Optimal trade-off control in machine learning-based library design, with application to adeno-associated virus (AAV) for gene therapy.

Adeno-associated viruses (AAVs) hold tremendous promise as delivery vectors for gene therapies. AAVs have been successfully engineered-for instance, for more efficient and/or cell-specific delivery to numerous tissues-by creating large, diverse starting libraries and selecting for desired properties. However, these starting libraries often contain a high proportion of variants unable to assemble or package their genomes, a prerequisite for any gene delivery goal. Here, we present and showcase a machine learning (ML) method for designing AAV peptide insertion libraries that achieve fivefold higher packaging fitness than the standard NNK library with negligible reduction in diversity. To demonstrate our ML-designed library's utility for downstream engineering goals, we show that it yields approximately 10-fold more successful variants than the NNK library after selection for infection of human brain tissue, leading to a promising glial-specific variant. Moreover, our design approach can be applied to other types of libraries for AAV and beyond.

Combined Multiharmonic Frequency Analysis for Improved Dynamic Energy Measurements and Accuracy in Charge Detection Mass Spectrometry.

The ability to determine ion energies in electrostatic ion-trap-based charge detection mass spectrometry (CDMS) experiments is important for the accurate measurement of individual ion m/z, charge, and mass. Dynamic energy measurements throughout the time an ion is trapped take advantage of the relationship between ion energy and the harmonic amplitude ratio (HAR) composed from the fundamental and second harmonic amplitudes in the Fourier transform of the ion signal. This method eliminates the need for energy-filtering optics in CDMS and makes it possible to measure energy lost in collisions and changes in ion masses due to dissociation. However, the accuracy of the energy measurement depends on the signal-to-noise ratio (S/N) of the amplitudes used to determine the HAR. Here, a major improvement to this HAR-based dynamic energy measurement method is achieved using HARs composed of higher-order harmonics in addition to the fundamental and second harmonic to determine ion energies. This combined harmonic amplitude ratios for precision energy refinement (CHARPER) method is applied to the analysis of a 103 nm polystyrene nanoparticle ion (359.7 MDa, m/z = 308,300) and the energy resolution (3140) and effective mass resolution (730) achieved are the best yet demonstrated in electrostatic ion-trap-based CDMS. The CHARPER method applied to an ensemble of several thousand adeno-associated virus ion signals also results in higher mass resolution compared to the basic HAR method, making it possible to resolve additional features in the composite mass histogram.

Dynamic Energy Measurements in Charge Detection Mass Spectrometry Eliminate Adverse Effects of Ion-Ion Interactions.

Ion-ion interactions in charge detection mass spectrometers that use electrostatic traps to measure masses of individual ions have not been reported previously, although ion trajectory simulations have shown that these types of interactions affect ion energies and thereby degrade measurement performance. Here, examples of interactions between simultaneously trapped ions that have masses ranging from ca. 2 to 350 MDa and ca. 100 to 1000 charges are studied in detail using a dynamic measurement method that makes it possible to track the evolution of the mass, charge, and energy of individual ions over their trapping lifetimes. Signals from ions that have similar oscillation frequencies can have overlapping spectral leakage artifacts that result in slightly increased uncertainties in the mass determination, but these effects can be mitigated by the careful choice of parameters used in the short-time Fourier transform analysis. Energy transfers between physically interacting ions are also observed and quantified with individual ion energy measurement resolution as high as ∼950. The mass and charge of interacting ions do not change, and their corresponding measurement uncertainties are equivalent to ions that do not undergo physical interactions. Simultaneous trapping of multiple ions in CDMS can greatly decrease the acquisition time necessary to accumulate a statistically meaningful number of individual ion measurements. These results demonstrate that while ion-ion interactions can occur when multiple ions are trapped, they have negligible effects on mass accuracy when using the dynamic measurement method.

Lighting Up at High Potential: Effects of Voltage and Emitter Size in Nanoelectrospray Ionization.

