Transcript of meeting in BioMOO, PPS Base 14th May '96 15:00 GMT: on the Discussion Topics of Formal Assignment 1

In this meeting the Discussion Topics of the first assessed assignment were addressed.

This transcript can currently be found on the tape 'discussion_tape' in the PPS Base, but will eventually have to be deleted to save disk space.


Jzt says, "testing discussion_tape 14th May 1996"

jzt turns the C-recorder off.

jzt turns the C-recorder on.

Jzt says, "There are 5 'discussion topics' in the assignment, 4 on page 1 and one on page 2 - shall we go through them in order or have a random mixture?"

RajG [guest] says, "Hello everyone"

Paolo says, "hello"

RajG [guest] says, "Since some people have not handed in the assignment"

JohnW [to RajG]: Would you like to go through them in order?

RajG [guest] says, "today we can only discuss general things about insulin"

JohnW nods

JohnW says, "Okay, I think its high time we began-"

RajG [guest] says, "does anyone have any interesting ideas or questions they had when doing the assignment ?"

JohnW says, "Just to reiterate that in this meeting we will be covering the 'discussion questions' of the Insulin Assignment only"

JohnW says, "-we cannot go over the actual assessed questions themselves at this stage"

RajG [guest] says, "does anyone have any comments on the assignment?"

RajG [guest] says, "was it hard ? fun? both?"

JohnW says, "So I think we should look at the first Discussion Question, at"

Paulyta says, "I have one that pertains to the first discussion question"

Paulyta says, "Why does the water not overlap the protein structure?"

JohnW nods

JohnW says, "It may help if you all have the relevant structure in a RasMol window at the moment."

RajG [guest] says, "how do you mean, because obviously one atom cannot overlap another, you must mean something else"

JohnW says, "I think What Tom means is that the solvent molecules are in a completely separate area of the asymmetric unit to the actual protein"

JohnW says, "Firstly, is everyone happy with the meaning of the term asymmetric unit?"

HorstJS finds his way in.

HorstJS waves.

Salim finds his way in.

RajG [guest] says, "I guess it is so that the protein structure is not complicated by all the waters, probably a crystallographic convention. The pdb file probably tells you the transformartion to make water and protein overlap"

CheeMV says, "Do you mean that the dimer is not perfectly symmetric?"

Salim . o O ( have bad connection, might have to just read the transcript.... )

JohnW says, "The point is, that due to crystal symmetry...."

RajG [guest] says, "true, in the structure you have investigated the symmetry is approximate but I dont think this explains why the waters are not given in their relative positions to the dimer"

JohnW says, " only need information on the coordinates of the smallest unit of symmetry, plus the symmetry operations which reproduce the entire unit cell..."

JohnW says, "remember, these PDB files describe the structure of entire crystal structures..."

JohnW says, "so the coordinates are only meaningful in that context..."

CheeMV says, "Is the experimental crystal data on the monomer, and the dimer structure is derived by symmetry operation?"

JohnW says, "when you apply the symmetry operations to this asymmetric unit, to restore the unit cell, which contains 3 copies of the asymm unit..."

JohnW says, "..the end result would be the same, whether the solvent coordinates are those of the waters around the protein itself, or are a symmetry copy..."

JohnW says, "If you were to reform one unit cell, you would get a hexamer of 3 dimers, with a separate group of waters above it- 3x what you have in the pdb file. "

JohnW says, "If you then stacked a load of unit cells together, the solvent would indeed overlap the protein. So this representation is I believe, as Raj said, a way of showing the protein uncluttered by solvent."

UlisesSanch finds his way in.

Ahotz finds her way in.

RajG [guest] says, "can anyone think of residues they could change in insulin so that it might crystallize as a monomer ?"

RajG [guest] says, "hint think about the dimerization interface between two insulin monomers"

Salim says, "Answer to Raj's Q - the residues at the dimer interphase, perhaps ??"

Paulyta says, "you could change these residues to polar ones"

RajG [guest] says, "paula is right but it still might not work, why ?"

Paolo says, "and residues involved in hexamerization, also"

Paulyta says, " may not crystallize or be biologically active"

RajG [guest] says, "des any body remember the dimerizing interaction not involving the side chain per se ?"

Salim . o O ( Bad connection, I'll have to leave, bye )

Salim has disconnected.

Iddo materializes out of thin air.

JohnW waves to Iddo

Paolo says, "yes, hbonds between the two mainchains at residues b24-26 "

UlisesSanch has disconnected.

RajG [guest] says, "we do not understand what is going on well enough. Novo made a change to a polar residue in the hope of getting a monomeric NMR structure, but it was dimeric to a greater stabilty than the native molecule !"

The counter-ion says, ''go up for the PPS '96 Office, down for the PPS Area Centre''.

The counter-ion finds its way home.

JohnW says, "Shall we go onto the next discussion topic?"

Iddo disappears suddenly for parts unknown.

Iddo materializes out of thin air.

