A video of our recent trip to the Pohang Accelerator Laboratory in Korea to shoot our protein crystals. The facility is great and the staff is very helpful. Great synchrotron facility!

We had such a great family time in Japan for both Christmas and new-year celebrations. We visited the beautiful Nara and miyajima (Itsukushima) along with Osaka, Okayama and Hiroshima with the peace park museum. Here a video of our trip.
 

Last year, we published in BMC bioinformatics a novel indicator for validating homology models of proteins namely the “H-factor”. I conceived the H-factor while working in Patrice Koehl’s lab at the University of California at Davis (UC Davis). The PDF can be downloaded here.

I am glad to report that Patrice and I, just submitted a new manuscript on the H-factor. I believe our bioinformatic tool has great potentials for helping biologists validating their models. In addition, the H-factor has great evolution capabilities to further increase its accuracy.

In a nutshell, the H-factor combines information of four scoring functions that evaluates (1) the template structure(s) (based on the corresponding PDB files);  (2) the sequence alignment between the template(s) and the target sequences; (3) the structural heterogeneity of the models built; and (4) the structural neighborhood within protein families.

The H-factor is meant to mimic the R-factor used in X-ray crystallography. We have developed a web service for computing the H-factor for models of a protein structure. This service is freely accessible at http://koehllab.genomecenter.ucdavis.edu/toolkit/h-factor

Lab projects are moving forward and I am really excited about that. Briefly, we are currently working lysine- and arginine-HMTases along with collaborative projects on membrane proteins. The primary objective of our projects is to understand the structure-function relationship of epigenetic modifiers and to perform a structure-based drug design to develop specific and selective novel inhibitors/modulators. A side project, involve the development of a novel indicator for validating homology models in bioinformatics.

Currently, we are working on four human HMTase (NSD1, NSD2/MMSET/WHSC1, and NSD3/WHSC1L1 and PRMT5) and thirteen Yeast HMTHases from both Saccharomyces cerevisiae and Schizosaccharomyces pombe. We have data we found exciting about the NSD family that we hope to it publish soon. In addition, we have been working on NSD inhibitors for a while. We finally overcame experimental difficulties regarding inhibitory enzymatic assays and we are working on a the second subset of molecules we identified that target specifically the histone binding area of the NSD’s SET domain. While the first subset of inhibitors (LEM-01 to LEM-05) did inhibit the NSDs at values I am not disclosing here, the molecules themselves weren’t drug-lead friendly. Exploiting the knowledge we gathered on the SET domain of the NSDs, we refined our screening and selection process to focus specifically on the inhibition of NSD2/MMSET/WHSC1. We are currently testing a second subset of molecules (namely LEM-06 to LEM-10) that are more drug-lead friendly. LEM-06 to LEM-10 do not violate the empirical Lipinski’s rule of five, have a wcLogP < 3 and a total polar surface area <90 A**2.

Here a video of our trip to New Caledonia in July. New Caledonia is a gorgeous island (French territory) located in the southwest Pacific Ocean, 1,210 kilometers east of Australia and 16,136 kilometers east of Metropolitan France. We had a great family time.

My lab along other groups have previously emphasized the role of NSD proteins (NSD1, NSD2/MMSET/WHSC1, and NSD3/WHSC1L1)  as oncogenes. A growing number of studies link the NSD proteins to a variety of cancers. NSD1 is associated with acute myeloid leukemia, multiple myeloma, and lung cancer; NSD2 with the prostate cancer and multiple myeloma; NSD3 with both lung and breasts cancers along with the acute myeloid leukemia. NSD1-NUP98 translocation is associated with childhood acute myeloid leukemia with the NUP98-NSD1 fusion protein being an active H3K36 methylase. NSD1 is amplified in multiple myeloma, lung cancer, neuroblastomas and glioblastomas NSD2 has been found associated with the prostate cancer and multiple myeloma. Furthermore, the amplification of either NSD1 or NSD2 trigger the cellular transformation, initiating carcinogenesis events . Increased NSD2 activity was reported in the tumor proliferation in glioblastoma multiforms. Overexpression of NSD2 in myeloma cells leads to aberrantly high global levels of H3K36 di-methylation, accompanied by a decrease in levels of H3K27 methylation. In myeloma cells, NSD2 contributes to disrupt the chromatin structure and function contributing to the cellular transformation. In addition, NSD2 is found overexpressed in 15 different cancers and is associated with tumor aggressiveness or prognosis in most types of cancers. NSD3 is found amplified in breast cancer cell lines and primary breast carcinomas. Moreover, NSD3 is involved in lung cancer and the acute myeloid leukemia where NSD3 is fused with NUP98, similarly as NSD1.

