High-throughput screening for high-efficiency small-molecule biosynthesis.

Systems metabolic engineering faces the formidable task of rewiring microbial metabolism to cost-effectively generate high-value molecules from a variety of inexpensive feedstocks for many different applications. Because these cellular systems are still too complex to model accurately, vast collections of engineered organism variants must be systematically created and evaluated through an enormous trial-and-error process in order to identify a manufacturing-ready strain. The high-throughput screening of strains to optimize their scalable manufacturing potential requires execution of many carefully controlled, parallel, miniature fermentations, followed by high-precision analysis of the resulting complex mixtures. This review discusses strategies for the design of high-throughput, small-scale fermentation models to predict improved strain performance at large commercial scale. Established and promising approaches from industrial and academic groups are presented for both cell culture and analysis, with primary focus on microplate- and microfluidics-based screening systems.

Faster, reduced cost calibration method development methods for the analysis of fermentation product using near-infrared spectroscopy (NIRS).

Recent innovations in synthetic biology, fermentation, and process development have decreased time to market by reducing strain construction cycle time and effort. Faster analytical methods are required to keep pace with these innovations, but current methods of measuring fermentation titers often involve manual intervention and are slow, time-consuming, and difficult to scale. Spectroscopic methods like near-infrared (NIR) spectroscopy address this shortcoming; however, NIR methods require calibration model development that is often costly and time-consuming. Here, we introduce two approaches that speed up calibration model development. First, generalized calibration modeling (GCM) or sibling modeling, which reduces calibration modeling time and cost by up to 50% by reducing the number of samples required. Instead of constructing analyte-specific models, GCM combines a reduced number of spectra from several individual analytes to produce a large pool of spectra for a generalized model predicting all analyte levels. Second, randomized multicomponent multivariate modeling (RMMM) reduces modeling time by mixing multiple analytes into one sample matrix and then taking the spectral measurements. Afterward, individual calibration methods are developed for the various components in the mixture. Time saved from the use of RMMM is proportional to the number of components or analytes in the mixture. When combined, the two methods effectively reduce the associated cost and time for calibration model development by a factor of 10.

Production of amorphadiene in yeast, and its conversion to dihydroartemisinic acid, precursor to the antimalarial agent artemisinin.

Malaria, caused by Plasmodium sp, results in almost one million deaths and over 200 million new infections annually. The World Health Organization has recommended that artemisinin-based combination therapies be used for treatment of malaria. Artemisinin is a sesquiterpene lactone isolated from the plant Artemisia annua. However, the supply and price of artemisinin fluctuate greatly, and an alternative production method would be valuable to increase availability. We describe progress toward the goal of developing a supply of semisynthetic artemisinin based on production of the artemisinin precursor amorpha-4,11-diene by fermentation from engineered Saccharomyces cerevisiae, and its chemical conversion to dihydroartemisinic acid, which can be subsequently converted to artemisinin. Previous efforts to produce artemisinin precursors used S. cerevisiae S288C overexpressing selected genes of the mevalonate pathway [Ro et al. (2006) Nature 440:940-943]. We have now overexpressed every enzyme of the mevalonate pathway to ERG20 in S. cerevisiae CEN.PK2, and compared production to CEN.PK2 engineered identically to the previously engineered S288C strain. Overexpressing every enzyme of the mevalonate pathway doubled artemisinic acid production, however, amorpha-4,11-diene production was 10-fold higher than artemisinic acid. We therefore focused on amorpha-4,11-diene production. Development of fermentation processes for the reengineered CEN.PK2 amorpha-4,11-diene strain led to production of > 40 g/L product. A chemical process was developed to convert amorpha-4,11-diene to dihydroartemisinic acid, which could subsequently be converted to artemisinin. The strains and procedures described represent a complete process for production of semisynthetic artemisinin.

High-throughput optofluidic screening for improved microbial cell factories via real-time micron-scale productivity monitoring.

