Mad Cow Home -- Best Links -- Search this site

Prion-specific epitope defined by a monoclonal antibody!
Summary table: who immuno-precipitates what?
DNA/protein sequence actually used to make bovine prion
Who/what is Prionics AG?
Quick commentary on new anti-prion antibody
Imagery of possible rogue prion dimer
Accuracy of Figure 3b, 3c questioned
Aguzzi/Weissmann on prion research frontiers
Researchers closer to test for mad cow disease
New test for early diagnosis of CJD
Conformational intermediates of purified human prion
Brain receptor role links Alzheimer's and smoking

Prion-specific epitope defined by a monoclonal antibody

Letter to Nature, Nov 6 issue, vol 390  pg 74 1997
C Korth, B Stierli, P Streit, M Moser, O Schaller, R Fischer, 
W Schultz-Schaeffer, H Kretzschmar, A Raeber, U Braun, 
F Ehrensperger, S Hornemann, R Glockshuber, R Riek, M Billeter,  K Wuthrich & B Oesch 
Rogue prion -specific epitope defined by a monoclonal antibody Prions are infectious particles causing transmissible spongiform encephalopathies (TSEs). They consist, at least in part, of a rogue isoform of the ubiquitous normal cellular prion protein. Conformational differences between conformers are evident from increased beta-sheet content and protease resistance in the bad conformer. Here the authors describe a monoclonal antibody, 15B3, that can discriminate between the normal and disease-specific forms of the prion protein.

Such an antibody has been long sought as it should be invaluable for characterizing the infectious particle as well as for diagnosis of TSEs such as BSE or CJD. 15B3 specifically precipitates bovine, murine or human rogue, but not normal conformer, suggesting that it recognizes an epitope common to rogue prions from different species.

Using immobilized synthetic peptides, they mapped three polypeptide segments responsible for the 15B3 epitope. In the NMR structure of recombinant mouse PrP, segments 2 and 3 of the 15B3 epitope are near neighbours in space, and segment 1 is located in a different part of the molecule. They discuss models for the PrPSc-specific epitope that ensure close spatial proximity of all three15B3 segments, either by intermolecular contacts in oligomeric forms of the prion protein or by intramolecular rearrangement.

15b3-1 involves residues 142-148 GSDYEDR(YY); 15b3-2 residues 162-170 YYRPVDQYS;15b3-3 residues 214-226 CITQYQRESQAYY; human numbering is used in the paper for bovine sequence.

Summary table for 15B3 asnd 6H4

8 Nov 97 webmaster
The summary table below, of what the two new anti-prion antibodies immuno-precipitate, is based on the highly compressed text, gel legends, illegible fax images, faulty memory of citation content, and guesswork. Please feel free to send a correction or two.


The fourth row reads, "Monoclonal antibody15B3, but not 6H4, recognizes something in brain homogenate post-treated with protease K from a BSE-afflicted cow, that gives rise, with significant losses, to a (possibly heterogeneous) gel band in the 27k-30k position where Prp(sc) is normally found, maybe removing or reducing infectious titre from the supernatant (along with some non-infectious aggregates or intermediates) while mostly leaving normal conformer, even though the antigen itself was non-infectious in mice over-expressing (by copy excess?) short incubation (?) mouse prion."

The sixth row reads, "Monoclonal antibody15B3 immuno-precipitates something in cerebellar homogenates from four cases of sporadic CJD with unspecified codon 129s and infective transmissivity, Collinge strain type 1 (equally glycosylated), that migrates, post-treated with protease K, within the 27k-30k band."

The eleventh row reads, "Post-treatment with protease K is more efficient than pre-treatment, for15B3 immuno-precipitation from brain homogenate from an unspecified strain of scrapie propagated for an unspecified number of generations probably in short incubation homozygous mouse genotype, possibly because pre-treatment by protease K generates fibrils that mask the epitopes."

row15B36H4reaction mix and protocol
1 no yes wildtype cow, human, sheep, mouse, antigen; western blot 33k-35k
2 yes yes BSE cow; western blot 33k-35k
3 yes no BSE+ ProK ImmP 27k-30k
4 yes- no BSE+ ImmP ProK 27k-30k
5 no no BSE- ImmP ProK 27k-30k
6 yes -- CJD-1; ImmP ProK 27k-30k
7 no -- non-CJD; ImmP ProK 27k-30k
8 yes -- scrapie-mouse; ProK ImmP 27k-30k
9 no -- null mouse; ProK ImmP 27k-30k
10 no -- wildtype mouse; ProK ImmP 27k-30k
11 yes+ -- scrapie-mouse; ImmP ProK 27k-30k
12 no -- null-mouse; ImmP ProK 27k-30k
13 no -- wildtype mouse; ImmP ProK 27k-30k

Who is Prionics AG?

