals are specific for the intended targets and are brighterand considerably more photostable than comparable or-ganic dyes. Using QDs with different emission spectra con-jugated to IgG and streptavidin, we simultaneously detectedtwo cellular targets with one excitation wavelength. The re-sults indicate that QD-based probes can be very effectivein cellular imaging and offer substantial advantages overorganic dyes in multiplex target detection.

Literature Search and Review 387

Commentary: Quantum dots are nanometer-sized crys-tals of inorganic semiconductors that have unusual flu-orescent properties. Unlike small organic fluorophoresand fluoresent proteins, these crystals can be excited atbroad wavelengths and have much narrowed emissionspectra. The dots can be successfully conjugated to pro-teins and nucleic acids, thus expanding the potential forfluorescent probes of cellular and biomolecular function.Because of the unusual fluorescent properties associatedwith these nanocrystals, novel optical setups (detectors)can be envisioned, and, at the very least, these moleculeswill add to the palette of fluorescent labeling tools usedin developing cellular and biochemical assays.

Quantum Dots

(A) Nuclear antigens in the nuclei of human epithelial cells were labeledwith ANA, anti-human IgG-biotin and QD 630-streptavidin. (B) Whennormal human IgGs were used in place of ANA, no detectable stain wasobserved. (C) The nuclei 3T3 cell was stained with ANA, anti-humanIgG-biotin, and QD 630-streptavidin (red). The microtubules were la-beled with mouse-anti-a-tubulin antibody, anti-mouse IgG-biotin, andQD 535-streptavidin (green). (D) Her2 on the surface of SK-BR-3 cellswas stained green with mouse anti-Her2 antibody and QD 535-IgG(green). Nuclear antigens were labeled with ANA, anti-human IgG-bi-otin and QD 630-streptavidin (red). Filter sets ex 420 nm/em 535-10 nmand ex 560. Reprinted from Wu X, et al: Immunofluorescent labeling ofcancer marker Her2 and other cellular targets with semiconductor quan-tum dots. Nature Biotechnology 21:41–46. © 2003.

Literature Search and Review388

Spark

Polacek N, Swaney S, Shinabarger D, Mankin AS:SPARK—a novel method to monitor ribosomal pep-tidyl transferase activity. Biochemistry 2002;41:11602–11610.

Abstract: The key enzymatic activity of the ribosome iscatalysis of peptide bond formation. This reaction is atarget for many clinically important antibiotics. However,the molecular mechanisms of the peptidyl transfer reac-tion, the catalytic contribution of the ribosome, and themechanisms of antibiotic action are still poorly under-stood. Here we describe a novel, simple, convenient, andsensitive method for monitoring peptidyl transferase ac-tivity (SPARK). In this method, the ribosomal peptidyltransferase forms a peptide bond between two ligands,one of which is tritiated whereas the other is biotin-tagged. Transpeptidation results in covalent attachmentof the biotin moiety to a tritiated compound. The amountof the reaction product is then directly quantified usingthe scintillation proximity assay technology: binding ofthe tritiated radioligand to the commercially availablestreptavidin-coated beads causes excitation of the bead-embedded scintillant, resulting in detection of radioac-tivity. The reaction is readily inhibited by known antibi-otics, inhibitors of peptide bond formation. The methodwe developed is amenable to simple automation, whichmakes it useful for screening for new antibiotics. The

method is useful for different types of ribosomal research.Using this method, we investigated the effect of muta-tions at a universally conserved nucleotide of the activesite of 23S rRNA, A2602 (Escherichia coli numbering),on the peptidyl transferase activity of the ribosome. Theactivities of the in vitro reconstituted mutant subunits,though somewhat reduced, were comparable with thoseof the subunits assembled with the wild-type 23S rRNA,indicating that A2602 mutations do not abolish the abil-ity of the ribosome to catalyze peptide bond formation.Similar results were obtained with double mutants car-rying mutations at A2602 and another universally con-served nucleotide in the peptidyl transferase center,A2451. The obtained results agree with our previous con-clusion that the ribosome accelerates peptide bond for-mation primarily through entropic rather than chemicalcatalysis.

Commentary: SPA technologies first made their impacton the world of HTS; however, one can envision this typeof technology being useful in the understanding of thebasic mechanisms of enzymes, which have been difficultto measure using older technologies. Polacek et al. pres-ent a novel method for easily measuring peptidyl trans-ferase activity using the scintillation proximity principle.The peptidyl transferase activity in the ribosome ofprokaryotic organisms is essential for protein synthesis

