Protein Structure answers37 - Revision history

Carol at 16:07, 9 October 2025

2025-10-09T16:07:22Z

← Older revision		Revision as of 18:07, 9 October 2025
Line 78:		Line 78:
	[[File:Question7a.png\|1000px]]		[[File:Question7a.png\|1000px]]

			''Commands:''

			:align chain C, chain A

	'''Question 8.''' Provide an hypothesis of why the cow profilin does not cross-react with the IgE antibody giving allergy to Hev b8.		'''Question 8.''' Provide an hypothesis of why the cow profilin does not cross-react with the IgE antibody giving allergy to Hev b8.
Line 90:		Line 92:

	[[File:Question7c.png\|700px]]		[[File:Question7c.png\|700px]]

			''Commands:''

			:select epitope2, br.(antibody around 4)
			:cmd.color_deep("orange", 'epitope2', 0)

			Here in "epitope2" (shown in orange) we can highlight the amino acids both in the allergen and in the cow profilin to compare them.

Carol at 16:03, 9 October 2025

2025-10-09T16:03:37Z

← Older revision		Revision as of 18:03, 9 October 2025
Line 77:		Line 77:

	[[File:Question7a.png\|1000px]]		[[File:Question7a.png\|1000px]]



			'''Question 8.''' Provide an hypothesis of why the cow profilin does not cross-react with the IgE antibody giving allergy to Hev b8.


			Here there is a zoom in of the region of interest, showing the epitope on the allergen in orange where you can see the differences to cow profilin residues of the protein that are in close contact with the antibody. It might happen that some of these loops missing in the cow protein prevent them for binding. However this is just a grasp of an hypothesis and more experiments could be done to prove if this is really the reason or maybe there are other factors that make this protein an allergen.

Carol at 16:01, 9 October 2025

2025-10-09T16:01:27Z

← Older revision		Revision as of 18:01, 9 October 2025
Line 82:		Line 82:


	[[File:Question7c.png\|~~300px~~]]		[[File:Question7c.png\|700px]]




	~~<!--~~




	As you can notice, all the 5 hits are X-ray structures with the same molecule name from the same organism. But if you inspect the "SCOP classification" panel you can notice that only 4 out of 5 hits contain the protein domain of interest ( for a definition of what a protein domain is look at the wikipedia page[[http://en.wikipedia.org/wiki/Protein_domain ]], but basically it is a part of a protein that has a well defined, stable structure that does not depend on other regions of the protein and that often performs a well defined function). The protein that is not classified as having the ''Rhamnogalacturonan acetylesterase'' domain is the 3C1U pdb entry, titled "D192N mutant of Rhamnogalacturonan acetylesterase". <br>
	So, even if this PDB entry is related to our protein, we might be careful since it has some -possibly important- mutations with respect to the wild type protein that might affect its structure/function and eventually our analyses. So, if we have other options, we might want to exclude this entry or to use it as a last resource.


	~~== Q3 Choose the best structure that has sulfate ions bound. Which one did you choose? Why? ==~~
	Both 1deo, 1k7c, and 2o14 have sulfate ions bound. However, the resolution of 1k7c is 1.12 Å, which is better than the 1.55 Å resolution of 1deo and the 2.1 Å of 2o14 (which, by the way, was among the non-relevant results). '''Note:''' 1.55 Å is very good under most circumstances and resolutions better than this – i.e. lower – are not common. The Rfree is 0.134 for 1k7c, which is also better than 0.200 for 1deo (again this is mostly a function of resolution although other factors such as data quality and refinement protocols also contribute).

	~~== Q4 What is the residue name for the sulfate ions? ==~~

	~~SO4~~

	~~== Q5 Click on H(ide) and select “waters”. What happened? ==~~
	~~When water molecules are hidden with the Hide command (button), they will simply be switched off. To turn them on again simply click on S(how) – nonbonded.~~

	~~== Q6 The active site of RGAE. ==~~
	The active site residues are: '''Ser9''', '''His195''' and '''Asp192'''. These numbers do not directly correspond to the information in Swiss-Prot entry. The reason is that residue numbering in the PDB file starts with the residues of the native protein, i.e. the mature protein sequence without the signal peptide. This means that all residue numbers in the PDB file are off by 17.

