Category Archives: Proteins
As a continuation of our previous work on estimating pKa values from DFT calculations for carboxylic acids, we now present the complementary pKb values for amino groups by the same method, and the coupling of both methodologies for predicting the isoelectric point -pI- values of amino acids as a proof of concept.
Analogously to our work on pKa, we now used the Minimum Surface Electrostatic Potentia, VS,min, as a descriptor of the availability of Nitrogen’s lone pair and correlated it with the experimental basicity of a large number of amines, separated into three groups: primary, secondary and tertiary amines.
Interestingly, the correlation coefficient between experimental and calculated pKb values decreases in the following order: primary (R2 = 0.9519) > secondary (R2 = 0.9112) > tertiary (R2 = 0.8172). This could be due to steric effects, the change in s-character of the lone pair or just plain old selection bias. Nevertheless, there is a good correlation between both values and the resulting equations can predict the pKb value of an amino group within less of a unit, which is very good for a statistical method that does not require the calculation of a full thermodynamic cycle.
We then took thirteen amino acids (those without titratable side chains) and calculated simultaneously VS,min and VS,max for the amino and the carboxyl group (this latter with the use of equation 2 from our previous work published in Molecules MDPI) and the arithmetical average of both gave us their corresponding pI values with an agreement of less than one unit.
This work is now available at the Journal of Chemical Information and Modeling (DOI: 10.1021/acs.jcim.9b01173); as always a shoutout is due to the people working on it: Leonardo “Leo” Lugo, Gustavo “Gus” Mondragón and leading the charge Dr. Jacinto Sandoval-Lira.
We’re always happy at the lab when a student defends their dissertation thesis and now it was the turn of Raúl Márquez-Avilés to do so with flying colors.
The title of his dissertation is “Molecular Dynamics Simulations of 5 potential entry inhibitors for HIV-1“. He performed 500 ns long molecular dynamics simulations of the CD4 – gp 120 proteins interacting with one or several molecules of various lead compounds with inhibitory properties. The leads were obtained previously in our group (by Durbis Castillo, now at McGill) from a massive docking library of ca. 16 million compounds, all having a central piperazine core (Fig1)
The protein gp120 is a surface glyco-protein located at the surface of the HIV virus which couples to the CD4 protein on lymphocytes-T, being this the first step in the infection process of a healthy cell; generating inhibitors of this coupling could help stop the infection from spreading systemically. Four systems were devised: (SB) The reference state for which only gp-120 and CD4 were considered, (S2) A single ligand molecule was placed in the Phe43 cavity of gp120 to assess their inhibitory capacity, (S3) the ligand was placed right outside the Phe43 cavity to assess their entry capacity, and (S4) five ligand molecules were placed outside the Phe43 cavity of gp120 to force their entry (Fig2). Their binding energies were calculated using MM-PBSA and although all five ligands show statistically similar results as inhibitors all five exhibit a stronger binding energy than the reference proving their efficacy in preventing the coupling of the virus to the healthy cell. As a bonus, his research on system S4 shed light on the existence of an allosteric site on gp120 that will warrant further research in our group.
This work is still pending publication.
Raúl Márquez has always proven to be a hard working person who is also very self-sufficient student, a very cheerful labmate, and, as I just learned yesterday, an avid chess player. I’m sure he has a bright future in whichever endeavor he chooses now. Congratulations Raúl Márquez-Avilés!
Recently, the journal ACS Central Science asked me to write a viewpoint for their First Reactions section about a research article by Prof. Alán Aspuru-Guzik from Harvard University on the evolution of the Fenna-Matthews-Olson (FMO) complex. It was a very rewarding experience to write this piece since we are very close to having our own work on FMO published as well (stay tuned!). The FMO complex remains a great research opportunity for understanding photosynthesis and thus the origin of life itself.