SPARK-PmnA

f-M e t

E

70S

P A

P m n

AU G

37 °Cf-M e t Pm n

f-M e t

E

70S

50S

P A

AU G A AA

37 °Cf-M et Lys

Lys

SPARK-2TbC

SPARK-2TaB

SPARK-50SD

E

70S

P A

P he

UU U U UU

37 °C 0 °C

M e O HP he

E P A

Ph e Ph e

Bio tin

[3 H ]-la b el

Ph e Ph e

Versions of SPARK. 3H-Labeled substrates are shown in red and biotinylated ligands in green. (A) SPARK-Pmn is preformed with 70S ribo-somes carrying a 3H-labeled formyl-Met-tRNA (f-Met) at the P-site and biotin-puromycin (Pmn) at the A-site. (B) In SPARK-2Ta, Lys-tRNA bio-tinylated at the «-amino group of lysine (Lysbiot-tRNA) serves as the acceptor (A-site) substrate, while formyl-[3H] Met-tRNA functions as the donorin the reaction performed by 70S ribosomes. (C) SPARK-2Tb employs unlabeled N-biotinylated Phe-tRNA and 3H-labeled Phe-tRNA as P- and A-site substrates for poly(U)-programmed 70S ribosomes. (D) The peptidyl transfer reaction in SPARK-50S is promoted by the large ribosomal subunitin the presence of N-biotinylated Phe-tRNA and 3H-labeled Phe-tRNA as donor and acceptor substrates, respectively. The reaction product, diphenyl-alanyl-tRNA, is the same in SPARK-2Tb and SPARK-50S. Reprinted with permission from Polacek N, Swaney S, Shinabarger D, Mankin AS: SPARK:A novel method to monitor ribosomal peptidyl transferase activity. Biochemistry 2002;41:11602–11610. © 2002 American Chemical Society.

and therefore survival of the microbe, thus making it anattractive target for anti-infective therapy. The activity ismeasured using a tritiated formyl methionine and a bio-tinylated Lys-tRNA or biotinylated puromycin. Forma-

Literature Search and Review 389

tion of a peptide bond between the two substrates resultsin energy transmitted to a streptavidin SPA bead. The as-say is demonstrated to be useful and consistent with priorstudies of enzyme mechanism.

SOPs

Mostacero JA, Taylor PB, Ramon F, Ashman S, Wil-son JM, Baddeley SM, Hertzberg RP, Bartram SL,Pope AJ, Battle CD, Macarron R, Bond BC, ClementsYM, Gaul NJ, McAllister WE. A standard operatingprocedure for assessing liquid handler performancein high-throughput screening. J Biomol Screen2002;6:554–569.

Abstract: The thrust of early drug discovery in recentyears has been toward the configuration of homogeneousminiaturized assays. This has allowed organizations tocontain costs in the face of exponential increases in thenumber of screening assays that need to be run to remaincompetitive. Miniaturization brings with it an increasingdependence on instrumentation, which over the past sev-eral years has seen the development of nanodispensingcapability and sophisticated detection strategies. Tomaintain confidence in the data generated from minia-turized assays, it is critical to ensure that both compoundsand reagents have been delivered as expected to the tar-get wells. The authors have developed a standard oper-ating procedure for liquid-handling quality control that

has enabled them to evaluate performance on 2 levels.The first level provides for routine daily testing on ex-isting instrumentation, and the second allows for morerigorous testing of new dispensing technologies. The pro-cedure has shown itself to be useful in identifying bothmethod programming and instrumentation performanceshortcomings and has provided a means to harmonizinginstrumentation usage by assay development and screen-ing groups. The goal is that this type of procedure be usedfor facilitating the exchange of liquid handler perfor-mance data across the industry.

Commentary: Automated laboratories are always in theprocess of evaluating new technologies, and centralamong these technologies are liquid handling apparatus.Poor liquid handling performance can lead to errors indata interpretation and result in false positives and falsenegatives in the screening campaign, therefore requiringthorough testing and calibration of liquid handling in-strumentation. Mostacero et al. present a very nice studyof methodologies to measure and calibrate such liquidhandling performance.

Drug Release

Shabbits JA, Chiu GN, Mayer LD: Development ofan in vitro drug release assay that accurately predictsin vivo drug retention for liposome-based delivery sys-tems. J Control Release 2002;84:161–170.

Abstract: The therapeutic activity of numerous drugs canbe dramatically improved by liposomal encapsulation.However, this requires that liposomes retain their encapsu-lated drugs following systemic administration. Often, invitro drug release assays do not accurately predict the li-posomal drug retention properties observed in vivo. Wepostulate that this discrepancy is due to the large membranepool present in blood cells and tissues, into which drugscan distribute after in vivo administration. Herein we de-scribe an in vitro drug release assay that more accuratelypredicts in vivo drug release from liposomes following sys-temic administration. Drug-encapsulated large unilamellarvesicles (LUVs) approximately 100 nm in diameter wereincubated with a 100–fold excess of multilamellar vesicles(MLVs) containing 300 mM sucrose, which served as “ac-

ceptors” for drug release and transfer from “donor” LUVs.Following incubation at 37°C, the donor and acceptor pop-ulations were separated with greater than 90% efficiencyby centrifugation at 1600xg for 10 min. The amount of drugin the MLV pellet reflects the degree of drug leakage fromthe donor liposomes. Drug release profiles using this in vitroassay were compared to those obtained using dialysis-basedassays and in vivo results following systemic administra-tion to mice. Our results indicate that this release assay isa better predictor of in vivo drug transfer than dialysis-basedsystems. We also demonstrate its utility in measuring ex-change of lipophilic components.

