Pedro Silva - Research interests

Pedro Silva - Research interests

In 1992, I enrolled in the "Licenciatura" in Biochemistry at the Faculty of Sciences, University of Oporto, Portugal. I finished this study in September 1996, after a 6-month training period under the supervision of Prof. Fred Hagen, at the Department of Biochemistry of the Wageningen Agricultural University, the Netherlands. In November 1996, I started my PhD research on the soluble hydrogenase of the hyperthermophilic archaeon Pyrococcus furiosus, under the joint supervision of Prof. Baltazar de Castro (Faculty of Sciences, University of Oporto, Portugal) and Prof. Fred Hagen (Wageningen Agricultural University, the Netherlands, and later at the Department of Biotechnology, Delft Technical University, the Netherlands). Eventually, the research focus broadened to include other metalloproteins from P. furiosus, which were successively purified and characterized by biochemical, electrochemical and spectroscopic methods (electron paramagnetic resonance, UV-Vis spectroscopy, cofactor analysis, bioinformatics, enzymology, etc.). In March 2001 I graduated as a PhD in Chemistry at the Faculty of Sciences, University of Porto, Portugal. After completing my PhD, I became an Assistant Professor at Universidade Fernando Pessoa (Porto, Portugal). My research focus then moved to the computational study of enzymatic and organic reaction mechanisms using quantum chemistry methods. I have been especially active in the determination of reaction mechanisms of cofactor-free oxygenases, enzymes in the heme biosynthesis pathway and Cu(I)-catalyzed reactions. Lately, I have also turned my attention to the development of enzyme inhibitors and to the critical examination of some alternatives to the chemiosmotic description of oxidative phosphorylation. I left my position at Universidade Fernando Pessoa in the end of August 2023. My new contact email is pedro_dot_dft_at_gmail_dot_com

2020-2023
2016-2019
2012-2015
2009-2011
2005-2008
Before 2005
Complete list of Scientific Publications

Current research
Tomasz M. Karpinski, Marcin Ozarowski, Pedro J. Silva , Mark Stasiewicz, Rahat Alam and Abdus Samad (2023) Discovery of terpenes as novel HCV NS5B polymerase inhibitors via molecular docking Pathogens, 12:842
Pedro J. Silva (2023) The reaction mechanism of the Cu(I) catalyzed alkylation of hetero-substituted alkynes. Catalysts, 13:17 [PDF]
Pedro J. Silva , Maria Osswald-Claro e Rosário Castro Mendonça (2022) How to tune the absorption spectrum of chlorophylls to enable better use of the available solar spectrum PeerJ Physical Chemistry, 4:e26 [PDF]
Pedro J. Silva (2022) Computational development of inhibitors of plasmid-borne bacterial dihydrofolate reductase Antibiotics, 11: 779 [PDF]
Pedro J. Silva and Qi Cheng (2022) An Alternative Proposal for the Reaction Mechanism of Light-Dependent Protochlorophyllide Oxidoreductase ACS Catalysis ,12:2589-2605 [PDF]
Pedro J. Silva and Carlos E. P. Bernardo (2021) Unveiling the reaction mechanism of the Das/Chechik/Marek synthesis of stereodefined quaternary carbon centers. Applied Sciences, 11: 5002 [PDF]
Pedro J. Silva (2020) Response to "Molecular-level understanding of biological energy coupling and transduction" Biophys. Chem 264, 106512     [PDF freely available through this link]
Pedro J. Silva (2020) New insights into the mechanism of Schiff base synthesis from aromatic amines in the absence of acid catalyst or polar solvents PeerJ Organic Chemistry, 2, e4 [PDF]
Sandra V.R.L. Silva and Pedro J. Silva (2020) Computational improvement of small-molecule inhibitors of Bacillus anthracis protective antigen activation through isostere-based substitutions. Journal of Biomolecular Structure and Dynamics DOI:10.1080/07391102.2020.1792987 FREE reprint
Pedro J. Silva (2020), Chemiosmotic misundertandings Biophys. Chem 264, 106424
2016-2019
Pedro J. Silva and Carlos E. P. Bernardo (2018), Influence of Alkyne and Azide Substituents on the Choice of the Reaction Mechanism of the Cu⁺-Catalyzed Addition of Azides to Iodoalkynes J. Phys. Chem. A, 122, 7497-7507    [free reprint upon registration]
Pedro J. Silva (2016), Refining the reaction mechanism of O₂ towards its substrate in cofactor-free dioxygenases PeerJ 4:e2805 [PDF]
Pedro J. Silva (2016), Will 1,2-dihydro-1,2-azaborine-based drugs resist metabolism by cytochrome P450 compound I? PeerJ 4:e2299 [PDF]
Carlos E. P. Bernardo and Pedro J. Silva (2016), Computational exploration of the reaction mechanism of the Cu⁺-catalysed synthesis of indoles from N-aryl enaminones Royal Society Open Science 3:150582 [PDF]
2012-2015
Pedro J. Silva and Viviana Rodrigues (2015), Mechanistic pathways of mercury removal from the organomercurial lyase active site. PeerJ 3:e1127 [PDF]
Pedro J. Silva (2014) With or without light: comparing the reaction mechanism of dark-operative protochlorophyllide oxidoreductase with the energetic requirements of the light-dependent protochlorophyllide oxidoreductase PeerJ, 2, e551 [PDF]
Carlos E. P. Bernardo and Pedro J. Silva (2014) Computational development of rubromycin-based lead compounds for HIV-1 reverse transcriptase inhibition PeerJ, 2, e470 [PDF]
Carlos E. P. Bernardo, Nicholas P. Baumann, Piotr Piecuch and Pedro J. Silva (2013) Evaluation of density functional methods on the geometric and energetic descriptions of species involved in Cu⁺-promoted catalysis. J. Mol. Model., 19, 5457-5467
Carla Sousa and Pedro J. Silva (2013) BBr₃-Assisted Cleavage of Most Ethers Does Not Follow the Commonly Assumed Mechanism. Eur. J. Org. Chem., 2013, 5195-5199. Correction in Eur. J. Org. Chem., 2013, 8048
Pedro J. Silva (2012) Unravelling the reaction mechanism of the reductive ring contraction of 1,2-pyridazines. J. Org. Chem., 77, 4653-4659
Pedro J. Silva, Marta A. S. Perez, Natércia F. Brás, Pedro A. Fernandes and Maria J. Ramos (2012) Improving the study of proton transfers between amino acid sidechains in solution: choosing appropriate DFT functionals and avoiding hidden pitfalls. Theoretical Chemistry Accounts, 131, 1179-1185
2009-2011

Accuracy of Density Functionals in the Prediction of Electronic Proton Affinities of Amino Acid Side Chains Successes and Failures of DFT Functionals in acid/base and redox reactions of organic and biochemical interest.

