9 new prion gene mutations include T188R and P238S
Normal prion protein has an activity like that of superoxide dismutase.
Protocadherin II: cell surface receptor for normal prion
A little more at Caprion's web site
Protocadherin backgrounder
Structure of the protocadherin extra-cellular domain
200-truncated PrP is present in mature sperm
Glycosylation differences between the normal and pathogenic prion
Role of Hsp70-related chaperone in yeast psi prion
Molecular genetics of human prion diseases in Germany Human Genetics, 105:244-252 1999 Windl O, Giese A, ..., Kretzschmar HAComment (webmaster): A complete set of published prion gene mutations organized by type are available at:
How do we justify the lack of a preprint server to families affected by inherited disorders? The clock is ticking for younger family members (as they are quick to tell you). If sporadic CJD is diagnosed in a grandparent, the parent has a 2.5% chance of dying from CJD [assuming autosomal dominance, half-penetrance, and 10% of sporadic familial] and the child 1.25 %. These are not reassuring numbers like meteorite hits or one per million.
This paper reports a highly significant methodological advance in high-quality screening for mutants in the prion gene in 578 patients. This advance can be safely inferred from the 5 new silent mutations they found (each in just one patient):
P68P (CCT-CCC) at the start of the third repeat Q212Q (CAG-CAA), which site also has Q212P disease elsewhere R228R (AGA-AGG), S230S (TCG-TCA), of CpG type +4 post stop codon (A-C), the first known 3' UTR mutation tga ggaA.The webmaster checked at GenBank dbEST and non-redundant to see if this 3' UTR human allele had been seen in those surveys but it had not: 0 in 14. There is a fair amount of reported phylogenetic variation early on 3': human tga ggaaggt is tga ggGaggC in mouse, tag gggcaac in cow, tag gggcaac in sheep and so on.
For known silent polymorphisms, 5.4% had the A117A (A-G), 0.5% had G124G (C-G), and 0.9% had an octapeptide deletion. The incidence of the first allele is often over-stated due to use of a primer that includes an associated 5' UTR allele. The second allele has no prior estimate of incidence. The incidence of repeat deletion is within the usual range of 1-2% so serves as a good internal control on sequencing quality.
This suggests that at a minimum 3-4 of the nvCJD victims have been incorrectly characterized.
While none of the silent mutations is a smoking gun, it is not at all safe to assume that they are neutral. There is the matter of upstream intron skipping to doppel, the internal conserved hairpin, mRNA stability, tRNA preferences, translation efficiency and termination, as well as issues we don't even know about.
Looking at codon use in humans, we see strong preference (2x) for the wildtype GCA alanine 117 codon over GCG. The wildtype codon is also favored at codon 124 (1.4x). Codon 68 and codon 228 have insignficant preferences; codon 212 favors wildtype by 2.9x; and codon 230 favors the mutant (2.5x). In other words, changing silent position could affect translations per mRNA lifetime.
Since the number of prion genes screened was only a small fraction of the worldwide total, many other studies are obviously completely missing silent polymorphisms or failing to report them. Missing coding mutations of course inflates the amount of 'sporadic' CJD.
The two new coding mutations are T188R (ACG-AGG so another CpG) and P238S (CCA-TCA): this was blood-based; neither patient had died. The diagnostic class was not given (probable CJD, possible, other) nor the status of codon 129. However, the authors note that these are highly non-conservative substitutions:
T188R cvnitikqhTvtttt at the end of helix 2 is indeed conserved in all mammals; not in bird or doppel. H187R was recently reported nearby.
P238S Ppvillisfliflivg-stop is well into the GPI-cleaved region but despite this is conserved in all mammals though not in bird and glutamine in doppel. M232R is an earlier mutation not found in the mature protein, though GPI addition may not go to completion.
The webmaster updated the point mutation collection to reflect these 7 additions to known point mutations, bringing the total to 28 different sites (7 silent, 3 neutral coding, 18 causative).
Five cases of extra repeat insertions were found. One of these, a 9x, had been reported previously. The other four cases were two new types of 5x found in two patients each. Iterated replication slippage giving rise to extra repeats is a constrained but somewhate out-of-control process that results in considerable microheterogeneity of the end product. However, accompaning point mutations are highly restricted by location within the 24bp repeat domain and only internal repeats participate.
R1 R2 R2 - R3g R3g R3g R2 R2 - R3 R4
R1 R2 R2 - R3 R2 R2 R2a R2 - R3 R4
Kindreds were not pursued here beyond the immediate family level, so given the unliklihood of a given repeat arising twice in a small sample, the repeat case multiplicities suggest single kindreds. Clinical details on the 3 cases of the first repeat showed hopeless phenotypic variability: spongiform, blurred and fleecy immunostaining, plaque, and punctate depositions.
The webmaster updated the extra repeat collection to reflect these 2 additions to known repeat mutations, bringing the total to 27 different kinds (from 1x to 9x; causative at 4x and beyond).
Five cases of single repeat deletion were found but not characterized. There are 5 known classes of these, determined by details of the slippage misalignment. The webmaster updated the repeat deletion collection to reflect these 5 additions to known repeat mutations, bringing the total to 5 different kinds in some 68 pedigrees.
This brings the grand total of known genetic events in the human gene to 62 (including the 2 UTR alleles). The combinatorics of 28 point, 5 deletion, 27 insertion, and 2 intronic alleles on two gene copies are such that 2.1 x 10 exp 37 genotypes could occur.
Deletional and insertional repeats are common in prion genes of other species of animals. (This phenomenon is also widespread in other genes with internal repeats.) In fact, 4 and 6 repeats have polymorphism status in some lineages. Recently 3 repeats in wild Siberian goats and 7 repeats in cow have been reported. It seems very likely that every year a number of cows will born worldwide that are within the pathological repeat zone. As most cows are recycled in part as feed for other mammals, this provides for feedback amplification of the better strains.
The webmaster checked in at OMIM and SwissProt to see how well they were tracking alleles in the human prion gene. Despite a 3 Nov 99 update, OMIM only carried 10 of 27 insertions, 14 of 21 known coding mutations, 1 of 8 of the silent mutations, and 0 of 5 of the deletion classes. See *176640 Prion Protein; Prnp
OMIM does not welcome reporting of errors and omissions and makes very sure there is no email address for an editor. Once the webmaster wrote them pointing out that they had confused thiamine (the vitamin) with thymidine (the nucleoside) site-wide but never heard back.
On the whole they do a good job -- for example, doppel is covered reasonably well. However Linux proved that the community-approach works better than single-editor.
SwisProt has 15 of the 21 known coding point mutations. However the attitude is much more constructive -- there is actually an established and encouraged mechanism by which additions/corrections can be funneled in and a dialogue established.
GenBank is approaching total meltdown: reporting of mutations is completely haphazard and sheep and cervids prions are in much better shape than humans. No one but the original wet lab can submit, comment on, correct sequences, or update annotations. If the principal investigator has died, the entry is in its final form. GenBank by policy refuses to carry email addresses to facilitate contact with authors of entries.
