Improving Biocatalyst Performance by Integrating Statistical Methods into Protein Engineering

Directed evolution and rational design were used to generate active variants of toluene-4-monooxygenase (T4MO) on 2-phenylethanol (PEA), with the aim of producing hydroxytyrosol, a potent antioxidant. Due to the complexity of the enzymatic system—four proteins encoded by six genes—mutagenesis is labor-intensive and time-consuming. Therefore, the statistical model of Nov and Wein (J. Comput. Biol. 12:247-282) was used to reduce the number of variants produced and evaluated in a lab. From an initial data set of 24 variants, with mutations at nine positions, seven double or triple mutants were identified through statistical analysis. The average activity of these mutants was 4.6-fold higher than the average activity of the initial data set. In an attempt to further improve the enzyme activity to obtain PEA hydroxylation, a second round of statistical analysis was performed. Nine variants were considered, with 3, 4, and 5 point mutations. The average activity of the variants obtained in the second statistical round was 1.6-fold higher than in the first round and 7.3-fold higher than that of the initial data set. The best variant discovered, TmoA I100A E214G D285Q, exhibited an initial oxidation rate of 4.4 ± 0.3 nmol/min/mg protein, which is 190-fold higher than the rate obtained by the wild type. This rate was also 2.6-fold higher than the activity of the wild type on the natural substrate toluene. By considering only 16 preselected mutants (out of ∼13,000 possible combinations), a highly active variant was discovered with minimum time and effort.Enzymes are remarkable biocatalysts that perform numerous chemical reactions. They have evolved in nature to do their task in an efficient and specific way, mostly under aqueous physiological conditions (12). However, the term “biocatalysis” refers to the use of enzymes as process catalysts under artificial conditions, and a major challenge today is to render biocatalysts suitable for the tough reaction conditions of an industrial process (11).A widely used approach for improving enzyme function is directed evolution, whereby protein sequences are repeatedly selected, mutated, or recombined in a process that mimics natural evolution to produce better and better “generations” of protein variants. Directed evolution has been successfully used in numerous studies, but since it requires generation, purification, and screening of large numbers of variants, it is typically expensive and labor-intensive (28, 35). An alternative to directed evolution is an approach termed rational design, whereby predictions are made as to how mutations in a protein will affect its structure and hence its interaction with the target molecule. Unfortunately, both the sequence-structure and the structure-activity relationships are extremely intricate, and while this approach proved to be fruitful in some cases, its practical use is still limited (15). The rational-design approach also requires knowledge of the three-dimensional structure of the protein, which, unlike with the protein''s sequence, is costly and time-consuming to decipher. It has been suggested lately that a combination of both methods may be the best tactic to obtain enzymes with desired activities and selectivities (15, 23).Yet a third approach for protein improvement, which is less often used, is based on statistical analysis. According to this approach, the activity of any protein variant is viewed as a random quantity, and statistical methods are used to predict from activity data which mutation combinations are likely to improve activity. The statistical approach does not require structural knowledge about the protein at hand and allows one to focus screening efforts on a few promising variants, thus reducing labor, time, and expenses. Earlier statistical models for the sequence-activity relationship include Kauffman''s NK model (14) and the “rough Mt. Fuji” model of Aita and Husimi (1). More recently, Fox et al. combined a machine learning technique termed ProSAR with directed evolution and rational design to significantly increase the catalytic function of a halohydrin dehalogenase in the production of the cholesterol-lowering drug atorvastatin (Lipitor) (9, 10). Liao et al. (18) employed eight machine learning algorithms to improve 20-fold the ability of proteinase K to hydrolyze a tetrapeptide substrate.The statistical model that lies at the center of this work is that of Nov and Wein (22). Unlike the statistical algorithms used in the work of Fox et al. (9, 10) and Liao et al. (18), which are generic methods from the machine learning literature, this model was devised specifically for the protein design problem to capture characteristics of the protein sequence-activity relationship. Barak et al. (3) employed a variation of this model in conjunction with directed evolution to greatly improve the oxidoreductase ChrR in reducing chromate and uranyl.In this work, we combine all three approaches—directed evolution, rational design, and statistical methods—to improve the capacity of toluene 4-monooxygenase (T4MO) to produce an important antioxidant, hydroxytyrosol (6). This phenol, which is naturally present in olives and olive oil, has the highest free radical scavenging capacity and has been shown to be beneficial in preventing various diseases, such as diabetes, atherosclerosis, and cancer (13, 19, 34). Developing a biotechnological process for the synthesis of this antioxidant is of interest to the food and cosmetics industries.T4MO from Pseudomonas mendocina KR1 is a soluble four-component enzyme belonging to the toluene monooxygenase family. T4MO is composed of six genes, designated tmoABCDEF, which are essential for the efficient catalysis and high regiospecificity of the enzyme. Genes tmoA, tmoB, and tmoE encode the α, β, and γ subunits, respectively, that comprise the (αβγ)₂ quaternary structure of the 212-kDa hydroxylase component. The α hydroxylase subunit contains the catalytically active diiron center (2, 8). The tmoC, tmoD, and tmoF genes encode the 12.5-kDa Rieske-type [2Fe-2S] ferredoxin, the 11.6-kDa effector protein, and the 36-kDa NADH oxidoreductase, respectively (2, 16, 21).Previously, a 35-fold improvement in T4MO activity on 2-phenylethanol (PEA) for the synthesis of hydroxytyrosol was reported for the TmoA I100A mutant (5). The goal of the present work was to further generate a better T4MO variant for the production of hydroxytyrosol. As the cloning steps associated with producing double and triple mutants of this enzyme are very laborious and time-consuming (e.g., QuikChange mutagenesis cannot be applied due to the large plasmid involved [9 kb], and a three-step PCR is needed for each mutant [5, 6]), the integration of the Nov and Wein statistical model was evaluated.     

