Prolonged Neuroepithelial Stem Cell Proliferation and Human Brain Size

In a previous blog post, I discussed the role of Notch signaling in regulating human brain development. Here in this blog post, I discuss evidence for a role of the ZEB2 transcription factor protein in regulating growth of the neuroepithelium during embryogenesis. In the 2021 article “An early cell shape transition drives evolutionary expansion of the human forebrain“, Silvia Benito-Kwiecinski et al suggested that regulated expression of ZEB2 (shifting its activity to later times, after more neural stem cells have formed) provides a way to produce a larger cerebral cortex in humans compared to other apes.

Image source: the Graphical Abstract from the 2021 article by Silvia Benito-Kwiecinski et al.

ZEB2 transcripts (Figure 6)

As shown in the image above, Silvia Benito-Kwiecinski et al proposed that the human cerebral cortex might grow particularly large in humans (compared to chimps and gorillas) in part due to delayed expression of the ZEB2 protein during human embryogenesis.

This research by Silvia Benito-Kwiecinski et al made use of cerebral organoids as a model system that allows investigation of growth and development of neuroepithelial stem cells in culture. In the image to the right, ZEB2 transcripts per million (TMP) total RNA transcripts are shown for human and gorilla cerebral organoids grown for the indicated number of days.

Similar results were found for ZEB2 protein levels (See the image below).

ZEB2 protein levels (Figure 6, panel B)

As shown in their Figure 6, panel B, ZEB2 protein levels peaked around day 10 in human organoid cultures. However, ZEB2 protein levels were already high by day 5 for gorilla organoids.

From Figure S2, Panel A, 10d organoids.

The image to the right illustrates the kind of morphological difference observed between human and gorilla cerebral organoids.

The human cerebral organoids are larger by day 10 in culture. The red immunostaining in Figure S2 is the tight junction protein ZO-1 which is concentrated at the apical side of the cells (the apical side faces the interior of the organoids).

Starting at about day 5 in culture, the cells in the gorilla cerebral organoids underwent a morphological change and had smaller apical membrane domains (see below, their Figure 2). Cells in the human organoids did not complete that morphological transition until about day 10 in culture.

Change of cell morphology by day 5 of gorilla cerebral organoid culture (Fig. 2).

As shown in their Figure 2, at about the time when some neuronal cells began to differentiate inside the organoids, there was a change in shape of many of the neuroepithelial cells.

As shown in the image above, green GFP labeling of cells allowed for visualization of the shape change in cell processes that accompanies the smaller apical domains.

As reviewed by Judith C. Birkhoff et al in 2021, the ZEB2 transcription regulatory protein had been linked to brain development in mammals. In humans, defects in the ZEB2 gene can cause Mowat-Wilson Syndrome which causes abnormal development of the central nervous system in humans.

Figure 7, panel E. Day 5 of organoid culture.

As shown in their Figure 7, by inducing expression of ZEB2 at earlier times in human cerebral organoids (+Dox), the organoids and stem cells took on the morphological properties of cells in the 5 day gorilla organoids.

These results suggest the hypothesis that ZEB2 functions slightly differently in humans and nonhuman apes. During early human nervous system development, delay in expression of ZEB2 seems to allow expansion of the neuroepithelium, resulting in larger brain structures.

ZEB2 interacts with SMAD proteins to regulate cell differentiation.

Additional experiments were performed in order to block some of the effects of ZEB2 expression in gorilla cerebral organoids. ZEB2 interacts with SMAD proteins and other studies (example) have suggested that signals downstream from BPM receptors can influence the ability of ZEB2 to regulate cell differentiation events (see the proposed signal transduction pathway in the image to the right).

Silvia Benito-Kwiecinski et al exposed gorilla cerebral organoids to BMP4 and observed a delay in the neuroepithelial cell morphological changes described above. They also exposed gorilla cerebral organoids to lysophosphatidic acid. Lysophosphatidic acid, acting via LPA receptors, is a known regulator of cell proliferation and neural differentiation (for example, see this article). They concluded: “Application of LPA on gorilla organoids at the onset of NE transition led to morphologies more closely mimicking the human phenotype”.

Related Reading: “A Human‐Specific De Novo Gene Promotes Cortical Expansion and Folding“.

Human Brain Evolution

Model of a role for Wnt in the brain that is disrupted in Alzheimer’s disease (source

One of the early posts to this blog was called “Entry into the age of genes” and I mentioned the Wnt cell-to-cell signaling molecule and its possible involvement in the disrupted synaptic function that is associated with Alzheimer disease. In the image shown to the right on this page, interaction of Wnt with its receptor is proposed to regulate abnormal production of the amyloid-β (Aβ) peptide.

Here, in this blog post, I want to explore the role for Notch signalling in the human brain and the idea that recent evolution of Notch-related genes may have played an important role in human brain evolution, contributing to our unique human cognitive abilities.

The Notch signaling mechanism (source).

In the figure shown to the left on this page, a Notch receptor protein is shown interacting with a ligand on an adjacent cell. Such Notch signaling is one of the fundamental mechanisms that controls cell differentiation (cell fate) during embryonic development, including the production of neurons in the human brain.

In 2018, two research teams [1], [2] made news by presenting evidence that Notch signal-modulating proteins encoded by genes on human chromosome 1 (1q21) have played an important role in the evolution of the unusually large human brain.

