Molecules and Memories

Abstract

MOLECULES AND MEMORIES BERNARD W. AGRANOFF* We believe we are studying the molecular basis of behavior. Perhaps this presentation will add strength to the thesis that neurochemistry is at the threshold of membership in that new sect, "molecular biology." Among our credentials are the following: (a) We use biochemical tools to study fundamental biological phenomena, (b) We find it exciting, (c) We cannot isolate reaction products. Ofcourse, the term "molecular" has already been used freely. There is, for example, ajournai concerned with "Molecular Pathology." Soon we may see on our bookshelves the "Molecular Psychopathology of Everyday Life," and perhaps "The Molecular Anatomy ofMelancholy." In this essay, the behavioral response will be considered as follows: An external stimulus is converted to electrical impulses in a sense organ. Interposed between this encoding process and the output, or "read out"— i.e., the electrical impulses leaving a central nerve apparatus to the muscles involved in the observed response—lies the associative mechanism. By a selection or instruction process, the brain makes decisions and records, in some as yet unknown form, information for a particular response which becomes, with time, preferred. We call this information memory. It has been compared to the biological information storage seen in immune body formation, enzyme induction, and the genetic process [i]. Most of our present knowledge of the nature of nerve impulses and interactions is based on electrophysiological observations. A single light flash can be detected electrically along the optic radiation and, by averaging techniques, evoked responses from repetitive external stimuli can be * Mental Health Research Institute and Department ofBiological Chemistry, The University of Michigan, Ann Arbor, Michigan 48104. My co-workers on the work with goldfish are Dr. Roger E. Davis and Dr.JohnJ. Brink (Interdisciplinary Training Fellows under USPHS training grant No. 5T7-MH-7417), Miss Patty Bright, and Mr. Paul Klinger. The research is sponsored by a National Science Foundation grant. 13 detected from the scalp ofan intact animal [2]. Experiments on the eye of limulus [3] andthefrog [4] indicatethat encoding ofstimuli can behighly selective. Whether learning can occur in sensory organs is not yet known. By comparison, chemical correlates ofnerve activity are crude. Increased temperature over the occipital cortex has been observed only after repeated massive visual stimulation [5]. Under similar conditions, increased local diffusion ofinert solutes has also been observed autoradiographically [6, 7]. Recently, increased incorporation of amino acids in the motor neurons ofthe rat spinal cord after prolonged exercise has been observed by autoradiography [8]. By means of microchemical measurements, changes in the RNA ofDeiter cells and the surrounding glia have been reported in rats following vestibular stimulation. Specific changes in the composition of RNA are also reported following training [9]. The chemical or electrical nature ofmemory has eluded clarification. It may not reside exclusively in the brain. Studies with transected planarians [10] suggest the existence ofa diffuse memory. Surgical lesions [11] as well as local electrical stimulation [12] and injections (see below) indicate some degree oflocalization in the temporal and hippocampal regions of higher animals. In what special way might chemical approaches be ofvalue in studies on memory? The answer lies in the hypothesis that memoryformation (the fixation ofexperience) involves theformation ofcovalent chemical bonds. It is supported by the following: i. The evolutionary experience ofa species is storedin DNA, the genetic "memory" molecule. DNA is assembled by formation of covalent chemical linkages and it does not appear to undergo metabolic turnover in the decoding process. The behavioral model would by analogy require molecules or structures which do not turn over and would thus accrete in the brain with time [13]. Some brain lipids fulfil these requirements. (It is known that brain cells do not contain on the average more DNA than other somatic cells [14].) Evidence for changes in the size ofthe cortex [15] is compatible with such an accretion process. We could classify the formation ofnew synapses as a type ofaccretion. A diminution process is also conceivable, in which irrelevant interneuronal connections are destroyedasafunctionofexperience . Suchapossibilityisratherdiscouraging from the standpoint ofchemical detection. It is easier to find the difference between 0 and 1 than between 100 and 101. 14 Bernard W. Agranoff · Molecules and Memories Perspectives in Biology and Medicine · Autumn 196...