Trapping a cross-linked lysine–tryptophan radical in the catalytic cycle of the radical SAM enzyme SuiB

The radical S-adenosylmethionine (rSAM) enzyme SuiB catalyzes the formation of an unusual carbon–carbon bond between the sidechains of lysine (Lys) and tryptophan (Trp) in the biosynthesis of a ribosomal peptide natural product. Prior work on SuiB has suggested that the Lys–Trp cross-link is formed via radical electrophilic aromatic substitution (rEAS), in which an auxiliary [4Fe-4S] cluster (AuxI), bound in the SPASM domain of SuiB, carries out an essential oxidation reaction during turnover. Despite the prevalence of auxiliary clusters in over 165,000 rSAM enzymes, direct evidence for their catalytic role has not been reported. Here, we have used electron paramagnetic resonance (EPR) spectroscopy to dissect the SuiB mechanism. Our studies reveal substrate-dependent redox potential tuning of the AuxI cluster, constraining it to the oxidized [4Fe-4S]²⁺ state, which is active in catalysis. We further report the trapping and characterization of an unprecedented cross-linked Lys–Trp radical (Lys–Trp•) in addition to the organometallic Ω intermediate, providing compelling support for the proposed rEAS mechanism. Finally, we observe oxidation of the Lys–Trp• intermediate by the redox-tuned [4Fe-4S]²⁺ AuxI cluster by EPR spectroscopy. Our findings provide direct evidence for a role of a SPASM domain auxiliary cluster and consolidate rEAS as a mechanistic paradigm for rSAM enzyme-catalyzed carbon–carbon bond-forming reactions.

The radical S-adenosylmethionine (rSAM) enzyme superfamily is the largest known in nature, with over 570,000 annotated and predominantly uncharacterized members spanning all domains of life (1–4). The uniting feature of rSAM enzymes is a [4Fe-4S] cluster, usually bound by a CX₃CX₂C motif that catalyzes reductive cleavage of SAM to form L-Met and a strongly oxidizing 5′-deoxyadenosyl radical (5′-dA•) (5–7). Recent studies on a suite of rSAM enzymes have revealed the presence of a previously unknown organometallic intermediate in this process, termed Ω, in which the 5′-C of 5′-dA• is bound to the unique iron of the [4Fe-4S] cluster (Fig. 1A) (8, 9). Homolysis of the Fe–C bond ultimately liberates 5′-dA•, which abstracts a hydrogen atom from substrate to initiate a profoundly diverse set of chemical reactions in both primary and secondary metabolism, including DNA, cofactor, vitamin, and antibiotic biosynthesis (5, 10–13).Open in a separate windowFig. 1.(A) Accepted scheme for radical initiation in rSAM enzymes. (B) X-ray crystal structure of SuiB (PDB ID: 5V1T). The RS domain, SPASM domain, and RiPP recognition element are rendered blue, green, and pink, respectively. [4Fe-4S] clusters are shown as spheres with the distances separating them indicated. (C) Lys–Trp cross-link formation (20) catalyzed by SuiB. The carbon–carbon bond installed by SuiB is shown in red. (D and E) Previously proposed EAS (D) and rEAS (E) mechanisms for SuiB-catalyzed Lys–Trp cross-link formation.Of the 570,000 rSAM enzyme superfamily members, over a quarter (∼165,000 genes from the Enzyme Function Initiative-Enzyme Similarity Tool) possess C-terminal extensions, called SPASM and twitch domains, which bind auxiliary Fe-S clusters (4, 14–19). The SPASM domain typically binds two auxiliary Fe-S clusters and is named after the rSAM enzymes involved in the synthesis of subtilosin, pyrroloquinoline quinone, anaerobic sulfatase, and mycofactocin. The twitch domain is a truncated SPASM domain and only binds one auxiliary cluster (15). Despite the wide prevalence of these domains and the characterization of several different SPASM/twitch rSAM enzymes by spectroscopic and structural studies, direct evidence for their catalytic function(s) has remained elusive.We previously performed functional and structural characterization on the SPASM rSAM enzyme SuiB (Fig. 1B), which is involved in the biosynthesis of a ribosomal peptide natural product in human and mammalian microbiome streptococci (14, 20–22). SuiB introduces an unusual carbon–carbon bond onto its substrate peptide, SuiA, between the sidechains of Lys2 and Trp6 (Fig. 1C). The mechanism for this transformation is of broader relevance, as a number of enzymes, such as RrrB, PqqE, and MqnC (2, 23, 24), are known to join unactivated aliphatic and aromatic carbons to generate sp³-sp² cross-links. A general mechanistic paradigm for this class of transformations is not yet available. For SuiB, two pathways have been proposed (20), one through a typical electrophilic aromatic substitution (EAS) mechanism, which is involved in other enzyme-catalyzed indole modifications, such as indole prenylation or flavin adenine dinucleotide (FAD)-enzyme-dependent indole chlorination (25–27). In this pathway, the 5′-dA• generates an alkyl radical, which upon a second one-electron oxidation, creates an α,β-unsaturated amide electrophile with which the indole sidechain reacts via Michael addition (Fig. 1D). Lanthionine cross-links observed in diverse lanthipeptides are built by this general scheme, though via heterolytic chemistry, with Cys acting as the nucleophile (28, 29). Alternatively, a radical electrophilic aromatic substitution (rEAS) reaction has been proposed, wherein the alkyl radical, formed by 5′-dA•, would react with the indole sidechain to generate a radical σ complex, a cross-linked Lys–Trp radical (Lys–Trp•), which upon oxidation and rearomatization would yield product (Fig. 1E). In both mechanisms, AuxI is proposed as an oxidant. Although this role for an rSAM auxiliary cluster has been previously suggested (30, 31), it has yet to be directly demonstrated experimentally. Mechanistic studies have favored the rEAS pathway (20); however, intermediates in the reaction of SuiB and enzymes that catalyze similar reactions have not yet been detected (15).In the current work, we sought to differentiate between the proposed mechanisms by trapping intermediates in the catalytic cycle of SuiB and characterizing them using electron paramagnetic resonance (EPR) spectroscopy. We report observation of three transient reaction intermediates, most importantly the sought-after Lys–Trp•, which is fundamentally different from previously characterized Trp radicals, as it is cross-linked and carries an indole tetrahedral center. We also provide evidence for AuxI as the oxidant of the Lys–Trp• intermediate as well as insights into redox potential changes of Fe-S clusters in SuiB that accompany SuiA binding. Together, our findings support the rEAS pathway for formation of the sp³-sp² cross-link and carry important implications for other enzymes that catalyze related transformations.     

Geometrical manipulation of complex supramolecular tessellations by hierarchical assembly of amphiphilic platinum(II) complexes

Here we report complex supramolecular tessellations achieved by the directed self-assembly of amphiphilic platinum(II) complexes. Despite the twofold symmetry, these geometrically simple molecules exhibit complicated structural hierarchy in a columnar manner. A possible key to such an order increase is the topological transition into circular trimers, which are noncovalently interlocked by metal···metal and π–π interactions, thereby allowing for cofacial stacking in a prismatic assembly. Another key to success is to use the immiscibility of the tailored hydrophobic and hydrophilic sidechains. Their phase separation leads to the formation of columnar crystalline nanostructures homogeneously oriented on the substrate, featuring an unusual geometry analogous to a rhombitrihexagonal Archimedean tiling. Furthermore, symmetry lowering of regular motifs by design results in an orthorhombic lattice obtained by the coassembly of two different platinum(II) amphiphiles. These findings illustrate the potentials of supramolecular engineering in creating complex self-assembled architectures of soft materials.

Tessellation in two dimensions (2D) is a very old topic in geometry on how one or more shapes can be periodically arranged to fill a Euclidean plane without any gaps. Tessellation principles have been extensively applied in decorative art since the early times. In natural sciences, there has been a growing attention on creating ordered structures with increasingly complex architectures inspired by semi-regular Archimedean tilings (ATs) and quasicrystalline textures on account of their intriguing physical properties (1–5) and biological functions (6). Recent advances in this regard have been achieved in various fields of supramolecular science, including the programmable self-assembly of DNA molecules (7), coordination-driven assembly (8–10), supramolecular interfacial engineering (11–13), crystallization of organic polygons (14, 15), colloidal particle superlattices (16), and other soft-matter systems (17–20). Moreover, tessellation in 2D can overcome the topological frustration to generate complex semi- or non-regular patterns by using geometrically simple motifs. As exemplified by the self-templating assembly of spherical soft microparticles (21), a vast array of 2D micropatterns encoding non-regular tilings, such as rectangular, rhomboidal, hexagonal, and herringbone superlattices were obtained by layer-by-layer strategy at a liquid–liquid interface. Tessellation principles have also been extended to the self-assembly of giant molecules in three dimensions (3D). Superlattices with high space-group symmetry (Im3¯m, Pm3¯n, and P4₂/mnm) were reported in dendrimers and dendritic polymers by Percec and coworkers (22–24). Recently, Cheng and coworkers identified the highly ordered Frank–Kasper phases obtained from giant amphiphiles containing molecular nanoparticles (25–28). Despite such advancements made in the field of soft matter, an understanding of how structural ordering in supramolecular materials is influenced by the geometric factors of its constituent molecules has so far remained elusive.In light of these developments and the desire to explore the supramolecular systems, square-planar platinum(II) (Pt^II) polypyridine complexes may serve as an ideal candidate for model studies not only because of their intriguing spectroscopic and luminescence properties (29, 30), but also because of their propensity to form supramolecular polymers or oligomers via noncovalent Pt···Pt and π–π interactions (31–39). Although rod-shaped and lamellar structures are the most commonly observed in the self-assembly of planar Pt^II complexes (34–39), 2D-ordered nanostructures, such as the hexagonally packed columns (31, 40) and honeycomb-like networks (41–43), were recently first demonstrated by us.Herein, we report a serendipitous discovery of a C_2h-symmetric Pt^II amphiphile (Fig. 1A) that can hierarchically self-assemble into a 3D-ordered nanostructure with hexagonal geometry. Interestingly, this structurally anisotropic molecule possibly undergoes topological transition and interlocks to form its circular trimer by noncovalent Pt···Pt and π–π interactions (Fig. 1B). The resultant triangular motif is architecturally stabilized and preorganized for one-dimensional (1D) prismatic assembly (Fig. 1C). Together with the phase separation of the tailored hydrophobic and hydrophilic sidechains, an unusual and unique 3D hexagonal lattice is formed (Fig. 1D), in which the Pt centers adopt a rare rhombitrihexagonal AT-like order. Finally, the nanoarchitecture develops in a hierarchical manner on the substrate due to the homogeneous nucleation (Fig. 1E).Open in a separate windowFig. 1.Hierarchical self-assembly of Pt^II amphiphile into hexagonal ordering. (A) Space-filling (CPK) model of a C_2h-symmetric Pt^II amphiphile (1). All of the hydrogen atoms and counterions are omitted for clarity. (B) CPK representations of possible models of regular triangular, tetragonal, pentagonal, and hexagonal motifs formed with Pt···Pt and π–π stacking. These motifs possess a hydrophilic core (red) with various diameters wrapped by a hydrophobic shell comprising long alkyl chains (gray). (C) CPK representation of a 1D prismatic structure consisting of circular trimers with long-range Pt···Pt and π–π stacking. (D) CPK representation of a 3D columnar lattice constructed by the prismatic assemblies adopting a rare rhombitrihexagonal AT-like order. With the assistance of the phase separation, the hydrophobic domain serves as a discrete column associated with six prismatic neighbors. (E) Schematic representation of the nanoarchitecture with homogeneous orientation.     