Effects of electrospray voltage on cluster size and abundance formed from aqueous CsI were investigated with emitter tip diameters between 260 ± 7 nm and 2.45 ± 0.30 µm. Cluster size increases with increasing voltage, increasing solution concentration and increasing emitter diameter consistent with formation of larger initial droplet sizes. For emitters with tip diameters above ∼1 µm, varying the voltage either up or down leads to reproducible voltage-dependent extents of cluster formation. In contrast, higher voltages with submicron diameter emitters can lead to only Cs+ and Cs(H2O)+ and no clusters. This change in ion formation reproducibly occurs at spray potentials >1.3 kV for 260 nm emitters and appears to be induced by a corona discharge and material build-up at the emitter tip. Under conditions where abundant Cs+ is observed and no clusters are formed, ions such as K+ and Cu1+ are also observed but ions with more negative solvation energies, such as Ba2+, are not. Similarly, ions from bradykinin and ubiquitin are observed predischarge but not post discharge. Ions with more positive solvation energies can desorb directly from the air-water interface that is created at the tip of these emitters, whereas ions with more negative solvation energies as well as peptide and protein ions do not. These results indicate that ion desorption directly from solution can occur, and similar experiments with even smaller emitters may lead to new insights into ion formation in electrospray ionization.

Accurate Sizing of Nanoparticles Using a High-Throughput Charge Detection Mass Spectrometer without Energy Selection.

The sizes and shapes of nanoparticles play a critical role in their chemical and material properties. Common sizing methods based on light scattering or mobility lack individual particle specificity, and microscopy-based methods often require cumbersome sample preparation and image analysis. A promising alternative method for the rapid and accurate characterization of nanoparticle size is charge detection mass spectrometry (CDMS), an emerging technique that measures the masses of individual ions. A recently constructed CDMS instrument designed specifically for high acquisition speed, efficiency, and accuracy is described. This instrument does not rely on an ion energy filter or estimates of ion energy that have been previously required for mass determination, but instead uses direct, in situ measurements. A standardized sample of ∼100 nm diameter polystyrene nanoparticles and ∼50 nm polystyrene nanoparticles with amine-functionalized surfaces are characterized using CDMS and transmission electron microscopy (TEM). Individual nanoparticle masses measured by CDMS are transformed to diameters, and these size distributions are in close agreement with distributions measured by TEM. CDMS analysis also reveals dimerization of ∼100 nm nanoparticles in solution that cannot be determined by TEM due to the tendency of nanoparticles to agglomerate when dried onto a surface. Comparing the acquisition and analysis times of CDMS and TEM shows particle sizing rates up to ∼80× faster are possible using CDMS, even when samples ∼50× more dilute were used. The combination of both high-accuracy individual nanoparticle measurements and fast acquisition rates by CDMS represents an important advance in nanoparticle analysis capabilities.

Apodization Specific Fitting for Improved Resolution, Charge Measurement, and Data Analysis Speed in Charge Detection Mass Spectrometry.

Short-time Fourier transforms with short segment lengths are typically used to analyze single ion charge detection mass spectrometry (CDMS) data either to overcome effects of frequency shifts that may occur during the trapping period or to more precisely determine the time at which an ion changes mass or charge, or enters an unstable orbit. The short segment lengths can lead to scalloping loss unless a large number of zero-fills are used, making computational time a significant factor in real-time analysis of data. Apodization specific fitting leads to a 9-fold reduction in computation time compared to zero-filling to a similar extent of accuracy. This makes possible real-time data analysis using a standard desktop computer. Rectangular apodization leads to higher resolution than the more commonly used Gaussian or Hann apodization and makes it possible to separate ions with similar frequencies, a significant advantage for experiments in which the masses of many individual ions are measured simultaneously. Equally important is a >20% increase in S/N obtained with rectangular apodization compared to Gaussian or Hann, which directly translates to a corresponding improvement in accuracy of both charge measurements and ion energy measurements that rely on the amplitudes of the fundamental and harmonic frequencies. Combined with computing the fast Fourier transform in a lower-level language, this fitting procedure eliminates computational barriers and should enable real-time processing of CDMS data on a laptop computer.

Effects of Molecular Size on Resolution in Charge Detection Mass Spectrometry.