Other than Prusiner's group, the webmaster haa yet to see a CJD researcher attempt a correct account of the current state of prion mutations (and of course reviewers could care less). It is appalling to see people ignore half the published literature and too lazy even to copy them down off a maintained web site where they are already collected along with the necessary cites.
Notice what happens now when someone writes a medical review article about disease alleles per gene or mutation distribution along genes. They go with what is at the repositories and Medline (no access to full text). The underlying data being flawed, so is the synthesis. The net result is error propagation and amplification.
Known point mutations were also found in the patient set: P102L, D178N, T183A, E200K, and V210I (15 cases total). Better ascertainment leads the authors to conclude that the percentage of CJD attributable to genetic factors is significantly higher than previously thought, from 12.8% to 17.8%, depending on how denominator total (definite, probable, ...) is chosen. Genetic cases tend to be atypical (ie, under-diagnosed) vis-a-vis conventional diagnostic protocols, as do valine homozygotes. France had more E200K but less D178N than Germany.
There is quite a bit of codon 129 this-and-that for which the article should be consulted. The main issue is cis-trans determination and reporting. The method, non-radioactive SSCP from multiple primers, does allow accurate determination of case genotype. In other words, each cell has two copies of chromosome 20 and two copies of the prion gene. One or more of several common polymorphisms may be present along with a serious mutation -- the question is how the variation is allocated across the two copies of the gene (and whether phenotype is affected).
A horrific situation has developed: Some groups determine only the values of codon 129 ignoring other sites, some groups do not even determine which of M/V was on the mutant chromosome, other groups (like here) make the determination but only partially report it (for lack of space in a print journal). No one is reporting the value of the doppel polymorphism. It would be best to report cis and trans alleles in a table, one line for each patient chromosome.
Thus we have only the dimmest idea of how much the genotype determines phenotype -- much of the reporting lumps different genotypes. This is not an academic exercise in prion disease, because the like-like principle governs recruitment to a considerable extent. And that principle may extend to doppel interactions.
No CJD genomic sample repository exists enabling a researcher to go back and resequence to the necessary standard. That standard should really include upstream exons, promoters, and doppel. It would really make more sense to send samples to a central cloning and sequencing facility. Searching OMIM for "consortium" turns up over 50 other rare diseases where a culture of cooperation has been established.
Kindreds are often poorly researched, either by interview or microsatellite markers. Note silent alleles are much more useful for founder finding and dating than microsatellites. This is not an idle exercise because if a therapy existed we need to get at the extended kindred to deliver it before the onset of symptoms. For example, if the 5x repeats had distant family connection, that extended kindred could be huge with half of them at risk (not to mention their blood and organ donations). These kindreds are really the only testing ground for therapies in non-dietary CJD -- no country will propose comprehensive brain biopsies on a healthy population.
Clearly, prion genotyping should be done from blood whenever there is an outside chance that the patient has CJD prior to invasive (or even expensive) diagnostic procedures.
Ann Neurol 1999 Nov;46(5):693-70 Skworc KH, Windl O, Schulz-Schaeffer WJ, Giese A, Bergk J, Nagele A, Vieregge P, Zerr I, Poser S, Kretzschmar HAThe clinical course, neuropathological features, and genetic findings in 3 members of a German family carrying a novel 120-bp insertion in the prion protein (PrP) gene are described. Genetic analysis of the mutated allele revealed a sequence of five extra octapeptide repeats, distinct from those of the two previously reported families with an insertion of this size.
There was distinctive variation in the clinical course and the onset and duration of the illness in the documented subjects. Neuropathological evaluation showed neuronal loss and gliosis in the neocortex of the 3 examined cases; spongiform degeneration was found in 2 of them. PrP immunoreactivity of unusual morphology and distinct distribution was present in the cerebellum and neocortex ("blurred staining") of 2 examined cases. One subject showed features usually found in sporadic Creutzfeldt-Jakob disease with a punctate type of PrP deposition in the cerebellum.
In addition, there were some plaque-like PrP aggregates morphologically similar to the other 2 cases in the molecular layer of the cerebellum, and unusual PrP immunoreactivity ("fleecy staining") was found in the neocortex. The clinicopathological heterogeneity in the documented family is in accordance with the phenotypic variability associated with previously reported insertions.
Biochem J 1999 Nov 15;344(Pt 1):1-5 Brown DR, Wong BS, Hafiz F, Clive C, Haswell SJ, Jones IMWe show here that mouse prion protein (PrP(C)) either as recombinant protein or immunoprecipitated from brain tissue has superoxide dismutase (SOD) activity. SOD activity was also associated with recombinant chicken PrP(C) confirming the evolutionary conserved phenotype suggested by sequence similarity. Acquisition of copper by PrP(C) during protein folding endowed SOD activity on the protein but the addition of copper following refolding did not. PrP(C) dependent SOD activity was abolished by deletion of the octapeptide-repeat region involved in copper binding. These results describe an enzymic function for PrP(C) consistent with its cellular distribution and suggest it has a direct role in cellular resistance to oxidative stress.
Prusiner's group reportedly also cannot get even-handed peer review from Nature. They gave up in 1992, some 181 papers ago. Here, there may have been failure to demonstrate a sufficiently respectful attitude to hardline viral theorists.
The Biochemistry Journal piece offers convincing evidence (with excellent controls) that normal prion has superoxide dismutase activity, but properly does not conclude SOD is all or even part of normal function (though it well may be; the authors favor this interpretation).
One residual concern concerns "escaped" activities for any copper binding protein perhaps not quite refolded to native conformation; indeed, the two covalently modified arginines in the pre-repeat region would missing from recombinant protein. Superoxide dismutation also procedes very rapidly on its own or with the help of small molecules [eg, EUK-134, a synthetic superoxide dismutase/catalase mimetic , PNAS 1999 Aug 17;96(17):9897-902], raising concerns about loose or non-specifically bound copper catalysis.
However, the specific activity on a per-peptide basis is comparable (15-30%) to conventional Zn/Cu superoxide dismutase according to figure 2c. This, along with numerous prior studies on copper and oxidative damage as well as localization on synapses and mature sperm, provides a somewhat persuasive picture for superoxide dismutase as a normal function.
Prion globular domain may have additional kinds of binding sites; doppel lacks an evident metal binding site. Once the repeat region is loaded with copper, it may fold back to bind the concavity of the globular domain, either of prion or -- interesting suggestion of Brown -- to doppel in a domain-swapped heterodimer of modulatory signifcance.
Other metals found in known superoxide dismutases (zinc, manganese, nickel, iron) were at inconsequential levels even in wildtype immunoprecipitant. Technical problems related to artefactual copper binding of the hexahistidine tag and non-repeat region binding were satisfactorily addressed. The webmaster proposed a geometry for an all-copper repeat for bird and mammal prions on 6 June 98 that does not need revision.