Assessment of enzyme active site positioning and tests of catalytic mechanisms through X-ray–derived conformational ensembles

How enzymes achieve their enormous rate enhancements remains a central question in biology, and our understanding to date has impacted drug development, influenced enzyme design, and deepened our appreciation of evolutionary processes. While enzymes position catalytic and reactant groups in active sites, physics requires that atoms undergo constant motion. Numerous proposals have invoked positioning or motions as central for enzyme function, but a scarcity of experimental data has limited our understanding of positioning and motion, their relative importance, and their changes through the enzyme’s reaction cycle. To examine positioning and motions and test catalytic proposals, we collected “room temperature” X-ray crystallography data for Pseudomonas putida ketosteroid isomerase (KSI), and we obtained conformational ensembles for this and a homologous KSI from multiple PDB crystal structures. Ensemble analyses indicated limited change through KSI’s reaction cycle. Active site positioning was on the 1- to 1.5-Å scale, and was not exceptional compared to noncatalytic groups. The KSI ensembles provided evidence against catalytic proposals invoking oxyanion hole geometric discrimination between the ground state and transition state or highly precise general base positioning. Instead, increasing or decreasing positioning of KSI’s general base reduced catalysis, suggesting optimized Ångstrom-scale conformational heterogeneity that allows KSI to efficiently catalyze multiple reaction steps. Ensemble analyses of surrounding groups for WT and mutant KSIs provided insights into the forces and interactions that allow and limit active-site motions. Most generally, this ensemble perspective extends traditional structure–function relationships, providing the basis for a new era of “ensemble–function” interrogation of enzymes.

The central role of enzymes in biology is embodied in the decades of effort spent to deeply investigate the origins of their catalysis (e.g., refs. 1–6). Enzyme studies now routinely identify the active-site groups that interact with substrates and reveal their roles in binding and in facilitating chemical transformations. Nevertheless, these so-called “catalytic groups” alone, outside of the context of a folded enzyme, do not account for the enormous rate enhancements and exquisite specificities exhibited by enzymes (4). Classic proposals for enzyme catalysis have invoked the importance of positioning of active-site groups within a folded enzyme and of substrates localized and positioned by binding interactions (6–15). While these proposals universally invoke restricted motion of catalytic groups, the amount of restriction and the amount of catalysis provided by that restriction has been the subject of much discussion and debate (16–20). Conversely, it is also clear that motions are inherent to enzymes, and that conformational transitions and structural rearrangements are important for enzyme function (e.g., refs. 11 and 21–23). Considering both positioning and motions, it has been recognized that: “For catalysis, flexible but not too flexible, as well as rigid but not too rigid, is essential. Specifically, the protein must be rigid enough to maintain the required structure but flexible enough to permit atomic movements as the reaction proceeds” (3).The importance of both positioning and motions to enzyme function suggests a nuanced view of enzyme catalysis and underscores the need for direct experimental measurements of positioning and motions within enzymes.As Feynman noted, “Everything that living things do can be understood in terms of the jigglings and wigglings of atoms” (24). But simply observing motions of active-site residues does not tell us how enzymes achieve catalysis. To understand enzymes, we want to know how much an enzyme dampens and alters the motions of catalytic residues. We want to know which increases or decreases in motion increase or decrease the reaction rate and what interactions and forces are most responsible for dampening motions. With this information we may be able to better design new enzymes. Additionally, to what extent are active-site residues positioned upon folding of the enzyme, or adjusted as the reaction proceeds, and are active-site residues more precisely positioned than residues throughout an enzyme?To address fundamental questions about how enzymes function and evolve, and how to ultimately design highly efficient enzymes, we need to obtain experimental information about enzyme conformation ensembles: The distribution of enzyme states dictated by their highly complex multidimensional energy landscapes over which conformational rearrangements occur. Observations of well-resolved electron densities from X-ray diffraction data indicate positioning of residues in and around the active site, but do not provide information on the extent and nature of that positioning. Crystallographic B-factors of residues are sometimes used to infer motions, but are only indirectly related to intrinsic motion and contain contributions from additional factors, such as crystallographic order (25, 26). NMR experiments identify groups with greater motional freedom and can provide temporal information, but these experiments typically lack information about the directions and extent of these motions (27). Molecular dynamics simulations provide atomic-level models for entire systems, but we currently lack the rigorous experimental tests needed to determine whether or not computational outputs reflect actual physical behavior, which prevents firm mechanistic conclusions from being inferred (28, 29).Two X-ray crystallographic approaches have recently emerged that can provide experimentally-derived conformational ensemble information: High-sequence similarity Protein Data Bank (PDB) structural ensembles (referred to as “pseudoensembles” herein) (30, 31) and multiconformer models from X-ray data obtained at temperatures above the protein’s glass transition (referred to as “room temperature” or ”RT” X-ray diffraction in the literature and herein) (22, 32, 33). These approaches are complementary. Pseudoensembles provide information about residues that move in concert (i.e., coupled motions) but require dozens of structures (see also SI Appendix, Supplementary Text 1). RT X-ray data from single crystals can provide multiconformer models, so that ensemble information about new complexes and mutants can more readily be acquired, but do not provide direct information about coupled motions. Furthermore, RT X-ray studies provide direct information about equilibrium distributions without cryocooling, which can alter and quench motions, and without assuming that different cryocooled crystals reproduce an equilibrium distribution of states (32, 34–36).Here we demonstrate consistency between these approaches and take advantage of the strengths of each: The ability to evaluate correlated side-chain rearrangements in and near the active site via pseudoensembles, and the ability to obtain new ensemble-type information of new states from single X-ray datasets at temperatures above the glass transition. Importantly, these analyses report on conformational heterogeneity and cannot give information about the timescales of motions and interconversions between states. Additionally, each traditional model within the pseudoensemble represents predominantly a single rather than average state and combining these states captures an ensemble distribution. Similarly, the alternate conformations in multiconformer models explicitly reduce bias toward average structures of multistate systems. Focusing on a model enzyme with very high-resolution data and with ligands representing steps along its reaction path has allowed us to obtain insights that would not be possible from static structures, from either ensemble approach alone or from less-extensive or lower-resolution data.We chose to investigate the enzyme ketosteroid isomerase (KSI) (Fig. 1) because of our ability to obtain high-resolution diffraction data, because of the accumulated wealth of structural and mechanistic information, and because of KSI’s use of catalytic strategies common to many enzymes. As a single-substrate enzyme, KSI allows structural information to be obtained with a bonified reactant bound. Furthermore, we obtained ensemble data for KSI from two species, which gave consistent results and allowed us to address unresolved questions from decades of KSI studies. We also used our ensembles from these KSI homologs to ask—and answer—more general questions. Our in-depth analyses of KSI bring an ensemble perspective to bear on traditional structure–function studies and provide the basis for a new era of ensemble–function studies.Open in a separate windowFig. 1.The KSI reaction. Reaction mechanism and schematic depiction of the active site (A) and its 3D organization (B) [PDB ID code 1OH0 (87)]. KSI catalyzes double bond isomerization of steroid substrates (shown for the substrate 5-androstene-3,17-dione) utilizing a general acid/base D40 (which we refer to herein as a general base, for simplicity), and an oxyanion hole composed of the side chains of Y16 and D103 (protonated); general base and oxyanion hole residues are colored in red and orange, respectively. The product in A, 4-androstene-3,17-dione, is the substrate of the reverse reaction and was used for RT X-ray crystallography herein. (C) Examples of oxyanion KSI TSAs used for the KSI TSA ensembles: Equilenin (Left) and a substituted phenolate (Right).     

