Paper Digest — May 9, 2026
Variant: B (Detail-First)
Papers: 3 | Neuroscience, Evolutionary Cell Biology
Paper 1: Acetylcholine demixes heterogeneous dopamine signals for learning and moving
Jang et al., Nature Neuroscience, 2026
DOI: 10.1038/s41593-026-02227-x
Abstract: Midbrain dopamine neurons promote reinforcement learning and movement vigor. An outstanding question is how dopamine-recipient neurons in the striatum parse these heterogeneous signals. Previous work suggests that cholinergic striatal interneurons may gate dopamine-dependent plasticity, but this has not been tested in behaving animals. Here we studied rats performing a decision-making task with reward-related and movement-related events. Optical measurement of dopamine and acetylcholine release in the dorsomedial striatum (DMS) revealed that reward cues evoked cholinergic pauses with different phase relationships relative to dopamine. When dopamine lagged cholinergic dips, dopamine predicted future behavior and DMS firing rates on subsequent trials. In contrast, when dopamine preceded cholinergic dips, there was no observable relationship between dopamine and learning. Finally, when dopamine was coincident with cholinergic bursts, it preceded and predicted the vigor of contralateral orienting movements. Our findings suggest that cholinergic dynamics determine whether dopamine promotes vigor or learning, depending on the instantaneous behavioral context.
Experiment
Task: Self-paced temporal wagering task in rats (n=16). Rats initiate trials by nose-poking in a central port. An auditory cue plays for ~1s indicating reward volume (5, 10, 20, 40, or 80 μl). After cue offset, reward is randomly assigned to one of two side ports (indicated by LED). Rats can wait for an unpredictable delay to obtain reward, or opt out by poking the other side port to start a new trial. Catch trials (15–25%) withhold rewards to measure willingness to wait.
Manipulation: Uncued blocks of low rewards (5, 10, 20 μl) vs. high rewards (20, 40, 80 μl), interleaved with mixed blocks offering all volumes.
Recordings: Fiber photometry in DMS measuring:
- Dopamine: GRABDA2h (n=6 rats) or red-shifted GRABrDA3m (n=4 rats)
- Acetylcholine: GRABACh4.3 (n=10 rats)
- Dual-color imaging: simultaneous DA + ACh in same rats (n=4 rats)
Results (Figure by Figure)
Figure 1 — Task and Behavior:
- Rats initiated trials faster in high vs. low blocks, indicating they tracked subjective value of the environment
- Willingness to wait (opt-out times) scaled with offered reward volume and block type
- DeepLabCut pose estimation revealed stereotyped orienting movements toward the reward port (contralateral to implant in some rats) and back to center after opt-outs
- Key insight: reward cues and orienting movements are separable in time, enabling independent analysis of RPE vs. movement signals
Figure 2 — Dopamine and Acetylcholine Dynamics:
- At offer cue: Dopamine showed phasic release scaling with reward volume (RPE signal). Acetylcholine showed dips (pauses) at the same time. The dip magnitude scaled modestly with reward volume.
- At reward cue (after delay): Dopamine scaled with delay (larger for longer delays = negative RPE). Again, acetylcholine showed dips.
- At movement events (LED on, opt-out): Dopamine showed larger phasic responses for contralateral orienting movements. Acetylcholine showed bursts (not dips) at contralateral movements.
- Dual-color confirmation: Same pattern when DA and ACh were measured simultaneously at the same site, ruling out location differences.
- Regional differences: DMS showed both RPE and movement signals. Ventral striatum showed RPE signals similar to DMS. Dorsolateral striatum (DLS) lacked RPE encoding at offer cue.
Figure 3 — Timing Gates Learning:
- Offer cue: Dopamine lagged acetylcholine dips by ~100 ms. Cross-correlogram peak at negative lag.
- Reward cue: Dopamine preceded acetylcholine dips by ~50 ms. The cholinergic dip was aligned to reward delivery (nose poke triggering solenoid), while dopamine RPE was aligned to the reward cue (light turning off).
- Learning prediction: When dopamine lagged cholinergic dips (offer cue), dopamine AUC predicted subsequent trial initiation times — a behavioral marker of learning. When dopamine preceded cholinergic dips (reward cue), there was no relationship between dopamine and learning.
- Model fit: A reinforcement learning model with reward-volume-modulated RPE captured trial initiation times well.