JohnW says, "Ok, the next discussion topic is stated as follows:"

JohnW says, "Use the RasMol command 'structure' to calculate the secondary structure according to the Kabsch and"

JohnW says, "Sander's hydrogen bond-based DSSP algorithm. How does this compare with the secondary structure designation in the PDB file? "

JohnW says, "So the best approach here is to get TWO RasMol windows, each with the same structure file..."

Jzt [to Iddo]: OK I will try to remember to cc BioMOO schedules to Gustavo. Thanks.

Abe arrives from far away places

The housekeeper arrives to cart Salim off to bed.

JohnW says, "then apply the 'structure' command to only one of them, and compare the two"

JohnW says, "If you have done that, does anyone notice any difference? If so, what?"

CrisCan says, "it doesn't appear the beta strand"

The housekeeper arrives to cart UlisesSanch off to bed.

JohnW nods to CrisCan

CrisCan says, "but the problem is: why? is there a different definition of strands in DSSp in comparison with PDB?"

JohnW says, "By now you will probably be aware that a good way of looking at secondary structure in RasMol is to select 'Ribbons' and then Colour the chains by 'Structure'"

Abe finds his way out.

BobJ finds his way in.

JohnW [to CrisCan]: Yes. The point is, there are different ways of defining secondary structure.

JohnW says, "Because there are cases where helices and sheets stray from their 'ideal' sforms, we can come across some ambiguity."

JohnW says, "For example, defining exactly where an alpha helix, or a beta strand, stops and starts. The authors of some protein structures in the past may have done this'by eye'."

JohnW says, "When you first load in a PDB file to RasMol, it uses the authors' description of secondary structure therein, unless you tell it to do otherwise."

JohnW says, "The K&S algorithm defines secondary structure by hydrogen bonding patterns- this in turn depends on the definition of allowable hydrogen bond geometry that you use."

BobJ has disconnected.

JohnW says, "When you apply the algorithm to the 4ins structure, the two strands described by the structure's authors simply are not designated as strands by DSSP."

JohnW says, "Note that there is no right or wrong answer- it simply depends on your definition."

JohnW says, "For example, another way you might define alpha helices or beta strands is by using the phi and psi values, but then you have to decide how strictly you are to apply the constraints of the 'ideal' structures."

JohnW says, "Anyway, shall we move onto the next topic?"

JohnW says, "This is, "Why are the three glycine residues with positive phi dihedral angles important for the insulin fold? ""

RajG [guest] says, "does any one know why gly can take +ve phi dihedral angles ?"

CrisCan says, "I think they are connected with the cysteine bridges"

RajG [guest] says, "(to criscan) please elaborate"

CrisCan says, "they can because they have not the side chain"

Iddo has disconnected.

RajG [guest] says, "thats right the reason is simply steric"

CrisCan says, "the presence og"

The housekeeper arrives to cart BobJ off to bed.

CrisCan says, "sorry the presence of glycine allow the main chain to turn in a way that could be impossible with other AA"

RajG [guest] says, "what would happen if such a gly residue were mutated to another amino acid ?"

CrisCan says, "I am afraid that the folding will change"

CheeMV says, "The insulin will not fold properly."

RajG [guest] says, "if you are interested look at the hystricomorph sequences where some of the gly have changed. they have weird CD spectra"

RajG [guest] says, "are glycines often found next to cysteines ? why ?"

CrisCan says, "they help the chain to turn bringing the cysteine in the right position"

RajG [guest] says, "why does the cys have to be in the right position (for example why does thuis not apply to all other AAs ?"

Paolo says, "in order to form a s-s bridge with another cys"

CheeMV says, "so that they can form disulfide bonds to stabilise the fold."

CrisCan says, "they have to be able to make the bridge, that is a reaction"

RajG [guest] says, "the answer is because disulphides stabilize a protein at the expense of creating a strain to line up two sulphydryls, and the strain in the backbone is alleviated to a greater extent by a gly"

The housekeeper arrives to cart Iddo off to bed.

RajG [guest] says, "I have an interesting question about a disulphide in insulin"

HorstJS says, ""but this may be true only for some SS bonds since there are SS Bonds which destabilize a protein"

RajG [guest] says, "when making a heavy atom derivative with mercury the x-ray structure showed the mercury atom had connected between the two sulphurs !!!!"

RajG [guest] says, "does this mean the disulphide was previously tense or floppy?"

CrisCan says, "what is about the other parts of the structure? does the presenze of this atom change the way of crystallization?"

RajG [guest] says, "(to criscan) no the rest is nearly the same i.e isomorphous, and native coordinates were used to phase the x-ray data"

HorstJS says, ""at least the protein I'm dealing with (a protein disulfide isomerase)"

HorstJS says, "has a destabilizing SSbond"

RajG [guest] says, "is it a tense disulphide which is waiting to snap open to catalyse difulide interchange with another ?"