Journal of Chemical Theory and Computation, 7, 3898-3908 Computational and Theoretical Chemistry, 966, 120-126

We have analyzed how well DFT functionals, often used to characterize complex and large models such as proteins, describe the zero-point-exclusive proton affinity at 0 K, PA_el^0K, for the ionizable side chains of lysine (Lys), histidine (His), arginine (Arg), and aspartate (Asp–) as well as the cysteine (Cys–), serine (Ser–), and tyrosine (Tyr–) anions. The reference values PA_el^0K were determined at the very accurate CCSD(T)/CBS level. Those values were obtained by the sum of the complete basis set limit of the MP2 energies plus a CCSD(T) correction term evaluated with the aug-cc-pVTZ basis set. The complete basis set limit of MP2 energies was determined using the Truhlar and Helgaker extrapolation schemes. A new, important, and consistent DFT benchmarking database for PA_el^0K and for proton transfer between two different ionizable side chains, ΔPA_el^0K, is provided, making this work relevant to all studies with ionizable amino acids side chains that use DFT.
The performance of 18 different DFT functionals in the prediction of absolute and relative energies of organic and biochemical acid/base and redox reactions was evaluated, using MP2 extrapolated to the complete basis set limit and CCSD(T)/aug-cc-pVTZ energies as benchmark. In the acid/base reaction test set, several hybrid-GGA and meta-hybrid GGA functionals (BHHLYP, B97-1, B97-2, X3LYP, M06-2X, as well as the popular B3LYP) yielded small errors (0.8-1.5 kcal.mol^-1) in the computation of relative energies across the tested model reactions. For most of the proton transfers between different acid/base pairs (or electron transfer between redox pairs) tested we could find at least one functional with very high accuracy (error <0.2 kcal.mol^-1). The reactions involving electron transfer between quinone (or disulfide) and other redox groups stand out as the clearest example of the shortcomings of DFT methods, as the best functionals are most often wrong by 1-3 kcal.mol^-1.

Computational Characterization of the Substrate-Binding Mode in Coproporphyrinogen III Oxidase A Tale Of Two Acids: When Arginine Is A More Appropriate Acid Than H₃O⁺

J. Phys. Chem. B, 115, 1903-1910[free reprint] J. Phys. Chem. B, 114, 8994–9001 [free reprint]

The molecular dynamics simulations described in this paper enabled the determination of a very promising substrate binding mode and the extensive characterization of the enzyme active site. The proposed binding mode is fully consistent with the known selectivity of the active site toward substituted tetrapyrroles and explains the lack of activity of the H131A, R135A, D274A, and R275A mutants and the reasons behind the nonoccurrence of catalysis on the C and D rings of the tetrapyrrole. An important role in this binding mode is fulfilled by G276, as its carbonyl oxygen intervenes in the substrate anchoring by hydrogen bonding its ring D pyrrole NH group. The presence of this interaction (which is only possible with the protonated NH pyrrole group) and the absence of positively charged side chains close to the pyrrole nitrogen (which might stabilize the N-deprotonated pyrrole postulated in some mechanistic proposals) show that the pyrrole ring is very unlikely to undergo deprotonation during the catalytic cycle and allow the discrimination between the previously postulated mechanistic proposals.

We have previously found that the rate-limiting step of uroporphyrinogen III decarboxylase is substrate protonation rather than the decarboxylation reaction. This protonation can be effected by an arginine residue (Arg37) in close proximity to the substrate. In this report we present evidence for the function of this arginine residue as a general acid catalyst. Although substrate protonation by H3O+ is both exergonic and very fast, our density functional calculations show that in the presence of a protonated Arg37 substrate decarboxylation becomes rate-limiting and the substrate spontaneously breaks upon protonation. These results suggest that the active site must be shielded from solvent protons. Consequently, H3O+ can be excluded from a role in both protonations proposed for the enzyme mechanism. In agreement with these conclusions a second arginine residue (Arg41) is uniquely positioned to act as donor of the second proton, with an activation barrier below 2 kcal mol-1. Generated mutant uroporphyrinogen III decarboxylase variants carrying amino acid exchanges in the position of both arginine residues (R41A, R41K, R37A and R37K) failed to produce coproporphyrinogen III.

Computational insights into the photochemical step of the reaction catalyzed by protochlorophylide oxidoreductase. Computational studies on the reactivity of substituted 1,2-dihydro-1,2-azaborines.

Int. J. Quantum Chem. , 111, 1472-1479 [PDF] J. Org. Chem, 74, 6120-6129 [free reprint]

The light-dependent enzyme protochlorophyllide oxidoreductase catalyzes the conversion of protochlorophyllide into chlorophyllide, during chlorophyll synthesis. The reaction has been proposed to proceed through light-induced weakening of the C17-C18 double bond in protochlorophyllide, which then facilitates hydride transfer from a NADPH co-substrate molecule. We have performed DFT and TDDFT computations on the reaction mechanism of this interesting enzyme. The results show that whereas in the ground state the reaction is strongly endergonic and has a very high activation free energy (38 kcal.mol-1), the first four excited states (corresponding to excitations within the conjugated porphyrin π-system) afford much lower activation free energies (<25 kcal.mol-1) and spontaneous (or only slightly endergonic) reaction paths. The sharp shape of the potential energy surface along the reaction coordinate in these excited states allows hydrogen tunneling to occur efficiently on the first excited state surface, lowering the barrier to values closer to experiment, in agreement with recent suggestions.

