The relationship between protein structure and low barrier hydrogen bonds

Jon Askey Author

08/04/2020 Added

35 Plays

Description

Summer 2020 UCARE-funded research project that uses statistical analysis to explore and classify any relationship that exists between protein characteristics and a special type of hydrogen bond that has not been proven to exist

Searchable Transcript

Search:

[00:00:00.960]Hello, and welcome to my research presentation
[00:00:04.000]for the summer 2020.
[00:00:05.860]My name is Jon Askey, and I conducted research
[00:00:08.360]at the University of Nebraska Lincoln
[00:00:10.210]with funding from UCARE under the mentorship
[00:00:12.920]of Dr. Mark Wilson from the Department of Biochemistry
[00:00:16.140]and Dr. Jennifer Clarke from the Department of Statistics.
[00:00:19.600]In my research, we explored the relationship
[00:00:23.750]between protein structure and low-barrier hydrogen bonds.
[00:00:28.440]As an introduction, it's natural to ask,
[00:00:31.180]what are low-barrier hydrogen bonds,
[00:00:33.210]also known as LBHBs.
[00:00:36.150]These are a special type of short hydrogen bond
[00:00:38.250]where a proton is delocalized between the donor
[00:00:40.520]and acceptor atoms from two different proteins
[00:00:42.910]or two regions of the same protein.
[00:00:45.220]The name low-barrier hydrogen bond is used
[00:00:46.990]to describe the low activation barrier of proton transfer
[00:00:50.330]between the proton donor and proton acceptor,
[00:00:52.760]where both protein residues have a similar pKa.
[00:00:55.930]Jumping down to figure 2,
[00:00:57.790]it's easier to visualize what this means.
[00:01:00.490]The proton would be represented by this H here,
[00:01:03.890]and the two atoms are that are connected
[00:01:06.920]by this hydrogen bond are these two oxygen atoms.
[00:01:10.990]In an LBHB, this proton is not necessarily more attracted
[00:01:15.540]to either one of these atoms,
[00:01:17.210]because they have a similar pKa,
[00:01:18.920]or level of attractivity toward this proton,
[00:01:23.020]so it can relatively freely go between these two atoms.
[00:01:28.970]Even though the existence
[00:01:29.910]of these special short hydrogen bonds is contested,
[00:01:32.830]they are thought to occur
[00:01:33.663]between carboxylate carboxylic acid bonds,
[00:01:35.700]such as bonds between glutamic acid
[00:01:36.777]and or aspartic acid residues.
[00:01:39.260]LBHBs provide a way to facilitate proton transfer reactions
[00:01:43.320]within enzymatically active sites,
[00:01:45.540]and they may also be important determinants
[00:01:47.260]of drug interactions.
[00:01:49.460]As we can see here from figure 1,
[00:01:51.180]we have the human DJ-1 protein, that is a dimer,
[00:01:56.500]and it has these two highlighted bonds.
[00:01:59.980]These are hydrogen bonds between glutamic and aspartic acid,
[00:02:03.700]and they contribute to the protein stability.
[00:02:08.040]While these bonds particularly have been proven
[00:02:10.120]to not be LBHBs, there is a prokaryotic homologue in YagL,
[00:02:15.070]and it's still uncertain
[00:02:16.450]as to whether or not those are LBHB bonds.
[00:02:20.520]So in order to statistically analyze and kind of decide
[00:02:24.900]whether or not there are any variables
[00:02:27.430]that would be good indicators of an LBHB,
[00:02:30.290]depending on the protein structure,
[00:02:32.410]that protein structure being asymmetric
[00:02:34.360]or symmetric proteins,
[00:02:36.750]we identified several of these variables.
[00:02:40.560]There are five of them.
[00:02:41.560]The first one would be bond distance,
[00:02:43.270]and this is defined as the distance
[00:02:44.870]between the two oxygen atoms of these hydrogen bonds.
[00:02:49.300]And then we have angle 1 and angle 2
[00:02:52.120]and torsion 1 and torsion 2.
[00:02:54.150]The numbering of angle 1 and angle 2,
[00:02:56.710]and torsion 1 and torsion 2,
[00:02:57.880]the difference would just be the residue that comes first,
[00:03:02.750]based off of conventional ordering of protein residues.
[00:03:08.220]As we can see here, angle is based
[00:03:11.960]or it's the angle centered at the oxygen
[00:03:15.390]before the hydrogen bond, and torsion is centered
[00:03:20.210]at the bond just before the hydrogen bonds.
[00:03:27.910]As a quick descriptor of the data,
[00:03:29.580]it was pulled from the Protein Data Bank,
[00:03:31.240]and it consists of four different sets,
[00:03:33.120]asymmetric and symmetric, asymmetric tight,
[00:03:35.620]and symmetric tight.
[00:03:37.290]Asymmetric and symmetric simply refer
[00:03:38.780]to the protein structure of the identified candidate LBHB,
[00:03:44.330]and that can be seen here in figure 1.
[00:03:45.810]This would be considered a symmetric protein
[00:03:47.990]because it's a dimer made of two symmetric monomers,
[00:03:53.029]and the asymmetric would be if they were not symmetric,
[00:03:57.010]the types set.
[00:03:58.240]So the tight versions of these asymmetric
[00:03:59.940]and symmetric are simply subsets.
[00:04:02.440]So asymmetric tight is a subset of asymmetric,
