Brain Metabolic DNA: A Long Story and Some Conclusions.

We have previously outlined the main properties of brain metabolic DNA (BMD) and its involvement in circadian oscillations, learning, and post-trial sleep. The presence of BMD in certain subcellular fractions and their behavior in cesium gradients have suggested that BMD originates from cytoplasmic reverse transcription and subsequently acquires a double-stranded configuration. More recently, it has been reported that some DNA sequences of cytoplasmic BMD in learning mice are different from that of the control animals. Furthermore, BMD is located in vicinity of the genes involved in different modifications of synaptic activity, suggesting that BMD may contribute to the brain's response to the changing environment. The present review outlines recent data with a special emphasis on reverse transcription of BMD that may recapitulate the molecular events at the time of the "RNA world" by activating mitochondrial telomerase and generating RNA templates from mitochondrial transcripts. The latter unexpected role of mitochondria is likely to promote a better understanding of mitochondrial contribution to cellular interactions and eukaryotic evolution. An initial step regards the role of human mitochondria in embryonic BMD synthesis, which is exclusively of maternal origin. In addition, mitochondrial transcripts involved in reverse transcription of BMD might possibly reveal unexpected features elucidating mitochondrial involvement in cancer events and neurodegenerative disorders.

Brain metabolic DNA: recent evidence for a mitochondrial connection.

This review highlights recent data concerning the synthesis of brain metabolic DNA (BMD) by cytoplasmic reverse transcription and the prompt acquisition of the double-stranded configuration that allows its partial transfer to nuclei. BMD prevails in the mitochondrial fraction and is present in presynaptic regions and astroglial processes where it undergoes a turnover lasting a few weeks. Additional data demonstrate that BMD sequences are modified by learning, thus indicating that the modified synaptic activity allowing proper brain responses is encoded in learning BMD. In addition, several converging observations regarding the origin of BMD strongly suggest that BMD is reverse transcribed by mitochondrial telomerase.

Sequences of Reverse Transcribed Brain DNA Are Modified by Learning.

Brain metabolic DNA (BMD) is continuously synthesized by reverse transcription in presynaptic synaptosomes and astroglia, and is partly transferred to nuclei after acquiring the double stranded configuration. Synthesis and turnover of BMD are markedly dependent on brain activity, as shown by circadian oscillations, environmental enrichment and impoverishment, and a variety of learning protocols. In rodents learning a two-way active avoidance task, BMD synthesis doubles, thus raising the possibility that sequences of learning BMD may differ from control BMD. The hypothesis has now been examined by sequencing cytoplasmic BMD. The present data indicate that most high-quality mapped BMD fragments hosting more than seven sequences are present in all mice. Three of them are exclusively present in learning BMD and four in control BMD. In addition, the annotated genes closest to them are mostly involved in modulating synaptic activity. The data support the conclusion that learning BMD sequences encode brain responses to the modified environment.

Deregulated Local Protein Synthesis in the Brain Synaptosomes of a Mouse Model for Alzheimer's Disease.

While protein synthesis in neurons is largely attributed to cell body and dendrites, the capability of synaptic regions to synthesize new proteins independently of the cell body has been widely demonstrated as an advantageous mechanism subserving synaptic plasticity. Thus, the contribution that local protein synthesis at synapses makes to physiology and pathology of brain plasticity may be more prevalent than initially thought. In this study, we tested if local protein synthesis at synapses is deregulated in the brains of TgCRND8 mice, an animal model for Alzheimer's disease (AD) overexpressing mutant human amyloid precursor protein (APP). To this end, we used synaptosomes as a model system to study the functionality of the synaptic regions in mouse brains. Our results showed that, while TgCRND8 mice exhibit early signs of brain inflammation and deficits in learning, the electrophoretic profile of newly synthesized proteins in their synaptosomes was subtly different from that of the control mice. Interestingly, APP itself was, in part, locally synthesized in the synaptosomes, underscoring the potential importance of local translation at synapses. More importantly, after the contextual fear conditioning, de novo synthesis of some individual proteins was significantly enhanced in the synaptosomes of control animals, but the TgCRND8 mice failed to display such synaptic modulation by training. Taken together, our results demonstrate that synaptic synthesis of proteins is impaired in the brain of a mouse model for AD, and raise the possibility that this deregulation may contribute to the early progression of the pathology.

Brain Metabolic DNA Is Reverse Transcribed in Cytoplasm: Evidence by Immunofluorescence Analysis.

