How to Build the Virtual Cell with Artificial Intelligence: Priorities and Opportunities.

The cell is arguably the most fundamental unit of life and is central to understanding biology. Accurate modeling of cells is important for this understanding as well as for determining the root causes of disease. Recent advances in artificial intelligence (AI), combined with the ability to generate large-scale experimental data, present novel opportunities to model cells. Here we propose a vision of leveraging advances in AI to construct virtual cells, high-fidelity simulations of cells and cellular systems under different conditions that are directly learned from biological data across measurements and scales. We discuss desired capabilities of such AI Virtual Cells, including generating universal representations of biological entities across scales, and facilitating interpretable in silico experiments to predict and understand their behavior using Virtual Instruments. We further address the challenges, opportunities and requirements to realize this vision including data needs, evaluation strategies, and community standards and engagement to ensure biological accuracy and broad utility. We envision a future where AI Virtual Cells help identify new drug targets, predict cellular responses to perturbations, as well as scale hypothesis exploration. With open science collaborations across the biomedical ecosystem that includes academia, philanthropy, and the biopharma and AI industries, a comprehensive predictive understanding of cell mechanisms and interactions has come into reach.

A cooperative network at the nuclear envelope counteracts LINC-mediated forces during oogenesis in C. elegans.

Oogenesis involves transduction of mechanical forces from the cytoskeleton to the nuclear envelope (NE). In Caenorhabditis elegans, oocyte nuclei lacking the single lamin protein LMN-1 are vulnerable to collapse under forces mediated through LINC (linker of nucleoskeleton and cytoskeleton) complexes. Here, we use cytological analysis and in vivo imaging to investigate the balance of forces that drive this collapse and protect oocyte nuclei. We also use a mechano-node-pore sensing device to directly measure the effect of genetic mutations on oocyte nuclear stiffness. We find that nuclear collapse is not a consequence of apoptosis. It is promoted by dynein, which induces polarization of a LINC complex composed of Sad1 and UNC-84 homology 1 (SUN-1) and ZYGote defective 12 (ZYG-12). Lamins contribute to oocyte nuclear stiffness and cooperate with other inner nuclear membrane proteins to distribute LINC complexes and protect nuclei from collapse. We speculate that a similar network may protect oocyte integrity during extended oocyte arrest in mammals.

ATM signaling modulates cohesin behavior in meiotic prophase and proliferating cells.

Cohesins are ancient and ubiquitous regulators of chromosome architecture and function, but their diverse roles and regulation remain poorly understood. During meiosis, chromosomes are reorganized as linear arrays of chromatin loops around a cohesin axis. This unique organization underlies homolog pairing, synapsis, double-stranded break induction, and recombination. We report that axis assembly in Caenorhabditis elegans is promoted by DNA-damage response (DDR) kinases that are activated at meiotic entry, even in the absence of DNA breaks. Downregulation of the cohesin-destabilizing factor WAPL-1 by ATM-1 promotes axis association of cohesins containing the meiotic kleisins COH-3 and COH-4. ECO-1 and PDS-5 also contribute to stabilizing axis-associated meiotic cohesins. Further, our data suggest that cohesin-enriched domains that promote DNA repair in mammalian cells also depend on WAPL inhibition by ATM. Thus, DDR and Wapl seem to play conserved roles in cohesin regulation in meiotic prophase and proliferating cells.

MJL-1 is a nuclear envelope protein required for homologous chromosome pairing and regulation of synapsis during meiosis in C. elegans.

Interactions between chromosomes and LINC (linker of nucleoskeleton and cytoskeleton) complexes in the nuclear envelope (NE) promote homolog pairing and synapsis during meiosis. By tethering chromosomes to cytoskeletal motors, these connections lead to processive chromosome movements along the NE. This activity is usually mediated by telomeres, but in the nematode Caenorhabditis elegans, special chromosome regions called "pairing centers" (PCs) have acquired this meiotic function. Here, we identify a previously uncharacterized meiosis-specific NE protein, MJL-1 (MAJIN-Like-1), that is essential for interactions between PCs and LINC complexes in C. elegans. Mutations in MJL-1 eliminate active chromosome movements during meiosis, resulting in nonhomologous synapsis and impaired homolog pairing. Fission yeast and mice also require NE proteins to connect chromosomes to LINC complexes. Extensive similarities in the molecular architecture of meiotic chromosome-NE attachments across eukaryotes suggest a common origin and/or functions of this architecture during meiosis.