No catalase activity -- another potential escape activity directed at hydrogen peroxide -- was detected. Singlet oxygen, superoxide reductase, and peroxynitrite dismutase activity were not assayed. Copper in enzymes in all cases binds small reactive di-oxygen. (A superoxide reductase using NADPH and rubredoxin was newly described by Jenny FE et al Science 1999 Oct 8;286(5438):306-309 -- peroxide is formed in place of oxygen.) The oft-proposed copper uptake role would also detoxify copper by taking it out of sensitive areas; the recent repeat structural study by Miura et al found a supporting pH-dependent conformational change in the copper repeat.
The paper found superoxide dismutase activity in both bird [M95404] and mammalian prion. This says the activity is quite old but does not clarify how it could have been present in the common ancestor of distinct hexapeptide and octapeptide repeat proteins. What function did the protein have in that era if not SOD?
Controls, too numerous to describe properly here, include prions with repeat deletions, copper sulfate vs cupric chloride, and effects of copper chelator (diethyldithiocarbamate, yes) and conventional SOD inhibitor (KCN, no). There are hints of other experiments for which there was no room in the paper: localization of active site with antibodies, pH variation encompassing the Miura effect, circular dichroism studies of copper-induced structural change, and correlation with observations in the earlier methionine sulfoxide paper [Wong BS et al. BBRC 1999 Jun 7;259(2):352-5].
Going by past interests, the next move will probably be to look at cell survival vis-a-vis superoxide in various mutants in the presence of various inhibitors.
27 Oct 99 CHI meeting in DC. Account by Alex Bossers + Proceedings Neil Cashman Univ of Toronto/Caprion PharmaceuticalsA high-affinity cell surface binding site for normal prion has been identified using PrP-alkaline phosphatase fusion proteins expressed in frog oocytes. The binding was 10x stronger than for a typical monoclonal antibody. The gene encoding this protein (called sPC2, a name already in use for 61 unrelated GenBank entries) is said cloned but is not at GenBank (though 522 other protocadherin sequences are). Protocadherin II is a 115k type I cell-surface receptor with one trans-membrane span whose structure is stabilized by calcium.
Using a library of in vitro transcription/translation, from pools and sub-pools they identified a single protein (sPC2) which bound to PrP. The N-terminus of PrP was important for recognition by the receptor and was the binding was dependent on copper. Recombinant bovine PrP bound as well as the fusion protein to sPC2. PrP-Sc bound even better (2x).
The protein sPC2 is not detectable (by FACS -- fluorescein activated cell sorter) on CD3 positive lymphocytes (T cells) while it was on B-cells. This surface recptor could thus play a role in oligomerization of PrP at the cell-surface.
Affinities for PrP were determined in an ELISA based test on brain homogenate without detergent (no deoxycholate even). Follicular dendritic cells (FDCs) have not yet been studied -- these have figured prominently in other studies of the interaction of prions with the immune system.
The richest source of sPC2 seems to be brain neurons for both mouse and humans (mouse, human). No infected/not-infected brains have been studied at this time. Mice overexpressing sPC2 (human) are being constructed; no knock-outs are available yet. There may be further receptors in addition to sPC2.
The meeting abstract states that it may be possible to base diagnostics and therapeutics upon this speicifc high-affinity cell surface receptor.
Comment (webmaster): The interaction of prion and receptor could be very useful in diagnostics but it does not necessarily imply a role in normal prion function despite the high binding constant. Blastp searches with the repeat region turn up hundreds of connective proteins with high glycine and proline content; protocalhedrin may recognize one of these extra-cellular matrix proteins instead (witah cross-over to prion repeat).
Nature Medicine Nov 99 pp 1308 - 1312 + N&V + press release K L Brown, K Stewart, D L Ritchie, N A Mabbott, A Williams, H Fraser, W I Morrison & M E BruceUnderstanding how prions, the infectious particles responsible for scrapie and bovine spongiform encephalopathy, or mad cow disease, take hold on the body is vital if a treatment for the disease is to be developed. Now a team of scientists at the institute for Animal Health, Edinburgh has identified the type of cells required for infection.
Moira Bruce and colleagues show that cells of the immune system called follicular dendritic cells (FDC), which present foreign material to other cells, and which depend on B cell signals for development, are required for scrapie prion replication. Using mice with an undeveloped immune system (SCID mice), the group found that only those with FDC cells in the spleen carrying a normal cellular prion could be infected. These cells may be a target for drug intervention.
In an accompanying News & Views article, Man-Sun Sy and Pierluigi Gambetti from Case Western Reserve University, discuss the plausibility and implications of the findings.
M Balter wrote in the 22 Oct 99 Science, that a " series of new experiments presented in Tübingen by Moira Bruce of the NPU in Edinburgh may help point to a solution of the riddle--one that implicates both B lymphocytes and FDCs in the spread of prions. Using a combination of gene knockout techniques and grafts of bone marrow--where the precursors of many immune cells develop--Bruce and her colleagues succeeded in creating two groups of chimeric mice with mismatches in the PrP status of their immune cells. One group had PrP-positive FDCs but PrP-negative B lymphocytes, whereas the other had PrP-negative FDCs but PrP-positive lymphocytes. When inoculated with scrapie, only the FDC-positive group could be infected."
3 Nov 99 CaprionThe Caprion website mentions a goal of "identifing relevant indications for compounds which disrupt prion binding to its receptor.... Caprionıs technology platform in prions include i) unique reagents to manipulate and measure prion protein binding, ii) the discovery of a prion receptor and, iii) a range of prion specific antibodies which are currently being developed to detect prions in Mad Cow Disease."
"Prion protein-alkaline phosphatase fusion proteins: hese novel detection reagents enable the stable manipulation of various species of the prion protein ("PrP") while maintaining their native conformation and enable the localization and quantification of PrP binding."
"Prion receptors: Caprionıs CPx is the first cell surface protein cloned that functionally binds to PrP with high affinity. This known cell surface receptor displays saturable and competable high affinity binding to the prion protein (recombinant PrP, mouse and human PrP-AP and brain homogenate) at low nanomolar range. This interaction, combined with the high sensitivity of the reagent, make it the ideal reagent for the detection of PrPSc in blood."
" CPx also serves as an ideal target for the development of therapeutic compounds. The interaction of PrP with CPx has been shown to affect several key biological phenomena, including neuronal apoptosis. Based on the specific binding domains of PrP to CPx, compounds have been identified which antagonize this interaction. These compounds, coupled with binding site modeling of CPx, will form the basis of rational drug design in parallel with high throughput drug screening with small molecule libraries."
"Prion antibodies: Caprion has developed a unique family of polyclonal and monoclonal antibodies against specific epitopes of the prion protein. Certain candidate antibodies may serve as secondary detection reagents in an ELISA-based diagnostic assay using CPx as the primary capture reagent. Other antibodies are currently being developed against unique PrPSc epitopes."
