Despite its importance in many industrial, geological and biological processes, the mechanism of crystallization from supersaturated solutions remains a matter of debate. Recent discoveries show that in many solution systems nanometre-sized structural units are already present before nucleation. Still little is known about the structure and role of these so-called pre-nucleation clusters. Here we present a combination of in situ investigations, which show that for the crystallization of calcium phosphate these nanometre-sized units are in fact calcium triphosphate complexes. Under conditions in which apatite forms from an amorphous calcium phosphate precursor, these complexes aggregate and take up an extra calcium ion to form amorphous calcium phosphate, which is a fractal of Ca 2 (HPO 4 ) 3 2 À clusters. The calcium triphosphate complex also forms the basis of the crystal structure of octacalcium phosphate and apatite. Finally, we demonstrate how the existence of these complexes lowers the energy barrier to nucleation and unites classical and non-classical nucleation theories.

To date, the knowledge of the multiple roles of additives in calcium carbonate crystallization results from empirical control of crystal growth. We present a crystallization assay generating controlled supersaturation, which allows the categorization of different modes of additive action. This facilitates a novel understanding of the action of crystallization additives.

A titration set-up coupling ion selective electrodes with pH adjustment was used to analyze the effects of additives present during precipitation of calcium carbonate. Besides industrially well-established antiscalants (sodium triphosphate, citrate, polyacrylate and poly(aspartic acid)), also functional polymers being active in morphosynthesis (polystyrene sulfonate and poly(styrene-alt-maleic acid)) were analyzed. Interestingly each additive acts in its specific way, suggesting the notation "fingerprinting" for a complex interplay of up to five "solution modes" of influencing CaCO(3) precipitation and crystallisation. The results provide new insights into the modes of additive controlled crystallisation, and in the long run, the insights may facilitate the design of precipitation systems that yield complex and tailor-made crystals.

The mollusk shell nacre layer integrates mineral phases with macromolecular components such as intracrystalline proteins. However, the roles performed by intracrystalline proteins in calcium carbonate nucleation and subsequent postnucleation events (e.g., organization of mineral deposits) in the nacre layer are not known. We find that AP7, a nacre intracrystalline C-RING protein, self-assembles to form amorphous protein oligomers and films on mica that further assemble into larger aggregates or phases in the presence of Ca2+. Using solution nuclear magnetic resonance spectroscopy, we determine that the protein assemblies are stabilized by interdomain interactions involving the aggregation-prone T31-N66 C-terminal C-RING domain but are destabilized by the labile nature of the intrinsically disordered D1-T19 AA N-terminal sequence. Thus, the dynamic, amorphous nature of the AP7 assemblies can be traced to the molecular behavior of the N-terminal sequence. Using potentiometric methods, we observe that AP7 protein phases prolong the time interval for prenucleation cluster formation but neither stabilize nor destabilize ACC clusters. Time-resolved flow cell scanning transmission electron microscopy mineralization studies confirm that AP7 protein phases delay the onset of nucleation and assemble and organize mineral nanoparticles into ring-shaped branching clusters in solution. These phenomena are not observed in protein-deficient assays. We conclude that C-RING AP7 protein phases modulate the time period for early events in nucleation and form strategic associations with forming mineral nanoparticles that lead to mineral organization.

A nacre protein, n16.3, forms phases that introduce textured mineral overgrowth and subsurface nanoporosities within calcite crystals.

Gold nanoparticles were observed to move at a liquid/solid interface 3 orders of magnitude slower than expected for the movement in a bulk liquid by Brownian motion. The nanoscale movement was studied with scanning transmission electron microscopy (STEM) using a liquid enclosure consisting of microchips with silicon nitride windows. The experiments involved a variation of the electron dose, the coating of the nanoparticles, the surface charge of the enclosing membrane, the viscosity, and the liquid thickness. The observed slow movement was not a result of hydrodynamic hindrance near a wall but instead explained by the presence of a layer of ordered liquid exhibiting a viscosity 5 orders of magnitude larger than a bulk liquid. The increased viscosity presumably led to a dramatic slowdown of the movement. The layer was formed as a result of the surface charge of the silicon nitride windows. The exceptionally slow motion is a crucial aspect of electron microscopy of specimens in liquid, enabling a direct observation of the movement and agglomeration of nanoscale objects in liquid.

Membrane proteins govern many important functions in cells via dynamic oligomerization into active complexes. However, analytical methods to study their distribution and functional state in relation to the cellular structure are currently limited. Here, we introduce a technique for studying single-membrane proteins within their native context of the intact plasma membrane. SKBR3 breast cancer cells were grown on silicon microchips with thin silicon nitride windows. The cells were fixed, and the epidermal growth factor receptor ErbB2 was specifically labeled with quantum dot (QD) nanoparticles. For correlative fluorescence- and liquid-phase electron microscopy, we enclosed the liquid samples by chemical vapor deposited (CVD) graphene films. Depending on the local cell thickness, QD labels were imaged with a spatial resolution of 2 nm at a low electron dose. The distribution and stoichiometric assembly of ErbB2 receptors were determined at several different cellular locations, including tunneling nanotubes, where we found higher levels of homodimerization at the connecting sites. This experimental approach is applicable to a wide range of cell lines and membrane proteins and particularly suitable for studies involving both inter- and intracellular heterogeneity in protein distribution and expression.

We report an interesting process whereby the formation of nanoparticle assemblies on and nanoporosities within calcite crystals is directed by an intrinsically disordered C-RING mollusk shell nacre protein, AP7. Under mineralization conditions, AP7 forms protein phases that direct the nucleation of ordered calcite nanoparticles via a repetitive protein phase deposition process onto calcite crystals. These organized nanoparticles are separated by gaps or spaces that become incorporated into the forming bulk crystal as nanoporosities. This is an unusual example of organized nanoparticle biosynthesis and mineral modification directed by a C-RING protein phase.