We have investigated important intermediates of electrophilic aromatic substitution reactions and one-electron oxidation of substituted 1,2-dihydro-1,2-azaborines with density-functional theory. The results show that electrophilic substitution reactions and one-electron oxidation of substituted 1,2-azaborines are generally much more favorable than those of the corresponding benzene derivatives. Both chlorination and nitration of several boron-unsubstituted 1,2-azaborines are expected to break the boron-hydrogen bond, yielding boron-chlorinated 1,2-azaborines and a novel class of boron-bound 1,2-azaborinyl nitrites, respectively. Comparison between the relative stabilities of C3-bound and C5-bound Wheland intermediates of different electrophilic substitution reactions of 1,2-azaborines further suggests that the preference of the C3- over C5-substitution decreases with decreasing electrophilicity of the attacking group.

Inductive and Resonance Effects on the Acidities of Phenol, Enols, and Carbonyl α-Hydrogens. Comparative Density Functional Study of Models for the Reaction Mechanism of Uroporphyrinogen III Synthase

J. Org. Chem., 74, 914-916 [PDF] J. Phys. Chem. B, 112, 3144-3148 [PDF]

The increased acidity of phenols and carboxylic acids over aliphatic alcohols has traditionally been attributed to resonance stabilization of their (deprotonated) anions. A competing explanation argues that this increase is rather due to destabilization of the acidic proton due to electrostatic effects of the adjacent C=O (in carboxylic acids) or phenyl ring (in phenols). A lively debate on the merits of this proposal has been carried out in the literature. Recent work has shown that in carboxylic acids inductive effects are responsible for 2/3 of the stabilization of the deprotonated anion, but for the case of phenol no solution to the controversy has been forthcoming. Meanwhile, the controversy regarding the origin of the increased acidity of phenols vs. aliphatic alcohols has entered the undergraduate education textbooks, with some arguing for the traditional view, whereas others downplay anion resonance stabilization. This study establishes that inductive effects, while more important than traditionally recognized, do not explain the majority of the observed acidity enhancement in phenol, enols or carbonyl α–hydrogens. Inductive effects account for 1/3 of the enhanced acidity of phenol vs. cyclohexanol, 2/5 of the enhanced acidity of enol vs. methanol, and l/4 of the enhanced acidity of carbonyl α-hydrogens vs. methane.

Uroporphyrinogen III synthase catalyzes the inversion of one of the four heterocyclic rings present in the hydroxymethylbilane. Two mechanisms have been proposed to explain this puzzling ring inversion, either through sigmatropic shifts or through the direct formation of a spirocyclic pyrrolenine intermediate. We performed the first high-level quantum mechanical calculations on model systems of this enzyme to analyze these contrasting reaction mechanisms. The results allow us to discard the sigmatropic shift mechanism and suggest that the D-ring of the hydroxymethylbilane substrate binds to the enzyme in a conformation that shields its terminal portion from reacting with ring A and prevents the formation of the biologically useless uroporphyrinogen I, whose accumulation (in individuals lacking functional uroporphyrinogen III synthase) leads to severe cutaneous dermatosis.

2005-2008

A comparative density-functional study of the reaction mechanism of the O₂-dependent coproporphyrinogen III oxidase. Reaction Mechanism of the Vitamin K-Dependent Glutamate Carboxylase: A Computational Study.

Bioorg. Med. Chem., 16, 2726-2733, [PDF] J. Phys. Chem. B, 111, 12883-12887 [PDF]

During heme biosynthesis, coproporphyrinogen III oxidase catalyzes the conversion of two propionate substituents from the highly reactive substrate coproporphyrinogen III into vinyl substituents, yielding protoporphyrinogen IX. We have performed DFT calculations on model systems in order to analyze several reaction mechanisms proposed for this enzyme. The results afford a full description of the different proposals and allow the rejection of a direct electron abstraction from the protonated substrate by dioxygen. We found that O₂ addition to the (preferentially deprotonated) pyrrole substrate (yielding a hydroperoxide, which then abstracts a proton from the reactive propionate substituent) is compatible with the observed experimental reaction rate, and that the reaction may then proceed through HO₂^- elimination, followed by decarboxylation.

In the reaction cycle of glutamate carboxylase, vitamin K epoxidation by O₂ has been proposed to generate a very strong base able to remove a proton from the carbon of a Glu residue, thus yielding a Glu-based carbanion that readily reacts with CO₂. We have used hybrid density functional theory to study this appealing mechanism. Our calculations show a very exergonic four-step mechanism with the reaction of (triplet) O₂ with the singlet vitamin K anion as the rate-limiting step, with a rate similar to the experimental value. Our study also establishes the need to apply continuum models when performing the optimization of minimum-energy crossing points between potential energy surfaces of different multiplicities for enzyme model systems.

Assessing the reliability of sequence similarities detected through hydrophobic sequence analysis A Density-Functional Study of Mechanisms for the Cofactor-Free Decarboxylation Performed by Uroporphyrinogen III Decarboxylase.

Proteins: Structure, Function and Bioinformatics, 70, 1588-1594 [PDF] Journal of Physical Chemistry B, 109, 18195-18200 [PDF]

Hydrophobic cluster analysis (HCA) has long been used as a tool to detect distant homologies between protein sequences, and to classify them into different folds. However, it relies on expert human intervention, and is sensitive to subjective interpretations of pattern similarities. I developed a novel algorithm to assess the similarity of hydrophobic amino acid distributions between two sequences. This algorithm correctly identifies as misattributions several HCA-based proposals of structural similarity between unrelated proteins present in the literature. Besides enabling a reliable identification of the correct fold of an unknown sequence and the choice of suitable templates, the new algorithm also shows that whereas most structural classes of proteins are very homogeneous in hydrophobic cluster composition, a tenth of the described families are compatible with a large variety of hydrophobic patterns. A browsable database of every major representative hydrophobic cluster pattern present in each structural class of proteins is present at http://homepage.ufp.pt/pedros/HCA_db/index.htm.