Recruitment of Polo-like kinase couples synapsis to meiotic progression via inactivation of CHK-2.

Meiotic chromosome segregation relies on synapsis and crossover (CO) recombination between homologous chromosomes. These processes require multiple steps that are coordinated by the meiotic cell cycle and monitored by surveillance mechanisms. In diverse species, failures in chromosome synapsis can trigger a cell cycle delay and/or lead to apoptosis. How this key step in 'homolog engagement' is sensed and transduced by meiotic cells is unknown. Here we report that in C. elegans, recruitment of the Polo-like kinase PLK-2 to the synaptonemal complex triggers phosphorylation and inactivation of CHK-2, an early meiotic kinase required for pairing, synapsis, and double-strand break (DSB) induction. Inactivation of CHK-2 terminates DSB formation and enables CO designation and cell cycle progression. These findings illuminate how meiotic cells ensure CO formation and accurate chromosome segregation.

Most animals, plants, and fungi reproduce sexually, meaning that the genetic information from two parents combines during fertilization to produce offspring. This parental genetic information is carried within the reproductive cells in the form of chromosomes. Reproductive cells in the ovaries or testes first multiply through normal cell division, but then go through a unique type of cell division called meiosis. During meiosis, pairs of chromosomes ­ the two copies inherited from each parent ­ must find each other and physically line up from one end to the other. As they align side-by-side with their partners, chromosomes also go through a mixing process called recombination, during which regions of one chromosome cross over to the paired chromosome to exchange information. Scientists are still working to understand how this process of chromosome alignment and crossing-over is controlled. If chromosomes fail to line up or cross over during meiosis, eggs or sperm can end up with too many or too few chromosomes. If these faulty reproductive cells combine during fertilization this can lead to birth defects and developmental problems. To minimize this problem, reproductive cells have a quality control mechanism during meiosis called "crossover assurance", which limits how often mistakes occur. Zhang et al. have investigated how cells can tell if their chromosomes have accomplished this as they undergo meiosis. They looked at egg cells of the roundworm C. elegans, whose meiotic processes are similar to those in humans. In C. elegans, a protein called CHK-2 regulates many of the early events during meiosis. During successful meiosis, CHK-2 is active for only a short amount of time. But if there are problems during recombination, CHK-2 stays active for longer and prevents the cell division from proceeding. Zhang et al. uncovered another protein that affects for how long CHK-2 stays switched on. When chromosomes align with their partners, a protein called PLK-2 sticks to other proteins at the interface between the aligned chromosomes. A combination of microscopy and test tube experiments showed that when PLK-2 is bound to this specific location, it can turn off CHK-2. However, if the chromosome alignment fails, PLK-2 is not activated to switch off CHK-2. Therefore, CHK-2 is only switched off when the chromosomes are properly aligned and move on to the next step in crossing-over, which then allows meiosis to proceed. Thus, PLK-2 and CHK-2 work together to detect errors and to slow down meiosis if necessary. Further experiments in mammalian reproductive cells will reveal how similar the crossover assurance mechanism is in different organisms. In the future, improved understanding of quality control during meiosis may eventually lead to improvements in assisted reproduction.

Phosphoregulation of DSB-1 mediates control of meiotic double-strand break activity.