The industrial synthetic biology sector has made huge investments to achieve relevant miniaturized screening systems for scalable fermentation. Here we present the first example of a high-throughput (>103 genotypes per week) perfusion-based screening system to improve small-molecule secretion from microbial strains. Using the Berkeley Lights Beacon® system, the productivity of each strain could be directly monitored in real time during continuous culture, yielding phenotypes that correlated strongly (r2 > 0.8, p < 0.0005) with behavior in industrially relevant bioreactor processes. This method allows a much closer approximation of a typical fed-batch fermentation than conventional batch-like droplet or microplate culture models, in addition to rich time-dependent data on growth and productivity. We demonstrate these advantages by application to the improvement of high-productivity strains using whole-genome random mutagenesis, yielding mutants with substantially improved (by up to 85%) peak specific productivities in bioreactors. Each screen of ∼5 × 103 mutants could be completed in under 8 days (including 5 days involving user intervention), saving ∼50-75% of the time required for conventional microplate-based screening methods.

High-throughput screening for improved microbial cell factories, perspective and promise.

High-throughput screening is a critical part of any industrial strain engineering effort, helping ensure the lowest cost product is produced in the shortest amount of time. Small-scale testing that correlates to manufacturing scale allows rapid strain development with confidence that engineering changes are relevant at-scale. In this review, the current state of high-throughput screening, the technological advances for the next generation strain screening pipeline, and options for implementation are reviewed. New technologies in cell culture (optofluidics) and measurement (acoustic mist ionization mass spectrometry) are highlighted, and special considerations (i.e. cost models, correlation between scales, data quality) are discussed.

Developing fermentative terpenoid production for commercial usage.

Terpenoids comprise a large (>55000) family of compounds, very few of which have been used commercially due to low and economically unpractical production in their native hosts (generally plants and microorganisms). Two examples of natural terpenoid production are described (rubber and astaxanthin), but the advent of metabolic engineering has allowed the development of fermentative production processes using heterologous microorganisms. The two biochemical pathways responsible for terpenoid production are described, along with manipulations that enable production of terpenoids at economically viable levels. Finally, this article reviews some terpenoids that are currently in commercial production or development, ranging from semisynthetic production of the antimalarial drug artemisinin, through fragrance molecules, to commodity chemicals such as isoprene and ß-farnesene.

Mass spectrometric strategy for the characterization of lipooligosaccharides from Neisseria gonorrhoeae 302 using FTICR.

The lipooligosaccharides (LOS) of Neisseria gonorrhoeae 302 were profiled using Fourier transform ion cyclotron resonance mass spectrometry (FTICR-MS). Using techniques developed in this laboratory, the topology and some of the linkages of the LOS were determined. Mass spectrometric analysis in the negative ion mode yielded a glycoform of the composition: Hex3 Hep2 Hxn1 PEA1 KDO2 DPLA. The composition was confirmed through exact mass measurements, which showed only a 2 ppm error between the exact mass and theoretical mass. Although the core structure has been postulated previously, the positioning of the three hexose moieties were in question for this particular strain of N. gonorrhoeae. Topology assignment was performed through collision-induced dissociation analysis of the O-deacylated glycoform in the negative ion mode followed by submission to the saccharide topology analysis tool (STAT) computer program, which confirmed the topology assignment. It was found that the three hexoses were added to the Hep[I] of the conserved core of N. gonorrhoeae in a linear fashion, while Hep[II] remains unbranched. Linkage position analysis was performed through application of a mild acid hydrolysis technique followed by collision-induced dissociation of the sodiated precursor ions, yielding a 1 --> 4 linkage between the terminating and penultimate hexoses.

Investigation of ion/molecule reactions as a quantification method for phosphorylated positional isomers. an FT-ICR approach.

