TALKING POINT Biochem. Sci. 20, 102-103 7 Doyle, D. A. et al. (1996) Cell 85, 1067-1076 8 Morais Cabral, J. H. et al. (1996) Nature 382, 649-652 9 Harrison,S. C. (1996) Cell 86, 341-343 10 Kim, E. et al. (1995) Nature 378, 85-88 11 Kornau, H-C., Schenker, L. T., Kennedy,M. B. and Seeburg, P. H. (1995) Science 269, 1737-1740 12 Kim, E., Cho, K-O., Rothschild,A. and Sheng, M. (1996) Neuron 17, 103-113 13 M(Jller,B. M. et al. (1996) Neuron 17, 255-265 14 Niethammer, M., Kim, E. and Sheng, M. (1996) J. Neurosci. 16, 2157-2163 15 Hemming, N. J. et al. (1995) J. Biol. Chem. 270, 5360-5366 16 Gumbiner, B. M. (1995) Curt. Opin. Cell Biol. 7, 634-640 17 Matsumine, A. et al. (1996) Science 272, 1020-1023 18 Perrimon, N. (1996) Cell 86, 513-516 19 Axelrod,J. D., Matsuno, K., Artavanis-Tsakonas,S. and Perrimon, N. (1996) Science 271, 1826-1832 20 Simske, J. S., Kaech, S. M., Harp, S. A. and Kim, S. K. (1996) Cell85, 195-204 21 Shieh, B-H. and Zhu, M. Y. (1996) Neuron 16, 991-998 TIBS 21 - DECEMBER 1996 22 Itoh, N. and Nagata, S. (1993) J. Biol. Chem. 268, 10932-10937 23 Sato, L, Irie, S., Kitada, S. and Reed, J. C. (1995) Science 268, 411-415 24 Gomperts, S. N. (1996) Cell 84, 659-662 25 Brenman,J. E. et al. (1996) Cell 84, 757-767 JAN SARAS AND CARL-HENRIKHELDIN Ludwig Institute for Cancer Research, Box 595, Biomedical Centre, S-751 24 Uppsala, Sweden Kinetic analysis of biosensor data: elementary tests for self-consistency Peter Schuck and Allen P. Minton The validity of the most common kinetic interpretation of biosensor data can be quickly assessed with the aid of two simple tests for self-consist- ency, requiring only back-of-the-envelope calculations. A search of the recent literature reveals that many published results fail these tests qualitatively. THE USE OF evanescent wave biosen- sors to study associations between sol- uble macroligands and immobilized ac- ceptors has increased greatly during the last five years, following the introduc- tion of commercially manufactured in- struments (BIAcore AB, Mfinity Sensors IAsys). The kinetic data obtained from the biosensor is most frequently inter- preted in the context of the simple binding model: k+ L+A ~ LA k_ where L denotes mobile ligand and A represents immobilized acceptor. We1,2 and others 3-6 have suggested a variety of chemical and instrumental reasons why it might not be valid to account for a particular set of data P. Schuck and A. P. Minton are at the Section of Physical Biochemistry, Laboratory of Biochemical Pharmacology, National Institute of Diabetes, Digestive and Kidney Diseases, National institutes of Health, Bethesda, MD 20892, USA. Emaih pschuck@helix.nih.gov or minton@helix.nih.gov 458 using this model, including neglect of mass transport, steric hindrance and/or the possibility of more-complex binding schemes. The purpose of the present communication is twofold: (1) to de- scribe two simple tests for the internal self-consistency of results obtained from analysis of biosensor data in the context of this elementary model; and (2) to argue that all such analyses be sub- jected to these tests before either sub- mission or acceptance for publication. Assume that the concentration of free ligand remains constant at a value of L0 throughout the time course of the association phase of the experiment, and at a value of zero throughout the time course of the dissociation phase of the experiment. With the additional conventional assumption that the dif- ference between the biosensor signal, R, and the baseline signal, R0, is propor- tional to the time-dependent concen- tration of LA, the reaction scheme above leads to the following descrip- tions of the time course of the observed signal R (Ref. 7): Association: R(t) = R0,a+ (R~a- R0,a) [1 - exp(-kobst)] (1) Published by Elsevier ScienceLtd Dissociation: R(t) = R~,d + (Ro, d - Roo,d)exp (-k_t) (2) where R0a denotes the signal at the start of tfie association experiment; R a denotes the signal at infinite time in th~ association experiment O.e. at attain- ment of association equilibrium); R0, d denotes the signal at the start of the dissociation experiment; and R, d de- notes the signal at infinite time in the dissociation experiment (i.e. when reversibly bound ligand has been fully dissociated) and: kobs = k+L 0 + k (3) Typically, association experiments are carried out at several different values of L 0. For each experiment, the values of kob s and R a can be evaluated either from linear regression of a plot of dR/dt vs R (Ref. 8) or by directly fitting an ex- pression of the form of Eqn 1 to the raw data 7. The values of k and k_ are then determined by linear regression of the dependence of kob s on L 0. The value of k_ might be obtained from the dissociation experiment by di- rectly fitting an expression of the form of Eqn 2 to the raw dissociation data 7, or by fitting a straight line to a logarith- mic first-order dissociation plot 8. If the simple binding model is correct, then the values of R a obtained at different values of L 0 ~hould be related by a simple Langmuir iso- therm s* . L0 R~,a(L0) = RO, a + [Rsa t - Ro,a]~ (4) K~"§ L o The values of K~q,the equilibrium dis- sociation constant, and Rsat, the signed *The Langmuirisotherm describes the dependence of reversiblybound ligand upon free ligand concen- tration at equilibrium that is expected for multiple identical sites binding ligand independently of each other. PII: S0968-0004(96)20025-8