In a previous study (Mol Neurobiol 55:7476-7486, 2017), newly synthesized brain metabolic DNA (BMD) from rat subcellular fractions has been shown to behave as a DNA-RNA hybrid when analyzed in cesium gradients at early [3H] thymidine incorporation times but to assume the double-stranded configuration at later times. Conversely, BMD from purified nuclei displayed the dsDNA configuration even at early incorporation times. The results were interpreted to support the BMD origin by reverse transcription in the cytoplasm and its later acquisition of the double-stranded configuration before the partial transfer to the nuclei. This interpretation has now been confirmed by immunofluorescence analyses of newly synthesized BrdU-labeled BMD from the mouse brain that demonstrates its cytoplasmic localization and colocalization with DNA-RNA hybrids. In addition, BrdU-labeled BMD has been shown to colocalize with astroglial anti-GFAP antibodies and with presynaptic anti-synaptophysin antibodies.

DNA in Squid Synaptosomes.

The synthesis of brain metabolic DNA (BMD) is modulated by learning and circadian oscillations and is not involved in cell division or DNA repair. Data from rats have highlighted its prevalent association with the mitochondrial fraction and its lack of identity with mtDNA. These features suggested that BMD could be localized in synaptosomes that are the major contaminants of brain mitochondrial fractions. The hypothesis has been examined by immunochemical analyses of the large synaptosomes of squid optic lobes that are readily prepared and identified. Optic lobe slices were incubated with 5-bromo-2-deoxyuridine (BrdU) and the isolated synaptosomal fraction was exposed to the green fluorescent anti-BrdU antibody. This procedure revealed that newly synthesized BrdU-labeled BMD is present in a significant percent of the large synaptosomes derived from the nerve terminals of retinal photoreceptor neurons and in synaptosomal bodies of smaller size. Synaptosomal BMD synthesis was strongly inhibited by actinomycin D. In addition, treatment of the synaptosomal fraction with Hoechst 33258, a blue fluorescent dye specific for dsDNA, indicated that native DNA was present in all synaptosomes. The possible role of synaptic BMD is briefly discussed.

Brain Metabolic DNA in Rat Cytoplasm.

Brain metabolic DNA (BMD) is not involved in cell division or DNA repair but is modulated by memory acquisition, sleep processing, and circadian oscillations. Using routine methods of subcellular fractionation, newly synthesized BMD from male rats is shown to be localized in crude nuclear, mitochondrial, and microsomal fractions and in two fractions of purified nuclei. Sub-fractionation of the mitochondrial fraction indicates the prevalent localization of BMD in free mitochondria and to a lesser degree in synaptosomes and myelin. Cesium density profiles of homogenate, subcellular fractions, and purified nuclei obtained after incorporation periods from 30 min to 4 h indicate that BMD synthesis takes place by reverse transcription in cytoplasmic organelles. Following the acquisition of the double-stranded structure, BMD is transferred to nuclei. Kinetic analyses lasting several weeks highlight the massive BMD turnover in subcellular fractions and purified nuclei and its dependence on age. Data are in agreement with the role of BMD as a temporary information store of cell responses of potential use in comparable forthcoming experiences.

Squid Giant Axons Synthesize NF Proteins.

Squid giant axon has been an excellent model system for studying fundamental topics in neurobiology such as neuronal signaling. It has been also useful in addressing the questions of local protein synthesis in the axons. Incubation of isolated squid giant axons with [35S]methionine followed by immunoprecipitation with a rabbit antibody against all squid neurofilament (NF) proteins demonstrates the local synthesis of a major 180 kDa NF protein and of several NF proteins of lower molecular weights. Their identification as NF proteins is based on their absence in the preimmune precipitates. Immunoprecipitates washed with more stringent buffers confirmed these results. Our data are at variance with a recent study based on the same experimental procedure that failed to visualize the local synthesis of NF proteins by the giant axon and thereby suggested their exclusive derivation from nerve cell bodies (as reported by Gainer et al. in Cell Mol Neurobiol 37:475-486, 2017). By reviewing the pertinent literature, we confute the claims that mRNA translation is absent in mature axons because of a putative translation block and that most proteins of mature axons are synthesized in the surrounding glial cells. Given the intrinsic axonal capacity to synthesize proteins, we stress the glial derivation of axonal and presynaptic RNAs and the related proposal that these neuronal domains are endowed with largely independent gene expression systems (as reported by Giuditta et al. in Physiol Rev 88:515-555, 2008).

Brain metabolic DNA in memory processing and genome turnover.

Sophisticated methods are currently used to investigate the properties of brain DNA and clarify its role under physiological conditions and in neurological and psychiatric disorders. Attention is now called on a DNA fraction present in the adult rat brain that is characterized by an elevated turnover and is not involved in cell division or DNA repair. The fraction, known as brain metabolic DNA (BMD), is modulated by strain, stress, circadian oscillations, exposure to enriched or impoverished environment, and notably by several training protocols and post-trial sleep. BMD is frequently localized in glial cells but is also present in neurons, often in the perinucleolar region. Its distribution in repetitive and non-repetitive DNA fractions shows that BMD differs from native DNA and that in learning rats its profile differs from that of control rats. More detailed knowledge of the molecular, cellular, and time-dependent BMD features will be necessary to define its role in memory acquisition and processing and in the pathogenesis of neurologic disorders.