Paul Schenker  10 Nov 97
"Prionics AG is a biotech start-up whose stock is privately held and not traded on any stock exchange. All shares of the company are still owned by exactly the same individuals who owned them from the beginning, from the time of incorporation earlier this year (i.e. mainly the three scientists who founded the company). There certainly were no stock sales to authors, friends, family, editors, publishers, reviewers, reporters or PR people in advance of publication of the article. So far there only was a lot of hard work getting the company started and even more to keep it going. More on Prionics AG on their web page."
Roland Heynkes
"We have the same situation in Germany and we also have such spin offs from our universities. Charles Weissmann and Detlef Riesner for example are cofounders of several successful biotech companies and their company EVOTEC Analytical Systems near Duesseldorf (Germany) tries to produce a TSE test. You can follow the connetions of Charles Weissmann, Detlef Riesner, Manfred Eigen and the former German research minister Heinz Riesenhuber with the companies Qiagen and Evotec on the internet."
Webmaster responds:
"This particular article seems scientifically sound vis-a-vis the antibodies. My problem is lack of a policy of disclosure from Nature/MacMillan. I don't see any disclosures on any of their articles. The New England Journal of Medicine or Journal of Biological Chemistry absolutely require full and total prior disclosure of financial stakes on the part of every player, from author through reviewer and pundit. They have a stake, so what, but the reader needs to know this to assess whether a given article might have to do with stock-kiting."

Commentary on new anti-prion antibody

Listserve 6 Nov 97 --  webmaster
This is really a breakthrough paper that will affect everyone's research. It builds on the ability to obtain nearly gram quantities of purified full-length matureprion protein from E.coli inclusion bodies, a crucial technical innovation that avoids endopeptidases of the periplasmic space.

On a quick read, basically 15B3 binds 142-148, 162-170, and 214-226 whereas 6H4 just binds in overlap at 144-152 in helix one. 6H4 is said to only bind normal prion conformer.

They had rather a triple miracle here in that they immunized with non-infectious bovine wildtype yet got out a monoclonal antibody specific to rogue conformers, got a second epitope to normal that didn't cross react despite the common domain, and it all worked across species, including to sporadic CJD strain-type 1.

They go to consider how remote region 1 could be brought into proximity to region 2 and region 3, which are already adjacent in the 3D mouse picture.

Mechanism 1 is head-to-tail oligomer without a major conformational shift. They didn't try docking two monomers for some unknown reason but instead, as I noted on October 28, found that hemoglobins show up among the best-fitting molecules in the 3D sense, of the 6,000-odd protein structures at PDB. [see below]

They fitted two prion monomers on two sickle cell beta monomers as associated in a crystal of sickle cell hemoglobin [PDB entry 1HBS]. Martin Billeter kindly posted a key graphic showing this all fully labelled.

The hemoglobin fits quite well, 2.4 A rmsd, and the sickle cell change, E6V, is said to line up with the anomalous trp 145 of prion protein, which extends most inappropriately into the solvent. Crystallographic axes do often identify weak dimer interfaces. On the other hand, sickle cell hemoglobin is a tetrameric molecule that does not form fibers like those of congophilic diseases, and there is no sequence homology at any significant level. And what happens to the beta-sheet? It is supposed to go from 43% helix, 3% sheet in normal prion to 34% helix, 43% sheet in rogue but here would stay about the same.

Get viewers first, then RasMol hemoglobin beta, two tetramers or Kinemage sickle cell, two sickled tetramers

The other issue, given that the 3D structure is far more conserved than primary structure, is where would the heme pocket be _if_ the structural homology is deep, does this imply an 02 or NO binding capability, allosteric shifts (so two normal conformers by design), and a clue to normal prion function.

My thinking is that the monomeric prion protein will not in fact bind heme, though the fold could be an ancient fragment from a globin duplication and a dimer of two half-pockets could bind conceivably bind a heme. Instead the heme pocket could accept the the tetra-repeat region viewed as essentially reinvented four delta-aminolevulinc acid units. In other words, it ligates a zinc, iron, or copper through the periodic histidines, and folds up into a planar ring that mimics in some structural/functional sense a heme group. It is this object that occupies the 'heme' pocket. Alternatively, the hemoglobin-like helix arrangement could have evolved convergently to bind this 'heme.'

