No abstract

This paper presents a method to synthetically tune atomically precise megamolecule nanobody–enzyme conjugates for prodrug cancer therapy. Previous efforts to create heterobifunctional protein conjugates suffered from heterogeneity in domain stoichiometry, which in part led to the failure of antibody–enzyme conjugates in clinical trials. We used the megamolecule approach to synthesize anti-HER2 nanobody–cytosine deaminase conjugates with tunable numbers of nanobody and enzyme domains in a single, covalent molecule. Linking two nanobody domains to one enzyme domain improved avidity to a human cancer cell line by 4-fold but did not increase cytotoxicity significantly due to lowered enzyme activity. In contrast, a megamolecule composed of one nanobody and two enzyme domains resulted in an 8-fold improvement in the catalytic efficiency and increased the cytotoxic effect by over 5-fold in spheroid culture, indicating that the multimeric structure allowed for an increase in local drug activation. Our work demonstrates that the megamolecule strategy can be used to study structure–function relationships of protein conjugate therapeutics with synthetic control of protein domain stoichiometry.

Phosphorylation is an important post‐translational modification on proteins involved in many cellular processes; however, understanding of the regulation and mechanisms of global phosphorylation remains limited. Herein, we utilize self‐assembled monolayers on gold for matrix‐assisted laser desorption/ionization mass spectrometry (SAMDI‐MS) with three phosphorylated peptide arrays to profile global phosphatase activity in cell lysates derived from five mammalian cell lines. Our results reveal significant differences in the activities of protein phosphatases on phospho‐ serine, threonine, and tyrosine substrates and suggest that phosphatases play a much larger role in the regulation of global phosphorylation on proteins than previously understood.

Accurate cancer detection and diagnosis is of utmost importance for reliable drug-response prediction. Successful cancer characterization relies on both genetic analysis and histological scans from tumor biopsies. It is known that the cytoskeleton is significantly altered in cancer, as cellular structure dynamically remodels to promote proliferation, migration, and metastasis. We exploited these structural differences with supervised feature extraction methods to introduce an algorithm that could distinguish cancer from non-cancer cells presented in high-resolution, single cell images. In this paper, we successfully identified the features with the most discriminatory power to successfully predict cell type with as few as 100 cells per cell line. This trait overcomes a key barrier of machine learning methodologies: insufficient data. Furthermore, normalizing cell shape via microcontact printing on self-assembled monolayers enabled better discrimination of cell lines with difficult-to-distinguish phenotypes. Classification accuracy remained robust as we tested dissimilar cell lines across various tissue origins, which supports the generalizability of our algorithm.