In said article, Aspuru-Guzik’s team climbed their way up a computationally generated phylogenetic tree for the FMO from different green sulfur bacteria by creating small successive mutations on the protein at a time while also calculating their photochemical properties. The idea is pretty simple and brilliant: perform a series of “educated guesses” on the structure of FMO’s ancestors (there are no fossil records of FMO so this ‘educated guesses’ are the next best thing) and find at what point the photochemistry goes awry. In the end the question is which led the way? did the photochemistry led the way of the evolution of FMO or did the evolution of FMO led to improved photochemistry?
Since both the article and viewpoint are both published as open access by the ACS, I wont take too much space here re-writing the whole thing and will instead exhort you to read them both.
Thanks for doing so!
For the fifth year in a row my research group has participated in this traditional meeting on theoretical and computational chemistry, now at the beautiful city of Merida in southeastern Mexico.
Several distinguished international guests included Profs. Jose Luis Mendoza (Florida State University), Adrián Roitberg (University of Florida), Vincent Ortiz (Auburn University) and Paul Ayers (McMaster U. Canada); Their contributions rounded up nicely those of household names like Drs. Alberto Vela, Gabriel Merino (CINVESTAV) (the latter was also the main organizer), Jesus Hernández-Trujillo (UNAM), Jose Luis Gazquez (UAM-I), Óscar Jimenez (Guanajuato), and so many others who were also present.
My students presented four posters summarized below:
1) Maru Sandoval and Gustavo Mondragón on Photosynthesis, particularly the search for exciton transference mechanisms in both natural and theoretical arrangements of photosynthetic pigments. Some very exciting results have been observed; their publication is really near.
2) Raúl Torres and Gustavo Mondragón presented their work on arsenic removing calixarenes, published earlier this year, and the extension of said work to As(III) acids. Graphene oxide is now considered in our simulations as per the experimental work of our colleagues, Prof. Reyes Sierra and Prof. Eddie Lopez-Honorato.
3) Marco Diaz, Guillermo Caballero, Gustavo Mondragón and Raúl Torres had this poster on the calculation of sigma holes as descriptors for predicting pka values in organic acids. Their +1600 calculations project has found the best levels of theory (and ruled out some like B3LYP, of course) with some nice correlations. Yet, much work is still to be done but we’re on the right track.
4) Durbis Castillo presented his work on molecular docking and dynamics of a large library of HIV-1 entry inhibitors for which he uses the suite MAESTRO as a continuation of another project of ours. His enormous library is now in the hundredths of thousands and although we’re facing some technical difficulties, Durbis is thriving in his search. This is our first serious attempt towards a more mature drug discovery project; a manuscript should be ready in the first part of next year.
This guys and the rest of the lab who weren’t present are the ones who make our research flourish and they’ve all earned a day or two at the beach!
Here’s to fifteen more years of RMFQT!
Thanks to Devang Sachdev from NVIDIA for bringing this webinar to my attention.
The future of computational chemistry seems to be written in CUDA for GPU’s specially when it comes to Molecular Dynamics; as such, NVIDIA has gone through great lengths into introducing scientific computing methods for GPU’s. I still have a pending review of a test drive that people at NVIDIA and EXXACTCORP kindly allowed me to run but that is the topic of the next post.
Next Thursday, April 4th, 2013 from 9:00 AM – 10:00 AM Pacific Standard Time there will be a webinar in which Dr. Erik Lindhal at Stockholm University and NVIDIA will discuss latest GPU-acceleration technologies available to GROMACS users; more specifically the latest accelerated version of GROMACS 4.6, which features are supported, it’s installation and use, and how it performs with latest NVIDIA Kepler GPUs.
Register here: http://goo.gl/0HtqJ
Please register and check your local timezone to avoid delays. I will register as soon as I finish typing this. Thanks once again to Devang Sachdev for all his help, patience and trust in this forum.