Heading perception depends on time-varying evolution of optic flow

There is considerable support for the hypothesis that perception of heading in the presence of rotation is mediated by instantaneous optic flow. This hypothesis, however, has never been tested. We introduce a method, termed “nonvarying phase motion,” for generating a stimulus that conveys a single instantaneous optic flow field, even though the stimulus is presented for an extended period of time. In this experiment, observers viewed stimulus videos and performed a forced-choice heading discrimination task. For nonvarying phase motion, observers made large errors in heading judgments. This suggests that instantaneous optic flow is insufficient for heading perception in the presence of rotation. These errors were mostly eliminated when the velocity of phase motion was varied over time to convey the evolving sequence of optic flow fields corresponding to a particular heading. This demonstrates that heading perception in the presence of rotation relies on the time-varying evolution of optic flow. We hypothesize that the visual system accurately computes heading, despite rotation, based on optic acceleration, the temporal derivative of optic flow.

James Gibson first remarked that the instantaneous motion of points on the retina (Fig. 1A) can be formally described as a two-dimensional (2D) field of velocity vectors called the “optic flow field” (or “optic flow”) (1). Such optic flow, caused by an observer’s movement relative to the environment, conveys information about self-motion and the structure of the visual scene (1–15). When an observer translates in a given direction along a straight path, the optic flow field radiates from a point in the image with zero velocity, or singularity, called the focus of expansion (Fig. 1B). It is well known that under such conditions, one can accurately estimate one’s “heading” (i.e., instantaneous direction of translation in retinocentric coordinates) by simply locating the focus of expansion (SI Appendix). However, if there is angular rotation in addition to translation (by moving along a curved path or by a head or eye movement), the singularity in the optic flow field will be displaced such that it no longer corresponds to the true heading (Fig. 1 C and D). In this case, if one estimates heading by locating the singularity, the estimate will be biased away from the true heading. This is known as the rotation problem (14).Open in a separate windowFig. 1.Projective geometry, the rotation problem, time-varying optic flow, and the optic acceleration hypothesis. (A) Viewer-centered coordinate frame and perspective projection. Because of motion between the viewpoint and the scene, a 3D surface point traverses a path in 3D space. Under perspective projection, the 3D path of this point projects onto a 2D path in the image plane (retina), the temporal derivative of which is called image velocity. The 2D velocities associated with all visible points define a dense 2D vector field called the optic flow field. (B–D) Illustration of the rotation problem. (B) Optic flow for pure translation (1.5-m/s translation speed, 0° heading, i.e., heading in the direction of gaze). Optic flow singularity (red circle) corresponds to heading (purple circle). (C) Pure rotation, for illustrative purposes only and not corresponding to any experimental condition (2°/s rightward rotation). (D) Translation + rotation (1.5 m/s translation speed, 0° heading, 2°/s rightward rotation). Optic flow singularity (red circle) is displaced away from heading (purple circle). (E) Three frames from a video depicting movement along a circular path with the line-of-sight initially perpendicular to a single fronto-parallel plane composed of black dots. (F) Time-varying evolution of optic flow. The first optic flow field reflects image motion between the first and second frames of the video. The second optic flow field reflects image motion between the second and third frames of the video. For this special case (circular path), the optic flow field evolves (and the optic flow singularity drifts) only due to the changing depth of the environment relative to the viewpoint. (G) Illustration of the optic acceleration hypothesis. Optic acceleration is the derivative of optic flow over time (here, approximated as the difference between the second and first optic flow fields). The singularity of the optic acceleration field corresponds to the heading direction. Acceleration vectors autoscaled for visibility.Computer vision researchers and vision scientists have developed a variety of algorithms that accurately and precisely extract observer translation and rotation from optic flow, thereby solving the rotation problem. Nearly all of these rely on instantaneous optic flow (i.e., a single optic flow field) (4, 9, 16–25) with few exceptions (26–29). However, it is unknown whether these algorithms are commensurate with the neural computations underlying heading perception.The consensus of opinion in the experimental literature is that human observers can estimate heading (30, 31) from instantaneous optic flow, in the absence of additional information (5, 10, 15, 32–34). Even so, there are reports of systematic biases in heading perception (11); the visual consequences of rotation (eye, head, and body) can bias heading judgments (10, 15, 35–37), with the amount of bias typically proportional to the magnitude of rotation. Other visual factors, such as stereo cues (38, 39), depth structure (8, 10, 40–43), and field of view (FOV) (33, 42–44) can modulate the strength of these biases. Errors in heading judgments have been reported to be greater when eye (35–37, 45, 46) or head movements (37) are simulated versus when they are real, which has been taken to mean that observers require extraretinal information, although there is also evidence to the contrary (10, 15, 33, 40, 41, 44, 47–50). Regardless, to date no one has tested whether heading perception (even with these biases) is based on instantaneous optic flow or on the information available in how the optic flow field evolves over time. Some have suggested that heading estimates rely on information accumulated over time (32, 44, 51), but no one has investigated the role of time-varying optic flow without confounding it with stimulus duration (i.e., the duration of evidence accumulation).In this study, we employed an application of an image processing technique that ensured that only a single optic flow field was available to observers, even though the stimulus was presented for an extended period of time. We called this condition “nonvarying phase motion” or “nonvarying”: The phases of two component gratings comprising each stationary stimulus patch shifted over time at a constant rate, causing a percept of motion in the absence of veridical movement (52). Phase motion also eliminated other cues that may otherwise have been used for heading judgments, including image point trajectories (15, 32) and their spatial compositions (i.e., looming) (53, 54). For nonvarying phase motion, observers exhibited large biases in heading judgments in the presence of rotation. A second condition, “time-varying phase motion,” or “time-varying,” included acceleration by varying the velocity of phase motion over time to match the evolution of a sequence of optic flow fields. Doing so allowed observers to compensate for the confounding effect of rotation on optic flow, making heading perception nearly veridical. This demonstrates that heading perception in the presence of rotation relies on the time-varying evolution of optic flow.     

From the Cover: Neurophysiological coordination of duet singing

Coordination of behavior for cooperative performances often relies on linkages mediated by sensory cues exchanged between participants. How neurophysiological responses to sensory information affect motor programs to coordinate behavior between individuals is not known. We investigated how plain-tailed wrens (Pheugopedius euophrys) use acoustic feedback to coordinate extraordinary duet performances in which females and males rapidly take turns singing. We made simultaneous neurophysiological recordings in a song control area “HVC” in pairs of singing wrens at a field site in Ecuador. HVC is a premotor area that integrates auditory feedback and is necessary for song production. We found that spiking activity of HVC neurons in each sex increased for production of its own syllables. In contrast, hearing sensory feedback produced by the bird’s partner decreased HVC activity during duet singing, potentially coordinating HVC premotor activity in each bird through inhibition. When birds sang alone, HVC neurons in females but not males were inhibited by hearing the partner bird. When birds were anesthetized with urethane, which antagonizes GABAergic (γ-aminobutyric acid) transmission, HVC neurons were excited rather than inhibited, suggesting a role for GABA in the coordination of duet singing. These data suggest that HVC integrates information across partners during duets and that rapid turn taking may be mediated, in part, by inhibition.

Animals routinely rely on sensory feedback for the control of their own behavior. In cooperative performances, such sensory feedback can include cues produced by other participants (1–8). For example, in interactive vocal communication, including human speech, individuals take turns vocalizing. This “turn taking” is a consequence of each participant responding to auditory cues from a partner (4–6, 9, 10). The role of such “heterogenous” (other-generated) feedback in the control of vocal turn taking and other cooperative performances is largely unknown.Plain-tailed wrens (Pheugopedius euophrys) are neotropical songbirds that cooperate to produce extraordinary duet performances but also sing by themselves (Fig. 1A) (4, 10, 11). Singing in plain-tailed wrens is performed by both females and males and used for territorial defense and other functions, including mate guarding and attraction (1, 11–16). During duets, female and male plain-tailed wrens take turns, alternating syllables at a rate of between 2 and 5 Hz (Fig. 1A) (4, 11).Open in a separate windowFig. 1.Neural control of solo and duet singing in plain-tailed wrens. (A) Spectrogram of a singing bout that included male solo syllables (blue line, top) followed by a duet. Solo syllables for both sexes (only male solo syllables are shown here) are sung at lower amplitudes than syllables produced in duets. Note that the smeared appearance of wren syllables in spectrograms reflects the acoustic structure of plain-tailed wren singing. (B and C) Each bird has a motor system that is used to produce song and sensory systems that mediate feedback. (B) During solo singing, the bird hears its own song, which is known as autogenous feedback (orange). (C) During duet singing, each bird hears both its own singing and the singing of its partner, known as heterogenous feedback (green). The key difference between solo and duet singing is heterogenous feedback that couples the neural systems of the two birds. This coupling results in changes in syllable amplitude and timing in both birds.There is a categorical difference between solo and duet singing. In solo singing, the singing bird receives only autogenous (hearing its own vocalization) feedback (Fig. 1B). The partner may hear the solo song if it is nearby, a heterogenous (other-generated) cue. In duet singing, birds receive both heterogenous and autogenous feedback as they alternate syllable production (Fig. 1C). Participants use heterogenous feedback during duet singing for precise timing of syllable production (4, 11). For example, when a male temporarily stops participating in a duet, the duration of intersyllable intervals between female syllables increases (4), showing an effect of heterogenous feedback on the timing of syllable production.How does the brain of each wren integrate heterogenous acoustic cues to coordinate the precise timing of syllable production between individuals during duet performances? To address this question, we examined neurophysiological activity in HVC, a nucleus in the nidopallium [an analogue of mammalian cortex (17, 18)]. HVC is necessary for song learning, production, and timing in species of songbirds that do not perform duets (19–24). Neurons in HVC are active during singing and respond to playback of the bird’s own learned song (25–27). In addition, recent work has shown that HVC is also involved in vocal turn taking (19).To examine the role of heterogenous feedback in the control of duet performances, we compared neurophysiological activity in HVC when female or male wrens sang solo syllables with syllables sung during duets. Neurophysiological recordings were made in awake and anesthetized pairs of wrens at the Yanayacu Biological Station and Center for Creative Studies on the slopes of the Antisana volcano in Ecuador. We found that heterogenous cues inhibited HVC activity during duet performances in both females and males, but inhibition was only observed in females during solo singing.     