Shape-shifter assays to prevent recruitment: Several targets which qualify as shape-shifters have been identified. These include some well known proteins which have been implicated in a specific disease etiology but which have not been previously identified as acting by a shape-shifter mechanism. Caprion is developing unique in vitro and in vivo assays which mimic the recruitment and conversion process of these proteins, thereby enabling the screening of compounds which would disrupt these processes and therefore retard and/or arrest disease. In contrast to other drug discovery paradigms, Caprionıs therapeutics are directed toward targets which are, by definition, fully validated. Indeed, Caprion has developed evidence which suggests that the degree of (shape-shifting) protein recruitment is proportional to disease progression, thus validating these targets as effective determinants of disease progression."
An August press release provides some further details:
WESTBROOK, Maine; August 30, 1999 IDEXX Laboratories, Inc. and Caprion Pharmaceuticals, Inc. announced today that they will collaborate on the development of Caprionıs proprietary reagents for the diagnosis of Mad Cow Disease and other Transmissible Spongiform Encephalopathies (TSEs).
IDEXX will receive exclusive global rights for veterinary applications in diagnostics and therapeutics, while Caprion retains all rights to human diagnostic and therapeutic applications. IDEXX has made an equity investment in Caprion, and will support research at Caprionıs Montreal-based laboratories. IDEXX will also conduct development work in its laboratory in Westbrook, Maine, contributing expertise in diagnostic test development. IDEXX will market veterinary products resulting from the collaboration and pay Caprion royalties on product sales.
The goal of the collaboration is development of a rapid diagnostic for detecting Bovine Spongiform Encephalopathy (BSE), commonly known as Mad Cow Disease, in live cattle. Current detection methods require brain tissue samples from slaughtered animals. A live animal diagnostic would allow governments and producers to implement cost-effective surveillance programs to protect and validate disease-free status. ..
"BSE and other TSEs are important health concerns," stated Dr. Erwin Workman, executive vice president and chief scientific officer of IDEXX. "The veterinary market currently has no proven diagnostic for BSE in live cattle. Caprionıs reagents and skills in the prion field position us well to take on this important challenge in veterinary health care." ..
Caprion Pharmaceuticals Inc. is a privately held Montreal-based biopharmaceutical company exploiting "shape-shifting" proteins as targets for diagnostics and therapeutic intervention. Prion proteins are the best-known shape-shifters and are the focus of Caprionıs earliest products. Caprion is using its expertise in protein conformational change as a key interventional target in the development of neurological and immunological therapies. Caprion Pharmaceuticals has established an alliance with IDEXX Laboratories, Inc., the worldıs leading veterinary diagnostics and therapeutics company. IDEXX has obtained a world-wide, exclusive license to Caprionıs technologies for diagnosis and treatment of Mad Cow Disease and its equivalents in veterinary markets. IDEXX has also acquired a 10% ownership stake in the company. Caprion will receive over $1 million in R&D support for product development, in addition to an equity investment, milestone payments and royalties on sales.
Caprion is in advanced stage discussions for collaborations on human blood screening and decontamination markets of its prion technologies, and is currently seeking partners for the commercialization of its technologies in the human blood market, as well as R&D partnerships for its therapeutic leads under development
Denise Sposato IDEXX Laboratories, Inc. One IDEXX Drive Westbrook, Maine 04092 (207) 856-0322 (207) 856-0319 (fax) www.idexx.com Lloyd M. SegalI Caprion Pharmaceuticals, Inc. 5375 Pare Street, Suite 201 Montreal, Qc, Canada H4P 1P7 (514) 940-3604 (514) 874-9077 (fax) www.caprion.com
2 Nov 99Comment (webmaster): The first question that comes up is why was this missed by yeast two-hybrid screens and similar methods tried earlier to find proteins that bound to prion protein? Possibly the transmembrane feature caused protocadherin not to be properly expressed in yeast. Little has been heard lately concerning the laminin receptor.
It is a mystery how the extra-cytoplasmic domain of protocadherin could specifically recognize prion protein (which is not to doubt their report). That mainly consists of 5-7 repeated calcium binding motifs. Is the prion repeat t presenting its copper or zinc (or calcium???) to the protocadherin? Recall the Japanese study showing prion-prion dimerization at lower pH through a shared metal ion.
Protocadherins at Medline:
Biochem Biophys Res Commun 1999 Jul 14;260(3):641-5 Hirano S, Ono T, Yan Q, Wang X, Sonta S, Suzuki ST SA.Using cDNA of human protocadherin 2A (pc2A; originally known as protocadherin 2) as a probe, we cloned a new member of the protocadherin 2 subfamily from mouse brain cDNA libraries and named it protocadherin 2C (pc2C). It was similar to pc2A throughout its entire coding region, and its C-terminal region was highly conserved. The locus of the pc2C gene was on the mouse chromosome 18C where the pc2A gene is located [probably corresponds to human 18q11.2], suggesting that genes of the pc2 subfamily form a gene cluster.
The expression of pc2C was restricted to the nervous system, and the expression started in the embryonic stage and increased up to the adult stage. The expression pattern was quite similar to that of OL-protocadherin, a distinct class of protocadherin, although the timing and relative strength of expression were different. These results suggest that pc2C may be involved in neural development along with other classes of protocadherins.
