Memory and Cellular Control Processes
237
bably, glia possibly), that mRNA being partly or wholly metabolically
stable and having been transcribed from inherited DNA. Third, memory
resides in the quantities or conformations of some cellular constituent
other than RNA, for example structural or enzymic proteins, but the process
oflaying it down involves the synthesis of some new, metabolically unstable,
mRNA. These last two hypotheses, or combinations of them, are perfectly
plausible, but there are also possible mechanisms for memory which use
only previously synthesized mRNA or protein (see, e.g., [29]) or even
none directly at all.
Then we have to decide how specific the transfer is [15]. For example,
suppose that memory is "set" by the release of some agent 5 whose quantity
depends on the degree of reinforcement [29]. Then an extract from a brain
in which much reinforcement had recently occurred might contain a large
amount of 5, which would therefore act as a "setting agent" in a brain into
which it was injected and thus improve memory. Such improvement would
presumably be primarily interpreted as "savings" rather than "retention",
but if there were several different setting agents, 5 1 ,5 2 , .•• , 5 m say, acting
specifically on different cells, or synapses, then effects mimicking learning,
and having some degree of specificity, could be produced.
Finally, if the essential effective material transferred is RNA, this fact
by itself still does not enable one to distinguish between the three hypotheses described above. For, even in the third, the transfer agent could be
RNA, namely that newly-synthesized, metabolically unstable, mRNA
mentioned there.
d) Axonal Protein Synthesis
If the laying down of long-term memory involves an action on the
genetic regulatory elements of cells, this would imply an influence at some
stage on the process of transcription from DNA to RNA or on protein
synthesis. It is natural to suppose that this must take place in or near the
nucleus of a cell, but this is not necessarily the case. It is now known that
mitochondria contain DNA and, almost certainly, a complete proteinsynthesizing apparatus, including ribosomes [2, 8, 45]. As each synaptic
knob usually contains a mitochondrion [48], it follows that it is possible
to envisage a control involving transcription or translation in the mitochondrion. The level of activity of the respiratory enzymes in the mitochondrion would be expected to affect the size and efficacy of its synaptic
knob and, therefore, inducers or repressors acting on the mitochondrion
would give a convenient method of control over the synapse. Whether
this is so or not remains, of course, to be seen, but the possibility should
be borne in mind in considering the effects on memory of puromycin and
the like.
237
bably, glia possibly), that mRNA being partly or wholly metabolically
stable and having been transcribed from inherited DNA. Third, memory
resides in the quantities or conformations of some cellular constituent
other than RNA, for example structural or enzymic proteins, but the process
oflaying it down involves the synthesis of some new, metabolically unstable,
mRNA. These last two hypotheses, or combinations of them, are perfectly
plausible, but there are also possible mechanisms for memory which use
only previously synthesized mRNA or protein (see, e.g., [29]) or even
none directly at all.
Then we have to decide how specific the transfer is [15]. For example,
suppose that memory is "set" by the release of some agent 5 whose quantity
depends on the degree of reinforcement [29]. Then an extract from a brain
in which much reinforcement had recently occurred might contain a large
amount of 5, which would therefore act as a "setting agent" in a brain into
which it was injected and thus improve memory. Such improvement would
presumably be primarily interpreted as "savings" rather than "retention",
but if there were several different setting agents, 5 1 ,5 2 , .•• , 5 m say, acting
specifically on different cells, or synapses, then effects mimicking learning,
and having some degree of specificity, could be produced.
Finally, if the essential effective material transferred is RNA, this fact
by itself still does not enable one to distinguish between the three hypotheses described above. For, even in the third, the transfer agent could be
RNA, namely that newly-synthesized, metabolically unstable, mRNA
mentioned there.
d) Axonal Protein Synthesis
If the laying down of long-term memory involves an action on the
genetic regulatory elements of cells, this would imply an influence at some
stage on the process of transcription from DNA to RNA or on protein
synthesis. It is natural to suppose that this must take place in or near the
nucleus of a cell, but this is not necessarily the case. It is now known that
mitochondria contain DNA and, almost certainly, a complete proteinsynthesizing apparatus, including ribosomes [2, 8, 45]. As each synaptic
knob usually contains a mitochondrion [48], it follows that it is possible
to envisage a control involving transcription or translation in the mitochondrion. The level of activity of the respiratory enzymes in the mitochondrion would be expected to affect the size and efficacy of its synaptic
knob and, therefore, inducers or repressors acting on the mitochondrion
would give a convenient method of control over the synapse. Whether
this is so or not remains, of course, to be seen, but the possibility should
be borne in mind in considering the effects on memory of puromycin and
the like.