Reducing NSDs activity through specific lysine-HMTase inhibitors appears promising to help suppressing cancer growth.

Over the last few months, we have been doing extensive virtually ligand screening of thousands of selected compounds along with a large number of HMTase assays. Here a short video of the top 10 molecules that tightly bind (inhibit) the histone lysine binding pocket of the SET domain of NSD2/MMSET/WHSC1, calculated in an “open-conformation (able to dock the histone lysine H3K36).

Here a video of my lovely baby girl Luna (9 months).

My lab is currently focused on epigenetic and cancers. Most particularly, we are interested in understanding the histone code and its implications in diseases, especially cancers. It is a really fascinating research areas.

Cancer initiation and progression are controlled by both genetic and epigenetic events. Both genetic and epigenetic alterations of transcriptional co-regulators are key features in carcinogenesis onset with aberrant gene functions and changes in gene expression levels.

Histone modifications along with other epigenetic mechanisms such as DNA methylation maintain gene activity states and are key in regulating a wide range of cellular processes. Alterations and deregulations in the function of enzymes that modify histones alter the array and levels of histone marks and ultimately affect the control of chromatin-based processes. It leads to dramatic changes in gene expression profiles, which eventually contribute to oncogenic transformation and the development of cancer.

Histones are the stage of multiple post-translational modifications. Specific residues on histones H2A / H2B, H3 and H4 can be modified by methylation (Lysine / Arginine), acetylation (Lysine), citrullination (Arginine), phosphorylation (Serine / Threonine), ubiquitination (Lysine), sumoylation (Lysine), ADP-ribosylation (Lysine), butyrylation (Lysine), propionylation (Lysine) and glycosylation (Serine / Threonine).

Amongst the array of covalent histone modifications, lysine methylation is one of the prominent signaling pathway in chromatin-regulatory mechanism. Lysine-histone methyltransferases (HMTases) are transcriptional co-regulators that target specific lysines on H3 and H4, and can transfer up to three methyl groups (Kme1, Kme2, and Kme3) on histone tails. Lysine methylation, or any of the other histone modifications, can have both activating and repressive functions on transcription events. All the covalent histone modifications contribute to finely regulating the diverse activities associated with the chromatin and may be referred as a language of covalent histone modifications or histone code that is still obscure.

One fascinating aspect of the regulation of the transcription lies in the ballet between histones modifiers “readers” and “writers”, both being regulated leading to various physiological output based on the cellular context. This appends complexity to our current limited understanding of gene transcription and its implication in human diseases.

I had great time attending the 2011 edition of the America Association for Cancer research satellite meeting “Molecular Targets and Cancer Therapeutics 2011″ held in San-Francisco. I am very pleased with the overall quality of the meeting and the data presented. It was a fantastic opportunity to expand my background, present our research and connect with various individuals for future collaborative work. Several companies have shown great interest in our work especially regarding our current studies on histone methyl transferase and their specific and selective inhibitors. The poster I presented can be downloaded here (PDF). Below, is the abstract submitted to the meeting.