MQRREDCKGLVFLGVLLGMLWKAGSTQIRYSVPEELEKGSRVGN ISKDLGLKSCELAERGVRIVSRGRAQLFALNPRSGSLVTAGRIDREELCMGSIKCQLN LEILMEDIGKIYGVEIEVRDINDNAPYFRESELEIKMSENAAAGMRFPLPHAWDPDIG KNSLQSYQLSPNAHFSLEVQNGADGNKYPELVLESSLDREEKAAHLLVLTASDAGDPV RTGTARIRVMVVDANDNAPAFARSEYRVSWENVAVGTQLLLVNATDPDEGANAEVIYS FRYVDDKAAKVFKLDSNLGTISTIGELNHEESGFYEMEVQATDNAGYSARAKVLVTVL DVNDNAPEVAITSLTNSVPENSPQGTLIALLNVNDQDSGENGQVICSIQENLPFKLEK SYGNYYRLVTDAVLDREEVPSYNITMTATDRGSPPLTTETHLALDIADTNDNSPVFLQ ASYWAYIPENNPRGASIASVTAHDPDSDKNAQVTYSLAEDTHQGVPLSSYVSINSDTG VLYALHSFDYEQFPDLQLQVIARDSGDPPLSSNVSLSLFVLDQNDNVPEILYPTLPTD GSTGVELAPRSAEPGYLVTKVVAVDRDSGQNAWLSYRLLKASEPGLFSVGLHTGEIST ARALMDRDALKQNLVVSVQDHGQPPLSATVTLTVAIADSIPDVLADLDNLESPSTSET SGLTLYLVVAVAAVSTVFLVFVLVLLALRLRRWHSLRLLRAGGPGLADVAASHFVGVD GVQAFLQTYSHEVSLTAGSRKSHLIFPQPNYADTLISQESCEKTEPLLPSGDSVFSKD NHALNQQAPPNTDWRFSQAQRPGTSGSQNGDETGTWPNNQFDTEMLQAMILASASEAA DGSSTLGGGAGTMGLSARYGPQFTLQHVPDYRQNVYIPGSNATLTNAAGKRDGKAPAG GNGNKKKSGKKEKK
The mouse sequence has 132 Blastp hits in humans: gb|AAD43713.1| (AF152319) protocadherin gamma A10 [Homo sap... 1388 0.0 gb|AAD43714.1| (AF152320) protocadherin gamma A11 [Homo sap... 1379 0.0 gb|AAD43718.1| (AF152324) protocadherin gamma A4 [Homo sapi... 1357 0.0 gb|AAD43719.1| (AF152325) protocadherin gamma A5 [Homo sapi... 1344 0.0 gb|AAD43767.1| (AF152506) protocadherin gamma A12 ... 1341 0.0 gb|AAD43712.1| (AF152318) protocadherin gamma A1 [Homo sapi... 1333 0.0 gb|AAD43723.1| (AF152329) protocadherin gamma A9 [Homo sapi... 1328 0.0 gb|AAD43717.1| (AF152323) protocadherin gamma A3 [Homo sapi... 1325 0.0 gb|AAD43721.1| (AF152327) protocadherin gamma A7 [Homo sapi... 1324 0.0 gb|AAD43720.1| (AF152326) protocadherin gamma A6 [Homo sapi... 1321 0.0 gb|AAD43716.1| (AF152322) protocadherin gamma A2 [Homo sapi... 1319 0.0 gb|AAD43722.1| (AF152328) protocadherin gamma A8 [Homo sapi... 1313 0.0 ....
J Neurosci 1999 Feb 1;19(3):995-1005 Hirano S, Yan Q, Suzuki STWe cloned a novel protocadherin cDNA, which we named OL-protocadherin (OL-pc), from mouse brain cDNA libraries. Its cytoplasmic region showed no similarities to other protocadherins, indicating that it belongs to a novel subfamily of protocadherins. The molecular mass of OL-pc was 140 kDa in the brain. Expression of OL-pc mRNA was specific to the nervous system, changing over time from the embryonic stage to the adult stage. The OL-pc expression seemed to be restricted to a subset of functionally related brain nuclei and regions such as the nuclei in the main olfactory system, the limbic system, and the olivocortical projection. There were at least two distinct patterns of distribution for the OL-pc protein. First, it was localized in particular brain nuclei or compartments, such as the stripes of the developing cerebellum. Second, it was found at the synapse in regions such as the glomeruli of the olfactory bulb. In addition, the OL-pc protein seemed not to be detected or was detected only weakly in some regions, such as hippocampus in which the mRNA was expressed at high levels. These results indicate that the expression of OL-pc is developmentally regulated in a subset of the functional systems and that it may be involved in the formation of the neural network by segregation of the brain nuclei and mediation of the axonal connections.
To clarify the structure of the proteins defined by the new clones, a putative entire coding sequence corresponding to one of the clones was determined. The overall structure is essentially the same as Pcdh2, indicating that the proteins defined by this clone, and probably by other clones, belong to the protocadherin family. Two PCR experiments and an RNase protection assay showed the existence of the corresponding mRNAs in rat brain preparations. Human and mouse cDNA clones with the same sequence properties were also isolated.... Since protocadherins encoded by these mRNAs are likely to have different specificity in their interaction and share a common activity at their cytoplasmic domains, these protocadherins may provide a molecular basis, in part, to support the complex cell cell interaction in brain.
Genomics 1998 Oct 1;53(1):81-9 Strehl S, Glatt K, Liu QM, Glatt H, Lalande MThe protocadherins are a subfamily of the calcium-dependent cell-cell adhesion and recognition proteins of the cadherin superfamily. In this study we describe the isolation and characterization of two novel protocadherins, PCDH8 and PCDH9, that constitute a new linkage group on human chromosome 13 and mouse chromosome 14. Like other protocadherins both genes are predominantly expressed in brain, but PCDH9 is also expressed in a broader variety of tissues, and the expression patterns appear to be developmentally regulated. We have determined the genomic organization of PCDH8, which differs significantly from that of the other cadherin subfamilies. In contrast to the classical and desmosomal cadherins, which in general consist of 15-17 exons and share a remarkable degree of conservation in intron position, PCDH8 consists of only three exons and lacks introns in the extracellular domain. The first exon encodes the extracellular domain, the transmembrane region, and part of the cytoplasmic tail. The second exon encodes the remainder of the cytoplasmic region and is partially untranslated. The differences in the genomic structure of cadherin subfamilies will be discussed in the context of the evolution of the cadherin superfamily.
J Cell Sci 1996 Nov;109 ( Pt 11):2609-11 or J Cell Biochem 1996 Jun 15;61(4):531-42 Suzuki STRecent cadherin studies have revealed that many cadherins and cadherin-related proteins are expressed in various tissues of different multicellular organisms. These proteins are characterized by the multiple repeats of the cadherin motif in their extracellular domains. The members of the cadherin superfamily are divided into two groups: classical cadherin type and protocadherin type [extracellular domains are composed of six or seven repeats; little cyotplasmic domain homology] .
The current cadherins appear to have evolved from a protocadherin type. Recent studies have proved the cell adhesion role of classical cadherins in embryogenesis. In contrast, the biological role of protocadherins is elusive. Circumstantial evidence, however, suggests that protocadherins are involved in a variety of cell-cell interactions. Since protocadherins, and many other new cadherins as well, have unique properties, studies of these cadherins may provide insight into the structure and biological role of the cadherin superfamily. ... Cadherin repeats roughly correspond to the folding unit of the extracellular domains. The common ancestor might have cadherin repeats similar to those of the current protocadherins, and to have common functional properties as Ca(2+)-dependent homophilic adhesion protein [ie, bind to a second monomer on the second cell.]