Extra neurons (brightly stained cells ) are produced in fruit fly larvae when Notch does not control cell fate choices (source). Top (Notch⁺): normal Notch signals. Bottom (Notch^–): no Notch activity. Insets: Notch protein detected by immunostaining.

The story of both Wnt signaling (1911) and Notch signaling began over a hundred years ago with the isolation of mutated fruit flies that showed interesting alterations to the normal developmental program of Drosophila. The Wnt and Notch signaling pathways can interact to coordinate neuronal differentiation (see this recent article). Notch signaling seems to have played an important role in the evolution of metazoans. Most metazoan clades have a single Notch gene, but vertebrates often have four related Notch genes. Drosophila only has one Notch gene. The “classical” phenotype of a fly with no Notch protein is an over-production of neurons during larval development (shown in the image, above).

Structures of Drosophila Notch and the four human Notch proteins (source).

Analysis of the human genome showed that humans have four Notch genes that can code for proteins with large extracellular domains that are similar to the single Notch receptor protein of Drosophila. The human Notch-2 gene (NOTCH2) is located on the p arm of chromosome 1 (1p12).

Locations of the NOTCH2 gene and the 1q21.1 region on human chromosome 1 (source).

Predicted domain structures of the NOTCH2NL proteins compared to the larger structure of the transmembrane Notch-2 protein. EGF-like domains are shown in pink. NOTCH2NLB has a conventional signal sequence, yellow (source).

At some point in the great ape lineage leading to humans (see the image, below), there was a partial duplication of the NOTCH2 gene which resulted in a non-functional gene. Later, uniquely in the human lineage, that defective gene was re-activated, creating the NOTCH2NL gene with a functional promoter that is similar to that of NOTCH2. In the human lineage there was subsequent duplication of the ancestral NOTCH2NL gene, creating a small gene family with three NOTCH2NL family members, located on the q arm of chromosome 1, at 1q21.1 (NOTCH2NLA) or 1q21.2 (NOTCH2NLB and NOTCH2NLC). It appears that the human NOTCH2NLB gene codes for a soluble, secreted protein (image to right) that may interact with Notch receptors during embryogenesis and modulate Notch-signaling and neuronal development [1], [2].

The origins of Notch2NL genes during human evolution (source).

NOTCH2NL seems to delay brain development, but ultimately results in more cortical neurons, possible accounting for human neotenic brain development [1].

There are other human-specific neuronally-expressed genes in the 1q21 region. In 2017 it was suggested that the HYDIN2 gene might play a special role in human evolution and brain development. However, in 2018 it was reported that the the apparent association of HYDIN2 with mutations in human patients who had either microcephaly or macrocephaly was cooincidental [1]. “Mapping of these atypical patients indicates that the NOTCH2NLB locus was fully duplicated in all three atypical duplication cases (all macrocephalic, 1q21 duplication syndrome), and fully deleted in all three atypical deletion cases (all microcephalic, 1q21 deletion syndrome).”

This year, Ian T. Fiddes was again first author on another published article about NOTCH2NL genes. In this 2019 article, they discuss the fact that each NOTCH2NL gene is closely paired with an adjacent NBPF gene.

Paired NOTCH2NL and NBPF genes on chromosome 1 (source).

NBPF genes are of interest because they are expressed in the brain and they contain multiple copies of a protein domain that has been known as the DUF1220 domain (recently renamed as the Olduvai domain).

It has been suggested that NBPF proteins can promote the proliferation of neuronal stem cells. A correlation has been found between the number of copies of the Olduvai domain present in a genome and the brain size of various primates (see the image below).

Wnt and Notch have been suggested to be part of a brain regulatory network that, if perturbed, can cause autism. There is also a reported correlation between 1q21 gene copy number and autism. The suggestion has been made that Olduvai domain copy number correlates with symptom severity in autism.

Primate brain size correlates with Olduvai domain copy number (source).

Linked NBPF and NOTCH genes (source).

In their May 2019 article, Fiddes et al note that the human-specific duplication of NOTCH2NL genes had greatly increased the Olduvai domain copy number in humans compared to non-human primates. As shown in the image below, there are several homologous Olduvai domains (CON1, CON2, CON3, HLS1, HLS2 and HLS3, see the image below). Most of the Olduvai domains in NBPF proteins are triplets containing one copy of each HLS domain. Very little is known about the function of NBPF proteins and Olduvai domains.

NBPF proteins and their Olduvai domains (source). The four NOTCh2NL-linked NBPF genes are outlined in red.

source

Sikela and Quick have proposed that during evolution there has been positive selection for more copies of the genes containing Olduvai domain. In their model (see the image below), gene duplications in the 1q21 region are a source of good fortune for humans, helping create our large brains, but also a source of misfortune. They propose that alterations in 1q21 genes can play a role in modulating the symptoms of both schizophrenia and autism. As shown in the image to the right, the number of Olduvai domain copies in the genomes of people with autism is quite variable. It has been reported that there are correlations between severity of autism symptoms and Olduvai domain copy number.

source

There are many other candidate genes that might also influence the unique pattern of human brain development. Future work should reveal the relative importance of NOTCH2NL, NBPF and other genes in producing the human brain.