Figure 4 — Acetylcholine Bursts Predict Movement Vigor:
- At contralateral orienting movements, acetylcholine bursts coincided with dopamine phasic responses
- Dopamine at these movement events predicted the vigor (speed/force) of the upcoming movement
- When dopamine and acetylcholine were coincident, dopamine promoted movement vigor, not learning
- This was tested by examining trials with different movement speeds: faster movements were preceded by larger dopamine and acetylcholine responses
Methods
Optical sensors: GRABDA2h (green, dopamine), GRABrDA3m (red, dopamine), GRABACh4.3 (green, acetylcholine). Viral delivery via AAV9-hSyn vectors. Fiber photometry with 470 nm excitation (green sensors) and 560 nm (red sensor). Motion correction using mCherry or isosbestic control.
Histology: Recording sites spanned ~3 mm across the anteroposterior axis of DMS, with some rats implanted in right hemisphere, others in left (counterbalanced).
Analysis: Cross-correlograms between dopamine and acetylcholine signals. Discriminability index for task variables. Reinforcement learning model fitting trial initiation times. z-scoring and baseline correction before pooling across rats.
Why It Matters
This paper resolves a fundamental puzzle: how can the same dopamine signal drive both learning and movement, when medium spiny neurons receive both types of signals? The answer is acetylcholine acts as a dynamic gate. When ACh dips, dopamine drives learning (RPE-dependent plasticity). When ACh bursts, dopamine drives movement vigor. The relative timing of DA and ACh determines which mode dominates.
This has immediate relevance for Parkinson’s disease (where both DA and cholinergic systems degenerate) and for understanding how the striatum multiplexes different functions. It also provides a mechanistic framework for why dopamine’s role in learning vs. movement has been so hard to disentangle — they’re not spatially segregated to different neurons or regions, but temporally segregated by ACh dynamics at the same synapses.
Paper 2: Breath-giving cooperation: critical review of origin of mitochondria hypotheses
Számadó & Szathmáry, Biology Direct, 2017
PMC: PMC5557255
Abstract: The origin of mitochondria is a unique and hard evolutionary problem, embedded within the origin of eukaryotes. The puzzle is challenging due to the egalitarian nature of the transition where lower-level units took over energy metabolism. Contending theories widely disagree on ancestral partners, initial conditions and unfolding of events. There are many open questions but there is no comparative examination of hypotheses. We have specified twelve questions about the observable facts and hidden processes leading to the establishment of the endosymbiont that a valid hypothesis must address. We have objectively compared contending hypotheses under these questions to find the most plausible course of events and to draw insight on missing pieces of the puzzle. Since endosymbiosis borders evolution and ecology, and since a realistic theory has to comply with both domains’ constraints, the conclusion is that the most important aspect to clarify is the initial ecological relationship of partners. Metabolic benefits are largely irrelevant at this initial phase, where ecological costs could be more disruptive. There is no single theory capable of answering all questions indicating a severe lack of ecological considerations. A new theory, compliant with recent phylogenomic results, should adhere to these criteria.
Experiment / Approach
This is a theoretical/review paper, not an empirical study. The authors systematically evaluate 8 major hypotheses for mitochondrial origin against 12 objective criteria:
Eight hypotheses evaluated:
- Hydrogen hypothesis (Martin & Müller)
- Photosynthetic symbiont theory
- Syntrophy hypothesis
- Phagocytosing archaeon theory
- Pre-endosymbiont hypothesis
- Sulfur-cycling hypothesis
- Origin-by-infection hypothesis
- Oxygen-detoxification hypothesis
Twelve evaluation questions:
- 6 “Observables” (readily testable facts): singularity of eukaryotic origin, lack of intermediate forms, chimaeric membranes, lack of membrane bioenergetics in host, lack of photosynthesis in symbiont, origin of MROs
- 6 “Historicals” (inferred processes): original host metabolism, original symbiont metabolism, initial ecological relationship, early selective advantage, mechanism of inclusion, mechanism of vertical transmission
Results (Figure by Figure)
Figure 1 — Schematic of Eight Hypotheses:
- All scenarios are depicted with topological changes showing membrane transformations
- Key distinction: syntrophic engulfment vs. phagocytosis vs. infection as inclusion mechanisms
- Archaea depicted with red membrane, bacteria with blue, purple = photosynthetic ability
- Dashed curves = degrading membranes
- Ultimate convergence: all end with mitochondria performing metabolic compartmentation and ATP production
Table 1 — Classification of Hypotheses:
- Host possibilities: primitive eukaryote, archaeon, bacterium
- Ecological relationships: syntrophy (+/+), predation (+/-), parasitism (-/+)
- Major schism: phagocytosis early/mitochondria late vs. mitochondria early/phagocytosis late
- No hypothesis scores perfectly across all dimensions
Key Findings from Comparative Evaluation:
-
Eukaryotic singularity: All eukaryotes have mitochondria or MROs; no second origin known. If mitochondria provided such an enormous advantage, why only once? Either there was a hard barrier, or parallel trials failed. Lane & Martin’s energetic argument (10× genome expansion) is questioned — amitochondriate eukaryotes exist (secondarily reduced) with ~10K genes, showing phagocytosis is feasible without mitochondria.