A solution to a sex ratio puzzle in Melittobia wasps

The puzzling sex ratio behavior of Melittobia wasps has long posed one of the greatest questions in the field of sex allocation. Laboratory experiments have found that, in contrast to the predictions of theory and the behavior of numerous other organisms, Melittobia females do not produce fewer female-biased offspring sex ratios when more females lay eggs on a patch. We solve this puzzle by showing that, in nature, females of Melittobia australica have a sophisticated sex ratio behavior, in which their strategy also depends on whether they have dispersed from the patch where they emerged. When females have not dispersed, they lay eggs with close relatives, which keeps local mate competition high even with multiple females, and therefore, they are selected to produce consistently female-biased sex ratios. Laboratory experiments mimic these conditions. In contrast, when females disperse, they interact with nonrelatives, and thus adjust their sex ratio depending on the number of females laying eggs. Consequently, females appear to use dispersal status as an indirect cue of relatedness and whether they should adjust their sex ratio in response to the number of females laying eggs on the patch.

Sex allocation has produced many of the greatest success stories in the study of social behaviors (1–4). Time and time again, relatively simple theory has explained variation in how individuals allocate resources to male and female reproduction. Hamilton’s local mate competition (LMC) theory predicts that when n diploid females lay eggs on a patch and the offspring mate before the females disperse, the evolutionary stable proportion of male offspring (sex ratio) is (n − 1)/2n (Fig. 1) (5). A female-biased sex ratio is favored to reduce competition between sons (brothers) for mates and to provide more mates (daughters) for those sons (6–8). Consistent with this prediction, females of >40 species produce female-biased sex ratios and reduce this female bias when multiple females lay eggs on the same patch (higher n; Fig. 1) (9). The fit of data to theory is so good that the sex ratio under LMC has been exploited as a “model trait” to study the factors that can constrain “perfect adaptation” (4, 10–13).Open in a separate windowFig. 1.LMC. The sex ratio (proportion of sons) is plotted versus the number of females laying eggs on a patch. The bright green dashed line shows the LMC theory prediction for the haplodiploid species (5, 39). A more female-biased sex ratio is favored in haplodiploids because inbreeding increases the relative relatedness of mothers to their daughters (7, 32). Females of many species adjust their offspring sex ratio as predicted by theory, such as the parasitoid Nasonia vitripennis (green diamonds) (82). In contrast, the females of several Melittobia species, such as M. australica, continue to produce extremely female-biased sex ratios, irrespective of the number of females laying eggs on a patch (blue squares) (15).In stark contrast, the sex ratio behavior of Melittobia wasps has long been seen as one of the greatest problems for the field of sex allocation (3, 4, 14–21). The life cycle of Melittobia wasps matches the assumptions of Hamilton’s LMC theory (5, 15, 19, 21). Females lay eggs in the larvae or pupae of solitary wasps and bees, and then after emergence, female offspring mate with the short-winged males, who do not disperse. However, laboratory experiments on four Melittobia species have found that females lay extremely female-biased sex ratios (1 to 5% males) and that these extremely female-biased sex ratios change little with increasing number of females laying eggs on a patch (higher n; Fig. 1) (15, 17–20, 22). A number of hypotheses to explain this lack of sex ratio adjustment have been investigated and rejected, including sex ratio distorters, sex differential mortality, asymmetrical male competition, and reciprocal cooperation (15–18, 20, 22–26).We tested whether Melittobia’s unusual sex ratio behavior can be explained by females being related to the other females laying eggs on the same patch. After mating, some females disperse to find new patches, while some may stay at the natal patch to lay eggs on previously unexploited hosts (Fig. 2). If females do not disperse, they can be related to the other females laying eggs on the same host (27–31). If females laying eggs on a host are related, this increases the extent to which relatives are competing for mates and so can favor an even more female-biased sex ratio (28, 32–35). Although most parasitoid species appear unable to directly assess relatedness, dispersal behavior could provide an indirect cue of whether females are with close relatives (36–38). Consequently, we predict that when females do not disperse and so are more likely to be with closer relatives, they should maintain extremely female-biased sex ratios, even when multiple females lay eggs on a patch (28, 35).Open in a separate windowFig. 2.Host nest and dispersal manners of Melittobia. (A) Photograph of the prepupae of the leaf-cutter bee C. sculpturalis nested in a bamboo cane and (B) a diagram showing two ways that Melittobia females find new hosts. The mothers of C. sculpturalis build nursing nests with pine resin consisting of individual cells in which their offspring develop. If Melittobia wasps parasitize a host in a cell, female offspring that mate with males inside the cell find a different host on the same patch (bamboo cane) or disperse by flying to other patches.We tested whether the sex ratio of Melittobia australica can be explained by dispersal status in a natural population. We examined how the sex ratio produced by females varies with the number of females laying eggs on a patch and whether or not they have dispersed before laying eggs. To match our data to the predictions of theory, we developed a mathematical model tailored to the unique population structure of Melittobia, where dispersal can be a cue of relatedness. We then conducted a laboratory experiment to test whether Melittobia females are able to directly access the relatedness to other females and adjust their sex ratio behavior accordingly. Our results suggest that females are adjusting their sex ratio in response to both the number of females laying eggs on a patch and their relatedness to the other females. However, relatedness is assessed indirectly by whether or not they have dispersed. Consequently, the solution to the puzzling behavior reflects a more-refined sex ratio strategy.     

From the Cover: Evidence from South Africa for a protracted end-Permian extinction on land

Earth’s largest biotic crisis occurred during the Permo–Triassic Transition (PTT). On land, this event witnessed a turnover from synapsid- to archosauromorph-dominated assemblages and a restructuring of terrestrial ecosystems. However, understanding extinction patterns has been limited by a lack of high-precision fossil occurrence data to resolve events on submillion-year timescales. We analyzed a unique database of 588 fossil tetrapod specimens from South Africa’s Karoo Basin, spanning ∼4 My, and 13 stratigraphic bin intervals averaging 300,000 y each. Using sample-standardized methods, we characterized faunal assemblage dynamics during the PTT. High regional extinction rates occurred through a protracted interval of ∼1 Ma, initially co-occurring with low origination rates. This resulted in declining diversity up to the acme of extinction near the Daptocephalus–Lystrosaurus declivis Assemblage Zone boundary. Regional origination rates increased abruptly above this boundary, co-occurring with high extinction rates to drive rapid turnover and an assemblage of short-lived species symptomatic of ecosystem instability. The “disaster taxon” Lystrosaurus shows a long-term trend of increasing abundance initiated in the latest Permian. Lystrosaurus comprised 54% of all specimens by the onset of mass extinction and 70% in the extinction aftermath. This early Lystrosaurus abundance suggests its expansion was facilitated by environmental changes rather than by ecological opportunity following the extinctions of other species as commonly assumed for disaster taxa. Our findings conservatively place the Karoo extinction interval closer in time, but not coeval with, the more rapid marine event and reveal key differences between the PTT extinctions on land and in the oceans.