Note that any metal or non-covalently attached heme would be lost during purification from 6M GuCl denaturation and use of distilled water and EDTA in the nmr mix [though I _told_ Wuthrich at a Dec 96 meeting not to do this!] The new RNA aptamer binds the pre-repeat region -- this supports persuasive evolutionary arguments that the Swiss reached an unfortunate conclusion about the structure-forming abilities of this non-random coil region.

Mechanism 2 [and they consider mixes of 1 and 2] is fairly radical and traditional conformational change to a common 'default' protein conformation, perhaps mediated by a chaperone on a denatured or fragmented prion moledule: say the greek key beta domain found in many unrelated proteins. There are other models along these lines already on the table, I guess they decided to place a bet on a position that was still available. Figure 3c would not allow an extended inter-molecular beta-sheet to form, so no fiber and no congo red binding. Better to simply swipe a transthyretin model amyloid out of a textbook, like the 'other' group did.

The epitopes of 15B3 are described in the new Nature paper relative to _mouse_ prion, 15B3 binds 142-148, 162-170, and 214-226 whereas 6H4 just binds 144-152 in helix one..

Unfortunately, there was a structual disagreement between mouse and hamster significantly impacting the structural description of these epitope regions -- they were significantly longer in hamster, attributed to stabilizing effects of extra residues 90-120. Hamster Helix A is two residues longer distally than mouse, 144-156 with the last turn distored; Helix B 172-194, is two turns longer (first irregular) than mouse 179-193; Helix C runs 3 full turns longer than mouse, so 200-227 versus 200-217, with distal irregularity.

* These same authors wrote a year ago that " a systematic search of the Brookhaven data bank with the program DALI did not lead to the identification of other proteins with folds similar to PrP (121-231), ..." [Nature 11 July 1996, pg 180].

A year later we have a prion dimer based "...on the observation of a structural similarity between and haemoglobins, which allows a superposition of the helices 1,2,and 3 of normal prion onto the helices 1, 6, and 7 of haemoglobin beta-subunit,...."

The reference sequence provided, 1HBS, was posted at Brookhaven on 29 October 1985, where it is given as fold rep 1babA from, surprise, Dali/FSSP (Families of Structurally Similar Proteins). Hemoglobin itself was the first oliomer to be worked out; the year was 1959, the structure a tetrahedral dimer of dimers. It would be very very amusing if Max Perutz could write a second prion News and Views commentary focusing on the alignment notions of this paper. Myoglobin has 8 alpha helices, 70% helix content, and no beta sheet. Let's see, helices 1,2, and 6 aren't really contiguous, are they?

Webmaster note of Oct 28 note concerning closest 3D relatives to the prion molecule:

     1bab-A  3.3  3.6   66   142    9  Hemoglobin thionville alpha chai
     2hbg    3.1  3.1   64   147    5  Hemoglobin (deoxy) 
     1maz    2.8  3.1   71   143   11  bcl-xl (apoptosis regulator bcl-
     1efu-B  2.7  3.0   49   282    6  elongation factor tu (elongation
     1ecm-A  2.7  3.8   60    91    7  endo-oxabicyclic transition stat
     1kay    2.6  4.9   67   378    9  70kd heat shock cognate protein 
     1rpo    2.5  6.6   56    61    2  Rop (cole1 repressor of primer) 
     1oxa    2.5  3.5   73   403    3  cytochrome p450 eryf 
     1hrd-A  2.5  3.9   64   449    8  glutamate dehydrogenase biologic
     1gtp-A  2.5  2.7   54   221    6  gtp cyclohydrolase i 
     1dnp-A  2.4  2.8   50   470    6  DNA photolyase (DNA cyclobutane 
     2gdm    2.3  4.9   70   153    9  leghemoglobin (oxy) 
     1grj    2.3  6.9   54   151    0  Grea transcript cleavage factor 
     1ytf-D  2.2  3.3   49   100    4  yeast tata-box binding protein f
     1lfb    2.2  3.7   57    77    7  Transcription factor lfb1 (homeo
     1ecr-A  2.2  7.0   53   305    6  replication terminator protein (
     2hpd-A  2.1  3.5   73   457    3  Cytochrome p450 (bm-3) (hemoprot
     1tfe    2.1  3.3   50   142    6  elongation factor ts fragment 
     1pma-P  2.1  4.2   50   203    4  proteasome (prosome, multicataly
     1mty-G  2.1  6.2   55   162    7  methane monooxygenase hydroxylas
     1hlb    2.1  4.1   64   157    3  Hemoglobin (sea cucumber) 
     1cnt-3  2.1  6.9   64   146    2  ciliary neurotrophic factor (cnt
     1ash    2.1  4.1   57   147    7  Hemoglobin (domain one) 
     1pth    2.0  4.8   50   551    4  prostaglandin h2 synthase-1 (cyc
     1ith-A  2.0  4.5   59   141    7  Hemoglobin (cyanomet) 
     1cfr    2.0  3.1   61   283    3  restriction endonuclease fragment