Commentary: Delivery of compounds using liposomeshas become an important tool for getting compounds intothe patient that are not normally orally bioavailable.However, predicting the performance of certain liposomecompositions has proven difficult. Shabbits et al. have de-vised a more accurate assay for predicting the deliveryof liposomes in vivo.

Holzenberger M, Dupont J, Ducos B, Leneuve P,Geloen A, Even PC, Cervera P, Le Bouc Y: IGF-1 re-ceptor regulates lifespan and resistance to oxidativestress in mice. Nature 2003;421:182–187.

Abstract: Studies in invertebrates have led to the identi-fication of a number of genes that regulate lifespan, someof which encode components of the insulin or insulin-likesignalling pathways. Examples include the related tyrosinekinase receptors InR (Drosophila melanogaster) and DAF-2 (Caenorhabditis elegans) that are homologues of themammalian insulin-like growth factor type 1 receptor(IGF-1R). To investigate whether IGF-1R also controlslongevity in mammals, we inactivated the IGF-1R genein mice (Igf1r). Here, using heterozygous knockout micebecause null mutants are not viable, we report thatIgf1r( 6 ) mice live on average 26% longer than theirwild-type littermates (p , 0.02). Female Igf1r( 6 ) micelive 33% longer than wild-type females (p , 0.001),whereas the equivalent male mice show an increase inlifespan of 16%, which is not statistically significant.Long-lived Igf1r( 6 ) mice do not develop dwarfism,

their energy metabolism is normal, and their nutrient up-take, physical activity, fertility and reproduction are un-affected. The Igf1r( 6 ) mice display greater resistanceto oxidative stress, a known determinant of ageing. Theseresults indicate that the IGF-1 receptor may be a centralregulator of mammalian lifespan.

Commentary: Mouse genetic manipulations have led toa better understanding of the genes and proteins involvedin the aging process. Many of these studies suggest thatoxidative stress is a major cause of reduced lifespan. Fur-thermore, mice deficient in certain growth factors showdwarfism, but also have prolonged lifespan. Holzen-berger et al. have made a heterozygous knockout in theIGF receptor and shown normal growth, metabolism, andreproduction, but significantly increased lifespan. Fur-thermore, the animals showed an increase in their resis-tance to oxidative stress demonstrated by resistance tointraperitoneal paraquot injection. This knowledge mightaid in the search for active compounds, for example, forpreventing tissue damage during inflammation.

Literature Search and Review390

Forever Young

Nuclear Receptors

Kassam A, Miao B, Young PR, Mukherjee R: RetinoidX receptor (RXR) agonist-induced antagonism of far-nesoid X receptor (FXR) activity due to absence ofcoactivator recruitment and decreased DNA binding.J Biol Chem 2003;278:10028–10032.

Abstract: The bile salt export pump (BSEP) plays an in-tegral role in lipid homeostasis by regulating the canalicu-lar excretion of bile acids. Induction of BSEP gene ex-pression is mediated by the farnesoid X receptor (FXR)which binds as a heterodimer with the retinoid X receptor(RXR) to the FXR response element (FXRE) located up-stream of the BSEP gene. RXR ligands mimic several part-ner ligands and show additive effects upon coadministra-tion. Using real-time quantitative PCR and cotransfectionreporter assays, we demonstrate that the RXR agonist,LG100268, antagonizes induction of BSEP expression me-diated by endogenous and synthetic FXR ligands, CDCAand GW4064, respectively. Moreover, this antagonism is ageneral feature of RXR agonists and is attributed to a de-crease in binding of FXR/RXR heterodimers to theBSEP-FXRE coupled with the inability of RXR agoniststo recruit coactivators to FXR/RXR. Our data suggest thatFXR/RXR is a conditionally permissive heterodimer andis the first example of RXR ligand-mediated antagonismof FXR activity. Since FXR agonists lower triglyceride

levels, our results suggest a novel role for RXR-mediatedantagonism of FXR activity in the development of hy-pertriglyceridemia observed with RXR agonists in ro-dents and humans.

Commentary: The nuclear receptors for farnesoids andretinoids are essential components for normal homeo-static regulation of lipid biosynthesis, catabolism, andabsorption. The receptors regulate the genes involved intransporting and metabolizing cholesterol and otherlipids and act by binding a cognate ligand, which inducesa complex with a coactivator protein and DNA in orderto modulate transcription of genes involved in lipid me-tabolism. Understanding the signal transduction mecha-nisms of such receptors and determining the appropriatemode of intervention of therapeutic agents is further com-plicated by the fact that these nuclear receptors act asdimers or heterodimers with other nuclear receptors, thusenabling a single ligand to have a variety of potential ef-fects depending on its selectivity, in combination with theparticular composition of the receptor dimer within thecell. Kassam et al. have nicely demonstrated that ago-nists of the RXR homodimer can act as antagonists of theRXR/FXR heterodimer, and that this works by decreas-ing the binding of the heterodimer to the FXRE of theDNA.