[00:04:04.890]and they just have more stringent cutoff values
[00:04:06.810]for the relevant variables, or more specifically,
[00:04:10.750]one of these would be the bond distance
[00:04:14.070]between the two oxygens of the carboxylic acid,
[00:04:17.200]carboxylate H bonds.
[00:04:19.730]Just taking a look at the descriptive statistics
[00:04:22.010]for these variables,
[00:04:23.570]One significant point would be that
[00:04:25.780]the asymmetric data set is significantly larger than
[00:04:29.420]the three other data sets.
[00:04:30.780]It's more than twice the size of its tight version.
[00:04:33.640]And the symmetric data set is just a fraction
[00:04:36.050]of the asymmetric dataset.
[00:04:39.820]For the data visualization,
[00:04:41.560]this is a technique used to glean more information
[00:04:44.740]from these variables and how the data looks,
[00:04:47.350]and then you can draw further conclusions based off of
[00:04:50.725]the visualizations.
[00:04:53.490]To visualize the data, we use programming software
[00:04:56.780]in Python 3.0 and packages such as Pandas,
[00:04:59.730]Matplotlib, and Seaborn.
[00:05:02.110]Here's a snippet of code used to create
[00:05:04.810]a univariate histogram,
[00:05:06.400]the variable distance and the set asymmetric.
[00:05:09.820]It's pretty standard code.
[00:05:12.660]One thing to point out is simply just choosing the bin size.
[00:05:15.850]This can be seen in graph 1.
[00:05:18.160]You want to choose bins that are,
[00:05:20.520]with a width that creates enough variations,
[00:05:23.410]you can appropriately kind of get an idea
[00:05:27.450]of how the data looks like.
[00:05:31.650]This is an example of an overlaid histogram
[00:05:35.580]from each of the data sets.
[00:05:37.963]What you would expect when looking
[00:05:40.970]at something like this is,
[00:05:43.390]or I guess what you can see
[00:05:44.780]from just looking at this is that most,
[00:05:46.810]the data seems to mostly follow a normal distribution,
[00:05:50.300]just kind of like a bell curve.
[00:05:52.060]However, there does seem to be some extra bonds over here
[00:05:56.930]for the asymmetric sets.
[00:06:00.070]And in further interpretation of data visualization,
[00:06:03.410]so to extract information from our graphs,
[00:06:05.980]you look at these for any abnormalities that may exist
[00:06:10.190]in the distributions compared
[00:06:11.820]to what you would just normally expect,
[00:06:13.820]and that would be called the null hypothesis.
[00:06:16.548]Further statistical techniques,
[00:06:18.300]like one-way ANOVA testing, are then used
[00:06:21.370]to see if these differences are statistically significant.
[00:06:25.500]Specifically, we tested and rejected the null hypothesis
[00:06:28.220]at an alpha level of less than 0.01,
[00:06:30.686]but the average bond distances of each group were equal.
[00:06:34.370]And this can be seen in this box plot where
[00:06:37.960]the mean distance for this tight sets is significant
[00:06:41.880]or looks a lot less than the asymmetric and symmetric tight.
[00:06:45.930]And, to be fair, this does make sense,
[00:06:48.490]considering that's how these sets are kind of defined.
[00:06:51.840]However, there are other variables involved,
[00:06:54.170]and it at least warrants further analysis
[00:06:59.140]to explain and characterize these differences.
[00:07:03.000]So ultimate conclusions from this work
[00:07:05.130]and the future of this project.
[00:07:08.350]The role of LBHBs play in protein stability
[00:07:12.040]and enzyme catalysis is still unclear,
[00:07:13.970]and it's has been hypothesized
[00:07:15.660]that LBHBs play important roles in switching elements
[00:07:19.020]in cooperativity pathways and multimeric enzymes.
[00:07:22.310]Such pathways play an important role
[00:07:23.810]in biological processes, such as antibiotic resistance.
[00:07:27.810]And these are key questions to answer
[00:07:29.310]in basic structural biology that are relevant
[00:07:31.490]to the study of disease, drug discovery,
[00:07:33.470]and protein engineering, and creating a filter
[00:07:36.080]or a set of criteria to better assess the likelihood
[00:07:38.350]of the presence of these LBHBs.
[00:07:41.580]It will enhance the scientist's ability to study them
[00:07:43.900]as potential important interactions in structural biology.
[00:07:48.460]Further work on this project
[00:07:49.500]could include similar ANOVA testing
[00:07:51.540]for the other LBHB characteristics,
[00:07:55.130]advanced data visualization with t-SNE plots
[00:07:57.650]and k-means clustering,
[00:07:59.030]and exploring classification techniques
[00:08:00.910]for predicting LBHB grouping.
[00:08:04.459]And finally, we just have our references.
[00:08:08.890]So thank you for listening to my research presentation,
[00:08:12.220]and have a nice day.

The screen size you are trying to search captions on is too small!

You can always jump over to MediaHub and check it out there.

Comments

0 Comments

The relationship between protein structure and low barrier hydrogen bonds

Description

Searchable Transcript

Comments icon comment

Related Channels

Comments