Lancet commentary: New reagent raises hopes for a diagnostic test for CJD

8 Nov 97
Volume 350, Number 9088 - Saturday 8 November 1997 
A monoclonal antibody, 15B3, has been developed that could be used to test for transmissible spongiform encephalopathies (TSEs). 15B3 recognises disease-specific forms of prion protein (PrPres) and should make it possible to test samples from living people and animals (Nature 1997; 390: 74?77) and to test blood for PrPres.

Prions, thought to be the infectious agents behind TSEs, contain an abnormal isoform of cellular PrP. Researchers based at Prionics AG (Zurich, Switzerland) say that 15B3 recognises three epitopes on PrPres from diseased sheep, cows, and people. These epitopes are not present on cellular prion protein.

The authors also discuss the properties of 6H4, another antibody which recognises bovine PrPres. Co-author Markus Moser says that a screening test based on 6H4 is in the final stage of development and could be in use for screening nervous tissue from all slaughtered Swiss cattle next year.

"A diagnostic test for humans using 15B3 might be around fairly quickly, but first we have to develop the procedure and then check that it is reliable", says Moser.
"A monoclonal antibody specific for the abnormal form of the prion protein is something that we have been wanting for a long time", notes John Pattison, chair of the UK Spongiform Encephalopathy Advisory Committee.

Accuracy of Figure 3b, 3c questioned

9 Nov 97 webmaster
My position is that Figure 3b and 3c of the Nature 390 74 1997 antibody article are simply cartoons that illustrate to the lay person how epitopes might be brought together in 3-space but have no relevence to the actual mechanism of rogue prion fiber formation.

The issue is, where are the articles with congo red binding to sickle cell hemoglobin? They are in short supply because, as noted, there is no beta structure in the picture, which congo red requires for the various reasons already I listed at atomic detail. Perhaps the 43% spectroscopic beta-sheet is pure artefact? I don't see the articles either showing these fancy fibers in the 16 congophilic disorders, none of which I recall are heterologous. Remember that beta hemoglobin has two much stronger evolved docking interfaces (beta-beta and beta-alpha) than val 6.

If trp 145 is the donor to the docking mechanism, then there needs to be an acceptor pocket that allows it to be buried in a hydrophobic milieu. My problem is that upon calculating for mouse which residues are exposed to solvent using SwissPdbViewer, I find tyr 169 [of epitope 2] as exposed at 52% while trp 145 is 48%. Additional hydrophobics at 35% or more exposed include leu125, met138, tyr155, val189, ala 224, and tyr226. In hamster, tyr 169 is marginally more exposed and trp marginally less.

Please don't get too carried away with trp because trp145 is conservatively changed to tyrosine 145 in human, cow, and sheep -- we need a more or less common structure. Tyr169 is invariant back to the earliest amniotes. This shows beyond any reasonable doubt that these residues are either not exposed in the native protein or exposed for some good reason -- there was plenty of time for mutational replacement by a polar amino acid.

The hydrogen-bonding indole nitrogen of trp 145 is fully exposed to water as is the pK 11.1 hydroxyl of Tyr 169. They have comparable hydrophilicities of 8.5 and 8.3 kcal/mol with hydrophobicities of 2.13 and 1.47kcal/mol, relative to phe at 2.27. Curiously, tyr 169 and trp 145 are at opposite ends of the molecule, perhaps what is needed for mutual burial, two birds with one stone and all that. Figure 3b has neither residue labelled; epitope 1 binds 3, not 2; both tyr 169 remain fully exposed (in a single strand fiber). I am not at all sure how they generated Figure 3b. To me it looks a little bit of a cover-up in its orientation choice.