Notch expression in hair follicles (source)

In this blog post, I have emphasized the role of Notch signaling in neural development and the evolution of the unique human brain. However, Notch signaling is used to control the development of many tissues. For example, Notch regulates stem cell proliferation and differentiation in the skin and Notch is needed for normal hair follicle development. NOTCH2NL is expressed in human skin so it will be interesting to see if other unique human phenotypic features such as our “naked” skin and our large number of sweat glands can be linked to the recent evolution of NOTCH2NL genes.

Build a bigger brain (source).

There are some folks who are not afraid to test the effects of human genes on brain development in non-human primates (example). It will be interesting to see if the insertion of NOTCH2NL genes into the genomes of other primates will lead to enlarged brains. We have to wonder: if three NOTCH2NL genes are good for humans, maybe 4 would be better?

Related Reading: The Driver of Extreme Human-Specific Olduvai Repeat Expansion Remains Highly Active in the Human Genome.

2022 UPDATE on the Notch2nl locus.

Also: The human-specific paralogs SRGAP2B and SRGAP2C differentially modulate SRGAP2A-dependent synaptic development

2023 UPDATE: Brain developmental and cortical connectivity changes in transgenic monkeys carrying the human-specific duplicated gene SRGAP2C.

The Chemiosmotic Theory

image source

This blog post is a celebration of the contribution to theoretical biology that was made by Peter D. Mitchell. He won a Nobel Prize in 1978 for “his contribution to the understanding of biological energy transfer through the formulation of the chemiosmotic theory“. Most of the cellular energy (ATP) in humans is generated by chemiosmotic coupling of electron transport to ATP synthesis. Many patients have diseases due to problems in their chemiosmotic coupling.

1978 was an important year for me. I had just begun my university studies and I was fascinated by the mystery of how human brains make use of “natural” languages such as English. I grew up as a native speaker of English and only knew a small amount of Spanish and Latin. I had recently learned several computer programming languages: APL, Basic and FORTRAN. Back in those days, running computer programs was frustrating because of the inherent slowness of computers. VLSI was just arriving; processor speeds were slow and fast memory storage was expensive.

Fictional machines that talk

Some artificial intelligence researchers such as Terry Winograd had shown that the primitive computers of the 1970s were too stupid to make use of natural languages like English. I dreamed of a future time when it might become possible to simulate human language behavior with computers, but that was the domain of science fiction. It wasn’t hard to decide that I was better off studying the biology of brains than trying to teach a clunky digital computer to speak. Maybe if we better understood how cellular and molecular components of brains make human language behavior possible then we might eventually find a way to make artificial lifeforms that could also use human language.

Brain Biology. I was intrigued by the fact that rapid transmission of information in brains depends on the spread of action potentials along the surface membrane of axons. Action potentials had been first explored and characterized by electrophysiologists, but it seemed obvious that they were eventually going to be studied in much finer detail by molecular biologists who would identify the genes that code for ion channel proteins.

Removing trans-membrane proteins from the lipid bilayer (image source)

model of the nicotinic acetylchioline receptor

In 1978, very little was known about membrane proteins such as ion channels. Researchers were just starting to learn how to efficiently use detergents to disrupt cell membranes, allowing the individual membrane proteins to be isolated and studied with nanoscale precision (example).

At my university, there were no research opportunities in labs studying neuronal ion channels, but I was able to do my senior research project in a lab that was studying a bacterium (Chloroflexus aurantiacus) with a flexible electron transport chain that could be powered by both light energy and respiration. It was the beauty of Mitchell’s chemiosmotic theory that attracted my interest to biological electron transport and its remarkable links to the transport of ions across cell membranes.

Back in the 1970s, the membrane protein components of the mitochondrial electron transport chain were being investigated (example), but progress was slow. Human mitochondrial disease conditions had been biochemically associated with some of the electron transport components (example). Finally, towards the end of the 20th century, the phenotypes of mitochondrial disease patients could finally be accounted for in terms of specific genetic defects (example). Similarly, many human neurological diseases such as some forms of epilepsy can now be understood in terms of molecular defects in ion channel proteins (see this recent article). In the past 40 years, much has been learned about trans-membrane transport; it provides an interesting contrast to reflect on just how little was known back in 1978 when Mitchell won the Nobel Prize.

Progress in understanding: the journey from 1978 to 2018

Electron transport chain block model – Mitchell – 1978

The role of ubiquinone in the electron transport chain was particularly mysterious. In the diagram above, Mitchell’s chemiosmotic model included a vague role for ubiquinone (Q) as an electron carrier. How a hydrophobic lipid-like molecule (ubiquinone) was involved in proton pumping by NADH:ubiquinone oxidoreductase (NADH deH in Mitchell’s diagram) was impossible to determine without detailed information about the protein components of the mitochondrial electron transport chain such as respiratory complex I (herafter just “complex I”). The diagram above shows how Mitchell thought about the electron transport chain at the time of his Nobel Prize award. The molecular details of the electron transport proteins were unknown, so Mitchell showed “complex I” as a box in the mitochondrial membrane. He knew the box contained protein subunits that had molecular components such as iron-sulfur complexes (FeS) and flavine mononucleotide (FMN), but that was about it. Most significant for the chemiosmotic model of oxidative phosphorylation was the hydrogen ion (H⁺) shown emerging from “complex I”. Mitchell was the first person to realize why electron transport chains are located in membranes. Electron transport chain components like “complex I” pump protons (H⁺) across the membrane, storing energy in a transmembrane electrochemical gradient that was somehow derived from the available energy arising from oxidation-reduction reactions.

