Reduction of therapeutic antibody self-association using yeast-display selections and machine learning.

Self-association governs the viscosity and solubility of therapeutic antibodies in high-concentration formulations used for subcutaneous delivery, yet it is difficult to reliably identify candidates with low self-association during antibody discovery and early-stage optimization. Here, we report a high-throughput protein engineering method for rapidly identifying antibody candidates with both low self-association and high affinity. We find that conjugating quantum dots to IgGs that strongly self-associate (pH 7.4, PBS), such as lenzilumab and bococizumab, results in immunoconjugates that are highly sensitive for detecting other high self-association antibodies. Moreover, these conjugates can be used to rapidly enrich yeast-displayed bococizumab sub-libraries for variants with low levels of immunoconjugate binding. Deep sequencing and machine learning analysis of the enriched bococizumab libraries, along with similar library analysis for antibody affinity, enabled identification of extremely rare variants with co-optimized levels of low self-association and high affinity. This analysis revealed that co-optimizing bococizumab is difficult because most high-affinity variants possess positively charged variable domains and most low self-association variants possess negatively charged variable domains. Moreover, negatively charged mutations in the heavy chain CDR2 of bococizumab, adjacent to its paratope, were effective at reducing self-association without reducing affinity. Interestingly, most of the bococizumab variants with reduced self-association also displayed improved folding stability and reduced nonspecific binding, revealing that this approach may be particularly useful for identifying antibody candidates with attractive combinations of drug-like properties.Abbreviations: AC-SINS: affinity-capture self-interaction nanoparticle spectroscopy; CDR: complementarity-determining region; CS-SINS: charge-stabilized self-interaction nanoparticle spectroscopy; FACS: fluorescence-activated cell sorting; Fab: fragment antigen binding; Fv: fragment variable; IgG: immunoglobulin; QD: quantum dot; PBS: phosphate-buffered saline; VH: variable heavy; VL: variable light.

Machine Learning Applied to Determine the Molecular Descriptors Responsible for the Viscosity Behavior of Concentrated Therapeutic Antibodies.

Predicting the solution viscosity of monoclonal antibody (mAb) drug products remains as one of the main challenges in antibody drug design, manufacturing, and delivery. In this work, the concentration-dependent solution viscosity of 27 FDA-approved mAbs was measured at pH 6.0 in 10 mM histidine-HCl. Six mAbs exhibited high viscosity (>30 cP) in solutions at 150 mg/mL mAb concentration. Combining molecular modeling and machine learning feature selection, we found that the net charge in the mAbs and the amino acid composition in the Fv region are key features which govern the viscosity behavior. For mAbs whose behavior was not dominated by charge effects, we observed that high viscosity is correlated with more hydrophilic and fewer hydrophobic residues in the Fv region. A predictive model based on the net charges of mAbs and a high viscosity index is presented as a fast screening tool for classifying low- and high-viscosity mAbs.

Machine Learning Feature Selection for Predicting High Concentration Therapeutic Antibody Aggregation.

Protein aggregation can hinder the development, safety and efficacy of therapeutic antibody-based drugs. Developing a predictive model that evaluates aggregation behaviors during early stage development is therefore desirable. Machine learning is a widely used tool to train models that predict data with different attributes. However, most machine learning techniques require more data than is typically available in antibody development. In this work, we describe a rational feature selection framework to develop accurate models with a small number of features. We applied this framework to predict aggregation behaviors of 21 approved monospecific monoclonal antibodies at high concentration (150 mg/mL), yielding a correlation coefficient of 0.71 on validation tests with only two features using a linear model. The nearest neighbors and support vector regression models further improved the performance, which have correlation coefficients of 0.86 and 0.80, respectively. This framework can be extended to train other models that predict different physical properties.

Protospacer Adjacent Motif-Induced Allostery Activates CRISPR-Cas9.

CRISPR-Cas9 is a genome editing technology with major impact in life sciences. In this system, the endonuclease Cas9 generates double strand breaks in DNA upon RNA-guided recognition of a complementary DNA sequence, which strictly requires the presence of a protospacer adjacent motif (PAM) next to the target site. Although PAM recognition is essential for cleavage, it is unknown whether and how PAM binding activates Cas9 for DNA cleavage at spatially distant sites. Here, we find evidence of a PAM-induced allosteric mechanism revealed by microsecond molecular dynamics simulations. PAM acts as an allosteric effector and triggers the interdependent conformational dynamics of the Cas9 catalytic domains (HNH and RuvC), responsible for concerted cleavage of the two DNA strands. Targeting such an allosteric mechanism should enable control of CRISPR-Cas9 functionality.

Theoretical Investigation of Water Oxidation on Fully Saturated Mn2O3 and Mn2O4 Complexes.