I have to say that I cannot see any basis whatsoever for a hydrophobic pocket in the region around ser 132 to receive trp 145 in their docking model, there is nothing to hydrogen bond to the indole, and trp 145 is still quite exposed, nowhere near val 215, tyr 225, and tyr 226. Uncertainties of mouse C-terminal coordinates make me wonder what docking means hear anyway -- unfortunately this involves epitope 3.

Anyone can play this game as long as the energetics of docking aren't considered. One sees many opportunities for strain-types in the docking site and screw pitch. End-to-tail being a no-brainer given the properties of 15B3, I made a model that forgets the heme and edge-face stacks the rings of trp 145 and tyr 169.

Asn 181 and asn 197 are safely out of the way for bulky glycosylation in both models. Helices don't form docking interfaces; hydrophobic patches, salt bridges, domain swaps, domain extensions do, so why use the hemoglobin screw axis? The real question is, what orientation _optimizes_ the binding. But you will say, this is normal prion, protease-sensitive, and oxidized disulphide, so why try to dock it at all, especially while holding it rigidly to normal conformer (while the globin has undergone a shift). Or you will see that it is a better fit to alpha hemoglobin which doesn't sickle.

Riek could have calculated with docking freeware the free energy of binding of two prion monomers, for figure 3b. Surely a big Unix lab can spare 15 seconds of free computer time. Are there steric hindrances perhaps? Interesting if they dock perfectly in both hamster and mouse despite big structural differences in the critical epitope regions. If the portions of 3 helices line up so well, why not thread the side chains too, anchoring on 6-145? Maybe they only intended Figure 3b as a heuristic model to exhibit how distant epitopes might be brought together in a fiber, with the hemoglobin fiber and its sticky hydrophobic patch a casual metaphor.

I like a rather different quickie default model to deal with sticky patches. And that would be chaparone based. Indeed, we read [cover story, 21 Aug 97 Nature pg 746] that "this switch in the chemical character of the [chaparone's] central cavity lining [from hydrophobic to hydrophilic] triggers dissociation of the non-native polypeptide from the wall of the cavity. The released polypeptide is now free to re-initiate folding as a isolated molecules in a much-enlarged cavity that has a hydophilic lining conducive to burial of the substrate polypeptide's hydrophocbic residues as it initiates folding into a native structure." In our case, no native structure exists for an off-membrane prion proteolytic fragment so that product would be a generic default folded low energy state, that of anti-parallel beta sheet. This gets rid of both trp 145 and tyr 169 problems. [Note too that the ternery complex between rogue, normal conformer and chaparone, as envisioned in some "protein X" models, is physically impossible in any characterized chaparone superfamily.]

Let's assume for a moment that they are making a serious proposal and go through the exact drill needed to distinguish happenstance from homology:

Recall that beta hemoglobin E6V is a lethal mutation highly dysfunctional to normal hemoglobin function, though it has value secondarily with malarial resistance (in the heterozygous state). It is commonly said to create a sticky (exposed hydrophobic) patch on the surface that favors fiber formation. Many proteins have the odd hydrophobic residue exposed yet don't form rogue fibers. In the short-lived rbc, the concentration of hemoglobin is enormous. I don't know if chaparone proteins even exist in mature erythrocytes, a lot has been pushed out.

Globin hemes have a histidine ligand below the plane of the heme arising in helix 8, lost upon cutting down to helices 1, 6, and 7. If not covalently attached as in cytochrome c via two cysteines, a heme group could be lost during denaturing purification in GuHCl.

Trp145, which is Swiss dialect for trp144, is not a mutation or polymorphism at all nor a lab mouse sequence artefact, but a conservative substitution found strictly in (murinid) rodents, where it has replaced tyrosine stably for 35 million years. (This region is oddly transposed in the amniotes, DEYR to YEDR). Tryptophan is strongly planar and lacks the rotamer library of the hemoglobin valine for docking purposes. Why should this codon be so conserved if it is such a structural depravity? Of course, the idea is that this residue is here for a good reason, such as formation of a natural prion dimer, interacting with the repeat region, other proteins, or the membrane.

We have no idea if it is on the surface or aqueous-solvent exposed in native prion because no one has found conditions in which the pre-repeat or repeat has the rigid structure expected from its primary structure, evolutionary conservation, GSS mutations, and aptamer binding. The mouse nmr structure refers not to a valid species but an unnatural lab mutant; however trp 145 is also positioned anomalously in the hamster nmr fragment. This is fine if rogue prion fiber is not a property of native protein, but a property of a proteolytic fragment. Indeed chaparones take in exactly this type of anomaly -- misfolded proteins with exposed bydrophobic residues.