A rapid and accurate method of quantifying positional isomeric mixtures of phosphorylated hexose and N-acetylhexosamine monosacchrides by using gas-phase ion/molecule reactions coupled with FT-ICR mass spectrometry is described. Trimethyl borate, the reagent gas, reacts readily with the singly charged negative ions of phosphorylated monosaccharides to form two stable product ions corresponding to the loss of one or two neutral molecules of methanol from the original adduct. Product distribution in the ion/molecule reaction spectra differs significantly for isomers phosphorylated in either the 1- or the 6-position. As a result, the percents of total ion current of these product ions for a mixture of the two isomers vary with its composition. In order to determine the percentage of each isomer in an unknown mixture, a multicomponent quantification method is utilized in which the percents of total ion current of the two product ions for each pure monosaccharide phosphate and the mixture are used in a two-equation, two-unknown system. The applicability of this method is demonstrated by successfully quantifying mock mixtures of four different isomeric pairs: Glucose-1-phosphate and glucose-6-phosphate; mannose-1-phosphate and mannose-6-phosphate; galactose-1-phosphate and galactose-6-phosphate; N-acetylglucosamine-1-phosphate and N-acetylglucosamine-6-phosphate. The effects of mixture concentrations and ion/molecule reaction conditions on the quantification are also discussed. Our results demonstrate that this assay is a fast, sensitive, and robust method to quantify isomeric mixtures of phosphorylated monosaccharides.

Strategy for selective chemical cross-linking of tyrosine and lysine residues.

Chemical cross-linking of proteins combined with mass spectral analysis is a powerful technique that can be utilized to yield protein structural information, such as the spatial arrangement of multi-protein complexes or the folding of monomeric proteins. The succinimidyl ester cross-linking reagents are commonly used to cross-link primary amine-containing amino acids (N-terminus and lysine). However, in this study they were used to react with tyrosines as well, which allowed for the formation of cross-links between two primary amines, one primary amine and one tyrosine, or two tyrosines. This result is extremely important to the chemical cross-linking community for two reasons: (1) all possible cross-linked residues must be considered when analyzing data from these experiments to generate correct distance constraints and structural information, and (2) utilizing the versatility of these cross-linking reagents allows more information content to be generated from a single cross-linking reagent, which may increase the number of cross-links obtained in the experiment. Herein, we study the reactivity of the succinimidyl ester labeling and cross-linking reagents with angiotensin I and oxidized insulin beta-chain. Using the succinimidyl acetate labeling reagent, the reactivity of the N-terminus was found to be greater than either lysine or tyrosine. However, a selectivity of the cross-linking reagent was observed for either tyrosine or lysine depending on the pH of the reaction solution. In acidic pH, it was observed that tyrosine was more reactive, while in alkaline pH lysine was more reactive. Exploiting this selectivity predominantly N-terminus-tyrosine or tyrosine-tyrosine cross-links were favored at acidic pH, while N-terminus-tyrosine or tyrosine-lysine cross-links were favored at alkaline pH.

Conformational studies of Zn-ligand-hexose diastereomers using ion mobility measurements and density functional theory calculations.

Ion mobility studies and density functional theory calculations were used to study the structures of [Zn/diethylenetriamine/Hexose/Cl]+ complexes in an effort to probe differences in the three-dimensional conformations. This information allows us to gain insight into the structure of these complexes before collisional activation, which is the first step in understanding the stereoselective dissociations observed under collisionally activated conditions. The collision cross sections obtained from the ion mobility measurements showed that the mannose structure is more compact than the galactose and glucose complexes, respectively. Using density functional theory, candidate structures for each of the experimentally observed complexes were generated. Two criteria were used to determine the most likely structures of these complexes before activation: (1) The allowed relative energies of the molecules (between 0-90 kJ/mol) and (2) collision cross section agreement (within 2%) between the theoretically determined structures and the experimentally determined cross section. It was found that the identity of the monosaccharide made a difference in the overall conformation of the metal-ligand-monosaccharide complex. For glucose and galactose, metal coordination to O(6) was found to be favorable, with the monosaccharide occupying the 4C1 chair conformation, while for mannose, O(2) metal coordination was found with the monosaccharide in a B3,0 conformation. Coordination numbers varied between four and six for the Zn(II) metal centers. Given these results, it appears that the stereochemistry of the monosaccharide influences the conformation and metal coordination sites of the Zn(II)/monosaccharide/dien complex. These differences may influence the dissociation products observed under collisionally activated conditions.