Mass extinctions are major perturbations of the biosphere resulting from a wide range of different causes including glaciations and sea level fall (1), large igneous provinces (2), and bolide impacts (3, 4). These events caused permanent changes to Earth’s ecosystems, altering the evolutionary trajectory of life (5). However, links between the broad causal factors of mass extinctions and the biological and ecological disturbances that lead to species extinctions have been difficult to characterize. This is because ecological disturbances unfold on timescales much shorter than the typical resolution of paleontological studies (6), particularly in the terrestrial record (6–8). Coarse-resolution studies have demonstrated key mass extinction phenomena including high extinction rates and lineage turnover (7, 9), changes in species richness (10), ecosystem instability (11), and the occurrence of disaster taxa (12). However, finer time resolutions are central to determining the association and relative timings of these effects, their potential causal factors, and their interrelationships. Achieving these goals represents a key advance in understanding the ecological mechanisms of mass extinctions.The end-Permian mass extinction (ca. 251.9 Ma) was Earth’s largest biotic crisis as measured by taxon last occurrences (13–15). Large outpourings from Siberian Trap volcanism (2) are the likely trigger of calamitous climatic changes, including a runaway greenhouse effect and ocean acidification, which had profound consequences for life on land and in the oceans (16–18). An estimated 81% of marine species (19) and 89% of tetrapod genera became extinct as established Permian ecosystems gave way to those of the Triassic. In the ocean, this included the complete extinction of reef-forming tabulate and rugose corals (20, 21) and significant losses in previously diverse ammonoid, brachiopod, and crinoid families (22). On land, many nonmammalian synapsids became extinct (16), and the glossopterid-dominated floras of Gondwana also disappeared (23). Stratigraphic sequences document a global “coral gap” and “coal gap” (24, 25), suggesting reef and forest ecosystems were rare or absent for up to 5 My after the event (26). Continuous fossil-bearing deposits documenting patterns of turnover across the Permian–Triassic transition (PTT) on land (27) and in the oceans (28) are geographically widespread (29, 30), including marine and continental successions that are known from China (31, 32) and India (33). Continental successions are known from Russia (34), Australia (35), Antarctica (36), and South Africa’s Karoo Basin (Fig. 1 and 37–40), the latter providing arguably the most densely sampled and taxonomically scrutinized (41–43) continental record of the PTT. The main extinction has been proposed to occur at the boundary between two biostratigraphic zones with distinctive faunal assemblages, the Daptocephalus and Lystrosaurus declivis assemblage zones (Fig. 1), which marks the traditional placement of the Permian–Triassic geologic boundary [(37) but see ref. 44]. Considerable research has attempted to understand the anatomy of the PTT in South Africa (38, 39, 45–52) and to place it in the context of biodiversity changes across southern Gondwana (53, 54) and globally (29, 31, 32, 44, 47, 55).Open in a separate windowFig. 1.Map of South Africa depicting the distribution of the four tetrapod fossil assemblage zones (Cistecephalus, Daptocephalus, Lystrosaurus declivis, Cynognathus) and our two study sites where fossils were collected in this study (sites A and B). Regional lithostratigraphy and biostratigraphy within the study interval are shown alongside isotope dilution–thermal ionization mass spectrometry dates retrieved by Rubidge et al., Botha et al., and Gastaldo et al. (37, 44, 80). The traditional (dashed red line) and associated PTB hypotheses for the Karoo Basin (37, 44) are also shown. Although traditionally associated with the PTB, the Daptocephalus–Lystrosaurus declivis Assemblage Zone boundary is defined by first appearances of co-occurring tetrapod assemblages, so its position relative to the three PTB hypotheses is unchanged. The Ripplemead member (*) has yet to be formalized by the South African Committee for Stratigraphy.Decades of research have demonstrated the richness of South Africa’s Karoo Basin fossil record, resulting in hundreds of stratigraphically well-documented tetrapod fossils across the PTT (37, 39, 56). This wealth of data has been used qualitatively to identify three extinction phases and an apparent early postextinction recovery phase (39, 45, 51). Furthermore, studies of Karoo community structure and function have elucidated the potential role of the extinction and subsequent recovery in breaking the incumbency of previously dominant clades, including synapsids (11, 57). Nevertheless, understanding patterns of faunal turnover and recovery during the PTT has been limited by the scarcity of quantitative investigations. Previous quantitative studies used coarsely sampled data (i.e., assemblage zone scale, 2 to 3 Ma time intervals) to identify low species richness immediately after the main extinction, potentially associated with multiple “boom and bust” cycles of primary productivity based on δ¹³C variation during the first 5 My of the Triassic (41, 58). However, many details of faunal dynamics in this interval remain unknown. Here, we investigate the dynamics of this major tetrapod extinction at an unprecedented time resolution (on the order of hundreds of thousands of years), using sample-standardized methods to quantify multiple aspects of regional change across the Cistecephalus, Daptocephalus, and Lystrosaurus declivis assemblage zones.     

Stable metal anodes enabled by a labile organic molecule bonded to a reduced graphene oxide aerogel

Metallic anodes (lithium, sodium, and zinc) are attractive for rechargeable battery technologies but are plagued by an unfavorable metal–electrolyte interface that leads to nonuniform metal deposition and an unstable solid–electrolyte interphase (SEI). Here we report the use of electrochemically labile molecules to regulate the electrochemical interface and guide even lithium deposition and a stable SEI. The molecule, benzenesulfonyl fluoride, was bonded to the surface of a reduced graphene oxide aerogel. During metal deposition, this labile molecule not only generates a metal-coordinating benzenesulfonate anion that guides homogeneous metal deposition but also contributes lithium fluoride to the SEI to improve Li surface passivation. Consequently, high-efficiency lithium deposition with a low nucleation overpotential was achieved at a high current density of 6.0 mA cm⁻². A Li|LiCoO₂ cell had a capacity retention of 85.3% after 400 cycles, and the cell also tolerated low-temperature (−10 °C) operation without additional capacity fading. This strategy was applied to sodium and zinc anodes as well.

Rechargeable batteries based on metal anodes including lithium (Li), sodium (Na), and zinc (Zn) show great promise in achieving high energy density (1–3). Unfortunately, the electrochemical interface of the metal anodes is not favorable for metal deposition. Metal nucleation is inhomogeneous at the surface, leading to the growth of metal dendrites (4–7) and the formation of an unstable solid–electrolyte interphase (SEI) that is incapable of protecting metals from the side reactions with the electrolyte (8–12).Substantial efforts have been devoted to stabilizing the interface of metal anodes, especially for Li metal. These include the design of artificial protective layers (13–17), alternative electrolytes (18–24), and sacrificial additives (25–30) to stabilize the metal–electrolyte interface, the development of mechanically robust coatings (31–34) to block Li dendrite growth, and the use of structured scaffolds to host dendrite-free Li deposition by reducing local current densities (35–43). However, the performance of metal anodes remains poor under high-current or low-temperature conditions. This is because the inhomogeneous Li nucleation and unstable SEI problems have not been well addressed, and these problems at the interface are even exacerbated under critical operating conditions, especially high-current densities and low temperatures (5, 6, 44).Toward this end, we report a simple molecular approach for regulating the electrochemical interface of metal anodes, which enables even Li deposition and stable SEI formation in a conventional electrolyte. This was realized by bonding a labile organic molecule, benzenesulfonyl fluoride (BSF), to a reduced graphene oxide (rGO) aerogel surface as the Li anode host (Fig. 1A). During Li deposition, BSF molecules electrochemically decompose at the interface and generate benzenesulfonate anions bonded to the rGO aerogel (Fig. 1B). The conjugated anions have a strong binding affinity for Li, serving as lithiophilic sites on the rGO surface to synergistically induce homogeneous Li nucleation of Li on the rGO surface. At the same time, BSF molecules contribute LiF to the SEI layer, which facilitates Li surface passivation (Fig. 1C). As a result, high-efficiency (99.2%) Li deposition was achieved at a Li deposition amount of 6.0 mAh cm⁻² and a current density of 6.0 mA cm⁻²; the barrier to Li nucleation was markedly reduced, as evidenced by the low nucleation overpotentials at high-current density (6.0 mA cm⁻²) or at a low temperature (−10 °C). A 400-cycle life with a capacity retention of 83.6% was achieved for a Li|LiCoO₂ (LCO) cell in a conventional carbonate electrolyte. Moreover, with the organic molecule-tuned interface, the Li|LCO cell can be stably cycled at a low operating temperature (−10 °C). This approach was applied to Na and Zn metal anodes as well.Open in a separate windowFig. 1.Illustration of a stable interface for Li deposition using a labile organic molecule, benzenesulfonyl fluoride (BSF). (A) Covalently bonded BSF on the rGO aerogel surface. (B) In situ generation of a lithiophilic conjugated anion (benzenesulfonate) and LiF on the surface during Li deposition. (C) Li nucleation preferentially occurs at the conjugated anion sites owing to the strong Li binding affinity, which leads to uniform Li deposition. In addition, the LiF that is formed is in the SEI layer and passivates the Li surface.     

A switch to feeding on cycads generates parallel accelerated evolution of toxin tolerance in two clades of Eumaeus caterpillars (Lepidoptera: Lycaenidae)

We assembled a complete reference genome of Eumaeus atala, an aposematic cycad-eating hairstreak butterfly that suffered near extinction in the United States in the last century. Based on an analysis of genomic sequences of Eumaeus and 19 representative genera, the closest relatives of Eumaeus are Theorema and Mithras. We report natural history information for Eumaeus, Theorema, and Mithras. Using genomic sequences for each species of Eumaeus, Theorema, and Mithras (and three outgroups), we trace the evolution of cycad feeding, coloration, gregarious behavior, and other traits. The switch to feeding on cycads and to conspicuous coloration was accompanied by little genomic change. Soon after its origin, Eumaeus split into two fast evolving lineages, instead of forming a clump of close relatives in the phylogenetic tree. Significant overlap of the fast evolving proteins in both clades indicates parallel evolution. The functions of the fast evolving proteins suggest that the caterpillars developed tolerance to cycad toxins with a range of mechanisms including autophagy of damaged cells, removal of cell debris by macrophages, and more active cell proliferation.

The genus Eumaeus Hübner (Lycaenidae, Theclinae) arguably contains the most aposematically colored caterpillars and butterflies among the ∼4,000 Lycaenidae in the world (1–6). The brilliant red and gold gregarious caterpillars (Fig. 1) sequester cycasin from the leaves of their cycad food plants (Zamiaceae), which deters predators (3–9). Other secondary metabolites in cycads (e.g., 10‒11) may also deter predators. Eumaeus adults have a bright orange-red abdomen and an orange-red hindwing spot (except for one species) (Fig. 2). Blue and green iridescent markings are especially conspicuous on a black ground color. Eumaeus adults are among the largest lycaenids and have more rounded wings and a slower, more gliding flight than most Theclinae (1). Cycads are among the most primitive extant seed-plants (9), and the “plethora of aposematic attributes suggests a very ancient association between Eumaeus and the cycad host plants” (3).Open in a separate windowFig. 1.Caterpillars and pupae of Theorema eumenia (Top) and Eumaeus godartii (Bottom) in Costa Rica. Clockwise from Upper Left, second or third instar (length, ∼13 mm), fourth (final) instar (∼20 mm), pupa (∼18 mm), pupa (∼24 mm), fourth (final) instar (∼27 mm), second or third instar (∼20 mm). (Images from authors W.H. and D.H.J.).Open in a separate windowFig. 2.Adult wing uppersides and undersides. Eumaeus childrenae (two Upper Left images), E. atala (two Upper Right images), Theorema eumenia (two Lower Left images), and Mithras nautes (two Lower Right images). Scale bar, 1 cm.Eumaeus has been classified as a separate family (12–14), a genus in the Riodinidae (15‒16), or a monotypic subfamily or tribe of the Lycaenidae (17–20). Alternatively, others called it a typical member of the Neotropical Lycaenidae (21‒22). The evolutionary question behind this discordant taxonomic history is whether Eumaeus is a phylogenetically isolated lineage long associated with cycads (3) or an embedded clade in which a recent food plant shift to cycads resulted in the rapid evolution of aposematism. Recent molecular evidence for a limited number of taxa suggested the latter (23). To answer this question definitively, we analyzed genomic sequences of Eumaeus and its relatives.To trace the evolution of cycad feeding, we report the caterpillar food plants of the genera most closely related to Eumaeus and illustrate their immature stages (Fig. 1 and SI Appendix). This natural history information combined with analyses of genome sequences is the foundation for investigating the subsequent evolutionary impact on the Eumaeus genome of the switch to eating cycads.     