You have to wonder about a membrane-associated protein with 6% mostly periodic tryptophan (9/144), several standard deviations away from the norm in composition when the amino acid in question and its alternate are stackable. The Swiss can repeat as many times as they please that the N terminalt is random coil, but I don't see any impact from this to the final outcome.

For homology here, suppose ancient globin gene experienced a duplication, this redundant copy then experienced a deletion bringing helices 1,6, and 7 into sequence adjacency, with some retention of the two independent docking sites. Possibly the half-exposed heme was bound by a second rotated copy using a residual dimer interface, burying it properly and setting the stage for acquisition of a new function .... that later became the normal prion function. Signal peptide and GPI anchors are easy to come by, and the ancestral proto-repeat surely gave rise to the current repeat structure in any scenario. However, we have no source for the missing but highly conserved region 91-138 -- the hemoglobin started at position 6.

We have here three helices with a low rmsd. Any situation with two helices with somewhat aligned axes (say 3,000 entries at PDB) will invariably give a high score because the low score is driven by the constraints of helices alone. So we are just talking about choosing a third helix (really just its axis) that lies across these two, from all the other helices the class of two helix proteins might possess since we are allowing non-contiguity. Now we are down to only 500 unrelated entries with good rmsd and we pick a few that have the best additional fine-tuning coincidences. And of course it automatically does fairly well on hydrophobicity pattern agreement since interior-exterior tendencies are standard and already forced here by the helical alignment.

The non-alignment of primary sequences could be blown off as a consequence of the time since divergence. The globin genes have 27 invariant residues and 40 almost-invariant residues; the prion gene has its share of these too and is evolving much slower overall. If these slow residues don't align, the divergence would be very ancient indeed and the valine 6 [sometimes asp] metaphor quite a stretch -- a lot can happen in 2 billion years.

Researchers closer to test for mad cow disease

 Reuters World Report  Wed, Nov 5, 1997 By Patricia Reaney
LONDON, Nov 5 - Scientists said on Wednesday they are a step closer to developing a diagnostic test for mad cow disease and its human equivalent. Researchers at Prionics AG, based at the University of Zurich, said they have discovered an antibody for the mutated prion brain protein that causes the brain-wasting illness.
"There are two forms of the prion protein -- the normal form and the disease-specific form. There are many antibodies that react with the normal form, however none until now which only reacted with the disease-specific form," Bruno Oesch said in a telephone interview
. Until now, cases of bovine spongiform encephalopathy (BSE) in animals or Creutzfeldt-Jakob disease (CJD) in humans could only be confirmed by examining brain tissue from victims. The antibody, which attaches to the mutated prion, could allow doctors to test for the disease at an early stage by isolating the tiny mutated prions. Their aim is to develop a diagnostic blood test.

In a report in the science journal Nature, Oesch described how he and his colleagues isolated the antibody, 15B3, that can distinguish between the good and bad brain prions.

"The question is whether you can inject animals or humans who have the disease and whether it is going to recess," Oesch added. "Our primary goal is to have a diagnosis as early as possible and then once you have that you can think about therapy."
CJD is an uncommon but rapidly progressive form of dementia in humans that normally occurs between the ages of 40 and 65. In 1996 scientists discovered a new strain of the deadly disease that hits young people. So far 21 people have been diagnosed with the new CJD variant. Two studies released in September confirmed that BSE causes the new strain of CJD and that eating infected beef is the likely cause.

Scientists believe an epidemic of BSE in British cattle that resulted in a European Union ban on beef products last year was caused by giving the animals protein feed which included the remains of sheep infected with scrapie, a similar brain disorder.

"It will be interesting to see whether 15B3 will be able to neutralise infectivity and thus be a potential therapeutic reagent...a prion test for living humans or animals is conceivable," Oesch added.

Brain receptor role links Alzheimer's and smoking

 November 4, 1997 Jones and Yakel: Journal of Physiology Nov 1997
WASHINGTON - U.S. government scientists said Tuesday they had found a new function for a brain receptor that could link smoking, Alzheimer's and epilepsy. They found the receptor, known as the nicotinic receptor, on a brain cell in rats and said this shed light on the function of chemicals in the brain.

The brain cells are known as interneurons and are found in the hippocampus, the part of the brain linked with learning and memory, Susan Jones and Jerrel Yakel of the National Institute of Environmental Health Sciences in Research Triangle Park, North Carolina, reported. Each interneuron can affect thousands of "excitatory" neurons -- the workhorses of the brain -- so Jones and Yakel theorize that finding the receptor on these means a broader role in the brain.