	~~[[File:1k7c.png\|800px]]~~

	~~== Q7 ==~~
	~~The active site residues of PAFA are: Ser 47, His 195 and Asp 192.~~
	~~The serine has a different numbering in the two structures, and actually the fact that the other residues have a similar numbering is accidental.~~
	~~This is caused by the small sequence/structure similarity between the two proteins.<br>~~

	~~As you can see from the structural superposition~~

	~~[[File:1k7c_1wab_pymol.png\|800px]]~~



	~~and from the corresponding sequence alignment (obtained from the structure superposition, see the note below )~~

	~~[[File:Struc_alig_1k7c_1wab.jpeg\|800px]]~~

	~~there is a poor structural similarity and an even lower sequence similarity. Look at the low number of identical residues and at the insertions/deletions in the alignment.<br>~~

	NB: the sequence similarity is actually so low that it is impossible to obtain a proper sequence alignment with basic tools. In this case, the sequence alignment was obtained using the structural superposition generated by jFATCAT_rigid on the pdb website.

	~~== PyMol magic (not part of the exercise, but it can make your life easier) ==~~
	~~Instead of manually looking for the catalytic site residues in the active site, you could use the following two commands in PyMOL:~~

	~~select triad_his, (byres resn his within 3 of resn asp) and (byres resn his within 3 of resn ser)~~

	This first command selects all those histidine residues (selection name: triad_his), which are simultaneously within 3 Å of any aspartic acid residue '''AND''' within 3 Å of any serine residue, which is the case for histidine residues found in catalytic triads of the kind we are looking for. The byres modifier ensures that we select entire residues and not just those atoms, which fulfill the distance requirement. It turns out that there is only one such residue in the 1k7c structure. To select the other residues of the catalytic triad, simply write:

	~~select triad, triad_his or (byres resn asp within 3 of triad_his) or (byres resn ser within 3 of triad_his)~~

	This second command selects the histidine residue found with the first command (again) along with aspartic acid and serine residues within 3 Å of that histidine. And voilà, you have found your catalytic triad =)

	~~To get the residue names and numbers, either click the residues in the viewer window or type the following three commands:~~

	~~triad_list = []~~
	~~iterate triad and name CA, triad_list.append((resn, resi)) for pair in triad_list: print pair[0], pair[1]~~

	~~This should print the following information in the command window:~~

	~~SER 9~~
	~~ASP 192~~
	~~HIS 195~~

	Note: In the general case, you can modify the commands above to look for other kinds of arrangements of residues, but you will of course need to know rather accurately how the residues of interest are positioned relative to each other.
	~~-->~~

Carol at 16:01, 9 October 2025

2025-10-09T16:01:01Z

← Older revision		Revision as of 18:01, 9 October 2025
Line 79:		Line 79:


	[[File:Question7b.png\|~~1000px~~]]		[[File:Question7b.png\|500px]]

Carol at 16:00, 9 October 2025

2025-10-09T16:00:40Z

← Older revision		Revision as of 18:00, 9 October 2025
Line 76:		Line 76:


	[[File:Question7a.png\|~~300px~~]]		[[File:Question7a.png\|1000px]]


	[[File:Question7b.png\|~~300px~~]]		[[File:Question7b.png\|1000px]]


	[[File:Question7c.png\|~~1000px~~]]		[[File:Question7c.png\|300px]]

Carol at 16:00, 9 October 2025

2025-10-09T16:00:13Z

← Older revision		Revision as of 18:00, 9 October 2025
Line 68:		Line 68:
	AlphaFill, is a model developed to fill the predicted structures from AlphaFold, with co-factors and ligands. This particular issue might be particularly relevant in some cases but not so in others.		AlphaFill, is a model developed to fill the predicted structures from AlphaFold, with co-factors and ligands. This particular issue might be particularly relevant in some cases but not so in others.
	In this case we have several protein structures of the same family determined with high accuracy, so all these models or predictions are of high accuracy. As a personal test I tend to like the alphaFold version that provides some confidence measures for the different regions of the protein. But with this answer we want to highlight that in real life biological questions, there is not always we clear or unique answer, but our bilogical question and constraints will determine which is the best option in our particular case.		In this case we have several protein structures of the same family determined with high accuracy, so all these models or predictions are of high accuracy. As a personal test I tend to like the alphaFold version that provides some confidence measures for the different regions of the protein. But with this answer we want to highlight that in real life biological questions, there is not always we clear or unique answer, but our bilogical question and constraints will determine which is the best option in our particular case.


			''' Question 7.''' Provide a screenshot of the alignment. Can you see the differences of the two Profilins?


			The differences of these two proteins are very small as these two proteins, as they share high homology, specially in the structural space. However there are some loops that tend to be longer in the rubber tree profilin and may explain why these protein can be bound by the IgE antibody while the cow profilin does not.