Brain synaptosomes harbor more than one cytoplasmic system of protein synthesis.

Synaptosomal protein synthesis from rat brain is selectively increased by learning and is massively enhanced during the recovery period from brain ischemia. To lay the groundwork for identification of the involved synaptic elements, we examined the effects induced by varying the concentrations of extracellular cations and endogenous calcium. Most of the recorded rate response curves exhibited biphasic profiles that suggested the presence of more than one translation system. Because comparable profiles were obtained by fully inhibiting mitochondrial translation, the data indicated the involvement of cytoplasmic translation systems present in different synaptosomal classes. Their properties may be individually investigated by exploiting the partially inhibited conditions we have described. The identification of the synaptic elements from which they originated and their newly synthesized proteins will significantly expand our understanding of the synaptic contribution to brain plastic events.

Local gene expression in nerve endings.

At the Nobel lecture for physiology in 1906, Ramón y Cajal famously stated that "the nerve elements possess reciprocal relationships in contiguity but not in continuity," summing up the neuron doctrine. Sixty years later, by the time the central dogma of molecular biology formulated the axis of genetic information flow from DNA to mRNA, and then to protein, it became obvious that neurons with extensive ramifications and long axons inevitably incur an innate problem: how can the effect of gene expression be extended from the nucleus to the remote and specific sites of the cell periphery? The most straightforward solution would be to deliver soma-produced proteins to the target sites. The influential discovery of axoplasmic flow has supported this scheme of protein supply. Alternatively, mRNAs can be dispatched instead of protein, and translated locally at the strategic target sites. Over the past decades, such a local system of protein synthesis has been demonstrated in dendrites, axons, and presynaptic terminals. Moreover, the local protein synthesis in neurons might even involve intercellular trafficking of molecules. The innovative concept of glia-neuron unit suggests that the local protein synthesis in the axonal and presynaptic domain of mature neurons is sustained by a local supply of RNAs synthesized in the surrounding glial cells and transferred to these domains. Here, we have reviewed some of the evidence indicating the presence of a local system of protein synthesis in axon terminals, and have examined its regulation in various model systems.

Sleep memory processing: the sequential hypothesis.

According to the sequential hypothesis (SH) memories acquired during wakefulness are processed during sleep in two serial steps respectively occurring during slow wave sleep (SWS) and rapid eye movement (REM) sleep. During SWS memories to be retained are distinguished from irrelevant or competing traces that undergo downgrading or elimination. Processed memories are stored again during REM sleep which integrates them with preexisting memories. The hypothesis received support from a wealth of EEG, behavioral, and biochemical analyses of trained rats. Further evidence was provided by independent studies of human subjects. SH basic premises, data, and interpretations have been compared with corresponding viewpoints of the synaptic homeostatic hypothesis (SHY). Their similarities and differences are presented and discussed within the framework of sleep processing operations. SHY's emphasis on synaptic renormalization during SWS is acknowledged to underline a key sleep effect, but this cannot marginalize sleep's main role in selecting memories to be retained from downgrading traces, and in their integration with preexisting memories. In addition, SHY's synaptic renormalization raises an unsolved dilemma that clashes with the accepted memory storage mechanism exclusively based on modifications of synaptic strength. This difficulty may be bypassed by the assumption that SWS-processed memories are stored again by REM sleep in brain subnuclear quantum particles. Storing of memories in quantum particles may also occur in other vigilance states. Hints are provided on ways to subject the quantum hypothesis to experimental tests.

Training old rats selectively modulates synaptosomal protein synthesis.

We have previously shown that the local synthesis of two synaptic proteins of 66.5-kDa and 87.6-kDa is selectively enhanced in male adult rats trained for a two-way active avoidance task. We report here that a comparable but not identical response occurs in 2-year-old male rats trained for the same task. In the latter age group, the local synthesis of the 66.5-kDa protein markedly increases in cerebral cortex, brainstem, and cerebellum, with a somewhat lower increment in synthesis of the 87.6-kDa protein. On the other hand, the newly synthesized 87.6-kDa protein correlates with avoidances and escapes and inversely correlates with freezings in cerebral cortex and brainstem, whereas the correlations of the newly synthesized 66.5-kDa protein remain below significance. These correlative patterns are sharply at variance with those present in trained adult rats. Our data confirm that the local system of synaptic protein synthesis is selectively modulated by training and show that the synaptic response of old rats differs from that of adult rats as reflected in behavioral responses.

Synaptosomal protein synthesis in P2 and Ficoll purified fractions.