High-throughput developability assays enable library-scale identification of producible protein scaffold variants

Proteins require high developability—quantified by expression, solubility, and stability—for robust utility as therapeutics, diagnostics, and in other biotechnological applications. Measuring traditional developability metrics is low throughput in nature, often slowing the developmental pipeline. We evaluated the ability of 10 variations of three high-throughput developability assays to predict the bacterial recombinant expression of paratope variants of the protein scaffold Gp2. Enabled by a phenotype/genotype linkage, assay performance for 10⁵ variants was calculated via deep sequencing of populations sorted by proxied developability. We identified the most informative assay combination via cross-validation accuracy and correlation feature selection and demonstrated the ability of machine learning models to exploit nonlinear mutual information to increase the assays’ predictive utility. We trained a random forest model that predicts expression from assay performance that is 35% closer to the experimental variance and trains 80% more efficiently than a model predicting from sequence information alone. Utilizing the predicted expression, we performed a site-wise analysis and predicted mutations consistent with enhanced developability. The validated assays offer the ability to identify developable proteins at unprecedented scales, reducing the bottleneck of protein commercialization.

A common constraint across diagnostic, therapeutic, and industrial proteins is the ability to manufacture, store, and use intact and active molecules. These protein properties, collectively termed developability, are often associated to quantitative metrics such as recombinant yield, stability (chemical, thermal, and proteolytic), and solubility (1–5). Despite this universal importance, developability studies are performed late in the commercialization pipeline (2, 4) and limited by traditional experimental capacity (6). This is problematic because 1) proteins with poor developability limit practical assay capacity for measuring primary function, 2) optimal developability is often not observed with proteins originally found in alternative formats [such as display or two-hybrid technologies (7)], and 3) engineering efforts are limited by the large gap between observation size (∼10²) and theoretical mutational diversity (∼10²⁰). Thus, efficient methods to measure developability would alleviate a significant bottleneck in the lead selection process and accelerate protein discovery and engineering.Prior advances to determine developability have focused on calculating hypothesized proxy metrics from existing sequence and structural data or developing material- and time-efficient experiments. Computational sequence-developability models based on experimental antibody data have predicted posttranslational modifications (8, 9), solubility (10, 11), viscosity (12), and overall developability (13). Structural approaches have informed stability (14) and solubility (10, 15). However, many in silico models require an experimentally solved structure or suffer from computational structure prediction inaccuracies (16). Additionally, limited developability information allows for limited predictive model accuracy (17). In vitro methods have identified several experimental protocols to mimic practical developability requirements [e.g., affinity-capture self-interaction nanoparticle spectroscopy (18) and chemical precipitation (19) as metrics for solubility]. However, traditional developability quantification requires significant amounts of purified protein. Noted in both fronts are numerous in silico and/or in vitro metrics to fully quantify developability (1, 5).We sought a protein variant library that would benefit from isolation of proteins with increased developability and demonstrate the broad applicability of the process. Antibodies and other binding scaffolds, comprising a conserved framework and diversified paratope residues, are effective molecular targeting agents (20–24). While significant progress has been achieved with regards to identifying paratopes for optimal binding strength and specificity (25, 26), isolating highly developable variants remains plagued. One particular protein scaffold, Gp2, has been evolved into specific binding variants toward multiple targets (27–29). Continued study improved charge distribution (30), hydrophobicity (31), and stability (28). While these studies have suggested improvements for future framework and paratope residues (including a disulfide-stabilized loop), a poor developability distribution is still observed (32) (Fig. 1 A and B). Assuming the randomized paratope library will lack similar primary functionality, the Gp2 library will simulate the universal applicability of the proposed high-throughput (HT) developability assays.Open in a separate windowFig. 1.HT assays were evaluated for the ability to identify protein scaffold variants with increased developability. (A and B) Gp2 variant expression, commonly measured via low-throughput techniques such as the dot blot shown, highlights the rarity of ideal developability. (C and D) The HT on-yeast protease assay measures the stability of the POI by proteolytic extent. (E and F) The HT split-GFP assay measures POI expression via recombination of a genetically fused GFP fragment. (G and H) The HT split β-lactamase assay measures the POI stability by observing the change in cell-growth rates when grown at various antibiotic concentrations. (I and J) Assay scores, assigned to each unique sequence via deep sequencing, were evaluated by predicting expression (Fig. 3). (K and L) HT assay capacity enables large-scale developability evaluation and can be used to identify beneficial mutations (Fig. 4).We sought HT assays that allow protein developability differentiation via cellular properties to improve throughput. Variations of three primary assays were examined: 1) on-yeast stability (Fig. 1 C and D)—previously validated to improve the stability of de novo proteins (33), antimicrobial lysins (34), and immune proteins (35)—measures proteolytic cleavage of the protein of interest (POI) on the yeast cell surface via fluorescence-activated cell sorting (FACS). We extend the assay by performing the proteolysis at various denaturing combinations to determine if different stability attributes (thermal, chemical, and protease specificity) can be resolved; 2) Split green fluorescent protein (GFP, Fig. 1 E and F)—previously used to determine soluble protein concentrations (36)—measures the assembled GFP fluorescence emerging from a 16–amino acid fragment (GFP₁₁) fused to the POI after recombining with the separably expressed GFP_1-10. We extend the assay by utilizing FACS to separate cells with differential POI expression to increase throughput over the plate-based assay; and 3) Split β-lactamase (Fig. 1 G and H)—previously used to improve thermodynamic stability (37) and solubility (38)—measures cell growth inhibition via ampicillin to determine functional lactamase activity achieved from reconstitution of two enzyme fragments flanking the POI. We expand assay capacity by deep sequencing populations grown at various antibiotic concentrations to relate change in cell frequency to functional enzyme concentration.In this paper, we determined the HT assays’ abilities to predict Gp2 variant developability. We deep sequenced the stratified populations and calculated assay scores (correlating to hypothesized developability) for ∼10⁵ Gp2 variants (Fig. 1I). We then converted the assay scores into a traditional developability metric by building a model that predicts recombinant yield (Fig. 1J). The assays’ capacity enabled yield evaluations for >100-fold traditional assay capacity (Fig. 1K, compared to Fig. 1B) and provide an introductory analysis of factors driving protein developability by observing beneficial mutations via predicted developable proteins (Fig. 1L).     

Free-energy changes of bacteriorhodopsin point mutants measured by single-molecule force spectroscopy

Single amino acid mutations provide quantitative insight into the energetics that underlie the dynamics and folding of membrane proteins. Chemical denaturation is the most widely used assay and yields the change in unfolding free energy (ΔΔG). It has been applied to >80 different residues of bacteriorhodopsin (bR), a model membrane protein. However, such experiments have several key limitations: 1) a nonnative lipid environment, 2) a denatured state with significant secondary structure, 3) error introduced by extrapolation to zero denaturant, and 4) the requirement of globally reversible refolding. We overcame these limitations by reversibly unfolding local regions of an individual protein with mechanical force using an atomic-force-microscope assay optimized for 2 μs time resolution and 1 pN force stability. In this assay, bR was unfolded from its native bilayer into a well-defined, stretched state. To measure ΔΔG, we introduced two alanine point mutations into an 8-amino-acid region at the C-terminal end of bR’s G helix. For each, we reversibly unfolded and refolded this region hundreds of times while the rest of the protein remained folded. Our single-molecule–derived ΔΔG for mutant L223A (−2.3 ± 0.6 kcal/mol) quantitatively agreed with past chemical denaturation results while our ΔΔG for mutant V217A was 2.2-fold larger (−2.4 ± 0.6 kcal/mol). We attribute the latter result, in part, to contact between Val²¹⁷ and a natively bound squalene lipid, highlighting the contribution of membrane protein–lipid contacts not present in chemical denaturation assays. More generally, we established a platform for determining ΔΔG for a fully folded membrane protein embedded in its native bilayer.