The first low-resolution molecular model of a transmembrane protein. Henderson & Unwin (1975)

During the 1980s, the structural details of membrane proteins began to be revealed. In 1988, a Nobel Prize was awarded for the high-resolution x-ray crystallography determination of a photosynthetic reaction center (see). The electron transport chains of photosynthetic organisms contain proton pump proteins just like the respiratory electron transport chain of mitochondria. Several parallel advances in available research methods made it possible to finally determine the atomic-level structure of transmembrane proteins.

In addition to advances in the chemical alchemy required to remove membrane proteins from their native membranes, keep them in their functional conformations and then coax them into forming crystals, better sources of x-rays became available (see). Computers became more powerful, allowing for digital x-ray reflection detectors (replacing the use of film) and computer-generated modeling of large protein complexes. Better gene sequencing methods made available the amino acid sequences of large membrane proteins and those sequences could be matched with the x-ray data to make molecular models.

2018 – molecular model of Complex I     (image source)

2018.  Forty years after Mitchell won his Nobel Prize, we have a much more detailed understanding of the structure of “complex I” (see the image above; source) than was possible during Mitchell’s lifetime in the previous century. In this model (above), the quinone (Q) molecule interacts with a specific and structurally well defined Q-binding pocket inside the ND1 protein subunit of “complex I”. Electrons are first handed from NADH to FMN and then sent along a chain of linked FeS centers to the quinone molecule. The reduced QH2 (ubiquinol ) can then exit from the quinone binding pocket, carrying electrons on to “complex III“, another proton pump of the electron transport chain.

In the model of Kaila, the movement of the reduced quinone molecule out of the Q-binding pocket is coupled to the pumping of protons across the membrane by means of conformational changes in several additional transmembrane protein subunits (ND2, ND4, ND5). With the help of these multiple proton-transporting protein subunits, “complex I” of the electron chain can maximize the number of transported protons according to the amount of available energy when electrons are sent from NADH to QH2. The molecular basis of human defects in the electron transport chain can now be taken to the level of understanding how specific amino acid changes disrupt the function of “complex I” (example) and the other parts of the mitochondrial electron transport chain.

ATP synthase (image source)

The energy stored in the proton gradient that is created by the mitochondrial electron transport chain is used to make ATP. Mitochondrial ATP synthesis is accomplished by the ATP synthase shown above (video of the enzyme in action). Transport of protons (H⁺) through the membrane part of the ATP synthase protein complex is coupled to production of ATP.

proton gradient-powered flagellum

Mitchell was the first person to realize that there need be no direct link between the proteins of an electron transport chain and the ATP synthase. The proton gradient can spread along a topologically closed membrane and also be used for other purposes such as powering the bacterial flagellum.

Since Mitchell’s time, it has been found that protons are not the only ions used for transmembrane ion gradients that power ATP synthesis. Some organisms have primary sodium or chloride ion pumps, so the more general term “ion-gradient-driven phosphorylation” needs to be used in the context of some life forms.

In the 1950s, when it was realized that the molecular components of the mitochondrial electron transport chain were mostly membrane-bound proteins and included the hydrophobic membrane quinone molecule, most of the enzymologists who were studying oxidative phosphorylation had no interest in membranes. Peter Mitchell was interested in the transport of molecular substances across cell membranes and as an “outsider” he was able to think more broadly and flexibly than the “insiders” who studied electron transport.

calcium channel structural model (source)

From our perspective here in this millennium, it is fun to look back through time and wonder how much Mitchell’s unorthodox chemiosmotic hypothesis sped up research into oxidative phosphorylation. That is hard to estimate, but there are other benefits from beautiful ideas in biology. The beauty of the DNA double helix and the idea of complementary base pairing provided inspiration and conceptual guidance for a new generation of molecular biologists. For me, Mitchell’s chemiosmotic theory provided inspiration for the exploration of how membrane transport proteins power information flow in the brain.

During the past 40 years much progress has been made in understanding the molecular details of the ion channel proteins that power axonal action potentials and that link action potential conduction to neurotransmitter release from synapses. Understanding the molecular complexities of synapses and how they make possible the storage of memories is the great remaining challenge for neurobiologists.

image source

ribosome structural model

With existing methods, it has been possible to reveal the molecular details of large multi-protein complexes such as ribosomes and electron transport chains. New tools might be needed to allow us to reveal and understand in detail how molecular complexes of synaptic proteins (example) change during learning and memory storage in the brain.

ion gradient-driven phosphorylation

There are remaining challenges at the level of 1) protein-protein interactions in individual synapses (example), 2) how all the synapses on a neuron interact to control the behavior of that neuron (example) and 3) how many spatially isolated neurons cooperate to make possible the learning of new behaviors and the storage of memories. Maybe there will be a few more theorists like Peter Mitchell who will be able to guide us through future exploration of these challenging research fields. Eventually we will understand how it is that children so easily learn to speak and maybe we will be able to make robots who can also learn to speak to us.

Related Reading: a rotating Na+-coupled ATP Synthase

Na+-pumping NADH-ubiquinone oxidoreductase from Vibrio cholerae

Synapse Disruption & Alzheimer Disease

Neurons make Aβ: presenilin (PS)-mediated proteolytic processing of APP

Brain dysfunction in Alzheimer Disease (AD) has been associated with abnormal production of the amyloid-β (Aβ) peptide. Aβ is generated by proteolytic processing from the amyloid precursor protein (APP). APP is a member of the APP gene family which codes for a small family of essential membrane proteins that are expressed at synapses in the brain.