Cytoplasmic protein synthesis of brain synaptosomes has generally been determined in the Ficoll purified fraction which contains fewer contaminating mitochondria, microsomes and myelin fragments than the parent P2 fraction. Using a highly selective assay of this activity we have compared the total translation activity and the specific activity of the proteins synthesized by either fraction in control rats and in rats trained for a two-way active avoidance task. In control rats the specific activity remained essentially the same in both fractions but in trained rats the value of the Ficoll fraction was markedly lower (38.5%) than in the P2 fraction. Furthermore, the total translation activity of the Ficoll fraction was 30% lower than in the P2 fraction in control rats and 62% lower in trained rats. These decrements indicate that a large proportion of active synaptosomes present in the P2 fraction is not recovered in the Ficoll fraction, notably in rats undergoing plastic brain changes. We conclude that cytoplasmic protein synthesis of brain synaptosomes is better preserved in the P2 fraction.

Protein synthesis in nerve terminals and the glia-neuron unit.

The progressive philogenetic lengthening of axonal processes and the increase in complexity of terminal axonal arborizations markedly augmented the demands of the neuronal cytoplasmic mass on somatic gene expression. It is proposed that in an adaptive response to this challenge, novel gene expression functions developed in the axon compartment, consisting of axonal and presynaptic translation systems that rely on the delivery of transcripts synthesized in adjacent glial cells. Such intercellular mode of gene expression would allow more rapid plastic changes to occur in spatially restricted neuronal domains, down to the size of individual synapses. The cell body contribution to local gene expression in well-differentiated neurons remains to be defined. The history of this concept and the experimental evidence supporting its validity are critically discussed in this article. The merit of this perspective lies with the recognition that plasticity events represent a major occurrence in the brain, and that they largely occur at synaptic sites, including presynaptic endings.

Synaptic mRNAs are modulated by learning.

We have recently demonstrated that brain plastic events significantly modify synaptic protein synthesis measured by the incorporation of [(35)S]methionine in brain synaptosomal proteins. Notably, in rats learning a two-way active avoidance task, the local synthesis of two synaptic proteins was selectively enhanced. Because this effect may be attributed to transcriptional modulation, we used reverse transcriptase-polymerase chain reaction methods to determine the content of discrete synaptosomal mRNAs in rats exposed to the same training protocol. Correlative analyses between behavioral responses and synaptosomal mRNA content showed that GAT-1 mRNA (a prevalent presynaptic component) correlates with avoidances and escapes in rat cerebellum, while glial fibrillary acid protein mRNA (an astrocytic component) correlates with freezings in cerebellum and cerebral cortex. These observations support the hypothesis that synaptic protein synthesis may be transcriptionally regulated. The cellular origin of synaptic transcripts is briefly discussed, with special regard to those present at large distances from neuron somas.

Natural computing and biological evolution: a new paradigm.

After a brief outline of the available hypotheses on the mechanism of biological evolution, attention is called on the global nature of the variations leading to the generation of new species. Integrated changes may hardly be attributed to beneficial random mutations of single traits even if assisted by a phylogenetic elimination of poorly adapted individuals. Rather, integrated variations are likely to reflect the outcome of cybernetic algorithms (natural computing) operating on organism's resources and impending environmental changes. As all organisms are endowed with computing capacities that modulate and integrate ontogenetic development and maintenance of biological functions, structures, and behaviors, these capacities are assumed to have moulded the evolutionary variations of organisms, and their transfer to the progeny.

Local gene expression in axons and nerve endings: the glia-neuron unit.

Neurons have complex and often extensively elongated processes. This unique cell morphology raises the problem of how remote neuronal territories are replenished with proteins. For a long time, axonal and presynaptic proteins were thought to be exclusively synthesized in the cell body, which delivered them to peripheral sites by axoplasmic transport. Despite this early belief, protein has been shown to be synthesized in axons and nerve terminals, substantially alleviating the trophic burden of the perikaryon. This observation raised the question of the cellular origin of the peripheral RNAs involved in protein synthesis. The synthesis of these RNAs was initially attributed to the neuron soma almost by default. However, experimental data and theoretical considerations support the alternative view that axonal and presynaptic RNAs are also transcribed in the flanking glial cells and transferred to the axon domain of mature neurons. Altogether, these data suggest that axons and nerve terminals are served by a distinct gene expression system largely independent of the neuron cell body. Such a local system would allow the neuron periphery to respond promptly to environmental stimuli. This view has the theoretical merit of extending to axons and nerve terminals the marginalized concept of a glial supply of RNA (and protein) to the neuron cell body. Most long-term plastic changes requiring de novo gene expression occur in these domains, notably in presynaptic endings, despite their intrinsic lack of transcriptional capacity. This review enlightens novel perspectives on the biology and pathobiology of the neuron by critically reviewing these issues.