Membrane proteins play a critical role in metabolism, transport, and signaling. Membrane proteins are also of intense biomedical interest, as they are the target for ∼50% of approved drugs (1). X-ray crystallography (2) and, more recently, cryo-electron microscopy (3–5) are yielding an accelerating number of high-resolution structures. Yet, predicting the folding and dynamics of membrane proteins remains an unmet need, in part, because it remains difficult to characterize the energetics that drive and stabilize a membrane protein’s folded structure (6–9). Membrane protein energetics are characterized in terms of the unfolding free energy (ΔG) and the change in that free energy upon introduction of a single amino acid point mutation (ΔΔG). Ideally, ΔG is interpretable as the sum of stabilizing interactions in the protein (e.g., van der Waals, hydrogen bonding), and ΔΔG isolates the energetic contribution of a single amino acid side chain to that stability. However, such an interpretation is only possible if the measurement technique reports the underlying molecular energetics with fidelity. For example, a measurement that is not made in the native bilayer may fail to accurately report stabilizing amino acid–lipid interactions. Here, we establish a single-molecule platform for measuring ΔΔG in bacteriorhodopsin (bR), a model membrane protein. The method is based upon mechanical unfolding of individual molecules in their native lipid bilayer (Fig. 1A), avoiding the nonnative detergent environment and the perturbative chemical denaturant used in traditional ensemble biochemical assays (Fig. 1B).Open in a separate windowFig. 1.Membrane protein energetics assays. (A) Illustration of the single-molecule mechanical unfolding assay where bR in its native bilayer is deposited onto a mica surface. An AFM cantilever then applies tension to the C-terminal tail, causing an eight-aa region (cyan) to reversibly unfold to a taut unfolded state (Inset). Unfolding is measured by a change in the polypeptide extension x. (B) Illustration of the traditional chemical denaturation assay showing reconstituted bR in a mixed micelle. Application of SDS then leads to a denatured state that retains ∼60% of its secondary α-helical structure (cylinders).Chemical denaturation of at least 85 bR point mutants (10–18) helps form the foundation for the current understanding of membrane protein energetics (8, 19) despite several key limitations. These measurements typically begin with reconstituted bR folded in a nonnative mixed micelle of phospholipid and detergent. Introduction of sodium dodecyl sulfate (SDS) then reversibly denatures the protein in a single step, disrupting the tertiary structure but leaving a significant fraction of secondary structure intact [∼60% as measured by circular dichroism (20)] (Fig. 1B). Analysis of the denaturant concentration dependence yields ΔG and, when repeated for point mutants, ΔΔG (21). However, this widely applied technique has four underlying limitations that can bias measured values. First, the bR is solvated in a nonnative mixed micelle, causing the free energy of the folded state to lack the energetic contributions from native protein–lipid interactions (22). Second, the measurements involve extrapolation from high denaturant concentration to zero, which assumes a linearity that steric-trapping experiments have called into question (23). Third, the free energy of the denatured state includes contributions from the ∼60% residual secondary structure (20, 24) and from nonspecific interactions with the denaturant (25). Fourth, the vast majority of membrane proteins—including all G-protein–coupled receptors—are not amenable to these measurements because there is no currently known condition under which their chemical denaturation is reversible (26).Force-induced mechanical unfolding represents an alternative means for measuring ΔΔG without the underlying limitations of chemical denaturation assays. In these atomic force microscopy (AFM) studies, the protein is initially folded in a lipid bilayer—native purple membrane in the case of bR—providing sensitivity to protein–lipid and quaternary interactions (27, 28). The cantilever of an AFM then pulls on one end of an individual membrane protein, causing it to unfold under force (F) to a stretched state (i.e., a highly extended polypeptide chain). This taut, unfolded state emerges from the bilayer into the surrounding buffer and thus contains no energetic contributions from residual secondary structure (29) or detergent interactions (Fig. 1A). As a result, both the folded and unfolded states are thermodynamically well defined, in contrast to the “unfolded” state induced by SDS denaturation (Fig. 1B). When probed with sufficient sensitivity, the unfolding process is shown to occur via a multitude of discrete states (30), each corresponding to a metastable unfolding intermediate. Reversible transitions between these states—some separated by the folding of as few as two to three amino acids (aa) (30, 31)—allow for local equilibrium energetic measurements of proteins that do not reversibly refold on larger scales.Many studies have reported AFM-based mechanical unfolding of bR (27, 30–40), including some reporting unfolding free energies (41–43); however, most prior studies lacked the ability to observe reversible transitions in the segments of bR that unfold first. In particular, assays that rely on nonspecific tip-sample attachment suffer from surface adhesion that prevents precise characterization of the first two of bR’s seven transmembrane helices to be extracted. These assays therefore cannot quantify the energetics of bR when the most native set of interactions are present. Nonspecific attachment is also not mechanically robust, preventing repeated unfolding and refolding over an extended period. Additionally, most prior AFM measurements of bR were far from equilibrium, did not account for the anomalous work of stretching the unfolded polypeptide, and had insufficient spatiotemporal resolution to observe occupancies of short-lived, closely spaced unfolding intermediates.Recent technological and methodological improvements have overcome the limitations of prior AFM-based force spectroscopy assays to allow the energetics of bR to be measured on the scale of a few aa. In particular, Yu et al. (31) showed that copper-free click chemistry can be used to form, in situ, a site-specific covalent bond between bR and an AFM tip. This chemistry allows for longer data collection and for smaller surface-contact forces, the latter reducing adhesion and enabling interpretation of the first eight aa to unfold. Additionally, Edwards et al. (44) developed focused ion beam (FIB)-modified cantilevers that provided a 100-fold improvement in time resolution and 10-fold improvement in force precision over prior studies of bR (27, 28). Recently, we combined these techniques to characterize the unfolding and refolding of the first eight aa at the C terminus of wild-type (wt) bR’s helix G, while the rest of the protein remained folded in native purple membrane. From these data, we deduced values of ΔG for these residues (∼2 kcal/mol per aa) that were consistent between three distinct analyses using both equilibrium and near-equilibrium protocols (45). Notably, this ΔG is ∼20-fold larger (on a per aa basis) than a prior chemical denaturation measurement (16). We attributed this difference in ΔG to the differing unfolded states between the two assays, with the SDS-denaturation measurement not accounting for the energy needed to disrupt the full α-helical secondary structure of bR and to solvate the unfolded protein into water (Fig. 1). Given this large difference in ΔG, it remained an open question whether consistent values of ΔΔG could be obtained between the two assays.Here, we applied these recently developed AFM techniques to measure ΔΔG of two alanine point mutants (L223A and V217A) that were previously characterized using SDS denaturation (16) and that probed different side-chain environments. Leu²²³ is oriented into the protein core of bR, whereas Val²¹⁷ is orientated outwards toward the lipid environment of purple membrane and contributes to the binding pocket of a crystallographically resolved squalene lipid (46). Ideally, ΔΔG values would agree between AFM unfolding and SDS denaturation, as both are meant to reflect the underlying contribution of the mutated side chain to the overall protein stability. Interestingly, our ΔΔG value for L223A agreed quantitatively with prior SDS-denaturation experiments despite the vastly different ΔG for the two assays. However, the AFM-based determination of ΔΔG for V217A was 2.2-fold larger than the SDS-based value, highlighting the strength of lipid–protein interactions and the importance of characterizing membrane protein energetics in the native bilayer. These measurements, therefore, constitute a practical demonstration of a widely applicable, single-molecule technique for measuring membrane protein thermodynamics that is not subject to the limitations of chemical denaturation experiments.     

A rate threshold mechanism regulates MAPK stress signaling and survival

Cells are exposed to changes in extracellular stimulus concentration that vary as a function of rate. However, how cells integrate information conveyed from stimulation rate along with concentration remains poorly understood. Here, we examined how varying the rate of stress application alters budding yeast mitogen-activated protein kinase (MAPK) signaling and cell behavior at the single-cell level. We show that signaling depends on a rate threshold that operates in conjunction with stimulus concentration to determine the timing of MAPK signaling during rate-varying stimulus treatments. We also discovered that the stimulation rate threshold and stimulation rate-dependent cell survival are sensitive to changes in the expression levels of the Ptp2 phosphatase, but not of another phosphatase that similarly regulates osmostress signaling during switch-like treatments. Our results demonstrate that stimulation rate is a regulated determinant of cell behavior and provide a paradigm to guide the dissection of major stimulation rate dependent mechanisms in other systems.