Cardinium in Plagiomerus diaspidis (Hymenoptera: Encyrtidae)

The bacterial symbiont Cardinium (Bacteroidetes) was previously implicated in the thelytokous reproduction of the parasitoid Plagiomerus diaspidis Crawford (Hymenoptera: Encyrtidae). Horizontal transmission of the symbiont among the cactus scale Diaspis echinocacti Bouché (Homoptera: Diaspididae) and its hymenopteran parasitoids has been suggested. In this study, the bacteria associated with D. echinocacti, its parasitoids P. diaspidis and Aphytis sp. (Hymenoptera: Aphelinidae), and the hyperparasitoid Marietta leopardina Motschulsky (Hymenoptera: Aphelinidae) were characterized using molecular fingerprinting techniques, and the localization of Cardinium in P. diaspidis was studied using fluorescence in situ hybridizations (FISH). Cardinium was the only bacterium found in P. diaspidis, but it could not be detected in any of the other insects tested. The symbiont was specifically located in the reproductive tissues of its P. diaspidis host.     

Coupling between neuronal firing rate, gamma LFP, and BOLD fMRI is related to interneuronal correlations

BACKGROUND: To what extent is activity of individual neurons coupled to the local field potential (LFP) and to blood-oxygenation-level dependent (BOLD) functional magnetic resonance imaging (fMRI)? This issue is of high significance for understanding brain function and for relating animal studies to fMRI, yet it is still under debate. RESULTS: Here we report data from simultaneous recordings of isolated unit activity and LFP by using multiple electrodes in the human auditory cortex. We found a wide range of coupling levels between the activity of individual neurons and gamma LFP. However, this large variability could be predominantly explained (r = 0.66) by the degree of firing-rate correlations between neighboring neurons. Importantly, this phenomenon occurred during both sensory stimulation and spontaneous activity. Concerning the coupling of neuronal activity to BOLD fMRI, we found that gamma LFP was well coupled to BOLD measured across different individuals (r = 0.62). By contrast, the coupling of single units to BOLD was highly variable and, again, tightly related to interneuronal-firing-rate correlations (r = 0.70). CONCLUSIONS: Our results offer a resolution to a central controversy regarding the coupling between neurons, LFP, and BOLD signals by demonstrating, for the first time, that the coupling of single units to the other measures is variable yet it is tightly related to the degree of interneuronal correlations in the human auditory cortex.     

Patterns of genetic diversity and candidate genes for ecological divergence in a homoploid hybrid sunflower, Helianthus anomalus

Natural hybridization accompanied by a shift in niche preference by hybrid genotypes can lead to hybrid speciation. Natural selection may cause the fixation of advantageous alleles in the ecologically diverged hybrids, and the loci experiencing selection should exhibit a reduction in allelic diversity relative to neutral loci. Here, we analyzed patterns of genetic diversity at 59 microsatellite loci associated with expressed sequence tags (ESTs) in a homoploid hybrid sunflower species, Helianthus anomalus. We used two indices, ln RV and ln RH, to compare variation and heterozygosity (respectively) at each locus between the hybrid species and its two parental species, H. annuus and H. petiolaris. Mean values of ln RV and ln RH were significantly lower than zero, which implies that H. anomalus experienced a population bottleneck during its recent evolutionary history. After correcting for the apparent bottleneck, we found six loci with a significant reduction in variation or with heterozygosity in the hybrid species, compared to one or both of the parental species. These loci should be viewed as a ranked list of candidate loci, pending further sequencing and functional analyses. Sequence data were generated for two of the candidate loci, but population genetics tests failed to detect deviations from neutral evolution at either locus. Nonetheless, a greater than eight-fold excess of nonsynonymous substitutions was found near a putative N-myristoylation motif at the second locus (HT998), and likelihood-based models indicated that the protein has been under selection in H. anomalus in the past and, perhaps, in one or both parental species. Finally, our data suggest that selective sweeps may have united populations of H. anomalus isolated by a mountain range, indicating that even low gene-flow species may be held together by the spread of advantageous alleles.     