For over a decade these meetings have gathered theoretical chemists every year to share and comment their current work and to also give students the opportunity to interact with experienced researchers, some of which in turn were even students of Prof. Robert Parr, Prof. Richard Bader or Prof. Per Olov Löwdin. This year the Mexican Meeting on Theoretical Physical Chemistry took place last weekend in Toluca, where CCIQS is located. You can find links to this and previous meetings here. We participated with a poster which is presented below (in Spanish, sorry) about our current research on the development of calixarenes and tia-calixarenes as drug carriers. In this particular case, we presented our study with the drug IMATINIB (Gleevec as branded by Novartis), a powerful tyrosinkynase inhibitor widely employed in the treatment of Leukaemia.
The International Journal of Quantum Chemistry is dedicating an issue to this reunion. As always, this meeting posed a great opportunity to reconnect with old friends, teachers, and colleagues as well as to make new acquaintances; my favourite session is still the beer session after all the seminars! Kudos to María Eugenia “Maru” Sandoval-Salinas for this poster and the positive response it generated.
What if you could convince people to help you doing your research on their spare time? What if you could convince a million people to contribute to a specific scientific effort without the need of recruiting them yourself? Even better if you can get all these exo-collaborators without making a huge dent in your budget, which sometimes is just impossible even if you are willing to do it. The world is an interconnected global one these days; millions of voices sound through the web which makes it hard for yours to stand out. As an interconnected entity, communicating to a large mass has become feasible but it comes with a price: you need to be attractive! That’s right, you can get a million people to help your scientific efforts, its called crowdsourcing (think about wikipedia for instance), but you have to make it rewarding in some way, and since you are trying to convince them to work for you in their free time you have to make it look like something they’d do on that free time; So why not making it a video game?
There are nowadays some serious games which are nothing more than an internet-based platform in which tons of data are loaded and accessed by many users who analyze them while being scored on their achievements in many different ways according to the rules of each game. Remember that famous NASA screensaver (SETI@home) which used the idle time on your computer to crunch data from the SETI (Search for Extraterrestrial Intelligence) project; that is a more passive example of exo-collaboration simply called distributed computing, since the user has to do nothing but allowing the entrance of data into their computers for further processing.
Fold.it is a videogame (that actually began as a distributed computing screensaver called Rosetta@home) that allows you to play around with a protein and fold it in many different ways while you score points according to the conformer’s plausibility. Obtaining the native tertiary (and even the secondary) structure of a protein from no other information than the primary structure is extremely difficult given the enormous amount of available degrees of freedom. Molecular dynamics alone is unable to predict the native tertiary structure of the protein; the number p of possible disulfide bonds present in a protein is p = n!/[(n/2)!2^n/2] where n is the number of cysteine residues available, plus computers know nothing about proteins or enzymatic catalysis so a hand from us fellow humans and our chemical insight is widely needed. Therefore our previous knowledge of chemistry, biochemistry and the nature of related proteins can help us help those programs in finding the best possible answer to ‘how does this protein look like in 3D space?’ but since this human-helped process is slow and cumbersome you need thousands of people working on it a great deal of time; almost as if every person playing with the same structure was a single core in your computer. Fold.it thus, is a sort of protein self docking, if you will, in which players are ranked according to their skills and rewarded according to how well your structure complies with three simple rules: 1) lack of voids (packing) 2) keeping the orange hydrophobic chains unexposed to the aqueous exterior and 3) avoiding clashes. Scoring functions for these three concepts are calculated and then yield a score for the player which is then ranked to other players folding the same protein (or to other players in their overall performance).
Fold.it has already collected some major success stories such as the one published on Nature Structural & Molecular Biology by David Baker (founder of Fold.it) et al. on September 2011 (doi:10.1038/nsmb.2119) in which players helped in solving the crystal structure of a protease from a retrovirus which causes AIDS in monkeys. The determination of this structure had already taken 15 years of work with only partial success; but the data was available in Fold.it for only three weeks when the appropriate match to the diffraction experiments was found! This case alone has stirred too much attention and for a beautifully written piece about it, you can check this article at the Discover Magazine by Ed Yong.