Understanding the factors that affect efficiency of manganese oxides as water oxidation catalysts is an essential step toward developing commercially viable electrocatalysts. It is important to understand the performance of the smallest versions of these catalysts, which will in return be advantageous with bottom up catalytic design. Density functional theory calculations have been employed to investigate water oxidation processes on Mn2(µ-OH)(µ-O)(H2O)3(OH)5 (Mn2O4·6H2O), Mn2(µ-OH)2(H2O)3(OH)4 (Mn2O3·6H2O), and Mn2(µ-OH)2(H2O)4(OH)4 (Mn2O3·7H2O) complexes. The effect of different oxidation states of manganese is considered in this study. Thermodynamically, the lowest energy pathway for the fully saturated Mn2O4·6H2O complex occurs through a nucleophilic attack of a solvent water molecule to a Mn(IV1/2)O oxo moiety. The lowest energy pathway on the Mn2O3·6H2O complex proceeds with an attack of Mn(VI)O group to the surface hydroxo group on the same manganese atom; this pathway is related to the third lowest energy pathway on the Mn2O4·6H2O complex. The water oxidation process on the fully saturated Mn2O3·7H2O complex also involves a nucleophilic attack from a solvent water molecule to a Mn(V)O moiety. The formation of these manganese oxo groups can be used as a descriptor for selecting a manganese-based water splitting catalyst due to the high electrochemical potentials required for the generation of these groups.

Insights from Theory and Experiment on the Photochromic spiro-Dihydropyrrolo-Pyridazine/Betaine System.

We elucidated the photochromic spiro-4a,5-dihydropyrrolo[1,2-b]pyridazine/betaine (DPP/betaine) system by comparing state-of-the-art density functional theory calculations with nanosecond/millisecond UV-vis absorption spectroscopy, as well as steady-state absorption and cyclization kinetics. Time-dependent density functional theory calculations are employed to examine the transformations occurring after photoexcitation. This study shows that the photochromic spiro-4a,5-dihydropyrrolo[1,2-b]pyridazine and spiro-1,8a-dihydroindolizine (DHI) systems react according to similar pathways. However, notable differences exist. Although photoexcitation of the spiro-DPP system also leads to cis-betaines, which then isomerize to trans-betaines, we found two distinct classes of cis isomers (cis-betaine rotamer-1 and cis-betaine rotamer-2), which do not exist in spiro-1,8a-dihydroindolizine. Similar to our previous study on the spiro-DHI/betaine system, a complicated potential-energy landscape between cis and trans isomers exists in the spiro-DPP system, consisting of a network of transition states and intermediates. Because the spiro-DPP/betaine is even more complicated than the spiro-DHI/betaine system, (substituted) photochromic systems featuring a 4a,5-dihydropyrrolo[1,2-b]pyridazine functional unit will require thorough in silico design to function properly as logical gates or in devices for information storage.

Ab initio electronic structure study of a model water splitting dimer complex.

A model manganese dimer electrocatalyst bridged by µ-OH ligands is used to investigate changes in spin states that may occur during water oxidation. We have employed restricted open-shell Hartree-Fock (ROHF), second-order Møller-Plesset perturbation theory (MP2), complete active space self-consistent field (CASSCF), and multireference second-order Møller-Plesset perturbation theory (MRMP2) calculations to investigate this system. Multiconfigurational methods like CASSCF and MRMP2 are appropriate methods to study these systems with antiferromagnetically-coupled electrons. Orbital occupations and distributions have been closely analyzed to understand the electronic details and contributions to the water splitting from manganese and oxygen atoms. The presence of Mn(IV)O˙ radical moieties has been observed in this catalytic pathway. Multiple nearly degenerate excited states were found close to the ground state in all structures. This suggests competing potential energy landscapes near the ground state may influence the reactivity of manganese complexes such as the dimers studied in this work.

Refined Insights in the Photochromic spiro-Dihydroindolizine/Betaine System.

We have revisited the photochromic spiro-dihydroindolizine/betaine system by comparing state-of-the-art density functional theory calculations with experimental data. Time-dependent density functional theory calculations are employed to examine the transformations occurring after photoexcitation. This study confirms that photoexcitation of the spiro-dihydroindolizine leads to the formation of the cis-betaine. However, isomerization to the trans-betaine follows through a complicated and formerly unknown potential energy landscape, which consists of a network of transition states and intermediates. The available pathways across this potential energy landscape will determine the kinetics of the forward reaction from the cis-betaine to the trans-betaine and then, even more importantly, the back-reaction. Virtually all practical applications of this optical switch rely on these reactions and, therefore, occur within this landscape. Predicting the network of transition states and intermediates for substituted spiro-dihydroindolizine/betaine systems will enable the in-silico design of optical switches with enhanced performance.