Cadherins at SwissProt:
"Cadherins are calcium dependent cell adhesion proteins. They preferentially interact with themselves in a homophilic manner in connecting cells; cadherins may thus contribute to the sorting of heterogeneous cell types. n-cadherin may be involved in neuronal recognition mechanisms."
signal 1 23 potential. propep 24 159 chain 160 906 neural-cadherin. domain 160 724 extracellular (potential). transmem 725 746 potential. domain 747 906 cytoplasmic (potential). repeat 160 267 cadherin 1. repeat 268 382 cadherin 2. repeat 383 497 cadherin 3. repeat 498 603 cadherin 4. repeat 604 714 cadherin 5. domain 863 878 ser-rich. carbohyd 190 190 potential. carbohyd 273 273 potential. carbohyd 325 325 potential. carbohyd 402 402 potential. carbohyd 572 572 potential. carbohyd 651 651 potential. carbohyd 692 692 potentialCadherins at OMIM:
"In the mouse the gene consists of 16 exons dispersed over more than 200 kb of genomic DNA. The large size of the N-cadherin gene, compared with its cDNA (4.3 kb), was ascribed to the fact that the first and second introns are 34.2 kb and more than 100 kb long, respectively. Human N-cadherin gene maps to a 250-kb region. The gene contains 16 exons and its sequence is highly similar to both the mouse NCAD gene (including the large first and second introns) and other cadherin genes....exon-intron boundaries are fully conserved between various cadherins , except that the first exon of P-cadherin includes the first and second exons of the other"
PDB xray structures for 6 calhedrinsA typical neural calhedrin calcium binding domain is shown as the dimer that occurs in the crystalographic unit. There are 5 repeat units in a full molecule. No structures are available for protocalhedrins but these are probably quite similar in the 3D fold.
J Biol Chem, Vol. 274, Issue 45, 32153-32158, November 5, 1999 Yuval Shaked, Hana Rosenmann, Galit Talmor, and Ruth GabizonWe show here that PrPC is present even in mature sperm cells, a polarized cell that retains only the minimal components required for DNA delivery, movement, and energy production. As opposed to PrPC in other cells, PrP in ejaculated sperm cells was truncated in its C terminus in the vicinity of residue 200. Sperm PrP, although membrane-bound, was not released by phosphatidylinositol phospholipase C as well as not localized in cholesterol-rich microdomains (rafts). Although no infertility was reported for PrP-ablated mice in normal situations, our results suggest that sperm cells originating from PrP-ablated mice were significantly more susceptible to high copper concentrations than sperm from wild type mice, allocating a protective role for PrP in specific stress situations related to copper toxicity. Since the functions performed by proteins in sperm cells are limited, these cells may constitute an ideal system to elucidate the function of PrPC.
Like most GPI-anchored proteins, both PrP isoforms are associated with cholesterol-rich membranal microdomains, denominated rafts. During spermiogenesis, spermatocytes differentiate into sperm in a process in which the Golgi develops into the large acrosome and mitochondria migrate from the rest of the cytoplasm to form a tightly wrapped sheet around the upper part of the flagellum. After the excess cell cytoplasm is pinched off, the sperm cell only carries with it an acrosome containing enzymes required for the penetration of the membranes covering the egg, as well as those other cellular components essential to provide energy for its movement. There is no de novo protein synthesis in mature sperm cells, and therefore changes occurring in protein patterns are due either to protein degradation or protein processing during sperm maturation.
During the process of sperm maturation, PrPC changes from its regular rafts-associated, GPI-anchored structure to a unique C-terminal-truncated peptide devoid of a GPI group and, as a result of that, is not released from the cell surface by phosphatidylinositol phospholipase C (PIPLC). In addition, mature sperm PrP was not longer associated with membranal rafts.
As opposed to the results with brain and with semi-mature sperm cells, the apparent molecular weight of deglycosylated PrP in mature human sperm from ejaculates, as detected by immunoblotting with mAb 3F4, was of about 17,000. Human mature sperm samples were immunoblotted with an antiserum raised against a synthetic peptide comprising residues 195-213 of the human and mouse PrP sequence (1E) and that has been shown to preferentially recognize the glutamate residue at position 200. [The authors failed to determine the actual C-terminus or even whether the second glycosylation site was present. -- webmaster]
PrP in platelets, another example of cells that do not synthesize new proteins and perform only specific unctions, cannot be released from the cell membrane by PIPLC (Holada, K.et al (1998) Br. J. Haematol. 103, 276-282).
As opposed to PrP, other GPI-anchored proteins are normally processed in sperm cells. Most GPI-anchored proteins have been shown to be targeted to special cholesterol-rich membranal microdomains denominated rafts. Rafts are insoluble in cold Triton X-100, and as such, proteins inserted in them can be separated from other membrane proteins as well as from soluble proteins by an assay in which the insoluble rafts will float to the top of density gradients and the soluble or detergent-solubilized proteins will remain in the lower fractions of these gradients . PrP is also a raft protein. ... It is possible that rafts are altogether absent from mature sperm cells as a result of changes in lipid membrane composition, which are known to occur during the different stages of sperm maturation.
Sperm cells of scrapie-infected hamsters are not more resistant to protease digestion than PrP from normal sperm. These results are consistent with the general notion that prion diseases are not transmitted sexually. Interestingly, it has been shown that male mice inoculated with prions loose their fertility before clinical signs of the disease are noticeable.
The possibility that PrP protects against copper damage was recently suggested from experiments showing that brain cells from PrP knock-out mice were more susceptible to copper toxicity than brains cells from wt mice (50). The results of the experiments described in Fig. 8 are also consistent with such a protective role for PrP against copper induced damage. Although no difference in motility was observed between wt mice sperm and PrP knock-out mice sperm at control conditions, the addition of copper was remarkable more toxic to PrP knock-out mice sperm cells than to wt sperm cells. Interestingly, a group of markers present in CSF and serum during prion disease (S100 proteins) have been shown to perform as copper-binding proteins and thereby protect against copper-induced cellular damage, suggesting a possible compensatory effect, which may be required during the last stages of prion infection.
...Once a reliable assay is developed for the function of PrPC, it will be possible to establish whether the loss of such function plays any role in the pathogenesis of prion diseases.
Proc. Natl. Acad. Sci. USA 1999 96(23): p. 13044-13049 Pauline M. Rudd, Tama Endo, Cristina Colominas,..., Hana Serban, Stanley B.Prusiner, Akira Kobata, and Raymond A. Dwek Submitted August 17, 1999Comment (webmaster): This is a very substantial article on the complexities of prion glycosylation in hamster that actually makes a specific proposal, downregulation of N-acetylglucosaminyltransferase III, for how one of the biosynthetic enzymes might give rise to differences seen in normal and disease states. The results here should also be viewed in the context of mouse glycans as reported in another quality paper by Burlingame, et al (1999) Biochemistry 38, 4885-4895, which is properly cited here but not compared-and-contrasted.
Their data strongly suggests that strain-typing based on simple-minded categories such as site occupation will not explain conformer differences. The notion that only specific partial glycosylation subset convert to rogue conformer is also not particularly consistent with the data. It is no wonder that CJD phenotyping has been so intractible.
Figure 4 provides a fairly serious molecular model of the prion molecule with glycans and GPI attached and to scale. It did not show any further detail at higher resolution -- was there really was an underlying coordinate structural file?
Doppel's glycans are not addressed Doppel would make an interesting study as to whether and how glycosylation diverged as the proteins diverged. Doppel is also somewhat of a control as to the novelty of the particular glycans found on prion protein.