-
Phylogenetic consensus: Strong support for archaeal host from TACK superphylum + alphaproteobacterial symbiont. LECA was already complex and mitochondriate. All MROs are monophyletic.
-
Initial ecology is the critical missing piece: Most theories assume mutualism from the start, but this is likely wrong. The transition from independent to dependent relationship was not instantaneous. Energy production and genome integration are conclusions, not conditions, of the merger.
-
No theory answers all 12 questions: Every hypothesis has “heavy shortcomings or debatable assumptions.” The analysis indicates a “severe lack of ecological considerations” in current theories.
-
Metabolic benefits are secondary at first: At the initial phase, ecological costs (conflict of interest, digestion risk, unregulated reproduction) could be more disruptive than metabolic benefits are supportive.
Methods
Review methodology: Systematic comparison modeled after Számadó & Szathmáry’s review of language evolution. Twelve objective criteria established a priori. Eight hypotheses selected based on current acceptance and phylogenetic plausibility. Evaluation scored in supplementary tables (S2, S3) with detailed justifications in Additional File 1.
Phylogenetic constraints: Host must be from TACK/Asgard archaea. Symbiont must be alphaproteobacterium. LECA must be complex, phagocytotic, aerobically respiring.
Why It Matters
This paper is a rare example of rigorous theoretical ecology applied to deep evolutionary history. The key insight — that we need to focus on initial ecological relationship rather than metabolic benefit — reframes the entire mitochondrial origin field. The authors show that assuming mutualism from the start is a fallacy; the early partners likely experienced conflict, exploitation risk, and ecological costs.
For Raghavendra’s interest in bioenergetics and the origin of complex cells (connected to Nick Lane’s work in The Vital Question), this is directly relevant. It challenges the “mitochondria-first energizes everything” narrative and replaces it with a more nuanced ecological co-evolution story. The 12-question framework itself is a methodological contribution that could be applied to other major evolutionary transitions.
Paper 3: On the origin of the nucleus: a hypothesis
Goldman et al., Microbiology and Molecular Biology Reviews, 2024
PMC: PMC10732040
Summary: In this hypothesis article, we explore the origin of the eukaryotic nucleus. In doing so, we first look afresh at the nature of this defining feature of the eukaryotic cell and its core functions—emphasizing the utility of seeing the eukaryotic nucleoplasm and cytoplasm as distinct regions of a common compartment. We then discuss recent progress in understanding the evolution of the eukaryotic cell from archaeal and bacterial ancestors, focusing on phylogenetic and experimental data which have revealed that many eukaryotic machines with nuclear activities have archaeal counterparts. In addition, we review the literature describing the cell biology of representatives of the TACK and Asgardarchaeota — the closest known living archaeal relatives of eukaryotes. Finally, bringing these strands together, we propose a model for the archaeal origin of the nucleus that explains much of the current data, including predictions that can be used to put the model to the test.
Experiment / Approach
This is a hypothesis/theory paper synthesizing:
- Nuclear architecture and function in modern eukaryotes
- Phylogenetic data on archaeal ancestry of nuclear machinery
- Cell biological studies of TACK and Asgard archaea
- The “inside-out” model of eukaryogenesis (building on original work by Baum & Baum)
The authors propose a distinct, more delimited model focused specifically on nuclear evolution rather than explaining the entire eukaryotic cell.
Results (Figure by Figure)
Figure 1 — Transcription-Translation Coupling:
- Compares spatial organization of gene expression in eukaryotic vs. bacterial cells
- In bacteria: transcription and translation are coupled (co-localized)
- In eukaryotes: transcription and translation are separate (nuclear vs. cytoplasmic)
- Key question: Are transcription and translation strongly coupled or partially uncoupled in archaeal relatives of eukaryotes? Partial uncoupling could enable local translation of some transcripts.