Free mycolic acid accumulation in the cell wall of the mce1 operon mutant strain of Mycobacterium tuberculosis.

The lipid-rich cell wall of Mycobacterium tuberculosis, the agent of tuberculosis, serves as an effective barrier against many chemotherapeutic agents and toxic host cell effector molecules, and it may contribute to the mechanism of persistence. Mycobacterium tuberculosis strains mutated in a 13-gene operon called mce1, which encodes a putative ABC lipid transporter, induce aberrant granulomatous response in mouse lungs. Because of the postulated role of the mce1 operon in lipid importation, we compared the cell wall lipid composition of wild type and mce1 operon mutant M. tuberculosis H37Rv strains. High resolution mass spectrometric analyses of the mce1 mutant lipid extracts showed unbound mycolic acids to accumulate in the cell wall. Quantitative analysis revealed a 10.7 fold greater amount of free mycolates in the mutant compared to that of the wild type strain. The free mycolates were comprised of alpha, methoxy and keto mycolates in the ratio 1:0.9:0.6, respectively. Since the mce1 operon is regulated in vivo, the free mycolates that accumulate during infection may serve as a barrier for M. tuberculosis against toxic products and contribute to the pathogen's persistence.

Structural characterization of a novel sulfated menaquinone produced by stf3 from Mycobacterium tuberculosis.

Mycobacterium tuberculosis, the causative agent of tuberculosis, produces unique sulfated metabolites associated with virulence. One such metabolite from M. tuberculosis lipid extracts, S881, has been shown to negatively regulate the virulence of M. tuberculosis in mouse infection studies, and its cell-surface localization suggests a role in modulating host-pathogen interactions. However, a detailed structural analysis of S881 has remained elusive. Here we use high-resolution, high-mass-accuracy, and tandem mass spectrometry to characterize the structure of S881. Exact mass measurements showed that S881 is highly unsaturated, tandem mass spectrometry indicated a polyisoprene-derived structure, and characterization of synthetic structural analogs confirmed that S881 is a previously undescribed sulfated derivative of dihydromenaquinone-9, the primary quinol electron carrier in M. tuberculosis. To our knowledge, this is the first example of a sulfated menaquinone produced in any prokaryote. Together with previous studies, these findings suggest that this redox cofactor may play a role in mycobacterial pathogenesis.

PapA1 and PapA2 are acyltransferases essential for the biosynthesis of the Mycobacterium tuberculosis virulence factor sulfolipid-1.

Mycobacterium tuberculosis produces numerous exotic lipids that have been implicated as virulence determinants. One such glycolipid, Sulfolipid-1 (SL-1), consists of a trehalose-2-sulfate (T2S) core acylated with four lipid moieties. A diacylated intermediate in SL-1 biosynthesis, SL(1278), has been shown to activate the adaptive immune response in human patients. Although several proteins involved in SL-1 biosynthesis have been identified, the enzymes that acylate the T2S core to form SL(1278) and SL-1, and the biosynthetic order of these acylation reactions, are unknown. Here we demonstrate that PapA2 and PapA1 are responsible for the sequential acylation of T2S to form SL(1278) and are essential for SL-1 biosynthesis. In vitro, recombinant PapA2 converts T2S to 2'-palmitoyl T2S, and PapA1 further elaborates this newly identified SL-1 intermediate to an analog of SL(1278). Disruption of papA2 and papA1 in M. tuberculosis confirmed their essential role in SL-1 biosynthesis and their order of action. Finally, the Delta papA2 and Delta papA1 mutants were screened for virulence defects in a mouse model of infection. The loss of SL-1 (and SL(1278)) did not appear to affect bacterial replication or trafficking, suggesting that the functions of SL-1 are specific to human infection.