Background: The nuclear receptor binding SET domain (NSD) protein is a family of three histone-lysine N-methyltransferase (HMTase), NSD1, NSD2/MMSET/WHSC1, and NSD3/WHSC1L1 that are critical in maintaining the chromatin integrity. NSD1 methylates H3K36 and H4K20 and is associated with acute myeloid leukemia, multiple myeloma, and lung cancer. The NSD1-NUP98 translocation plays a significant role in childhood acute myeloid leukemia with NUP98-NSD1 being an active H3K36 methylase. NSD1 is amplified in multiple myeloma, lung cancer, neuroblastomas and glioblastomas. NSD2 methylates H3K36 and is linked to prostate cancer and multiple myeloma. Over expression of NSD2 in myeloma cells leads to aberrantly high levels of H3K36 di-methylation, accompanied by a decrease in H3K27 methylation. NSD2 is found over expressed in fifteen different cancers and is associated with tumor aggressiveness or prognosis in most types of cancers. NSD3 methylates H3K36 and is associated with both lung and breast cancers along with the acute myeloid leukemia. The amplification of either NSD1 or NSD2 triggers the cellular transformation. NSD3 is found amplified in breast cancer cell lines and primary breast carcinomas. Reducing NSDs activity through specific and selective lysine-HMTase inhibitors appears promising to help suppressing cancer growth.

Little is known about the NSD pathways and our understanding of the histone Lysine-HMTase mechanism is partial. The SET domain of NSD1 has specific mechanisms to recognize histone marks unlike other HMTase. The precise catalytic activities of the NSDs are obscure and discrepancies exist hindering progress in understanding their exact biological functions and pathways in cancer pathogenesis. In this study, we explored the in vitro catalytic activities on histone substrates to understand the substrate recognition and to pave the way for the design of selective and specific NSD inhibitors usable in cancer therapies.

Methods: We used both biochemical and computational methods to understand the substrates recognition by the NSDs and to investigate the structural mechanisms happening in the SET domain during the binding of histone tails.

Results: A key regulatory and a recognition mechanism is driven by the flexibility of a loop at the interface of the SET and postSET region who rotates ~45° and translated 7Å opening the SET domain for the binding of the peptide ligand. This regulatory loop acts as a seat belt for the ligand and contributes to the discrimination and the substrate specificity. In vitro, The SET domain of the NSDs favor H3 recognition and are able to methylate a range of substrate. To reconcile with the in vivo activities previously reported on H3K36 and H4K20, we propose a cross-talk mechanism controlling the substrate recognition.

My lab is having a terrific summer so far. I’m not saying that to make a statement or massage my ego but I am happy to see progress. My two Dutch students (Daan and Lennart) made significant contributions to the projects. With Masayo in maternity leave, I was anxious about slowdowns in the lab. Lennart took some of Masayo duties in managing the day to day lab-life. I must say Lennart is doing a great job. As for myself, I’m stuck in front of my computers writing papers and grant proposals. However, I’m always pleased to discuss results, propose futur experiments, and explain projects anytime with my lab members. My office door is always open, and I like to keep it that way.

A photo of Lennart’s crystal photos of PRMT5, a 72Kda human argine-methyltransferase. This is one of the hits we obtained and we are currently optimizing the crystallization conditions.

Delicate poetry and chef-d’œuvre by Ryan J. Woodward.

Thought of You from Ryan J Woodward on Vimeo.

I won the second place at the art of science image contest during the 2011 annual meeting of the Biophysical Society at Baltimore. Thank you all for voting for my image and congratulations to the winner of the contest, Jorge Bernardino!

My image is the hexameric assembly of the Quinolinate phosphoribosyl transferase enzyme (BNA6) from S. cerevisiae. Quinolinic acid phosphoribosyl transferase (QAPRTase, EC 2.4.2.19) is a 32 kDa enzyme encoded by the BNA6 gene in yeast and catalyzes the  formation of nicotinate mononucleotide from quinolinate and 5-phosphoribosyl-1-pyrophosphate (PRPP). QAPRTase plays a key role in the tryptophan degradation pathway via kynurenine, leading to the de novo biosynthesis of NAD and clearing the neurotoxin quinolinate. The image shows the hexameric surface with electrostatic properties. The structure was solved by APBS and visualized and rendered with VMD 1.8.7.  The electrostatic field lines are displayed and are “emerging” out of the 3 active sites.