Figure 2 — Information Flow in Eukaryotes:
- DNA → transcription in nucleus → RNA processing → export through NPCs → cytoplasmic translation
- Some mRNAs remain silent (bound by RNA-binding proteins) and are trafficked to peripheral locations
- Proteins with signal peptides: translation at rough ER → ER/Golgi modification → vesicle trafficking to periphery
- Shows the directed outward flow of information from nucleus to cell periphery
Figure 3 — Nuclear Pore Complex Structure and Assembly:
- (A) Structure of nucleus and connected ER. NPCs sit at highly curved membrane junctions between inner and outer nuclear membranes
- (B) Single NPC viewed from side — 8-fold symmetric, ~50 nm central channel
- (C) NPC insertion pathways: (i) at mitotic exit into gaps in reforming nuclear envelope, (ii) interphase insertion via out-folding of inner nuclear envelope, membrane bulbing, and fusion with outer membrane
Key Architectural Insights:
-
Nucleoplasm and cytoplasm are one compartment: The inner and outer nuclear membranes are physically continuous via pores. The ER lumen and perinuclear space form a single continuous network. Nucleoplasm and cytoplasm freely exchange via NPCs. They should be viewed as “sub-domains or discrete phases of a single compartment” rather than truly separate compartments.
-
Nuclear sub-regionalization without membranes: The nucleolus, Cajal bodies, replication foci, transcription/silencing domains, chromosome territories — all established by weak multivalent interactions (biomolecular condensates), not membrane boundaries.
-
NPCs as gatekeepers: ~500 proteins in yeast, ~1000 in humans. Mass: 50 MDa (yeast) to 120 MDa (human). Central channel ~50 nm diameter. FG-repeat mesh acts as selective filter — small molecules diffuse freely, large complexes (>30 kDa) require importins/exportins.
-
Ran-GTP gradient: Powers directed nuclear transport. Ran-GTP loaded in nucleus (by chromatin-bound GEF), hydrolyzed in cytoplasm (by GAPs). This gradient can polarize space even without a membrane — used during mitosis and in oocytes.
-
NPC assembly from the inside-out: Interphase insertion begins with out-folding of inner nuclear envelope → accumulation of nucleoporins at membrane neck → fusion of inner and outer membranes → new pore. Torsin AAA-ATPase powers membrane remodeling.
Proposed Model for Nuclear Origin:
- Builds on inside-out model but more focused
- Many nuclear activities have archaeal counterparts (histones, transcription machinery, some NPC components)
- TACK/Asgard archaea show intermediate features
- Predictions for testing: specific archaeal features that should/should not exist if model is correct
Methods
Synthesis approach: Literature review across nuclear cell biology, archaeal cell biology, phylogenomics, and membrane biophysics. Distinction from full inside-out model: this model focuses only on nucleus, not attempting to explain all eukaryotic features.
Key sources: TACK and Asgard archaea cell biology studies; NPC structure work (Beck lab, others); nuclear envelope biogenesis studies; phylogenomic analyses of nuclear protein homologs in archaea.
Why It Matters
This paper challenges the common “nucleus as sealed compartment” intuition. The nucleoplasm and cytoplasm are not separate compartments — they’re sub-regions of one compartment, connected by thousands of massive pores. This reframes how we think about nuclear function: it’s about regulating flux and creating sub-domains (via condensates and selective gating), not about sealing off the genome.
For understanding eukaryogenesis, the model makes testable predictions about what features we should find in Asgard archaea. If the nucleus evolved gradually from archaeal ancestors, we should see partial uncoupling of transcription and translation, proto-NPC components, and membrane remodeling machinery.
The connection to The Vital Question and Breath-giving cooperation critical review of origin of mitochondria hypotheses is clear: eukaryogenesis wasn’t a single miracle but a series of gradual transitions — mitochondrial endosymbiosis and nuclear compartmentalization as coupled but separable problems.
Digest Notes
Figure extraction: No figures could be extracted for this digest — the Nature Neuroscience paper is paywalled (PDF inaccessible), and PMC PDF downloads were blocked by anti-bot measures. Figure descriptions were reconstructed from the article text. For full figure access, PDFs would need to be provided manually.
Papers covered:
- Jang et al. (2026) — Acetylcholine gating of dopamine signals (Nature Neuroscience)
- Számadó & Szathmáry (2017) — Mitochondrial origin hypotheses review (Biology Direct)
- Goldman et al. (2024) — Origin of the nucleus hypothesis (MMBR)