In the first meiotic cell division, proper segregation of chromosomes in most organisms depends on chiasmata, exchanges of continuity between homologous chromosomes that originate from the repair of programmed double-strand breaks (DSBs) catalyzed by the Spo11 endonuclease. Since DSBs can lead to irreparable damage in germ cells, while chromosomes lacking DSBs also lack chiasmata, the number of DSBs must be carefully regulated to be neither too high nor too low. Here, we show that in Caenorhabditis elegans, meiotic DSB levels are controlled by the phosphoregulation of DSB-1, a homolog of the yeast Spo11 cofactor Rec114, by the opposing activities of PP4PPH-4.1 phosphatase and ATRATL-1 kinase. Increased DSB-1 phosphorylation in pph-4.1 mutants correlates with reduction in DSB formation, while prevention of DSB-1 phosphorylation drastically increases the number of meiotic DSBs both in pph-4.1 mutants and in the wild-type background. C. elegans and its close relatives also possess a diverged paralog of DSB-1, called DSB-2, and loss of dsb-2 is known to reduce DSB formation in oocytes with increasing age. We show that the proportion of the phosphorylated, and thus inactivated, form of DSB-1 increases with age and upon loss of DSB-2, while non-phosphorylatable DSB-1 rescues the age-dependent decrease in DSBs in dsb-2 mutants. These results suggest that DSB-2 evolved in part to compensate for the inactivation of DSB-1 through phosphorylation, to maintain levels of DSBs in older animals. Our work shows that PP4PPH-4.1, ATRATL-1, and DSB-2 act in concert with DSB-1 to promote optimal DSB levels throughout the reproductive lifespan.

Robust, versatile DNA FISH probes for chromosome-specific repeats in Caenorhabditis elegans and Pristionchus pacificus.

Repetitive DNA sequences are useful targets for chromosomal fluorescence in situ hybridization. We analyzed recent genome assemblies of Caenorhabditis elegans and Pristionchus pacificus to identify tandem repeats with a unique genomic localization. Based on these findings, we designed and validated sets of oligonucleotide probes for each species targeting at least 1 locus per chromosome. These probes yielded reliable fluorescent signals in different tissues and can easily be combined with the immunolocalization of cellular proteins. Synthesis and labeling of these probes are highly cost-effective and require no hands-on labor. The methods presented here can be easily applied in other model and nonmodel organisms with a sequenced genome.

How and Why Chromosomes Interact with the Cytoskeleton during Meiosis.

During the early meiotic prophase, connections are established between chromosomes and cytoplasmic motors via a nuclear envelope bridge, known as a LINC (linker of nucleoskeleton and cytoskeleton) complex. These widely conserved links can promote both chromosome and nuclear motions. Studies in diverse organisms have illuminated the molecular architecture of these connections, but important questions remain regarding how they contribute to meiotic processes. Here, we summarize the current knowledge in the field, outline the challenges in studying these chromosome dynamics, and highlight distinctive features that have been characterized in major model systems.

Complete genomic and epigenetic maps of human centromeres.

Existing human genome assemblies have almost entirely excluded repetitive sequences within and near centromeres, limiting our understanding of their organization, evolution, and functions, which include facilitating proper chromosome segregation. Now, a complete, telomere-to-telomere human genome assembly (T2T-CHM13) has enabled us to comprehensively characterize pericentromeric and centromeric repeats, which constitute 6.2% of the genome (189.9 megabases). Detailed maps of these regions revealed multimegabase structural rearrangements, including in active centromeric repeat arrays. Analysis of centromere-associated sequences uncovered a strong relationship between the position of the centromere and the evolution of the surrounding DNA through layered repeat expansions. Furthermore, comparisons of chromosome X centromeres across a diverse panel of individuals illuminated high degrees of structural, epigenetic, and sequence variation in these complex and rapidly evolving regions.

Analysis of meiosis in Pristionchus pacificus reveals plasticity in homolog pairing and synapsis in the nematode lineage.

Meiosis is conserved across eukaryotes yet varies in the details of its execution. Here we describe a new comparative model system for molecular analysis of meiosis, the nematode Pristionchus pacificus, a distant relative of the widely studied model organism Caenorhabditis elegans. P. pacificus shares many anatomical and other features that facilitate analysis of meiosis in C. elegans. However, while C. elegans has lost the meiosis-specific recombinase Dmc1 and evolved a recombination-independent mechanism to synapse its chromosomes, P. pacificus expresses both DMC-1 and RAD-51. We find that SPO-11 and DMC-1 are required for stable homolog pairing, synapsis, and crossover formation, while RAD-51 is dispensable for these key meiotic processes. RAD-51 and DMC-1 localize sequentially to chromosomes during meiotic prophase and show nonoverlapping functions. We also present a new genetic map for P. pacificus that reveals a crossover landscape very similar to that of C. elegans, despite marked divergence in the regulation of synapsis and crossing-over between these lineages.