Prion protein contains two conserved N-glycosylation sites (Asn-181 and Asn-197 in Syrian hamster). The functions of the sugars are not yet established, although transgenic mice expressing prions with deletions of one or both glycosylation sites show unusual patterns of PrPSc deposition (3). It has also been proposed that specific SDS/PAGE banding patterns, which represent the protease-resistant core of PrPSc (designated PrP 27-30) with zero, one, or two occupied glycosylation sites, correlate with different strains of prions (4). This hypothesis has proved controversial and the subcellular trafficking of PrP argues against strain-enciphered properties being specified by sugar chains (5-9). At present, no detailed comparisons of the glycosylation between PrPC and PrPSc have been reported although earlier studies of PrP 27-30 from Syrian hamsters (10-14) and mice (15) have indicated that glycosylation is complex. A recent study of murine PrPSc glycans by MS (15) has shown that both sites are partially occupied and that they contain different but overlapping sets of sugars.
The glycan pools from PrPC and PrP 27-30 [brains of scrapie-infected hamsters] were released by hydrazinolysis and the glycans were analyzed by MALDI MS (Fig. 1; Table 1) and, after fluorescent labeling with 2-AB (18), by weak anion exchange chromatography (data not shown) and normal phase (NP) HPLC (Fig. 2; Table 1). Structures were assigned from the MS compositions, NP HPLC elution positions expressed in glucose units, previously assigned incremental values for the addition of monosaccharide residues to glycan cores (19) and the results of digestions of each glycan pool by using arrays of exoglycosidases.
Weak anion exchange chromatography (20) and MALDI MS analysis of the intact glycan pools (data not shown) both revealed that there was no significant difference in the relative proportions of the populations of sialylated structures in PrPC and PrP 27-30. In each case, 48% of the sugars were neutral, 20% contained one sialic acid residue, 18% were disialylated, 10% were trisialylated, 3.5% were tetrasialylated, and 0.5% contained five sialic acids.
The desialylated glycan pools also were analyzed by MALDI MS (Fig. 1 A and B). In both the PrPC and the PrP 27-30 spectra, 54 ion peaks with masses consistent with oligosaccharide structures were present. Many of these peaks contained isomeric glycans (Table 1). NP-HPLC resolved at least 10 of these isomeric structures and confirmed that there are remarkably few differences between the two pools (Fig. 2 Aii and Bii). The digestion products resulting from incubating the glycan pools with sialidase and almond meal fucosidase (specificity for fucose linked 1-3/4 to GlcNAc) were consistent with the presence of outer-arm fucose forming part of a LewisX- or sialyl LewisX-type structures (Fig. 2 Aiii and Biii). Analysis of the glycan pool following digestion with an enzyme array containing bovine testes -galactosidase (Fig. 2 Aiv and Biv) showed losses of galactose consistent with the presence of bi-, tri-, and tetraantennary structures containing core fucosylation (confirmed by incubation with bovine kidney -fucosidase (Fig. 2 Av and Bv). Incubation of the glycans in this fraction with Streptococcus pneumoniae -N-acetylhexosaminidase at arm-specific concentration (5 milliunits/ml, data not shown) confirmed the presence of bisected and unbisected bi-, tri- and tetraantennary structures and showed that the majority of triantennary glycans were branched on the 1,3 arm of the trimannosyl core.
The glycans that were identified on both PrPC and PrP 27-30 were typical of brain sugars in that they were heavily sialylated and contained a high proportion of structures with one or more outer arm fucose residues (59%) and with bisecting GlcNAc (25). Ninety-five percent of the structures contained core fucose (Fig. 2 Aiv and Biv). The nonbisected structures were found both with and without core fucose, whereas bisected structures were generally core fucosylated. After desialylation the major structure (9.5% on PrPC and 7.9% on PrP 27-30) was the monogalactosylated biantennary glycan containing an outer-arm fucose residue [A2G1FBF(1,3 to GlcNAc) see footnote to Table 1 for structure designation]. Fig. 2 highlights the digestion of this oligosaccharide by the enzyme arrays.
The extent of the heterogeneity and the conservation of structures makes it unlikely that a subset of PrPC molecules carrying particular sugars are predisposed to convert to PrPSc. It is also unlikely that the protein structure around the glycosylation sites is significantly different in these two hypothetical sets of PrPC molecules when the glycan processing takes place in the Golgi apparatus, because the local three-dimensional structure is known to play a role in directing its own glycosylation (26). Previous work by Endo et al. (12) concludes that both glycosylation sites on SHa PrP 27-30 are occupied. Our preliminary data (data not shown) based on SDS/PAGE analysis of SHa PrP 27-30 and PrPC suggest that there are no major differences in site occupancy between the two forms and that most SHa PrP glycoforms contain two occupied glycosylation sites. Therefore selective site occupancy and site-specific processing does not appear to explain the current data, nor is the level of site occupancy in itself a requirement for PrPSc formation. However, variations in site occupancy may affect neuronal targeting of PrPC and thus selective deposition of PrPSc (3).
Although the range of oligosaccharides in PrPC and PrP 27-30 was identical, the relative proportions of some glycans were different in the two forms. In particular, in the final digest, where all the outer arm sugars have been digested to reveal the structures of the glycan cores (Fig. 2 Av and Bv; Table 2), it became evident that the proportion of bisected glycans was reduced by 11% in PrP 27-30 compared with that in PrPC, whereas the proportion of tri- and tetraantennary glycans were increased by 5% and 10%, respectively. The ratio of bisected to nonbisected glycans in PrP 27-30 was 1.3:1 compared with 2:1 in PrPC. MALDI MS confirmed this finding. The difference spectrum (Fig. 1, lower trace) highlights the changes in concentration of all glycans from PrP 27-30 compared with those in PrPC and shows the reduction in the abundance of bisected sugars and an increase in the proportion of tri- and tetraantennary glycans in PrP 27-30.
Glycans were released in two separate hydrazinolysis reactions from two different preparations of SHa PrP 27-30 and PrPC and by protein N-glycosidase-F from SDS/PAGE gels following the procedure described in ref. 17. No significant differences in the proportions of the glycoform populations were seen in the final HPLC (error less than ħ1%) or in the MALDI profiles. Many enzyme digests were performed by using enzymes in arrays and singly to confirm findings. The data shown are from two representative analyses.
These results are most simply explained by a down-regulation of N-acetylglucosamine transferase-III (GnTIII) activity in the cells from which PrPSc was formed. At any stage in the glycan processing, further branching is possible provided that a bisecting GlcNAc residue has not been added in 1-4 linkage to the conserved pentasaccharide core (Fig. 3). Our data argue that a substantial portion of PrPC is converted into PrPSc in cells in which the glycosylation machinery has been perturbed.