All cells employ signal transduction pathways to respond to physiologically relevant changes in extracellular stressors, nutrient levels, hormones, morphogens, and other stimuli that vary as functions of both concentration and rate in healthy and diseased states (1–7). Switch-like “instantaneous” changes in the concentrations of stimuli in the extracellular environment have been widely used to show that the strength of signaling and overall cellular response are dependent on the stimulus concentration, which in many cases needs to exceed a certain threshold (8, 9). Previous studies have shown that the rate of stimulation can also influence signaling output in a variety of pathways (10–17) and that stimulation profiles of varying rates can be used to probe underlying signaling pathway circuitry (4, 18, 19). However, it is still not clear how cells integrate information conveyed by changes in both the stimulation rate and concentration in determining signaling output. It is also not clear if cells require stimulation gradients to exceed a certain rate in order to commence signaling.Recent investigations have demonstrated that stimulation rate can be a determining factor in signal transduction. In contrast to switch-like perturbations, which trigger a broad set of stress-response pathways, slow stimulation rates activate a specific response to the stress applied in Bacillus subtilis cells (10). Meanwhile, shallow morphogen gradient stimulation fails to activate developmental pathways in mouse myoblast cells in culture, even when concentrations sufficient for activation during pulsed treatment are delivered (12). These observations raise the possibility that stimulation profiles must exceed a set minimum rate or rate threshold to achieve signaling activation. Although such rate thresholds would help cells decide if and how to respond to dynamic changes in stimulus concentration, the possibility of signaling regulation by a rate threshold has never been directly investigated in any system. Further, no study has experimentally examined how stimulation rate requirements impact cell phenotype or how cells molecularly regulate the stimulation rate required for signaling activation. As such, the biological significance of any existing rate threshold regulation of signaling remains unknown.The budding yeast Saccharomyces cerevisiae high osmolarity glycerol (HOG) pathway provides an ideal model system for addressing these issues (Fig. 1A). The evolutionarily conserved mitogen-activated protein kinase (MAPK) Hog1 serves as the central signaling mediator of this pathway (20–22). It is well established that instantaneous increases in osmotic stress concentration induce Hog1 phosphorylation, activation, and translocation to the nucleus (18, 21, 23–30). Activated Hog1 governs the majority of the cellular osmoadaptation response that enables cells to survive (23, 31, 32). Multiple apparently redundant MAPK phosphatases dephosphorylate and inactivate Hog1, which, along with the termination of upstream signaling after adaptation, results in its return to the cytosol (Fig. 1A) (23, 25, 26, 33–39). Because of this behavior, time-lapse analysis of Hog1 nuclear enrichment in single cells has proven an excellent and sensitive way to monitor signaling responses to dynamic stimulation patterns in real time (18, 27–30, 40, 41). Further, such assays have been readily combined with traditional growth and molecular genetic approaches to link observed signaling responses with cell behavior and signaling pathway architecture (27–29).Open in a separate windowFig. 1.Hog1 signaling and cell survival are sensitive to the rate of preconditioning osmotic stress application. (A) Schematic of the budding yeast HOG response. (B) Preconditioning protection assay workflow indicating the first stress treatments to a final concentration of 0.4 M NaCl (Left), high-stress exposure (Middle), and colony formation readout (Right). (C) High-stress survival as a function of each first treatment relative to the untreated first stress condition. Bars and errors are means and SD from three biological replicates. *Statistically significant by Kolmogorov–Smirnov test (P < 0.05). NS = not significant. (D) Treatment concentration over time. (E) Treatment rate over time for quadratic and pulse treatment. The rate for the pulse is briefly infinite (blue vertical line) before it drops to 0. (F) Hog1 nuclear localization during the treatments depicted in D and E. (Inset) Localization pattern in the quadratic-treated sample. Lines represent means and shaded error represents the SD from three to four biological replicates.Here, we use systematically designed osmotic stress treatments imposed at varying rates of increase to show that a rate threshold condition regulates yeast high-stress survival and Hog1 MAPK signaling. We demonstrate that only stimulus profiles that satisfy both this rate threshold condition and a concentration threshold condition result in robust signaling. We go on to show that the protein tyrosine phosphatase Ptp2, but not the related Ptp3 phosphatase, serves as a major rate threshold regulator. By expressing PTP2 under the control of a series of different enhancer–promoter DNA constructs, we demonstrate that changes in the level of Ptp2 expression can alter the stimulation rate required for signaling induction and survival. These findings establish rate thresholds as a critical and regulated component of signaling biology akin to concentration thresholds.     

Developmental influence on evolutionary rates and the origin of placental mammal tooth complexity

Development has often been viewed as a constraining force on morphological adaptation, but its precise influence, especially on evolutionary rates, is poorly understood. Placental mammals provide a classic example of adaptive radiation, but the debate around rate and drivers of early placental evolution remains contentious. A hallmark of early dental evolution in many placental lineages was a transition from a triangular upper molar to a more complex upper molar with a rectangular cusp pattern better specialized for crushing. To examine how development influenced this transition, we simulated dental evolution on “landscapes” built from different parameters of a computational model of tooth morphogenesis. Among the parameters examined, we find that increases in the number of enamel knots, the developmental precursors of the tooth cusps, were primarily influenced by increased self-regulation of the molecular activator (activation), whereas the pattern of knots resulted from changes in both activation and biases in tooth bud growth. In simulations, increased activation facilitated accelerated evolutionary increases in knot number, creating a lateral knot arrangement that evolved at least ten times on placental upper molars. Relatively small increases in activation, superimposed on an ancestral tritubercular molar growth pattern, could recreate key changes leading to a rectangular upper molar cusp pattern. Tinkering with tooth bud geometry varied the way cusps initiated along the posterolingual molar margin, suggesting that small spatial variations in ancestral molar growth may have influenced how placental lineages acquired a hypocone cusp. We suggest that development could have enabled relatively fast higher-level divergence of the placental molar dentition.

Whether developmental processes bias or constrain morphological adaptation is a long-standing question in evolutionary biology (1–4). Many of the distinctive features of a species derive from pattern formation processes that establish the position and number of anatomical structures (5). If developmental processes like pattern formation are biased toward generating only particular kinds of variation, adaptive radiations may often be directed along developmental–genetic “lines of least resistance” (2, 4, 6, 7). Generally, the evolutionary consequences of this developmental bias have been considered largely in terms of how it might influence the pattern of character evolution (e.g., refs. 1, 2, 8–10). But development could also influence evolutionary rates by controlling how much variation is accessible to natural selection in a given generation (11).For mammals, the dentition is often the only morphological system linking living and extinct species (12). Correspondingly, tooth morphology plays a crucial role in elucidating evolutionary relationships, time calibrating phylogenetic trees, and reconstructing adaptive responses to past environmental change (e.g., refs. 13–15). One of the most pervasive features of dental evolution among mammals is an increase in the complexity of the tooth occlusal surface, primarily through the addition of new tooth cusps (16, 17). These increases in tooth complexity are functionally and ecologically significant because they enable more efficient mechanical breakdown of lower-quality foods like plant leaves (18).Placental mammals are the most diverse extant mammalian group, comprising more than 6,000 living species spread across 19 extant orders, and this taxonomic diversity is reflected in their range of tooth shapes and dietary ecologies (12). Many extant placental orders, especially those with omnivorous or herbivorous ecologies (e.g., artiodactyls, proboscideans, rodents, and primates), convergently evolved a rectangular upper molar cusp pattern from a placental ancestor with a more triangular cusp pattern (19–21). This resulted from separate additions in each lineage of a novel posterolingual cusp, the "hypocone'''' [sensu (19)], to the tritubercular upper molar (Fig. 1), either through modification of a posterolingual cingulum (“true” hypocone) or another posterolingual structure, like a metaconule (pseudohypocone) (19). The fossil record suggests that many of the basic steps in the origin of this rectangular cusp pattern occurred during an enigmatic early diversification window associated with the divergence and early radiation of several placental orders (20, 21; Fig. 1). However, there remains debate about the rate and pattern of early placental divergence (22–24). On the one hand, most molecular phylogenies suggest that higher-level placental divergence occurred largely during the Late Cretaceous (25, 26), whereas other molecular phylogenies and paleontological analyses suggest more rapid divergence near the Cretaceous–Paleogene (K–Pg) boundary (21, 24, 27–29). Most studies agree that ecological opportunity created in the aftermath of the K–Pg extinction probably played an important role in ecomorphological diversification within the placental orders (30, 31). But exactly how early placentals acquired the innovations needed to capitalize on ecological opportunity remains unclear. Dental innovations, especially those which facilitated increases in tooth complexity, may have been important because they would have promoted expansion into plant-based dietary ecologies left largely vacant after the K–Pg extinction event (32).Open in a separate windowFig. 1.Placental mammal lineages separately evolved complex upper molar teeth with a rectangular cusp pattern composed of two lateral pairs of cusps from a common ancestor with a simpler, triangular cusp pattern. Many early relatives of the extant placental orders, such as Eritherium, possessed a hypocone cusp and a more rectangular primary cusp pattern. Examples of complex upper molars are the following: Proboscidea, the gomphothere Anancus; Rodentia, the wood mouse Apodemus; and Artiodactyla, the suid Nyanzachoerus.Mammalian tooth cusps form primarily during the “cap” and “bell” stage of dental development, when signaling centers called enamel knots establish the future sites of cusp formation within the inner dental epithelium (33, 34). The enamel knots secrete molecules that promote proliferation and changes in cell–cell adhesion, which facilitates invagination of the dental epithelium into an underlying layer of mesenchymal cells (34, 35). Although a range of genes are involved in tooth cusp patterning (36–38), the basic dynamics can be effectively modeled using reaction–diffusion models with just three diffusible morphogens: an activator, an inhibitor, and a growth factor (39–41). Candidate activator genes in mammalian tooth development include Bmp4, Activin A, Fgf20, and Wnt genes, whereas potential inhibitors include Shh and Sostdc, and Fgf4 and Bmp2 have been hypothesized to act as growth factors (38, 40–43). In computer models of tooth development, activator molecules up-regulated in the underlying mesenchyme stimulate differentiation of overlying epithelium into nondividing enamel knot cells. These in turn secrete molecules that inhibit further differentiation of epithelium into knot cells, while also promoting cell proliferation that creates the topographic relief of the cusp (40). Although many molecular, cellular, and physical processes have the potential to influence cusp formation, and thereby tooth complexity (35, 37), parameters that control the strength and conductance of the activator and inhibitor signals, the core components of the reaction–diffusion cusp patterning mechanism (39, 40) are likely to be especially important.Here, we integrate a previous computer model of tooth morphogenesis called ToothMaker (41), with simulations of trait evolution and data from the fossil record (Fig. 2), to examine the developmental origins of tooth complexity in placental mammals. Specifically, we ask the following: 1) What developmental processes can influence how many cusps form? 2) How might these developmental processes influence the evolution of tooth cusp number, especially rates? And 3) what developmental changes may have been important in the origins of the fourth upper molar cusp, the hypocone, in placental mammal evolution?Open in a separate windowFig. 2.Workflow for simulations of tooth complexity evolution. (A) Tooth shape is varied for five signaling and growth parameters in ToothMaker. (B) From an ancestral state, each parameter is varied in 2.5% increments up to a maximum of ± 50% of the ancestral state. (C) Tooth complexity and enamel knot (EK) pattern were quantified for each parameter combination. Tooth complexity was measured using cusp number/EK number and OPC. ToothMaker and placental upper second molars were classified into categories based on EK/cusp pattern. (D) The parameter space was populated with pattern and tooth complexity datums to build a developmental landscape. (E) Tooth complexity evolution was simulated on each developmental landscape. (F) Resulting diversity and pattern of tooth complexity was compared with placental mammal molar diversity.