A degron-based strategy reveals new insights into Aurora B function in C. elegans.

The widely conserved kinase Aurora B regulates important events during cell division. Surprisingly, recent work has uncovered a few functions of Aurora-family kinases that do not require kinase activity. Thus, understanding this important class of cell cycle regulators will require strategies to distinguish kinase-dependent from independent functions. Here, we address this need in C. elegans by combining germline-specific, auxin-induced Aurora B (AIR-2) degradation with the transgenic expression of kinase-inactive AIR-2. Through this approach, we find that kinase activity is essential for AIR-2's major meiotic functions and also for mitotic chromosome segregation. Moreover, our analysis revealed insight into the assembly of the ring complex (RC), a structure that is essential for chromosome congression in C. elegans oocytes. AIR-2 localizes to chromosomes and recruits other components to form the RC. However, we found that while kinase-dead AIR-2 could load onto chromosomes, other components were not recruited. This failure in RC assembly appeared to be due to a loss of RC SUMOylation, suggesting that there is crosstalk between SUMOylation and phosphorylation in building the RC and implicating AIR-2 in regulating the SUMO pathway in oocytes. Similar conditional depletion approaches may reveal new insights into other cell cycle regulators.

A compartmentalized signaling network mediates crossover control in meiosis.

During meiosis, each pair of homologous chromosomes typically undergoes at least one crossover (crossover assurance), but these exchanges are strictly limited in number and widely spaced along chromosomes (crossover interference). The molecular basis for this chromosome-wide regulation remains mysterious. A family of meiotic RING finger proteins has been implicated in crossover regulation across eukaryotes. Caenorhabditis elegans expresses four such proteins, of which one (ZHP-3) is known to be required for crossovers. Here we investigate the functions of ZHP-1, ZHP-2, and ZHP-4. We find that all four ZHP proteins, like their homologs in other species, localize to the synaptonemal complex, an unusual, liquid crystalline compartment that assembles between paired homologs. Together they promote accumulation of pro-crossover factors, including ZHP-3 and ZHP-4, at a single recombination intermediate, thereby patterning exchanges along paired chromosomes. These proteins also act at the top of a hierarchical, symmetry-breaking process that enables crossovers to direct accurate chromosome segregation.

Superresolution microscopy reveals the three-dimensional organization of meiotic chromosome axes in intact Caenorhabditis elegans tissue.

When cells enter meiosis, their chromosomes reorganize as linear arrays of chromatin loops anchored to a central axis. Meiotic chromosome axes form a platform for the assembly of the synaptonemal complex (SC) and play central roles in other meiotic processes, including homologous pairing, recombination, and chromosome segregation. However, little is known about the 3D organization of components within the axes, which include cohesin complexes and additional meiosis-specific proteins. Here, we investigate the molecular organization of meiotic chromosome axes in Caenorhabditis elegans through STORM (stochastic optical reconstruction microscopy) and PALM (photo-activated localization microscopy) superresolution imaging of intact germ-line tissue. By tagging one axis protein (HIM-3) with a photoconvertible fluorescent protein, we established a spatial reference for other components, which were localized using antibodies against epitope tags inserted by CRISPR/Cas9 genome editing. Using 3D averaging, we determined the position of all known components within synapsed chromosome axes to high spatial precision in three dimensions. We find that meiosis-specific HORMA domain proteins span a gap between cohesin complexes and the central region of the SC, consistent with their essential roles in SC assembly. Our data further suggest that the two different meiotic cohesin complexes are distinctly arranged within the axes: Although cohesin complexes containing the kleisin REC-8 protrude above and below the plane defined by the SC, complexes containing COH-3 or -4 kleisins form a central core, which may physically separate sister chromatids. This organization may help to explain the role of the chromosome axes in promoting interhomolog repair of meiotic double-strand breaks by inhibiting intersister repair.