Lipidomics reveals control of Mycobacterium tuberculosis virulence lipids via metabolic coupling.

Mycobacterium tuberculosis synthesizes specific polyketide lipids that interact with the host and are required for virulence. Using a mass spectrometric approach to simultaneously monitor hundreds of lipids, we discovered that the size and abundance of two lipid virulence factors, phthiocerol dimycocerosate (PDIM) and sulfolipid-1 (SL-1), are controlled by the availability of a common precursor, methyl malonyl CoA (MMCoA). Consistent with this view, increased levels of MMCoA led to increased abundance and mass of both PDIM and SL-1. Furthermore, perturbation of MMCoA metabolism attenuated pathogen replication in mice. Importantly, we detected increased PDIM synthesis in bacteria growing within host tissues and in bacteria grown in culture on odd-chain fatty acids. Because M. tuberculosis catabolizes host lipids to grow during infection, we propose that growth of M. tuberculosis on fatty acids in vivo leads to increased flux of MMCoA through lipid biosynthetic pathways, resulting in increased virulence lipid synthesis. Our results suggest that the shift to host lipid catabolism during infection allows for increased virulence lipid anabolism by the bacterium.

Fatty acid analysis tool (FAAT): An FT-ICR MS lipid analysis algorithm.

Electrospray ionization mass spectrometry is becoming an established tool for the investigation of lipids. As the methods for lipid analysis become more mature and their throughput increases, computer algorithms for the interpretation of such data will become a necessity. Toward this end, an algorithm dedicated to the analysis of Fourier transform mass spectral data from lipid extracts has been developed. The algorithm, Fatty Acid Analysis Tool, termed FAAT, has been successfully used to investigate complex lipid extracts containing thousands of components, from various species of mycobacteria including M. tuberculosis and M. abscessus. FAAT is rapid, generally taking tens of seconds to interpret multiple spectra, and accessible to most users as it is implemented in Microsoft Excel Visual Basic Software. In the reduction of data, FAAT begins by scaling spectra (i.e., to account for dilution factors), identifying monoisotopic ions, and assigning isotope packets. Unique features of FAAT include the following: (1) overlapping saturated and unsaturated lipid species can be distinguished, (2) known ions are assigned from a user-defined library including species that possess methylene heterogeneity, (3) and isotopic shifts from stable isotope labeling experiments are identified and assigned (up to a user-defined maximum). In addition, abundance differences between samples grown under normal and stressed conditions can be determined. In the analysis of mycobacterial lipid extracts, FAAT has successfully identified isotopic shifts from incorporation of 15N in M. abscessus. Additionally, FAAT has been used to successfully determine differences in lipid abundances between M. tuberculosis wild-type and mutant strains.

A sulfated metabolite produced by stf3 negatively regulates the virulence of Mycobacterium tuberculosis.

Sulfated molecules have been shown to modulate isotypic interactions between cells of metazoans and heterotypic interactions between bacterial pathogens or symbionts and their eukaryotic host cells. Mycobacterium tuberculosis, the causative agent of tuberculosis, produces sulfated molecules that have eluded functional characterization for decades. We demonstrate here that a previously uncharacterized sulfated molecule, termed S881, is localized to the outer envelope of M. tuberculosis and negatively regulates the virulence of the organism in two mouse infection models. Furthermore, we show that the biosynthesis of S881 relies on the universal sulfate donor 3'-phosphoadenosine-5'-phosphosulfate and a previously uncharacterized sulfotransferase, stf3. These findings extend the known functions of sulfated molecules as general modulators of cell-cell interactions to include those between a bacterium and a human host.

Unambiguous assignment of intramolecular chemical cross-links in modified mammalian membrane proteins by Fourier transform-tandem mass spectrometry.