Other such examples of crowdsourcing in scince, more specifically in astronomy, are Galaxy Zoo and Moon Zoo in which thousands of images from the Hubble telescope and numerous moon probes are made available for users to sort and classify. The aim of Moon Zoo is to study the amount, shape and occurrence of craters, which basically never erode unlike those on Earth. This analysis will let us know more about the origin of our natural satellite and ultimately about the origins of our solar system.
To the participants in this specific kind of scientific crowdsourcing the term Citizen Science is applied and even publications such as the Scientific American magazine host a section where you can call out for volunteers in your projects. Some sort of classified ads for the lonely scientists in their labs in search for “idle” hands that can make a significant contribution to science. Some Citizen Science projects are intended for kids and teenagers as a way to get more people interested in scientific disciplines by engaging them directly in activities with a measurable progress of their own contributions. It is worth mentioning that projects like Fold.it, Moon Zoo and Galaxy Zoo are developed in a way that can be used by people with no expertise in the field in order to recruit as many people as possible just to perform a very specific task, proving thus that the human brain is a powerful and beautiful machine whose insight isn’t equaled by any artificial system, yet.
Well, it is now time to go back to work before I’m deemed a permanent exo-collaborator by my bosses. Just a final thought: What were our mothers saying about us playing too much with our video games?
2011, International Year of Chemistry
As usual please share your thoughts in the comments section, rate this post and let me know that you are out there reading this.
Is the C atom in methane sp3 hybridized because it’s tetrahedral or is it tetrahedral because it’s sp3 hybridized? It’s funny how many students think to this date that the correct answer is the latter; specially those working in inorganic chemistry. I ignore the reason for such trend. What is true is that most chemistry teachers seem to have lost links to certain historical facts that have shaped our scientific discipline; most of those lay in the realm of physics, maybe that’s why.
What Linus Pauling, in a very clever way, stated was that once you have a set of eigenvectors (orbitals) of the atomic Hamiltonian any combination of them will also be an eigenvector (which is normal since one of the properties of Hermitian operators is that they are linear); so why not making a symmetry adapted one? Let’s take the valence hydrogenoid orbitals (hydrogenoid being the keyword here) and construct a linear combination of them, in such a way that the new set transforms under the irreducible representations of a given point group. In the case of methane, the 2s and 2p orbitals comprise the valence set and their symmetry-adapted-linear-combination under the Td point group constitutes a set of new orbitals which now point into the vertexes of a tetrahedron. Funny things arise when we move to the next period of the table; it has been a controversy for a number of years the involvement of empty d orbitals in pentacoordinated P(V) compounds. Some claim that they lay too high in energy to be used in bond formation; while others claim that their involvement depends on the nature (electronegativity mainly) of the surrounding substituents.
In many peer reviewed papers authors are still making the mistake of actually assigning a type of hybridization to set of valence orbitals of an atom based on the bond angles around it. Furthermore, it is not uncommon to find claims of intermediate hybridizations when such angles have values in between those corresponding to the ideal polyhedron. Symmetry is real, orbitals are not; they are just a mathematical representation of the electron density distribution which allows us to construct mind images of a molecule.
Linus Pauling is one of my favourite scientific historical figures. Not only did he build a much needed at the time bridge between physics and chemistry but he also ventured into biochemistry (his model of an alpha-helix for the alanine olygopeptide became the foundation to Watson & Cricks later double helix DNA model), X-ray diffractometry, and humanities (his efforts in reducing/banning the proliferation of nuclear weapons got him the Nobel Peace Prize long after he had already received the Nobel Price in Chemistry). He was a strong believer of ortho-molecular nutrition, suggesting that most illnesses can be related to some sort of malnutrition. Linus Pauling and his book On the Chemical Bond will remain a beacon in our profession for the generations to come.
Disclaimer: The question above, with which I opened this post, was taken from an old lecture by Dr. Raymundo Cea-Olivares at UNAM back in the days when I was an undergraduate student.