Instrumental resolution of Fourier transform-charge detection mass spectrometry instruments with electrostatic ion trap detection of individual ions depends on the precision with which ion energy is determined. Energy can be selected using ion optic filters or from harmonic amplitude ratios (HARs) that provide Fellgett's advantage and eliminate the necessity of ion transmission loss to improve resolution. Unlike the ion energy-filtering method, the resolution of the HAR method increases with charge (improved S/N) and thus with mass. An analysis of the HAR method with current instrumentation indicates that higher resolution can be obtained with the HAR method than the best resolution demonstrated for instruments with energy-selective optics for ions in the low MDa range and above. However, this gain is typically unrealized because the resolution obtainable with molecular systems in this mass range is limited by sample heterogeneity. This phenomenon is illustrated with both tobacco mosaic virus (0.6-2.7 MDa) and AAV9 (3.7-4.7 MDa) samples where mass spectral resolution is limited by the sample, including salt adducts, and not by instrument resolution. Nevertheless, the ratio of full to empty AAV9 capsids and the included genome mass can be accurately obtained in a few minutes from 1× PBS buffer solution and an elution buffer containing 300+ mM nonvolatile content despite extensive adduction and lower resolution. Empty and full capsids adduct similarly indicating that salts encrust the complexes during late stages of droplet evaporation and that mass shifts can be calibrated in order to obtain accurate analyte masses even from highly salty solutions.

Gas-Phase Protonation Thermodynamics of Biological Lipids: Experiment, Theory, and Implications.

Phospholipids are important to cellular function and are a vital structural component of plasma and organelle membranes. These membranes isolate the cell from its environment, allow regulation of the internal concentrations of ions and small molecules, and host diverse types of membrane proteins. It remains extremely challenging to identify specific membrane protein-lipid interactions and their relative strengths. Native mass spectrometry, an intrinsically gas-phase method, has recently been demonstrated as a promising tool for identifying endogenous protein-lipid interactions. However, to what extent the identified interactions reflect solution- versus gas-phase binding strengths is not known. Here, the "Extended" Kinetic Method and ab initio computations at three different levels of theory are used to experimentally and theoretically determine intrinsic gas-phase basicities (GB, ΔG for deprotonation of the protonated base) and proton affinities (PA, ΔH for deprotonation of the protonated base) of six lipids representing common phospholipid types. Gas-phase acidities (ΔG and ΔH for deprotonation) of neutral phospholipids are also evaluated computationally and ranked experimentally. Intriguingly, it is found that two of these phospholipids, sphingomyelin and phosphatidylcholine, have the highest GB of any small, monomeric biomolecules measured to date and are more basic than arginine. Phosphatidylethanolamine and phosphatidylserine are found to be similar in GB to basic amino acids lysine and histidine, and phosphatidic acid and phosphatidylglycerol are the least basic of the six lipid types studied, though still more basic than alanine. Kinetic Method experiments and theory show that the gas-phase acidities of these phospholipids are high but less extreme than their GB values, with phosphatidylserine and phosphatidylglycerol being the most acidic. These results indicate that sphingomyelin and phosphatidylcholine lipids can act as charge-reducing agents when dissociated from native membrane protein-lipid complexes in the gas phase and provide a straightforward model to explain the results of several recent native mass spectrometry studies of protein-lipid complexes.

Imidazole Derivatives Improve Charge Reduction and Stabilization for Native Mass Spectrometry.

Noncovalent interactions between biomolecules are critical to their activity. Native mass spectrometry (MS) has enabled characterization of these interactions by preserving noncovalent assemblies for mass analysis, including protein-ligand and protein-protein complexes for a wide range of soluble and membrane proteins. Recent advances in native MS of lipoprotein nanodiscs have also allowed characterization of antimicrobial peptides and membrane proteins embedded in intact lipid bilayers. However, conventional native electrospray ionization (ESI) can disrupt labile interactions. To stabilize macromolecular complexes for native MS, charge reducing reagents can be added to the solution prior to ESI, such as triethylamine, trimethylamine oxide, and imidazole. Lowering the charge acquired during ESI reduces Coulombic repulsion that leads to dissociation, and charge reduction reagents may also lower the internal energy of the ions through evaporative cooling. Here, we tested a range of imidazole derivatives to discover improved charge reducing reagents and to determine how their chemical properties influence charge reduction efficacy. We measured their effects on a soluble protein complex, a membrane protein complex in detergent, and lipoprotein nanodiscs with and without embedded peptides, and used computational chemistry to understand the observed charge-reduction behavior. Together, our data revealed that hydrophobic substituents at the 2 position on imidazole can significantly improve both charge reduction and gas-phase stability over existing reagents. These new imidazole derivatives will be immediately beneficial for a range of native MS applications and provide chemical principles to guide development of novel charge reducing reagents.