HorstJS says, "actually this has nothing to do with a sterical strain but with the"

HorstJS says, "electrostatic properties of an alpha helix (helix dipole)"

RajG [guest] says, "very interesting, which PDI is it ?"

HorstJS says, "so the point is,"

HorstJS says, " sorry - its a bacterial PDI from E.coli called DsbA"

HorstJS says, "no actually the disulfide bridge does not contain any strain"

HorstJS says, "but there are electrostatic effects in the protein"

HorstJS says, "which lead to a deprotonation of the SH-Group"

HorstJS says, ""its that additional charge that causes the destabilization"

JohnW says, "Ok, lets move on to the next question: 'How could insulin delivery by infusion pumps be improved?'"

JohnW . o O ( hmmm )

RajG [guest] says, "does any one have any ideas? because infusion pumps were exciting about 10 years ago for brittle diabetics but therte have been many problems owing to the nature of insulin"

RajG [guest] says, "hint : how do you wetbobs prevent protein aggregation in solution ?"

Paulyta says, "I add some detergent"

RajG [guest] says, "correct, they tried that, can you suggest one which wont harm a diabetic ?"

RajG [guest] says, "SDS? tritonX? glycerol?"

CheeMV says, "Add albumin."

RajG [guest] says, "albumin is good idea since it is already in your body and cheap"

RajG [guest] says, "remember that you must also consider the mechanical properties of the pump and syringe"

CrisCan says, "sorry, but there is something I dont understand. why they adhere? glass is hydrophilic or hydrophobic?"

CheeMV says, "but its expensive in the quality required for human therapeutic use."

JohnW apologises for having to leave the meeting for a while

CheeMV says, "It interacts with the polymer in plastic by hydrophobic interaction."

JohnW goes up.

RajG [guest] says, "a lot of these problems were overcome but insulin self-aggregation over the weeks in a pump remains a problem"

CrisCan says, "I thought it was hydroplilic and we use silanization to have it hydrophobic"

RajG [guest] says, "cheemv is right. what sort of polymer would be best?"

Paulyta says, "teflon"

RajG [guest] says, "do you guys know the answer to this one, I would like to learn"

RajG [guest] says, "I guess you would have to make sure the abrasion from the plunger does not give you a dose of teflon !"

RajG [guest] says, "before moving on, how about suggesting mutations to the molecule ?"

Paolo says, "we could mutate some aa on the di- and hexamerization surfaces"

RajG [guest] says, "incidently the self aggregated insulin builds up in the kidneys and liver of test dogs (amyloidosis) so a lot of work to be done here"

RajG [guest] says, "shall we move on ?"

Paolo says, "yes because some of them are involved in receptor binding"

RajG [guest] says, "Why is insulin made as proinsulin?"

Paolo says, "in order to carry outside of the cell the two chains together, maybe?"

Paulyta says, "so that it folds correctly"

CrisCan says, "perhaps, to help folding"

RajG [guest] says, "yes there are so many good ideas"

RajG [guest] says, "I will elaborate"

RajG [guest] says, "it is true that in this way you get equimolar quantities of A and B chain in the right place at the right time for folding"

RajG [guest] says, "also unfolded insulin only refolds bak to native fold about 10% but proinsulin"

RajG [guest] says, "folds back to 70%"

RajG [guest] says, "is the c-peptide hydrophobic or hydrophillic ?"

CrisCan says, "it is hydrophillic"

RajG [guest] says, "yes and it has been proposed that this aids translocation in the ER from ribosome to secretory granule"

RajG [guest] says, "any interesting questions o proinsulin ?"

Paulyta says, " I thoght it was cleaved before secretion"

RajG [guest] says, "yes, but about 10% finds its way uncleaved into the bloodstream"

RajG [guest] says, "proinsulin has only 10% of the receptor affinity of insulin"

RajG [guest] says, "can anyone think why proinsulin might be considered for some diabetics?"

RajG [guest] says, "hint: immunological response to injected insulin"

RajG [guest] says, "ok i will answer"

CheeMV says, "Proinsulin does not have the epitope for anti-insulin antibody."

RajG [guest] says, "about 10% of diabetics make antibodies to injected insulin, not so much these days with humulin"

RajG [guest] says, "they needed to take larger and larger doses"

RajG [guest] says, "since the C-peptide covers the end of the B chain which is an antigenic determinant, Eli Lilly and Co proposed to use proinsulin in these patients (recombinant)"

RajG [guest] says, "but it gave the test mice cancer !"

RajG [guest] says, "if you have no further questions shall we end the meeting?"

CheeMV says, "Will the proinsulin be taken up by cells and the C-peptide cleaved?"

Paulyta says, "yes"

Paulyta waves

CrisCan says, "ok thank you very much"

RajG [guest] says, "just before you all go"

Paulyta has disconnected.

Paolo says, "Bye and thanks!"

Jzt says, "Thanks a lot Raj! I'll stop the recorder now"

jzt turns the C-recorder off.