Embodied intelligence via learning and evolution.

The intertwined processes of learning and evolution in complex environmental niches have resulted in a remarkable diversity of morphological forms. Moreover, many aspects of animal intelligence are deeply embodied in these evolved morphologies. However, the principles governing relations between environmental complexity, evolved morphology, and the learnability of intelligent control, remain elusive, because performing large-scale in silico experiments on evolution and learning is challenging. Here, we introduce Deep Evolutionary Reinforcement Learning (DERL): a computational framework which can evolve diverse agent morphologies to learn challenging locomotion and manipulation tasks in complex environments. Leveraging DERL we demonstrate several relations between environmental complexity, morphological intelligence and the learnability of control. First, environmental complexity fosters the evolution of morphological intelligence as quantified by the ability of a morphology to facilitate the learning of novel tasks. Second, we demonstrate a morphological Baldwin effect i.e., in our simulations evolution rapidly selects morphologies that learn faster, thereby enabling behaviors learned late in the lifetime of early ancestors to be expressed early in the descendants lifetime. Third, we suggest a mechanistic basis for the above relationships through the evolution of morphologies that are more physically stable and energy efficient, and can therefore facilitate learning and control.

Resonance assignment of coiled-coil 3 (CC3) domain of human STIM1.

The protein stromal interaction molecule 1 (STIM1) plays a pivotal role in mediating store-operated calcium entry (SOCE) into cells, which is essential for adaptive immunity. It acts as a calcium sensor in the endoplasmic reticulum (ER) and extends into the cytosol, where it changes from an inactive (tight) to an active (extended) oligomeric form upon calcium store depletion. NMR studies of this protein are challenging due to its membrane-spanning and aggregation properties. Therefore follow the divide-and-conquer approach, focusing on individual domains first is in order. The cytosolic part is predicted to have a large content of coiled-coil (CC) structure. We report the 1H, 13C, 15N chemical shift assignments of the CC3 domain. This domain is crucial for the stabilisation of the tight quiescent form of STIM1 as well as for activating the ORAI calcium channel by direct contact, in the extended active form.

Interhelical interactions within the STIM1 CC1 domain modulate CRAC channel activation.

The calcium release activated calcium channel is activated by the endoplasmic reticulum-resident calcium sensor protein STIM1. On activation, STIM1 C terminus changes from an inactive, tight to an active, extended conformation. A coiled-coil clamp involving the CC1 and CC3 domains is essential in controlling STIM1 activation, with CC1 as the key entity. The nuclear magnetic resonance-derived solution structure of the CC1 domain represents a three-helix bundle stabilized by interhelical contacts, which are absent in the Stormorken disease-related STIM1 R304W mutant. Two interhelical sites between the CC1α1 and CC1α2 helices are key in controlling STIM1 activation, affecting the balance between tight and extended conformations. Nuclear magnetic resonance-directed mutations within these interhelical interactions restore the physiological, store-dependent activation behavior of the gain-of-function STIM1 R304W mutant. This study reveals the functional impact of interhelical interactions within the CC1 domain for modifying the CC1-CC3 clamp strength to control the activation of STIM1.