"We are not exactly sure what role it plays," Yakel said in a telephone interview. "This class of neuron previously was not thought to have functional receptors."
Receptors are a kind of chemical doorway on the outside of a cell. This particular receptor is designed to let acetylcholine -- a neurotransmitter or message-carrying chemical -- deliver its information into the cell. It is also the receptor that nicotine acts on -- thus its name and its role in smoking. Yakel said the receptor could shed light on studies that show nicotine can sometimes help the memories of people with Alzheimer's. Alzheimer's patients have been shown to have fewer nicotinic receptors in their brains.
"There is pretty good evidence that some of the symptoms of Alzheimer's could be mitigated a bit by nicotine or nicotinic-type drugs," Yakel added. "That's a pretty good clue."
A rare form of epilepsy known as autosomal dominant nocturnal frontal lobe epilepsy is known to result from a mutation in the gene for one part of this receptor -- thus the link to epilepsy. Nicotine is also known to enhance cognition -- the basic brain function involved in learning and recognition.

New test for early diagnosis of CJD

 Nov 6 1997   BY NIGEL HAWKES Times
EARLIER diagnosis of CJD may be possible in a new test developed by Swiss scientists. It identifies the rogue protein that is the hallmark of the disease, and may be sensitive enough to pick it up in the bloodstream, and to check for the infection in beef.

Bruno Oesch of Prionics AG, based at the University of Zurich, reports in Nature that an antibody, called 15B3, has been found to bind with the aberrant form of the prion protein which is the common feature of diseases such as CJD. The antibody, a natural product made by the body as part of its defence mechanism, could even be used as a treatment.

"Our primary goal is to have a diagnosis as early as possible and then once you have that you can think about therapy," Dr Oesch said.
So far it has been used only for detecting abnormal prions in brain material, and improvement in sensitivity will be needed to test beef.

Nearly 300,000 tonnes of rendered-down cattle remains are being stored at about five sites around Britain, says the Government's Intervention Board. More than 1.5 million cattle have been destroyed since May 1996, because of potential infection with "mad cow" disease, and 13,000 are slaughtered each week. Talks are being held with power companies about using the waste as fuel in old coal-fired plants. One disposal company is burning waste at Fawley, Hampshire.

pH-dependent Stability and Conformation of the Recombinant Human Prion

J Biol Chem 1997 Oct 31;272(44):27517-27520 
Swietnicki W, Petersen R, Gambetti P, Surewicz WK
A recombinant protein corresponding to the human prion protein domain encompassing residues 90-231 was expressed in Escherichia coli in a soluble form and purified to homogeneity. Spectroscopic data indicate that the conformational properties and the folding pathway are strongly pH-dependent. Acidic pH induces a dramatic increase in the exposure of hydrophobic patches on the surface of the protein. At pH between 7 and 5, the unfolding ) in guanidine hydrochloride occurs as a two-state transition. This contrasts with the unfolding curves at lower pH values, which indicate a three-state transition, with the presence of a stable protein folding intermediate. While the secondary structure of the native structure is largely alpha-helical, the stable intermediate is rich in beta-sheet structure. These findings have important implications for understanding the initial events on the pathway toward the conversion of the normal into the pathological forms of prion protein.

Sequence sent in response to webmaster query by Dr. Bruno Oesch:

protein fragment:

partial nucleotide:

Prion research, the next frontiers

Nature 1 Nov 1997
The article in last week's Nature, "Prion research, the next frontiers', as the StarTrek title and post-prize stardate suggest, is not a review article per se, but an assessment of what still needs research clarification and what specific experiments might get us there. Some of these are quite clever and -- at least to me -- fresh ideas. There are many other excellent points (not discussed further here) that make the full article well worth reading. My comments here are organized to make little parodies of their section headings:

The generals are always fighting the last war.

The alt.theories do not warrant a diversion of mainstream research. There is plenty to do without catering to cranks, let's see where we are when the dust from 15B3 and 6H4 settles. AA/CW speak of using 'clean' eukaryote cells to satisfy virus people -- good luck with this and bon voyage. (There are 500,000 dispersed and diverged copies of ALU alone in the human genome.) They cite Karl Popper, but William of Occam already had the situation covered by 1347.

The not-invented-here syndrome.