We have recently published an investigation of the reaction mechanism of uroporphyrinogen III decarboxylase (UroD), the enzyme whose malfunctioning is the cause of porphyria cutanea tarda, the most prevalent form of porphyria. Uroporphyrinogen III decarboxylase catalyzes the fifth step in heme biosynthesis: the elimination of carboxyl groups from the four acetate side chains of uroporphyrinogen III to yield coproporphyrinogen III. The enzyme acts by successively protonating each of the four pyrrole rings present in the substrate, thereby allowing decarboxylation of their side chains, but the identity of the proton donors has not been established yet. Tyr164 has been suggested as a proton donor, and Asp86 has been proposed to act either as a proton donor or as an intermediate-stabilizing residue. We have performed density-functional calculations to study this reaction mechanism, and found that the rate-limiting step is substrate protonation, rather than decarboxylation. Surprisingly, whereas Tyr164 is unable to protonate the substrate, this protonation can be effected by a nearby arginine residue (Arg37), with a free energy barrier of 21.4 kcalÂ·mol-1, in remarkable agreement with the experimental value of 19.5 kcalÂ·mol-1. The central positioning of this residue in close proximity to all four pyrrole rings in the substrate may play a key role in the sequential activation of each of these moieties.

Earliest research (before 2005)
Prostaglandin H synthase (cyclooxygenase)

Fuller version, with interactive models of every intermediate in the reaction cycle of cyclooxygenase
Prostaglandin H synthase catalyzes the oxygenation of arachidonic acid into the cyclic endoperoxide, prostaglandin G2 (PGG2), and the subsequent reduction of PGG2 to the corresponding alcohol, prostaglandin H₂ (PGH₂), the precursor of all prostaglandins and thromboxanes. Both radical abstraction by a neighbouring tyrosyl radical and combined radical/carbocationic models have been proposed to explain the cyclooxygenase part of this reaction. We have used DFT calculations to study the mechanism of the formation of the cyclooxygenated product PGG2. We found an activation free energy for the initial hydrogen abstraction by the tyrosine radical of 15.6 kcal/mol, and of 14.5 kcal/mol for peroxo bridge formation, in remarkable agreement with the experimental value of 15.0 kcal/mol, whereas A combined radical/carbocation mechanism proceeding through a sigmatropic H-shift has a much larger activation free energy of 36.5 kcal/mol. Our study of prostaglandin H synthase (cyclooxygenase) has definitely shown the mechanism to be fully radical-based, rather than including the intermediate formation of carbocation species.

Spectroscopic investigations of metalloproteins

The syntrophic propionate-oxidizing bacterium Syntrophobacter fumaroxidans contains two formate dehydrogenases (CO2-reductases) (FDH-1 and FDH-2). Both enzymes were produced in axenic fumarate-grown cells as well as in cells which were grown syntrophically on propionate with Methanospirillum hungatei as the H₂ and formate scavenger. The purified enzymes exhibited extremely high formate-oxidation and CO₂-reduction rates, and low Km values for formate. For the enzyme designated FDH-1, a specific formate oxidation rate of 700 U.mg-1 and a Km for formate of 0.04 mm were measured when benzyl viologen was used as an artificial electron acceptor. The enzyme designated FDH-2 oxidized formate with a specific activity of 2700 U.mg-1 and a Km of 0.01 mm for formate with benzyl viologen as electron acceptor. The specific CO₂-reduction (to formate) rates measured for FDH-1 and FDH-2, using dithionite-reduced methyl viologen as the electron donor, were 900 U.mg-1 and 89 U.mg-1, respectively. Both enzymes contained tungsten and selenium, while molybdenum was not detected. EPR spectroscopy suggested that FDH-1 contains at least four [2Fe-2S] clusters per molecule and additionally paramagnetically coupled [4Fe-4S] clusters. FDH-2 contains at least two [4Fe-4S] clusters per molecule.⁵

The membrane-bound nitrate reductase (NaR) from Pseudomonas chlororaphis DSM 50135 contains, like othe membrane-bound NaRs, three subunits, of 129, 66 (64) and 24 kDa, referred to in the literature as α, β and γ, respectively. Electrocatalytic studies revealed that only the membrane-bound, not the solubilized form of the enzyme, can accept electrons from a menaquinone analog, menadione, whereas both forms can accept electrons from methylviologen. The isolated enzyme possesses several iron–sulfur clusters and a molybdopterin guanine dinucleotide active center. The iron–sulfur clusters can be grouped in two classes according to their redox properties, the high-potential and low-potential clusters. In the as-isolated enzyme, two forms of the molybdenum center, high- and low-pH, are detectable by electron paramagnetic resonance spectroscopy. The low-pH form shows a hyperfine splitting due to a proton, suggesting the presence of an –OH_x ligand. Dithionite reduces the Mo(V) center to Mo(IV) and subsequent reoxidization with nitrate originates a new Mo(V) signal, identical to the oxidized low-pH form but lacking its characteristic hyperfine splitting. The isolated preparation also contains heme c (in a sub-stoichiometric amount) with the ability to relay electrons to the molybdenum center, suggesting that this nitrate reductase may contain heme c instead of the heme b usually found in this class of enzymes.⁷.