Intact proteins of the APP family might function at synapses, for example as cell adhesion proteins. The Aβ fragment of APP might also have normal functions as a synaptic signal molecule, so there have been extensive efforts to identify synaptic receptors that can bind Aβ. One such receptor for Aβ is the cellular prion protein (PrP^C).

Among the human gene variants that can influence the incidence and progression of AD are:

1. presenilin (PS) mutations that alter the rate of production of Aβ and cause AD
2. an APP mutation that alters processing of APP by PS and protects against AD
3. a common PrP^C variant that is associated with rapid progression of AD (see)

image source

Results from experiments that were published in 1998 (see this article) suggested a role for FYN tyrosine kinase in the neurotoxic effects of  pathogenic Aβ oligomers. There is also evidence (2009) from experiments using PrP^C knockout mice that suggest pathogenic oligomers of Aβ can disrupt normal synaptic function by binding to membrane complexes containing the PrP^C protein. Since 2014 (see this article), the lab of  Stephen Strittmatter has been investigating the idea that Aβ oligomers can regulate FYN kinase activity by binding to PrP^C (or, possibly, other proteins that interact with PrP^C) and altering the function of mGluR5 receptors (see the image, above).

image source

2017. A recent article from the Strittmatter lab (Silent Allosteric Modulation of mGluR5 Maintains Glutamate Signaling while Rescuing Alzheimer’s Mouse Phenotypes) presents evidence for the ability of a drug (a silent allosteric modulator, SAM, BMS-984923) to bind mGluR5 receptors and inhibit neurotoxicity arising from  Aβ oligomers.

The mGluR5 receptor-binding drug BMS-984923 is called a “silent allosteric modulator” because it does not block the normal function of mGluR5 receptors to respond to the neurotransmitter glutamate (Glu).

Intracellular tau aggregation may be related to Aβ-induced neurotoxicity and AD pathology.

In contrast, conventional inhibitors of mGluR5 receptors (NAMs, see the image, above) disrupt normal glutaminergic synaptic neurotransmission. Results from Strittmatter et al suggest that BMS-984923 might prevent disregulation of FYN kinase, hyperphosphorylation of tau protein and neurotoxicity caused by Aβ oligomers.

Related Reading: Targeting Fyn Kinase in Alzheimer’s Disease.
Tau passive immunization inhibits not only tau but also Aβ pathology.

Protein Kinase M and LTP

My previous blog post was also about LTP: long-term potentiation of synapses. Here, the topic is what has been called late-LTP (L-LTP), a strengthening of synapses that can last for months in the hippocampus of laboratory animals. Molecular and cellular mechanisms of L-LTP are prized among neurobiology researchers because LTP seems to be one form of synaptic change that helps animals store long-lasting memories. Does Protein kinase M play an important role in making L-LTP and long-lasting memories?

Francis Crick

When Francis Crick was in his later years, he dabbled in theoretical neurobiology. One mechanism for memory that Crick discussed was the possibility that permanent changes in synaptic connections might involve enzymes (such as protein kinases) that would remain active and continually lock a particular neuronal synapse in a state that is required for storing a memory (see “Memory and molecular turnover“). Protein kinase M is proposed to provide such a molecular memory mechanism.

My favorite neurobiology research article from 2016 is “Compensation for PKMζ in long-term potentiation and spatial long-term memory in mutant mice“. The “compensation” that is explored in this article concerns a second protein kinase called PKCiota/lambda (PKCι/λ). [The same orthologous gene is called lambda in mice and iota in humans.] The authors of this 2016 article present evidence in support of the idea that in the absence of protein kinase M (PKM), PKCι/λ can act as a substitute kinase and provide mammals with some capacity to store memories. However, they believe that PKM is normally the protein kinase that is most important for L-LTP and long-lasting memories. In this model, PKCι/λ can provide a type of emergency backup memory system that is not as effective as the PKM system. The “mutant mice” in this study are transgenic mice that have been engineered to lack PKM.

Protein Kinase M

PKM is the c-terminal end of PKC (source)

What is protein kinase M? Protein kinase M got its name back in the 1970s when Yasutomi Nishizuka and his co-workers were studying a new protein kinase that came to be called protein kinase C (PKC). They noticed that brain tissue contained a protein kinase that only required magnesium for its activity, and they called it PKM. In fresh brain tissue, there was very little PKM, but PKM could be produced from PKC by proteolysis.

Synthesis of PKM (source)

Later, it was found that there are several different human genes that code for members of the PKC family of enzymes. Protein kinase C was so named because some forms (the “classical” members of the family) of the enzyme require calcium for activation. In most cases, the N-terminal regulatory domain inhibits the substrate-binding  catalytic domain unless activators such as diacylglycerol (DAG) bind to the regulatory domain and relieve the inhibition.

One of the genes in the PKC gene family codes for a version of the enzyme that is called PKCzeta (PKCζ). In 1993, Todd Sacktor reported that a form of this enzyme (that was called PKMζ) could be detected in brain tissue, particularly following the induction of LTP (see “Persistent activation of the ζ isoform of protein kinase C in the maintenance of long-term potentiation“).