From my 2008 paper “Comprehensive x-ray structural studies of the quinolinate phosphoribosyl transferase (BNA6) from Saccharomyces cerevisiae”; E. di Luccio, and D.K. Wilson, (2008) Biochemistry, 47(13); 4039-50. Epub Mar 6, 2008.

Link to Biophysical Society website.

The last 4 months have been really exciting. The work in the lab has been terrific. Thanks to Masayo’s skills, the projects have been really taking off. We have been doing a bunch of cloning and building numerous constructs of various histone methyltransferase (HMTase). In addition, we started working on some putative regulatory partners. One of our goals is to find highly selective and specific HMTase inhibitors, especially for the NSD family. To do so, we are heavily using virtual ligand screening methods and come-up with a short list of molecules to try in the lab. The latest incarnation of AutoDock, AutoDock Vina, have been very useful and efficient to us. Here an animation of the best candidate we found so far, out of over 10,000 molecules of potential interest.

NSD1 SET domain virtual ligand screening

Recently, Jonathan Eisen at UC Davis tweeted a comment made by a structural biologist at the 18th Annual International Meeting on Microbioal Genomics in California. The comment was along this line “if you don’t study purified enzymes in tube it isn’t science”. Although I understand what the structural biologist meant, I was shocked to hear that type of comment in 2010.

Unfortunately, the heated debate between “experimentalists” and bioinformaticians is far from being over. In a nutshell, the biology world can be roughly divided into 3 worlds. The experimentalists doing only wet biology occupy the first category. They trust only what is inside an Eppendorf tube. At the opposite, we found the second category: the bioinformatics world. The bioinformaticians put into equations what the experimentalists found. Then the bioinformaticians try to understand, and model with an ever-increasing accuracy any biological elements. Their work can be a genomic analysis or predicting the 3D structure of a protein among others. The third category of scientist sits in the middle. They believe that using both wet and dry biology is the right combo to better understand everything.

A fringe of hardcore experimentalist does not believe in all the bioinformatics tools and their power to answer some fundamental questions. Of course, that leads to endless debates amongst scientists. *Yawn* The main argument of experimentalists is “A computer program can’t explain or predict a biological phenomena because there are too many unknowns. If you don’t do wet biology/biochemistry in a tube, it isn’t science. ” Well…sure…Last time I checked, there is a descent number of BS analysis coming from experimentalists. Right? Also, there are numbers of wrong analysis using only bioinformatics too. *Facepalm right here*

As for myself, I am a strong advocate of using the best of both worlds. For instance in structural biology, an X-ray structure gives an accurate snapshot of a 3D structure…but that’s it. What about the flexibility during functions? What about the domain motions? How to accurately understand the mechanism of an enzyme? How does the substrate enter the active site? How the product leaves? How does a protein interact with a partner?

There is a huge gap between the world of known structures and the universe of known protein sequences. Structural genomic projects are unable to keep up with newly discovered genes. How to accurately gain access to nearly all of the 3D protein structure of the whole proteome? Those are just few questions among many more.

Bioinformatics offer many great tools to answers all those questions as long as they are properly used and validated. Unfortunately, some scientists don’t agree with that.  Well! This is 2010, people. It is time to embrace the 30+ years of steady developments in both computer science and bioinformatics and to apply them to biology, broadly define. In science, most of the “easy” stuff has been already unravel. What is left? More difficult and complex problems. In order to make significant contributions in Science, one has to cross boundaries with mixed methods approach. Inter-disciplinary is the way to go!

There are many cool things that can be done when mixing wet “classical” biochemistry and computational biology/biophysics. The virtual ligand screening is one of those things that fall into the major cool category. Nowadays computer have plenty of horsepower that can be put into good use to simulate the binding of libraries of small molecules onto an active site of an enzyme for instance. Following the in silico simulations, the *best* molecules are assessed in the lab for their *experimental* binding/inhibitory properties. In the following video, I used AutodockVina to dock a small subset of 943 compounds into the human SETD1 (NSD1): 20 best docking solutions for each compounds = 18860 docking solutions!