**1. Introduction**

The study of fundamental processes in cell biology, as well as biomedical and technological cell-based applications, require the culturing of cells ex vivo. For this purpose, culture systems that mimic the native extracellular matrix (ECM) microenvironment by providing biochemical (cell adhesion sites, growth factors) and biophysical (mechanical stiffness) cues are in high demand. Cell culture substrates exist that are derived from natural sources as well as produced synthetically. Prominent examples of naturally derived substrates are Matrigel™ and Geltrex™. The well-studied Matrigel™ is a heterogeneous glycoprotein mixture secreted by mouse sarcoma cells whose composition is comparable to that of an embryonic basement membrane, containing a variety of ECM proteins (particularly laminin-111, collagen IV, entactin, and heparan sulfate proteoglycan) and active growth factors (e.g., fibroblast growth factor, epidermal growth factor, transforming growth factor-β, insulin-like growth factor and platelet-derived growth factor) [1,2]. Although Matrigel™ is widely used for in vitro and in vivo applications [3], it is xenogenic and has a poorly defined, complex composition that suffers from batch-to-batch variability and offers limited experimental control. Simpler functional substrates of better defined composition are constituted by isolated ECM proteins. To this effect, full-length human

fibronectin and vitronectin can either be extracted from plasma as native proteins or recombinantly produced (R&D Systems), while laminin isoforms can be expressed recombinantly in human cell lines (BioLamina). However, the production yield of these proteins is limited and their use can be troubled by the presence of impurities, uncontrollable degradation, and, in in vivo applications, possible immunogenicity [4]. Fragments from ECM proteins can also be bioactive, recapitulating the activity of the full-length protein with reasonable efficiency, to the point of supporting the long-term culturing of demanding cell types such as human pluripotent stem cells (hPSCs). Examples of bioactive ECM-protein fragments are: Laminin-511 (a combination of α5, β1 and γ1 laminin chains) produced recombinantly in human embryonic kidney cells [5]; the N-terminal somatomedin B domain of vitronectin expressed in *Escherichia coli* [6]; and the Fn7-Fn14 fragment from fibronectin expressed in mouse myeloma cells [7]. Fragments have the advantage of easing recombinant production and yielding substrates of improved purity. However, they are incomplete mimics of the ECM and proper fragment choice is critical to achieve suitable performance [7].

There is considerable interest in overcoming the limitations of cell substrates based on ECM-components through the engineering of materials that are xeno-free, feeder-free, of controlled composition, and with tailored biological functionalities. To this effect, synthetic substrates have been developed, largely in the form of peptide conjugates [8,9]. A leading example of a synthetic cell substrate is Synthemax™. In this, an acrylate base carries carboxylic acid groups to which short linear peptides (instead of folded proteins) are conjugated using chemical linkers [10]. Synthemax™ fulfills the requirements listed above and successfully supports cell proliferation. However, its usage for long-term cell culturing is not widely spread as some concerns exist related to a possible higher propensity for spontaneous cell differentiation [11] and the potential induction of karyotypic abnormalities [12].

Fulfilling the requirement of improved biodegradability and biocompatibility, materials based on self-assembling peptides have also been developed [13–17]. Peptides self-assemble to form hydrogels and can be designed to be responsive to different physical parameters, like pH and temperature. In addition, peptides produced by chemical synthesis or recombinant technologies offer optimized homogeneity as required to standardize cell-based applications. The molecular design of self-assembling peptides is usually inspired on naturally occurring, self-assembling fibrous proteins, such as collagens, elastin, silk, keratins, amyloids, and coiled-coils. Self-assembling peptides carrying bioactive sequence motifs have yielded excellent achievements in cell-based applications and are now well established [9]. Despite their demonstrated potential, the functionalization limits of peptidic systems in regards to bulky, three-dimensional bioactive components are still uncertain, as the peptidic building block is small and largely invested in mediating the polymeric assembly of the scaffold.

Compared to existing substrates, biomaterials based on self-assembling, full-length proteins can offer economy of production via recombinant methodologies, controlled composition, scalable yields, high purity with low batch variability, biodegradability, full encodability, and ease of "bottom-up" functionalization through genetic engineering—thereby bridging the advantages of synthetic designer peptides and those of natural systems based on ECM-components. Importantly, protein-based materials are ideally suited to exploit modularity by combining multiple protein domains with different functions into a single, multi-block polymer. The prospect of independent tunability of individual domain functions holds high promise for achieving fine control over multiple material properties, which is expected to lead to the development of complex multifunctional matrices of clinical significance [18–20]. Yet, designing controlled self-assembly in full-length globular proteins is highly challenging and, currently, designer protein polymers are mostly based on small and well-characterized protein subunits, where a rational molecular design is feasible. A notable example is the consensus tetratricopeptide repeat protein (CTPR), a de novo designed protein system composed of small, independently folded super-secondary helical motifs that can form a range of supramolecular assemblies, including nanofibers, nanotubes, films, and ordered monolayers [21]. Interestingly, strategies for the fusion of symmetric globular proteins and the re-design of proteins that naturally assemble into nano-objects (e.g., viral capsids) are also proving successful, with a variety of assemblies in the form of filaments, molecular layers and 3D-crystals being

possible [22,23]. The resulting molecular designs are directed to applications in encapsulation and drug delivery, multivalent epitope display and in synthetic biology. However, such multi-domain protein systems are rarely conceived as microenvironments for cell culturing applications.

We have developed a functionalized, modular biomaterial formed by the controlled self-assembly of two proteins. The polymer, termed ZT, efficiently supports the long-term self-renewal of pluripotent stem cells. Here, we review the molecular design of this new cell substrate, the principles of its assembly and functionalization, and its current application to cell culturing. Its robustness, ease and economy of recombinant production and its high versatility make it a promising system to support complex functionalities in biological applications.