PhD thesis: The role of hydrogen in the metabolism of Pyrococcus furiosus

[HTML summary] PhD thesis [PDF] [Table of contents]
P. furiosus is a strictly anaerobic Archaeon that grows near 100ºC by fermentation of carbohydrates through a modified Embden-Meyerhoff pathway, producing acetate, alanine, CO₂ and H₂. Two soluble H₂-producing enzymes are known in this organism: sulfhydrogenase I and sulfhydrogenase II. During metabolism of carbohydrates and amino acids, reducing equivalents are transferred to an iron-sulfur protein, ferredoxin. The oxidized form of ferredoxin must be regenerated in order to sustain the continuing operation of these metabolic pathways. However, the sulfhydrogenases from P. furiosus cannot use reduced ferredoxin as an electron donor, which prevents them from directly disposing of the reducing equivalents produced by the metabolism of P. furiosus. Instead, they are able to oxidize NADPH to NADP⁺ with concomitant production of H₂. It was therefore suggested that these reducing equivalents may be transferred to NADP⁺ (thus generating NADPH) by a sulfide dehydrogenase, which can also function as a ferredoxin:NADP⁺ oxidoreductase . The produced NADPH will be used to reduce hydrogenase. It has been shown that the reduction, catalyzed by enzymes, of protons to H₂ by pyridine nucleotides is possible only when the partial pressure of H₂ is kept below 0.1 kPa. P. furiosus is however able to grow under H₂ accumulating conditions (pH₂ > 30 kPa). In order to understand the pathway of electron disposal in P. furiosus each of the enzymes involved (sulfide dehydrogenase and sulfhydrogenase) was purified and characterized. The physiological activities of these proteins in cell extracts were also investigated. The results of this study are succintly described here, and in my PhD thesis. (A table of contents is also available)