			[[File:Question7a.png\|300px]]


			[[File:Question7b.png\|300px]]


			[[File:Question7c.png\|1000px]]

Carol at 15:58, 9 October 2025

2025-10-09T15:58:20Z

← Older revision		Revision as of 17:58, 9 October 2025
Line 53:		Line 53:
	:select epitope, br.(antibody around 4)		:select epitope, br.(antibody around 4)
	:cmd.color_deep("red", 'epitope', 0)		:cmd.color_deep("red", 'epitope', 0)


			''' Question 5.''' For that purpose, we will need to search for an homologous sequence to the one in our crystal structure, and guess which software we will use for that?


			BLASTP


			''' Question 6.''' Can you explain what is the main difference of the crystal structures, the predicted structure and the modelled structure with SWISS-MODEL? Which one do you think would be more reliable? Here you are encouraged to use a chatbot to ask for hints, like chatGPT, but you will have to justify your selection.

			All of them seem to provide a very similar structure however they are based on different approaches to predict the structure. In this case, all the structures seem to be very homologous to the others and of good quality as eveidenced by the AlphaFold quality scores, however we can still compare them as they have been obtained by different methods.
			AlphaFold is a machine learning method that has been trained on all PDB structures to predict 3D structures. It gives very accurate predictions when the protein we want to predict is somewhat related or share some domains with crystally determined and known structures.
			SwissMODEL uses homology modelling, wich is normally slow but very accurate when we have some structure that is highly homologous to the one we want to predict.
			AlphaFill, is a model developed to fill the predicted structures from AlphaFold, with co-factors and ligands. This particular issue might be particularly relevant in some cases but not so in others.
			In this case we have several protein structures of the same family determined with high accuracy, so all these models or predictions are of high accuracy. As a personal test I tend to like the alphaFold version that provides some confidence measures for the different regions of the protein. But with this answer we want to highlight that in real life biological questions, there is not always we clear or unique answer, but our bilogical question and constraints will determine which is the best option in our particular case.

Carol at 15:46, 9 October 2025

2025-10-09T15:46:27Z

@@ Line 35: / Line 35: @@
 In the allergen side is mostly alpha helices and loops.
 If you want to analyse the antibody surface too (which is called paratope) we might need to go back to the cartoon view again. There are three well known loops called CDR1, CDR2 and CDR3, both in the light (L) chain and in the heavy (H) chain.
 <!--

Carol at 15:41, 9 October 2025

2025-10-09T15:41:35Z

← Older revision		Revision as of 17:41, 9 October 2025
Line 20:		Line 20:
	:color blue, chain H		:color blue, chain H
	:color yellow, chain L		:color yellow, chain L
	~~:select antibody, chain H+L~~
	~~:cmd.show("surface" ,"antibody")~~


	''' Question 3.''' Provide a screenshot of the interaction surface. Which secondary structures of the allergen is the antibody interacting with: alpha helices, beta-sheets or loops?		''' Question 3.''' Provide a screenshot of the interaction surface. Which secondary structures of the allergen is the antibody interacting with: alpha helices, beta-sheets or loops?


	[[File:Question3.png\|1000px]]		[[File:Question3.png\|1000px]]


			''Commands:''

			:select antibody, chain H+L
			:cmd.show("surface" ,"antibody")

	In the allergen side is mostly alpha helices and loops.		In the allergen side is mostly alpha helices and loops.

Carol at 15:40, 9 October 2025

2025-10-09T15:40:09Z

← Older revision		Revision as of 17:40, 9 October 2025
Line 15:		Line 15:
	''Commands:''		''Commands:''

	fetch 7SBD		:fetch 7SBD
	remove hetatm		:remove hetatm
	util.color_chains("(all)",_self=cmd)		:util.color_chains("(all)",_self=cmd)
	color blue, chain H		:color blue, chain H
	color yellow, chain L		:color yellow, chain L
	select antibody, chain H+L		:select antibody, chain H+L
	cmd.show("surface" ,"antibody")		:cmd.show("surface" ,"antibody")

← Older revision		Revision as of 18:00, 9 October 2025
Line 76:		Line 76:


	[[File:Question7a.png\|~~300px~~]]		[[File:Question7a.png\|1000px]]


	[[File:Question7b.png\|~~300px~~]]		[[File:Question7b.png\|1000px]]


	[[File:Question7c.png\|~~1000px~~]]		[[File:Question7c.png\|300px]]