The ratio of bisected to nonbisected glycans in PrP 27-30 was 1.3:1 compared with 2:1 in PrPC. The relative amounts of bi-, tri-, and tetraantennary glycans in PrPC are 51, 32, and 17%, and in PrPSc are 37, 36, and 27%, respectively. In mammalian cells the GlcNAc transferases are a family of five enzymes that first initiate the conversion of oligomannose to complex or hybrid glycans by attaching GlcNAc in a 1-2 linkage to the Man1-3Man1-4GlcNAc2-Asn core (GnTI). This step is essential for the action of further processing enzymes including that of the core 6-fucosyltransferase and of GnTII, -IV, and -V, which mediate the substitution of additional GlcNAc residues in the trimannosyl core leading to the formation of bi-, tri-, and tetraantennary glycans respectively. GnTV cannot operate until GnTII has added GlcNAc to the 2 position of the mannose (Man) residue linked 1,6 to the core. GnTIII is the enzyme which links a GlcNAc residue to position 4 of the conserved pentasaccharide core. This substitution, which can occur at any point in the pathway, inhibits further branching of the sugars by preventing processing by GnTII, IV, or V.
The observed changes in glycosylation between PrPC and PrPSc probably reflect biochemical changes occurring within cells, perhaps those involved in the pathogenesis of neurodegeneration. The Syrian hamsters, from which PrP 27-30 was isolated, exhibited clinical signs of central nervous system dysfunction. Earlier studies showed that such hamsters also show frank neuropathologic changes (27). Associations between the perturbation of glycosylation machinery and pathology have been well documented in a number of disease states. For example, in rheumatoid arthritis (RA) there is a decrease in the population of serum IgG sugars containing terminal galactose, which correlates with disease activity (28) and poor prognosis (29). In PrPSc, as with RA, there are no glycans associated with the diseased state that are not present in the normal glycoprotein. In the case of RA the changes in the glycosylation pattern of IgG may reflect a decrease in galactosyl transferase (GTase) activity (30, 31). Interestingly, an alteration in GnTIII activity also was noted in RA where there was a Fab-specific increase in oligosaccharides containing a bisecting N-acetylglucosamine (32). It remains to be determined whether the reduction in the activity of GnTIII in cells producing PrPSc correlates with disease severity and whether the glycosylation of other glycoproteins in PrPSc-producing cells are similarly affected. Likewise, the underlying cause of the down-regulation of GnTIII activity remains to be determined.
The prevalence of GnTIII in brain tissue compared with many other tissues has been demonstrated in mice (33) and many studies have shown an association of altered levels of bisecting GlcNAc with disease processes (17, 31-35). Regulation of the human GnTIII gene (Mgat3) is complex and it has multiple promoters (34), suggesting that there are many points at which gene expression could be affected by cellular events. GnTIII-deficient mice exhibited normal cellularity and morphology in a range of organs including brain (35).
A molecular model for the region of PrPC (residues 90-231), corresponding to the protease-resistant core, is shown in Fig. 4. Both of the N-glycans are attached to the region of the protein with the best defined secondary structure, the disulfide-bridged helix-loop-helix motif at the C terminus (36, 37). This face of the protein has a net negative charge (37), and thus the negatively charged glycans are unlikely to be involved in long-range interactions with the protein surface. The glycans will sterically hinder any direct protein-protein interactions with this face of the molecule, either intramolecularly, involving residues 1-90, or intermolecularly. On average, the sialic acids from each glycan contribute three negative charges to the glycoprotein; therefore, two glycans will increase the negative charge in this region significantly, and hence increase the overall dipole moment of the protein. Tri- and tetraantennary glycans, of which there are a higher proportion in PrPSc, are both larger and more rigid than biantennary structures (38). Two possible orientations are shown in Fig. 4, one with the protein extending as far as possible from the membrane and one with the protein oriented so that its positive face is toward the membrane (37). In both cases, the region of the protein facing away from the membrane is covered by a glycan and protected from outside influences.
A molecular model of PrPC, residues 90-231, based on one of the NMR structures of the recombinant protein (36), the sequence of the GPI anchor (39), and the N-glycan analysis. The glycan structures (A3G3FBS3 at site 181 and A4G4FBS4 at site 197) were built by using the database of average crystallographic linkages (40). The torsion angles around the Asn C-C and C-C bonds were adjusted to eliminate unfavorable steric interactions between the glycans and the protein surface. The GPI anchor is shown in two orientations with respect to the protein, (a) the GPI extends directly away >from the protein maximizing the distance between the negatively charged glycans and the membrane surface and (b) with the positively charged face of the protein oriented toward the membrane surface (37). There is likely to be considerable dynamic freedom at all the attachment points for the glycans.
Both normal and pathogenic prion proteins display extensive heterogeneity in their glycan populations. The glycan analysis demonstrates that the same set of glycans are present on both the normal and diseased molecules and that the differences in the relative proportions of bi-, tri-, and tetraantennary glycans as well as the decrease in bisected structures in PrPSc can be accounted for by a decrease in the activity of one enzyme, GnTIII. These findings suggest that some cells forming PrPSc undergo a change that diminishes the activity of an enzyme in the glycosylation pathway. The cause and significance of this change, the impact of which might not be confined to the glycosylation pathway, remains to be determined.
Abbreviations in footnote to table 1:
H, hexose (mannose, galactose);
N, HexNAc (GlcNAc);
F, deoxyhexose (fucose).
A, number of antennae;
G, number
of galactose residues (numbers in parentheses indicate the antennae to which the
galactose is attached);
F, number of fucose residues;
B, presence of a bisecting GlcNAc;
PolyLac, structure must contain at least one N-acetyllactosamine extension.
Molecular and Cellular Biology, Dec 1999, p8103-8112, Vol. 19, No. 12 $4.00 Yury O. Chernoff, Gary P. Newnam, Jaijit Kumar, Kim Allen, and Amy D. ZinkPropagation of the yeast protein-based non-Mendelian element [PSI], a prion-like form of the release factor Sup35, was shown to be regulated by the interplay between chaperone proteins Hsp104 and Hsp70. While overproduction of Hsp104 protein cures cells of [PSI], overproduction of the Ssa1 protein of the Hsp70 family protects [PSI] from the curing effect of Hsp104.
Here we demonstrate that another protein of the Hsp70 family, Ssb, previously implicated in nascent polypeptide folding and protein turnover, exhibits effects on [PSI] which are opposite those of Ssa. Ssb overproduction increases, while Ssb depletion decreases, [PSI] curing by the overproduced Hsp104. Both spontaneous [PSI] formation and [PSI] induction by overproduction of the homologous or heterologous Sup35 protein are increased significantly in the strain lacking Ssb.
This is the first example when inactivation of an unrelated cellular protein facilitates prion formation. Ssb is therefore playing a role in protein-based inheritance, which is analogous to the role played by the products of mutator genes in nucleic acid-based inheritance. Ssb depletion also decreases toxicity of the overproduced Sup35 and causes extreme sensitivity to the [PSI]-curing chemical agent guanidine hydrochloride. Our data demonstrate that various members of the yeast Hsp70 family have diverged from each other in regard to their roles in prion propagation and suggest that Ssb could serve as a proofreading component of the enzymatic system, which prevents formation of prion aggregates.