SCIENTIFIC PUBLICATIONS:
[1] Pedro J. Silva, Baltazar de Castro and Wilfred R. Hagen (1999) On the prosthetic groups of the NiFe sulfhydrogenase from Pyrococcus furiosus: topology, structure, and temperature-dependent redox chemistry. Journal of Biological Inorganic Chemistry, 4, 284-291.
[2] Pedro J. Silva, M. João Amorim, Peter-Leon Hagedoorn, Hans Wassink, Huub Haaker and Wilfred R. Hagen (1999) Effects of temperature on the electron transfer between Pyrococcus furiosus hydrogenase and its redox partners. Journal of Inorganic Biochemistry, 74, 297.
[3] Wilfred R. Hagen, Pedro J. Silva, M.A. Amorim, Peter-Leon Hagedoorn, Hans Wassink, Huub Haaker and Frank T. Robb (2000) Novel structure and redox chemistry of the prosthetic groups of the iron-sulfur flavoprotein sulfide dehydrogenase from Pyrococcus furiosus; evidence for a [2Fe-2S] cluster with Asp(Cys) ₃ ligands. Journal of Biological Inorganic Chemistry, 5, 527-534.
[4] Pedro J. Silva, Eyke C. D. van den Ban, Hans Wassink, Huub Haaker, Baltazar de Castro, Frank T. Robb and Wilfred R. Hagen (2000) Enzymes of hydrogen metabolism in Pyrococcus furiosus. European Journal of Biochemistry, 267, 6541-6551.
[5] Frank A.M. de Bok, Peter-Leon Hagedoorn, Pedro J. Silva, Wilfred R. Hagen, Emile Schiltz, Kathrin Fritsche and Alfons J.M. Stams (2003) Two W-containing formate dehydrogenases (CO₂-reductases) involved in syntrophic propionate oxidation by Syntrophobacter fumaroxidans. European Journal of Biochemistry, 270, 2476-2485 .
[6] Pedro J. Silva, Pedro A. Fernandes and Maria J. Ramos (2003) A theoretical study of radical-only and combined radical/carbocationic mechanisms of arachidonic acid cyclooxygenation by prostaglandin H synthase. Theoretical Chemistry Accounts, 110, 345-351.
[7] Dora Pinho, Stéphane Besson, Pedro J. Silva, Baltazar de Castro e Isabel Moura (2005) Isolation and spectroscopic characterization of the membrane-bound nitrate reductase from Pseudomonas chlororaphis DSM 50135. Biochimica et Biophysica Acta - General Subjects, 1723, 151-162.
[8] Pedro J. Silva and Maria J. Ramos (2005) A Density-Functional Study of Mechanisms for the Cofactor-Free Decarboxylation Performed by Uroporphyrinogen III Decarboxylase. Journal of Physical Chemistry B, 109, 18195-18200
[9] Pedro J. Silva and Maria J. Ramos (2007) Reaction Mechanism of the Vitamin K-Dependent Glutamate Carboxylase: A Computational Study. J. Phys. Chem. B, 111, 12883-12887
[10] Pedro J. Silva (2008) Assessing the reliability of sequence similarities detected through hydrophobic sequence analysis. Proteins: Structure, Function and Bioinformatics, 70, 1588-1594
[11] van Haaster DJ, Silva PJ, Hagedoorn PL, Jongejan JA, Hagen WR. (2008) A re-investigation of the steady-state kinetics and the physiological function of the soluble NiFe-hydrogenase-I of Pyrococcus furiosus. J. Bacteriol., 190, 1584-1587
[12] Pedro J. Silva and Maria J. Ramos (2008) A comparative density-functional study of models for the reaction mechanism of uroporphyrinogen III synthase. J. Phys. Chem. B, 112, 3144-3148.
[13] Pedro J. Silva and Maria J. Ramos (2008) A comparative density-functional study of the reaction mechanism of the O₂-dependent coproporphyrinogen III oxidase. Bioorg. Med. Chem., 16, 2726-2733.
[14] Pedro J. Silva (2009) Inductive and Resonance Effects on the Acidities of Phenol, Enols, and Carbonyl α-Hydrogens. J. Org. Chem, 74, 914-916
[15] Pedro J. Silva and Maria J. Ramos (2009) Computational studies on the reactivity of substituted 1,2-dihydro-1,2-azaborines. J. Org. Chem, 74, 6120-6129 [free reprint]
[16] Pedro J. Silva and Maria J. Ramos (2011) Computational insights into the photochemical step of the reaction catalyzed by protochlorophylide oxidoreductase. Int. J. Quantum Chem., 111, 1472-1479
[17] Pedro J. Silva, Claudia Schulz, Martina Jahn, Dieter Jahn and Maria J. Ramos (2010) A Tale Of Two Acids: When Arginine Is A More Appropriate Acid Than H₃O⁺ J. Phys. Chem. B., 114, 8994–9001 [free reprint]
[18] Pedro J. Silva and Maria J. Ramos (2011) Computational Characterization of the Substrate-Binding Mode in Coproporphyrinogen III Oxidase J. Phys. Chem. B, 115, 1903-1910[free reprint]
[19] Pedro J. Silva and Maria J. Ramos (2011) Successes and Failures of DFT Functionals in acid/base and redox reactions of organic and biochemical interest. Computational and Theoretical Chemistry, 966, 120-126
[20] Natércia F. Brás, Marta A. S. Perez, Pedro A. Fernandes,Pedro J. Silva and Maria J. Ramos (2011) Accuracy of Density Functionals in the Prediction of Electronic Proton Affinities of Amino Acid Side Chains Journal of Chemical Theory and Computation, 7, 3898-3908
[21] Pedro J. Silva, Marta A. S. Perez, Natércia F. Brás, Pedro A. Fernandes and Maria J. Ramos (2012) Improving the study of proton transfers between amino acid sidechains in solution: choosing appropriate DFT functionals and avoiding hidden pitfalls. Theoretical Chemistry Accounts, 131, 1179-1185
[22] Pedro J. Silva (2012) Unravelling the reaction mechanism of the reductive ring contraction of 1,2-pyridazines. J. Org. Chem., 77, 4653-4659
[23] Carla Sousa and Pedro J. Silva (2013) BBr₃-Assisted Cleavage of Most Ethers Does Not Follow the Commonly Assumed Mechanism. Eur. J. Org. Chem., 2013, 5195-5199. Correction in Eur. J. Org. Chem., 2013, 8048
[24] Carlos E. P. Bernardo, Nicholas P. Baumann, Piotr Piecuch and Pedro J. Silva (2013) Evaluation of density functional methods on the geometric and energetic descriptions of species involved in Cu⁺-promoted catalysis. J. Mol. Model., 19, 5457-5467
[25] Carlos E. P. Bernardo and Pedro J. Silva (2014) Computational development of rubromycin-based lead compounds for HIV-1 reverse transcriptase inhibition PeerJ, 2, e470 [PDF]
[26] Pedro J. Silva (2014) With or without light: comparing the reaction mechanism of dark-operative protochlorophyllide oxidoreductase with the energetic requirements of the light-dependent protochlorophyllide oxidoreductase PeerJ, 2, e551 [PDF]
[27] Pedro J. Silva and Viviana Rodrigues (2015), Mechanistic pathways of mercury removal from the organomercurial lyase active site. PeerJ 3:e1127 [PDF]
[28] Carlos E. P. Bernardo and Pedro J. Silva (2016), Computational exploration of the reaction mechanism of the Cu⁺-catalysed synthesis of indoles from N-aryl enaminones Royal Society Open Science 3:150582 [PDF]
[29] Pedro J. Silva (2016), Will 1,2-dihydro-1,2-azaborine-based drugs resist metabolism by cytochrome P450 compound I? PeerJ 4:e2299 [PDF]
[30] Pedro J. Silva (2016), Refining the reaction mechanism of O₂ towards its substrate in cofactor-free dioxygenases PeerJ 4:e2805 [PDF]
[31] Pedro J. Silva and Carlos E. P. Bernardo (2018), Influence of Alkyne and Azide Substituents on the Choice of the Reaction Mechanism of the Cu⁺-Catalyzed Addition of Azides to Iodoalkynes J. Phys. Chem. A    [free reprint upon registration]
[32]Pedro J. Silva (2020), Chemiosmotic misundertandings Biophys. Chem 264, 106424
[33] Sandra V.R.L. Silva and Pedro J. Silva (2020) Computational improvement of small-molecule inhibitors of Bacillus anthracis protective antigen activation through isostere-based substitutions. Journal of Biomolecular Structure and Dynamics DOI:10.1080/07391102.2020.1792987 FREE reprint
[34] Pedro J. Silva (2020) New insights into the mechanism of Schiff base synthesis from aromatic amines in the absence of acid catalyst or polar solvents PeerJ Organic Chemistry, accepted
[35] Pedro J. Silva (2020) Response to "Molecular-level understanding of biological energy coupling and transduction"  Biophysical Chemistry, accepted
[36] Pedro J. Silva and Carlos E. P. Bernardo (2021) Unveiling the reaction mechanism of the Das/Chechik/Marek synthesis of stereodefined quaternary carbon centers. Applied Sciences, 11: 5002 [PDF]
[37] Pedro J. Silva and Qi Cheng (2022) An Alternative Proposal for the Reaction Mechanism of Light-Dependent Protochlorophyllide Oxidoreductase ACS Catalysis ,12:2589-2605 [PDF]
[38] Pedro J. Silva (2022) Computational development of inhibitors of plasmid-borne bacterial dihydrofolate reductase Antibiotics, 11: 779 [PDF]
[39]Pedro J. Silva , Maria Osswald-Claro e Rosário Castro Mendonça (2022) How to tune the absorption spectrum of chlorophylls to enable better use of the available solar spectrum PeerJ Physical Chemistry, 4:e26 [PDF]
[40] Pedro J. Silva (2023) The reaction mechanism of the Cu(I) catalyzed alkylation of hetero-substituted alkynes. Catalysts, 13:17 [PDF]