The research directions could have been illuminated by consideration of the context of prion diseases within molecular biology, which is settling down to quaternary conformational disorders, or rather a subset of these, the congophilic cross-beta disorders, that includes 16 or more unrelated genes that seemingly shift to a universal default extensible architecture in the bad conformer, the inter-monomer anti parallel sheet.

Been there, done that (years ago) -- that's what transthyretin researchers are saying about prion in vitro conversion. Transthyretin research probably provides the best introduction to prion diseases, but who has the time? Jeffrey W Kelly has an excellent new review out, Structure 5(5) 595-600 1997.

Given the structures of these disparate proteins, can we extract principles that would allow a SwissProt screen of all proteins that have a propensity to form amyloid fiber? The tricky part is accommodating a proteolytically clipped [deglycosylatation is easier] product that actually participates in disease, as in Alzheimer or prion disease. This is a more contemporary genomics approach to human disease, it may be more efficient than grinding them out one at a time.

There is already fair success in automated bulk prediction of secondary structure, Chou and Fasman etc. etc. Some amino acids, notably gly and pro, simply don't like beta sheet, so we can safely ignore the suggestions of AA/CW that the repeat or pre-repeat region is the source of rogue beta or that it is rich in alternative conformers.

There is a lesson here too for normal prion function. The gene functions of some amyloid dementias are already known and they have nothing whatsoever to do with the nervous system. I have some problem grasping the role of buffy coat or spleen in vertebrate neurology. It seems the brain is just a wonderfully sensitive detector -- neurons accumulates damage over decades (unlike, say, a red blood cell) and we are immediately aware of the slightest change in behavior in a spouse, whereas a wayward liver protein might just be a vague malaise if that.

Given that the prion GPI anchor has been in steady use for 310 million years and more, that we have a complete inventory of GPI proteins in the yeast genome, that we can search GenBank for all GPI proteins (many of known roles and common domains), why bring up signalling and transport for normal prion function?

The good, the bad, and the ugly:

U1: No discussion of animal suffering [evidently irrelvent] and perversion of agriculture in the proposal to breed knockout livestock. We get fast food, the cow gets fatal familial insomnia. AA/CW see only two issues here: can gene ablation be carried out technically, and is it practical/useful. Will more high-tech agriculture get us out of the mess that high-tech agriculture got us into?

Does NIH really need to feed BSE to chimpanzees in 1998 -- what could this possibly prove? As far as I am concerned, they have killed far too many chimpanzees already. What comes next -- feeding BSE to prisoners in China?

B1: Yeast prions do not form a 'non-functional aggregate' as far as anybody knows -- those are fighting words to a protein chemist. Aggregates are the complete antithesis of oligomer, fiber, or crystal -- you get aggregates by boiling an egg or going to pH 2. What I see published are good clean electron micrographs of two highly ordered 2D crystals representing stable conformational strain types and efficient rescue of amber mutants.

B2: Intramolecular crosslinks are 'unorthodox approaches' to structure determination? -- they have been used to good effect for decades by thousands of labs on thousands of proteins. Consider transthyretin for example.

G1: The observation that the prion gene product may be functionally replaceable if it is absent at conception but not if it is knocked out abruptly in adulthood. By spotting onto microchips, it might actually be possible to compare mRNAs, to implement the authors' suggestion for identifying non-orthologous replacements.

G2: "It is sobering to realize that very little is understood about the mechanisms by which prions elicit brain damage." So true, this is where the story breaks down. This area seems more important to me for prevention/therapy than how the immune or peripheral nervous system are involved in transmission -- amounts to locking the barn door after the horse is gone, for BSE and nvCJD.

G3: The observation that Prp(SC), protease K resistance, Prp*, conformational intermediates (and dead ends) are totally scrambled in practical use and hardly synonymous with the ultimate infective bad actor. I don't see how to get everyone on the same page terminologically. There could be many, many protein conformational states in equilibria with each other.

G4: Good to see some older cites to Cuille and Chelle [scrapie transmission, 1939], Alper, Griffith. There is inevitably a certain amount of Whig history in selecting papers that look good later, this was brought home to me by some 30-year old reprints kindly sent in by Peter Lewin, now at Sante Fe. He proposed a protein-only theory that reversed the last step of the central dogma. The situation was quite murky at the time Prusiner got behind protein-only. I also got an interesting sugars-rule glycosylation driven strain type suggestion more recently.

G5-G25: Read the full text. It is an excellent treatment overall of a tough assignment.

Mad Cow Home -- Best Links -- Search this site