The synaptonemal complex has liquid crystalline properties and spatially regulates meiotic recombination factors.

The synaptonemal complex (SC) is a polymer that spans ~100 nm between paired homologous chromosomes during meiosis. Its striated, periodic appearance in electron micrographs led to the idea that transverse filaments within this structure 'crosslink' the axes of homologous chromosomes, stabilizing their pairing. SC proteins can also form polycomplexes, three-dimensional lattices that recapitulate the periodic structure of SCs but do not associate with chromosomes. Here we provide evidence that SCs and polycomplexes contain mobile subunits and that their assembly is promoted by weak hydrophobic interactions, indicative of a liquid crystalline phase. We further show that in the absence of recombination intermediates, polycomplexes recapitulate the dynamic localization of pro-crossover factors during meiotic progression, revealing how the SC might act as a conduit to regulate chromosome-wide crossover distribution. Properties unique to liquid crystals likely enable long-range signal transduction along meiotic chromosomes and underlie the rapid evolution of SC proteins.

Meiotic recombination and the crossover assurance checkpoint in Caenorhabditis elegans.

During meiotic prophase, chromosomes pair and synapse with their homologs and undergo programmed DNA double-strand break (DSB) formation to initiate meiotic recombination. These DSBs are processed to generate a limited number of crossover recombination products on each chromosome, which are essential to ensure faithful segregation of homologous chromosomes. The nematode Caenorhabditis elegans has served as an excellent model organism to investigate the mechanisms that drive and coordinate these chromosome dynamics during meiosis. Here we focus on our current understanding of the regulation of DSB induction in C. elegans. We also review evidence that feedback regulation of crossover formation prolongs the early stages of meiotic prophase, and discuss evidence that this can alter the recombination pattern, most likely by shifting the genome-wide distribution of DSBs.

The auxin-inducible degradation (AID) system enables versatile conditional protein depletion in C. elegans.

Experimental manipulation of protein abundance in living cells or organisms is an essential strategy for investigation of biological regulatory mechanisms. Whereas powerful techniques for protein expression have been developed in Caenorhabditis elegans, existing tools for conditional disruption of protein function are far more limited. To address this, we have adapted the auxin-inducible degradation (AID) system discovered in plants to enable conditional protein depletion in C. elegans. We report that expression of a modified Arabidopsis TIR1 F-box protein mediates robust auxin-dependent depletion of degron-tagged targets. We document the effectiveness of this system for depletion of nuclear and cytoplasmic proteins in diverse somatic and germline tissues throughout development. Target proteins were depleted in as little as 20-30 min, and their expression could be re-established upon auxin removal. We have engineered strains expressing TIR1 under the control of various promoter and 3' UTR sequences to drive tissue-specific or temporally regulated expression. The degron tag can be efficiently introduced by CRISPR/Cas9-based genome editing. We have harnessed this system to explore the roles of dynamically expressed nuclear hormone receptors in molting, and to analyze meiosis-specific roles for proteins required for germ line proliferation. Together, our results demonstrate that the AID system provides a powerful new tool for spatiotemporal regulation and analysis of protein function in a metazoan model organism.

The Chromosome Axis Mediates Feedback Control of CHK-2 to Ensure Crossover Formation in C. elegans.

CHK-2 kinase is a master regulator of meiosis in C. elegans. Its activity is required for homolog pairing and synapsis and for double-strand break formation, but how it drives and coordinates these pathways to ensure crossover formation remains unknown. Here we show that CHK-2 promotes pairing and synapsis by phosphorylating a family of zinc finger proteins that bind to specialized regions on each chromosome known as pairing centers, priming their recruitment of the Polo-like kinase PLK-2. This knowledge enabled the development of a phospho-specific antibody as a tool to monitor CHK-2 activity. When either synapsis or crossover formation is impaired, CHK-2 activity is prolonged, and meiotic progression is delayed. We show that this common feedback circuit is mediated by interactions among a network of HORMA domain proteins within the chromosome axis and generates a graded signal. These findings reveal conserved regulatory mechanisms that ensure faithful meiotic chromosome segregation in diverse species.