Accuracy of Density Functionals in the Prediction of Electronic Proton Affinities of Amino Acid Side Chains	Successes and Failures of DFT Functionals in acid/base and redox reactions of organic and biochemical interest.
Journal of Chemical Theory and Computation, 7, 3898-3908	Computational and Theoretical Chemistry, 966, 120-126
We have analyzed how well DFT functionals, often used to characterize complex and large models such as proteins, describe the zero-point-exclusive proton affinity at 0 K, PA_el^0K, for the ionizable side chains of lysine (Lys), histidine (His), arginine (Arg), and aspartate (Asp–) as well as the cysteine (Cys–), serine (Ser–), and tyrosine (Tyr–) anions. The reference values PA_el^0K were determined at the very accurate CCSD(T)/CBS level. Those values were obtained by the sum of the complete basis set limit of the MP2 energies plus a CCSD(T) correction term evaluated with the aug-cc-pVTZ basis set. The complete basis set limit of MP2 energies was determined using the Truhlar and Helgaker extrapolation schemes. A new, important, and consistent DFT benchmarking database for PA_el^0K and for proton transfer between two different ionizable side chains, ΔPA_el^0K, is provided, making this work relevant to all studies with ionizable amino acids side chains that use DFT.	The performance of 18 different DFT functionals in the prediction of absolute and relative energies of organic and biochemical acid/base and redox reactions was evaluated, using MP2 extrapolated to the complete basis set limit and CCSD(T)/aug-cc-pVTZ energies as benchmark. In the acid/base reaction test set, several hybrid-GGA and meta-hybrid GGA functionals (BHHLYP, B97-1, B97-2, X3LYP, M06-2X, as well as the popular B3LYP) yielded small errors (0.8-1.5 kcal.mol^-1) in the computation of relative energies across the tested model reactions. For most of the proton transfers between different acid/base pairs (or electron transfer between redox pairs) tested we could find at least one functional with very high accuracy (error <0.2 kcal.mol^-1). The reactions involving electron transfer between quinone (or disulfide) and other redox groups stand out as the clearest example of the shortcomings of DFT methods, as the best functionals are most often wrong by 1-3 kcal.mol^-1.
Computational Characterization of the Substrate-Binding Mode in Coproporphyrinogen III Oxidase	A Tale Of Two Acids: When Arginine Is A More Appropriate Acid Than H₃O⁺
J. Phys. Chem. B, 115, 1903-1910[free reprint]	J. Phys. Chem. B, 114, 8994–9001 [free reprint]
The molecular dynamics simulations described in this paper enabled the determination of a very promising substrate binding mode and the extensive characterization of the enzyme active site. The proposed binding mode is fully consistent with the known selectivity of the active site toward substituted tetrapyrroles and explains the lack of activity of the H131A, R135A, D274A, and R275A mutants and the reasons behind the nonoccurrence of catalysis on the C and D rings of the tetrapyrrole. An important role in this binding mode is fulfilled by G276, as its carbonyl oxygen intervenes in the substrate anchoring by hydrogen bonding its ring D pyrrole NH group. The presence of this interaction (which is only possible with the protonated NH pyrrole group) and the absence of positively charged side chains close to the pyrrole nitrogen (which might stabilize the N-deprotonated pyrrole postulated in some mechanistic proposals) show that the pyrrole ring is very unlikely to undergo deprotonation during the catalytic cycle and allow the discrimination between the previously postulated mechanistic proposals.	We have previously found that the rate-limiting step of uroporphyrinogen III decarboxylase is substrate protonation rather than the decarboxylation reaction. This protonation can be effected by an arginine residue (Arg37) in close proximity to the substrate. In this report we present evidence for the function of this arginine residue as a general acid catalyst. Although substrate protonation by H3O+ is both exergonic and very fast, our density functional calculations show that in the presence of a protonated Arg37 substrate decarboxylation becomes rate-limiting and the substrate spontaneously breaks upon protonation. These results suggest that the active site must be shielded from solvent protons. Consequently, H3O+ can be excluded from a role in both protonations proposed for the enzyme mechanism. In agreement with these conclusions a second arginine residue (Arg41) is uniquely positioned to act as donor of the second proton, with an activation barrier below 2 kcal mol-1. Generated mutant uroporphyrinogen III decarboxylase variants carrying amino acid exchanges in the position of both arginine residues (R41A, R41K, R37A and R37K) failed to produce coproporphyrinogen III.
Computational insights into the photochemical step of the reaction catalyzed by protochlorophylide oxidoreductase.	Computational studies on the reactivity of substituted 1,2-dihydro-1,2-azaborines.
Int. J. Quantum Chem. , 111, 1472-1479 [PDF]	J. Org. Chem, 74, 6120-6129 [free reprint]
The light-dependent enzyme protochlorophyllide oxidoreductase catalyzes the conversion of protochlorophyllide into chlorophyllide, during chlorophyll synthesis. The reaction has been proposed to proceed through light-induced weakening of the C17-C18 double bond in protochlorophyllide, which then facilitates hydride transfer from a NADPH co-substrate molecule. We have performed DFT and TDDFT computations on the reaction mechanism of this interesting enzyme. The results show that whereas in the ground state the reaction is strongly endergonic and has a very high activation free energy (38 kcal.mol-1), the first four excited states (corresponding to excitations within the conjugated porphyrin π-system) afford much lower activation free energies (<25 kcal.mol-1) and spontaneous (or only slightly endergonic) reaction paths. The sharp shape of the potential energy surface along the reaction coordinate in these excited states allows hydrogen tunneling to occur efficiently on the first excited state surface, lowering the barrier to values closer to experiment, in agreement with recent suggestions.	We have investigated important intermediates of electrophilic aromatic substitution reactions and one-electron oxidation of substituted 1,2-dihydro-1,2-azaborines with density-functional theory. The results show that electrophilic substitution reactions and one-electron oxidation of substituted 1,2-azaborines are generally much more favorable than those of the corresponding benzene derivatives. Both chlorination and nitration of several boron-unsubstituted 1,2-azaborines are expected to break the boron-hydrogen bond, yielding boron-chlorinated 1,2-azaborines and a novel class of boron-bound 1,2-azaborinyl nitrites, respectively. Comparison between the relative stabilities of C3-bound and C5-bound Wheland intermediates of different electrophilic substitution reactions of 1,2-azaborines further suggests that the preference of the C3- over C5-substitution decreases with decreasing electrophilicity of the attacking group.
Inductive and Resonance Effects on the Acidities of Phenol, Enols, and Carbonyl α-Hydrogens.	Comparative Density Functional Study of Models for the Reaction Mechanism of Uroporphyrinogen III Synthase
J. Org. Chem., 74, 914-916 [PDF]	J. Phys. Chem. B, 112, 3144-3148 [PDF]
The increased acidity of phenols and carboxylic acids over aliphatic alcohols has traditionally been attributed to resonance stabilization of their (deprotonated) anions. A competing explanation argues that this increase is rather due to destabilization of the acidic proton due to electrostatic effects of the adjacent C=O (in carboxylic acids) or phenyl ring (in phenols). A lively debate on the merits of this proposal has been carried out in the literature. Recent work has shown that in carboxylic acids inductive effects are responsible for 2/3 of the stabilization of the deprotonated anion, but for the case of phenol no solution to the controversy has been forthcoming. Meanwhile, the controversy regarding the origin of the increased acidity of phenols vs. aliphatic alcohols has entered the undergraduate education textbooks, with some arguing for the traditional view, whereas others downplay anion resonance stabilization. This study establishes that inductive effects, while more important than traditionally recognized, do not explain the majority of the observed acidity enhancement in phenol, enols or carbonyl α–hydrogens. Inductive effects account for 1/3 of the enhanced acidity of phenol vs. cyclohexanol, 2/5 of the enhanced acidity of enol vs. methanol, and l/4 of the enhanced acidity of carbonyl α-hydrogens vs. methane.	Uroporphyrinogen III synthase catalyzes the inversion of one of the four heterocyclic rings present in the hydroxymethylbilane. Two mechanisms have been proposed to explain this puzzling ring inversion, either through sigmatropic shifts or through the direct formation of a spirocyclic pyrrolenine intermediate. We performed the first high-level quantum mechanical calculations on model systems of this enzyme to analyze these contrasting reaction mechanisms. The results allow us to discard the sigmatropic shift mechanism and suggest that the D-ring of the hydroxymethylbilane substrate binds to the enzyme in a conformation that shields its terminal portion from reacting with ring A and prevents the formation of the biologically useless uroporphyrinogen I, whose accumulation (in individuals lacking functional uroporphyrinogen III synthase) leads to severe cutaneous dermatosis.