Fourier transform tandem mass spectrometry (FT-MS/MS) can be used to unambiguously assign intramolecular chemical cross-links to specific amino acid residues even when two or more possible cross-linking sites are adjacent in the cross-linked protein. Bovine rhodopsin (Rho) in its dark-adapted state was intramolecularly cross-linked with lysine-cysteine (K-C) or lysine-lysine (K-K) cross-linkers to obtain interatomic distance information. Large, multiply charged, cross-linked peptide ions containing adjacent lysines, corresponding to Rho(50-86) (K(66) or K(67)) cross-linked to Rho(310-317) (C(316)) or Rho(318-348) (K(325) or K(339)), were fragmented by collision-induced dissociation (CID), infrared multiphoton dissociation (IRMPD), and electron capture dissociation (ECD). Complementary sequence-specific information was obtained by combining cross-link assignments; however, only ECD revealed full palmitoylation of adjacent cysteines (C(322) and C(323)) and cross-linking of K(67) (and not K(66)) to C(316), K(325), and K(339). ECD spectra contained crucial c- and z-ions resulting from cleavage of the bond between K(66) and K(67). To our knowledge, this work also presents the first demonstration that ECD can be used to characterize S-linked fatty acid acylation on cysteines. The comprehensive fragmentation of large peptides by CID, IRMPD, and particularly ECD, in conjunction with the high resolution and mass accuracy of FT-MS/MS, is shown to be a valuable means of characterizing mammalian membrane proteins with both chemical and posttranslational modifications.

Solvent effects in tandem mass spectrometry: mechanistic studies indicating how a change in solvent conditions and pH can dramatically alter CID spectra.

Dramatically different CID (collision-induced dissociation) spectra are obtained when the complex [Zn(dien-glucose)](+) is electrosprayed from acidic and basic solutions. To understand this peculiar phenomenon, an in-depth mechanistic study was performed on one of the product ions that is present when the initial complex is diluted in basic solution but absent when the complex is diluted with acidic solution. On the basis of the results of this study, the differences in the CID spectra can be rationalized by the fact that the complex electrosprayed from basic solution was kinetically trapped, with the deprotonation site distal from the metal center. Under acidic conditions, the deprotonation site is at a hydroxyl group coordinated to the metal ion. A variety of experiments support this hypothesis. The studies herein underscore the importance of using identical solvent conditions when comparing sets of CID spectra. The data also highlight a very interesting phenomenon involving deprotonation of a hydroxyl group, which was several atoms away from the Zn(II) metal center.

Screening and identification of acidic carbohydrates in bovine colostrum by using ion/molecule reactions and Fourier transform ion cyclotron resonance mass spectrometry: specificity toward phosphorylated complexes.

A screening method was developed for the identification of acidic saccharides from biological mixtures utilizing gas-phase derivatization and mass spectrometry. Phosphorylated compounds were differentiated from other acidic species by exploiting the selective reactivity of chlorotrimethylsilane with the phosphate ions (phosphorylated compounds shift by 72 Da, allowing rapid compound detection). A 13-component mock mixture was used to demonstrate the viability of the method, and a detection limit of 600 nM (30 fmol) was determined. This method was applied to the identification of acidic compounds from bovine colostrum. To further verify the selectivity of the ion/molecule reaction, exact mass measurements were used to determine the elemental composition of 14 compounds. Eight novel acidic carbohydrate species were observed in bovine colostrum, six of which have never been reported previously in milks. Tandem mass spectrometric experiments allowed compound characterization for two of these components.

Analysis of phosphate position in hexose monosaccharides using ion-molecule reactions and SORI-CID on an FT-ICR mass spectrometer.

Through the use of ion-molecule reactions and SORI-CID, the phosphate position in hexose phosphate monosaccharides has been determined in the negative ion mode. Trimethyl borate was used as a reagent gas and was found to react readily with the phosphorylated hexose monosaccharides. After reaction of the reagent gas with the hexose phosphate, ion activation of the precursor by SORI-CID yielded different MS/MS spectra. Different diagnostic ions were generated for the two isomers, thus enabling differentiation and linkage position determination of the phosphate moiety.