PKCzeta is an atypical protein kinase C (image source). PS: the pseudosubstrate domain

In 2003, it was shown that there is a PKCzeta primary RNA transcript that can undergo alternate splicing in the brain (see “Protein Kinase Mζ Synthesis from a Brain mRNA Encoding an Independent Protein Kinase Cζ Catalytic Domain“). They obtained evidence for a special mRNA that can directly code for just the c-terminal catalytic domain of the PKCzeta enzyme, thus allowing for the synthesis of PKMζ as a translation product, and generating a persistently-activated form of PKC that is free of any need for stimulation of the enzyme by second messengers. Sacktor et al concluded: “These results indicated that brain PKMζ was not formed by a proteolytic mechanism but perhaps as a distinct ζ gene product.” In their model, synthesis of a persistently-activated PKMζ  is an important molecular mechanism for L-LTP and long-lasting memories.

Model for activation of a Classical PKC (image source)

ZIP
Support for Sacktor’s model of memory storage came from inhibitor studies. The zeta inhibitory peptide (ZIP) is a potent competitive inhibitor of PKMζ in neurons (see “Matching biochemical and functional efficacies confirm ZIP as a potent competitive inhibitor of PKMζ in neurons“). ZIP can block L-LTP and also block memories (see the ZIP literature reviewed in this recent article).

Knockout
Confidence in the importance of PKMζ for memory was shaken when it was observed that mice engineered to lack PKCzeta still have some capacity to learn and form memories (see this and this). These results for PKCzeta knockout mice were not very alarming for me because other important protein kinases in the brain display overlap/redundancy in their functions (example).

In 2015, Sacktor et al published a theoretical model in which PKCι/λ could possibly take the place of  PKMζ  in mice that  lack a functional PKCzeta gene (see “Atypical PKCs in memory maintenance: the roles of feedback and redundancy“). Are there data to support this model?

2016 Data

ZIP inhibits PKCι/λ (source)

In their 2016 article, Sacktor et al present several types of data:

1. ZIP blocks L-LTP in both normal mice and mice lacking PKMζ.
2. ZIP inhibits both the purified PKMζ enzyme and purified PKCι/λ.
3. PKMζ -antisense oligonucleotides block L-LTP in normal mice but not in mice lacking PKMζ.
4. An antagonist for PKCι/λ (ICAP) blocks the maintenance of L-LTP in mice lacking PKMζ.
5. ICAP blocks spatial long-term memory maintenance in mice lacking PKMζ  but not in normal mice.
6. PKCι/λ is persistently up-regulated in brain tissue lacking PKMζ when L-LTP is induced, unlike the situation for normal mice.
7. They also show that memory storage in mice lacking PKMζ  is not entirely normal. Mice that depend on PKCι/λ for memory storage do not learn as efficiently as normal mice.

image source

Sacktor et al suggest that for some types of memory, such as remembering the location of danger in the environment, PKCι/λ normally functions in short-term memory formation while PKMζ  is normally most important for the persistence of long-lasting memories. In mice that lack PKMζ, PKCι/λ can apparently function for both purposes, but its on-going involvement in short-term memory processes and responding to second messengers might partially disrupt the long-term memory storage function. For an interesting evolutionary perspective on the origins of PKM see: “Memory maintenance by PKMζ–an evolutionary perspective“.

Related reading:
1. Distinct Roles of PKCι/λ and PKMζ in the Initiation and Maintenance of Hippocampal Long-Term Potentiation and Memory.
2. commentary by Richard Morris.
3. Persistent increased PKMζ in long-term and remote spatial memory.
4. Role of Atypical Protein Kinases in Maintenance of Long-Term Memory and Synaptic Plasticity.
5. Memory Erasure Experiments Indicate a Critical Role of CaMKII in Memory Storage

Presynaptic Endocannabinoid-initiated LTP

presynaptic cannabinoid receptors (source)

“LTP” is long-term potentiation, a type of synaptic plasticity found in the hippocampus that is involved in episodic memory storage. A recent research article reports on the mechanism by which endocannabinoid signaling affects the induction of hippocampal long-term potentiation (LTP).

The new study concerns the role of presynaptic cannabinoid receptors (CB1R, see the image to the right) that respond to the endogenous cannabinoid 2-arachidonoyl-sn-glycerol (2AG). Two arachidonic acid derivatives, N-arachidonoylethanolamine (anandamide, AEA) and 2AG are the best studied endogenous cannabinoids. Both Δ⁹-tetrahydrocannabinol (THC), the main psychoactive molecule from cannabis and these two endogenous cannabinoids activate the CB1R.

The endogenous cannabinoid 2AG can be made in dendrites by two enzymes, DGL or phospholipase C (PLC). The new research from the laboratory of Gary Lynch concerns production of LTP in the lateral perforant path (LPP), one of the cortical inputs to the hippocampus. Endogenous cannabinoids such as  2AG play normal roles in brain function and the actions of THC at CB1R are probably involved in the influence of cannabinoid drugs on orderly thought.

One way of activating PLC and 2AG synthesis in dendrites is via the action of glutamate on GluR5, a G protein-coupled receptor (see the diagram at the top right of this blog post). Glutamate is a common neurotransmitter at synapses in the hippocampus. In some parts of the brain, glutamate actions to activate NMDA receptors has been shown to lead to LTP, often via postsynaptic mechanisms involving altered levels of postsynaptic membrane receptors. Lynch et al provide evidence concerning how glutamate actions on GluR5, production of 2AG and activation of presynaptic CB1R can cause another form of LTP.

synthesis and degradation pathways for 2AG (source)

proposed model (source)

The image to the right illustrates the proposed presynaptic mechanism of CB1R-mediated LTP at the glutamanergic synapses formed by the LPP with granule cells of the dentate gyrus (DG, an important input region of the hippocampal formation).