Reconstructing visual experiences from brain activity evoked by natural movies

Quantitative modeling of human brain activity can provide crucial insights about cortical representations [1, 2] and can form the basis for brain decoding devices [3-5]. Recent functional magnetic resonance imaging (fMRI) studies have modeled brain activity elicited by static visual patterns and have reconstructed these patterns from brain activity [6-8]. However, blood oxygen level-dependent (BOLD) signals measured via fMRI are very slow [9], so it has been difficult to model brain activity elicited by dynamic stimuli such as natural movies. Here we present a new motion-energy [10, 11] encoding model that largely overcomes this limitation. The model describes fast visual information and slow hemodynamics by separate components. We recorded BOLD signals in occipitotemporal visual cortex of human subjects who watched natural movies and fit the model separately to individual voxels. Visualization of the fit models reveals how early visual areas represent the information in movies. To demonstrate the power of our approach, we also constructed a Bayesian decoder [8] by combining estimated encoding models with a sampled natural movie prior. The decoder provides remarkable reconstructions of the viewed movies. These results demonstrate that dynamic brain activity measured under naturalistic conditions can be decoded using current fMRI technology.     

Aromatic aldehyde and hydrazine activated peptide coated quantum dots for easy bioconjugation and live cell imaging

We present a robust scheme for preparation of semiconductor quantum dots (QDs) and cognate partners in a conjugation ready format. Our approach is based on bis-aryl hydrazone bond formation mediated by aromatic aldehyde and hydrazinonicotinate acetone hydrazone (HyNic) activated peptide coated quantum dots. We demonstrate controlled preparation of antibody--QD bioconjugates for specific targeting of endogenous epidermal growth factor receptors in breast cancer cells and for single QD tracking of transmembrane proteins via an extracellular epitope. The same approach was also used for optical mapping of RNA polymerases bound to combed genomic DNA in vitro.     

Robust control of nitrogen assimilation by a bifunctional enzyme in E. coli

Bacteria regulate the assimilation of multiple nutrients to enable growth. How is balanced utilization achieved, despite fluctuations in the concentrations of the enzymes that make up the regulatory circuitry? Here we address this question by studying the nitrogen system of E. coli. A mechanism based on the avidity of a bifunctional enzyme, adenylyltransferase (AT/AR), to its multimeric substrate, glutamine synthetase, is proposed to maintain a robust ratio between two key metabolites, glutamine and α-ketoglutarate. This ratio is predicted to be insensitive to variations in protein levels of the core circuit and to the rate of nitrogen utilization. We find using mass spectrometry that the metabolite ratio is robust to variations in protein levels and that this robustness depends on the bifunctional enzyme. Moreover, robustness carries through to the bacteria growth rate. Interrupting avidity by adding a monofunctional AT/AR mutant to the native system abolishes robustness, as predicted by the proposed mechanism.     

Glycoside hydrolases as components of putative carbohydrate biosensor proteins in <Emphasis Type="Italic">Clostridium thermocellum</Emphasis>

The composition of the cellulase system in the cellulosome-producing bacterium, Clostridium thermocellum, has been reported to change in response to growth on different carbon sources. Recently, an extensive carbohydrate-sensing mechanism, purported to regulate the activation of genes coding for polysaccharide-degrading enzymes, was suggested. In this system, CBM modules, comprising extracellular components of RsgI-like anti-σ factors, were proposed to function as carbohydrate sensors, through which a set of cellulose utilization genes are activated by the associated σ^I-like factors. An extracellular module of one of these RsgI-like proteins (Cthe_2119) was annotated as a family 10 glycoside hydrolase, RsgI6-GH10, and a second putative anti-σ factor (Cthe_1471), related in sequence to Rsi24, was found to contain a module that resembles a family 5 glycoside hydrolase (termed herein Rsi24C-GH5). The present study examines the relevance of these two glycoside hydrolases as sensors in this signal-transmission system. The RsgI6-GH10 was found to bind xylan matrices but exhibited low enzymatic activity on this substrate. In addition, this glycoside hydrolase module was shown to interact with crystalline cellulose although no hydrolytic activity was detected on cellulosic substrates. Bioinformatic analysis of the Rsi24C-GH5 showed a glutamate-to-glutamine substitution that would presumably preclude catalytic activity. Indeed, the recombinant module was shown to bind to cellulose, but showed no hydrolytic activity. These observations suggest that these two glycoside hydrolases underwent an evolutionary adaptation to function as polysaccharide binding agents rather than enzymatic components and thus serve in the capacity of extracellular carbohydrate sensors.