A comparative density-functional study of the reaction mechanism of the O₂-dependent coproporphyrinogen III oxidase.	Reaction Mechanism of the Vitamin K-Dependent Glutamate Carboxylase: A Computational Study.
Bioorg. Med. Chem., 16, 2726-2733, [PDF]	J. Phys. Chem. B, 111, 12883-12887 [PDF]
During heme biosynthesis, coproporphyrinogen III oxidase catalyzes the conversion of two propionate substituents from the highly reactive substrate coproporphyrinogen III into vinyl substituents, yielding protoporphyrinogen IX. We have performed DFT calculations on model systems in order to analyze several reaction mechanisms proposed for this enzyme. The results afford a full description of the different proposals and allow the rejection of a direct electron abstraction from the protonated substrate by dioxygen. We found that O₂ addition to the (preferentially deprotonated) pyrrole substrate (yielding a hydroperoxide, which then abstracts a proton from the reactive propionate substituent) is compatible with the observed experimental reaction rate, and that the reaction may then proceed through HO₂^- elimination, followed by decarboxylation.	In the reaction cycle of glutamate carboxylase, vitamin K epoxidation by O₂ has been proposed to generate a very strong base able to remove a proton from the carbon of a Glu residue, thus yielding a Glu-based carbanion that readily reacts with CO₂. We have used hybrid density functional theory to study this appealing mechanism. Our calculations show a very exergonic four-step mechanism with the reaction of (triplet) O₂ with the singlet vitamin K anion as the rate-limiting step, with a rate similar to the experimental value. Our study also establishes the need to apply continuum models when performing the optimization of minimum-energy crossing points between potential energy surfaces of different multiplicities for enzyme model systems.
Assessing the reliability of sequence similarities detected through hydrophobic sequence analysis	A Density-Functional Study of Mechanisms for the Cofactor-Free Decarboxylation Performed by Uroporphyrinogen III Decarboxylase.
Proteins: Structure, Function and Bioinformatics, 70, 1588-1594 [PDF]	Journal of Physical Chemistry B, 109, 18195-18200 [PDF]
Hydrophobic cluster analysis (HCA) has long been used as a tool to detect distant homologies between protein sequences, and to classify them into different folds. However, it relies on expert human intervention, and is sensitive to subjective interpretations of pattern similarities. I developed a novel algorithm to assess the similarity of hydrophobic amino acid distributions between two sequences. This algorithm correctly identifies as misattributions several HCA-based proposals of structural similarity between unrelated proteins present in the literature. Besides enabling a reliable identification of the correct fold of an unknown sequence and the choice of suitable templates, the new algorithm also shows that whereas most structural classes of proteins are very homogeneous in hydrophobic cluster composition, a tenth of the described families are compatible with a large variety of hydrophobic patterns. A browsable database of every major representative hydrophobic cluster pattern present in each structural class of proteins is present at http://homepage.ufp.pt/pedros/HCA_db/index.htm.	We have recently published an investigation of the reaction mechanism of uroporphyrinogen III decarboxylase (UroD), the enzyme whose malfunctioning is the cause of porphyria cutanea tarda, the most prevalent form of porphyria. Uroporphyrinogen III decarboxylase catalyzes the fifth step in heme biosynthesis: the elimination of carboxyl groups from the four acetate side chains of uroporphyrinogen III to yield coproporphyrinogen III. The enzyme acts by successively protonating each of the four pyrrole rings present in the substrate, thereby allowing decarboxylation of their side chains, but the identity of the proton donors has not been established yet. Tyr164 has been suggested as a proton donor, and Asp86 has been proposed to act either as a proton donor or as an intermediate-stabilizing residue. We have performed density-functional calculations to study this reaction mechanism, and found that the rate-limiting step is substrate protonation, rather than decarboxylation. Surprisingly, whereas Tyr164 is unable to protonate the substrate, this protonation can be effected by a nearby arginine residue (Arg37), with a free energy barrier of 21.4 kcalÂ·mol-1, in remarkable agreement with the experimental value of 19.5 kcalÂ·mol-1. The central positioning of this residue in close proximity to all four pyrrole rings in the substrate may play a key role in the sequential activation of each of these moieties.