Direct Visualization Reveals Kinetics of Meiotic Chromosome Synapsis.

The synaptonemal complex (SC) is a conserved protein complex that stabilizes interactions along homologous chromosomes (homologs) during meiosis. The SC regulates genetic exchanges between homologs, thereby enabling reductional division and the production of haploid gametes. Here, we directly observe SC assembly (synapsis) by optimizing methods for long-term fluorescence recording in C. elegans. We report that synapsis initiates independently on each chromosome pair at or near pairing centers-specialized regions required for homolog associations. Once initiated, the SC extends rapidly and mostly irreversibly to chromosome ends. Quantitation of SC initiation frequencies and extension rates reveals that initiation is a rate-limiting step in homolog interactions. Eliminating the dynein-driven chromosome movements that accompany synapsis severely retards SC extension, revealing a new role for these conserved motions. This work provides the first opportunity to directly observe and quantify key aspects of meiotic chromosome interactions and will enable future in vivo analysis of germline processes.

The chromosome axis controls meiotic events through a hierarchical assembly of HORMA domain proteins.

Proteins of the HORMA domain family play central, but poorly understood, roles in chromosome organization and dynamics during meiosis. In Caenorhabditis elegans, four such proteins (HIM-3, HTP-1, HTP-2, and HTP-3) have distinct but overlapping functions. Through combined biochemical, structural, and in vivo analysis, we find that these proteins form hierarchical complexes through binding of their HORMA domains to cognate peptides within their partners' C-terminal tails, analogous to the "safety belt" binding mechanism of Mad2. These interactions are critical for recruitment of HIM-3, HTP-1, and HTP-2 to chromosome axes. HTP-3, in addition to recruiting the other HORMA domain proteins to the axis, plays an independent role in sister chromatid cohesion and double-strand break formation. Finally, we find that mammalian HORMAD1 binds a motif found both at its own C terminus and at that of HORMAD2, indicating that this mode of intermolecular association is a conserved feature of meiotic chromosome structure in eukaryotes.

The C. elegans DSB-2 protein reveals a regulatory network that controls competence for meiotic DSB formation and promotes crossover assurance.

For most organisms, chromosome segregation during meiosis relies on deliberate induction of DNA double-strand breaks (DSBs) and repair of a subset of these DSBs as inter-homolog crossovers (COs). However, timing and levels of DSB formation must be tightly controlled to avoid jeopardizing genome integrity. Here we identify the DSB-2 protein, which is required for efficient DSB formation during C. elegans meiosis but is dispensable for later steps of meiotic recombination. DSB-2 localizes to chromatin during the time of DSB formation, and its disappearance coincides with a decline in RAD-51 foci marking early recombination intermediates and precedes appearance of COSA-1 foci marking CO-designated sites. These and other data suggest that DSB-2 and its paralog DSB-1 promote competence for DSB formation. Further, immunofluorescence analyses of wild-type gonads and various meiotic mutants reveal that association of DSB-2 with chromatin is coordinated with multiple distinct aspects of the meiotic program, including the phosphorylation state of nuclear envelope protein SUN-1 and dependence on RAD-50 to load the RAD-51 recombinase at DSB sites. Moreover, association of DSB-2 with chromatin is prolonged in mutants impaired for either DSB formation or formation of downstream CO intermediates. These and other data suggest that association of DSB-2 with chromatin is an indicator of competence for DSB formation, and that cells respond to a deficit of CO-competent recombination intermediates by prolonging the DSB-competent state. In the context of this model, we propose that formation of sufficient CO-competent intermediates engages a negative feedback response that leads to cessation of DSB formation as part of a major coordinated transition in meiotic prophase progression. The proposed negative feedback regulation of DSB formation simultaneously (1) ensures that sufficient DSBs are made to guarantee CO formation and (2) prevents excessive DSB levels that could have deleterious effects.