In this model, the 2AG synthesizing enzyme diacylglycerol lipase (DGL) requires calcium for activity and is localized to dendritic spines, where it forms a multimolecular complex with mGluR5. In the model, glutamate (green triangles) activates mGluR5 and NMDA receptors leading to increases in postsynaptic calcium levels, activation of DGL and the production of 2AG (pink circles).

latrunculin A

The 2AG then binds to presynaptic CB1R triggering a long-lasting increase in glutamate release. Block of actin filament reorganization with latrunculin A inhibited this LTP, suggesting that reorganization of the actin cytoskeleton in LPP axon terminals is involved in facilitating glutamate release during LTP. They propose that activated presynaptic CB1 receptors can promote cytoskeletal reorganization via a signaling cascade that involves a  small GTPases (RhoA, Rac or Rap) that can regulate actin polymerization.

actin filament regulation via CB1R interactions with the WAVE1 complex (source)

Related reading: A cannabinoid link between mitochondria and memory

Circadian Pacemaker

constant dark (source)

In my previous blog post, I mentioned that the GHSR1 ghrelin receptor has a high level of activity independent of ghrelin binding.  Recently, an “orphan” G Protein-coupled Receptor (GPCR) (Gpr176) has been implicated in the function of the circadian pacemaker of the suprachiasmatic nucleus (SCN). Gpr176 might not have a ligand and it might function simply by having daily swings in the level of its expression.

Gpr176 in the SCN (source)

In gene knock-out mice without this particular GPCR (Gpr176^-/-) the circadian period is slower than in normal mice. This function of Gpr176 was discovered by searching through the many (about 140) orphan GPCRs for those that have high levels of expression in the suprachiasmatic nucleus. Gpr176 expression was shown to be regulated with a circadian pattern.

circadian expression of Gpr176 mRNA [cry1] (source)

It was previously known that vasoactive intestinal peptide (VIP) regulates the circadian rhythm.

The Vipr2 GPCR for VIP stimulates adenylate cyclase and  increased levels of cAMP in neurons of the SCN circadian pacemaker. Gpr176 inhibits adenylate cyclase through the inhibitory G-protein, Gz.

High cAMP in Gpr176 knockout mice (source)

In gene knockout mice that do not express Gpr176, the measured levels of cAMP in the SCN are higher than for control mice.

(source)

.

Ghrelin

Known actions of ghrelin on various target tissues (source).

Ghrelin is a vertebrate peptide hormone that has been most heavily studied for its role as a signalling molecule that is made in the gastrointestinal tract and which can modulate the behavior of ghrelin receptors on neurons, both in the peripheral and the central nervous system. One of the most well-studied functions of ghrelin in the nervous system is for regulation of eating behavior. Soon after its discovery (2001), exogenous ghrelin was shown to increase food consumption in humans.

acylation of ghrelin by GOAT (source)

One of the interesting features of ghrelin as a peptide hormone is that acylation of the ghrelin peptide is required for binding of the hormone to its receptor, GHSR1.

The acylation of ghrelin seems to be regulated. An hypothesis was proposed (2009) that ghrelin might function as part of a regulatory system for supporting and enhancing food consumption under conditions where high-energy sources are available in the environment.

The regulation of ghrelin secretion is complex and poorly understood. There seems to be a learned pattern of ghrelin secretion (see). There might be higher average ghrelin levels following consumption of high-calorie foods and even carbonated beverages (see), leading to faster weight gain when certain foods are available in the environment.

How ghrelin acts through binding to the GHSR1 G Protein-coupled receptor is being studied in both peripheral sensory neurons and neurons of the CNS. Some of ghrelin’s action to increase feeding behavior might be due to blockage of appetite suppressing signals that reach the brain by way of sensory neurons that transmit signals from the gut.

Two CNS targets for ghrelin (image source).

In addition to effects on neurons of the peripheral nervous system, ghrelin influences the activity of CNS neurons such as populations of neurons in the hypothalamus that regulate feeding behavior (2015a). In a study of hypothalamic neurons (2015), GHSR1 was shown to inhibit pre-synaptic calcium channel activity and inhibit GABA release. It is hypothesized that this could contribute to ghrelin’s ability to regulate parts of the hypothalamus that are involved in the regulation of eating behavior.

Ghrelin receptors in other parts of the brain in addition to the hypothalamus seem to be involved in the regulation of learned eating behaviors and calorie-rich food seeking behaviors. Since 1) the GHSR1 receptor has high levels of activation even in the absence of ghrelin binding (2009) and 2) GHSR1 can form complexes with other receptors (2013), it is not clear that all the functions of ghrelin receptors in the brain rely on the transport of ghrelin across the blood-brain barrier. Neurons in various brain regions such as the mesolimbic reward system display altered electrical activity in response to administered ghrelin under conditions where GHSR1 receptors interact in complex ways with other neurotransmitter systems (2016).

2018 UPDATE. Recent article on how variations in meal content can influence ghrelin levels: “A high carbohydrate, but not fat or protein meal attenuates postprandial ghrelin, PYY and GLP-1 responses in Chinese men“.

Model for action of ghrelin in the CNS (source)

Measles

From the World Health Organization’s Measles Fact sheet (Updated November 2014):

From the World Health Organization

Measles information from the U.S. Centers for Disease Control:

Manual for the Surveillance of Vaccine-Preventable Diseases, Chapter 7: Measles

The CDC measles outbreak webpage.

California Department of Public Health measles webpage with information for the current outbreak.

From the California Department of Public Health

I recently turned 56 years of age and I was surprised to see that the majority of measles cases in California for the current outbreak are among older people. I’ve read that one measles patient in the current outbreak is 57 and another is 70 years old.

Vaccine Gap “You’re squarely in that gap if you were born between 1957 and 1971”

I was vaccinated for measles back in the 1960s, before the current two-dose-vaccination protocol became recommended, so today I went to the SCNM clinic and got a MMR (measles, mumps, and rubella) vaccination shot.

Looking to the future: what could happen in the United States?

2008–2011 measles outbreak, France

The figure above is from Measles Elimination Efforts and 2008–2011 Outbreak, France. In 2011 there were more than 20,000 measles cases in France and almost 5,000 patients were hospitalized, including 1,023 for severe pneumonia and 27 for encephalitis/myelitis; 10 patients died.

Measles and the brain

Studies indicate that about 1 of every 1000 people infected by the measles virus will show evidence of the spread of virus to the central nervous system (infectious virus can be isolated from the cerebrospinal fluid or brain). The measles virus has a lipid envelope with two glycoproteins that have been shown to be important for infection of human cells: the fusion protein (F) and the receptor-binding protein (H).

Since some people with compromised immune system function cannot be vaccinated and since some vaccinated individuals develop encephalitis, additional treatments for measles are being sought. For example, it has been demonstrated that for laboratory animals, peptides with sequences corresponding to parts (such as the C-terminal heptad repeat region; HRC) of the measles virus fusion protein can be conjugated to cholesterol and upon injection into animals can protect against fatal measles virus infection.

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3838268/figure/F7/

Related Reading:

Feb. 2 2015 article about measles by Dr. Matthew Baral.

Feb. 3 2015 AZCentral.com “Gilbert pediatrician speaks out after kids’ measles exposure”

Invitation to Outbreak: Arizona does not enforce its own rules for documenting school children with non-medical reasons for not vaccinating (story at azcentral.com).

Outbreak timeline

December 17-20 2014 – Disneyland visitors infected (n=39, 5 from Arizona, 4 Pinal Co. residents (P0) & 1 Mericopa Co. resident (case M0))

January 11, 2015 – 1 additional Mericopa County woman (M1) was infected by P0 at a clinic in Mesa

Jan 20 & 21 – additional people (including unvaccinated children <1 year old) were exposed at a clinic in Mesa by M1

Jan 22 – Announcement of case M0

Jan 27 – Announcement that a total of 2 more people (P1, P2) had been infected in Pinal Co.

No additional cases were reported: summary of the outbreak

Narrated PowerPoint Presentations

PowerPoint animation options

I previously blogged about using ProfCast to add audio narration to PowerPoint presentations. I had better results recording audio using the internal microphone of a Macintosh than the internal microphone of an HP laptop, but I was left wanting to try a better microphone. Also, when I lecture, I tend to spend a significant amount of time discussing rather busy slides and doing a lot of pointing at specific features in the slide, so I needed a good way to automate all that pointing.

I finally got a headset with a noise-canceling microphone, which, amazingly, can be used right next to a running fan. I have also taken the time to start learning how to use the PowerPoint animation features, which can replace my habit of pointing at details on the slides.

I tried recording directly into PowerPoint using the Insert > Audio > Record Audio… option. I ran into the problem (discussed here) of the default audio quality for PowerPoint 2010 being set too low for me to get a good recording.

Another issue is that I like to be able to edit my audio. In the past I’ve used GarageBand, but I wanted to record the audio on the same HP laptop that has PowerPoint 2010. I’m now using Audacity to record audio from my headset’s microphone. I export the sound files from Audacity and then use PowerPoint’s “Insert> Audio > Audio from File…” option.

PowerPoint vs Video. I prefer to provide students with PowerPoint presentations that contain narration and automated pointing because I often insert hypertext links in my slides. However, it is possible to save a PowerPoint presentations as video and some students might prefer a video if they are using a hand-held device that will not play PowerPoint presentations. Select “Save and Send > Create a Video” from the file menu:

Here is an example of a video made from PowerPoint: http://youtu.be/9Veh6s3lbH8

If you use PowerPoint’s default size for presentation slides then you will never really be happy with how videos of your PowerPoint presentations look on YouTube. YouTube uses 16:9 (width:height) video, but the PowerPoint default is a 4:3 ratio. I suggest using PowerPoint >Display menu > Page Setup >Slides sized for: Custom > Width 13.33 inches, Height 7.5 inches. The PowerPoint option to save your videos as “Internet & DVD 852 x 480” will work well with YouTube’s 853 x 480 video format.

I’d like to try making media files available to students  via iTunes U.

I discovered that if you have a slide with animated elements and you add a video to the slide, the video will not play automatically even if you select Video Tools > Playback > Start Automatically. In order for this to work, the video has to be moved to the top (shown below for “gif to movie.wmv”) of the list of animated elements in the Animation Pane where all of the animated elements of the page are listed:

Related Reading about PowerPoint