Review Fire and ICE: The role of -containing in and

D.L. Gumucio1, A. Diaz2, P. Schaner1, N. Richards1, C. Babcock1, M. Schaller1, T. Cesena1

1Department of Cell and Developmental ABSTRACT The term “ a u t o i n fl a m m at o ry ” wa s Biology, 2Department of Internal Medicine, The genetic bases for several human recently proposed to describe a set of University of Michigan, Ann Arbor, Michi- autoinflammatory syndromes have re - heritable human diseases characterized gan, USA. cently been identified, and the mutated by inflammation, without evidence of DLG acknowledges support from the proteins responsible for these diseases high titer antibodies or ap p a re n t Arthritis Foundation; AD is grateful for are rapidly being characterized. Here, involvement of antigen-specific T cells support from NIH K08 AI50092. we examine two of these newly identi - (1). The hereditary periodic fever syn- Please address correspondence to: fied pro t e i n s , py rin (also called ma - d romes constitute a subset of these Deborah L. Gumucio, PhD, Department of renostrin, product of the familial Medi - a u t o i n fl a m m at o ry diseases, a n d Cell and Developmental Biology, Universi- terranean fever locus, MEFV) and cry - i n cl u d e : familial Mediterranean feve r ty of Michigan Medical School, 5704 Med- ical Science II, Ann Arbor, MI 48109-0616, o py rin (product of the CAIS1 locus, (FMF); tumor necrosis factor receptor- USA. and mutated in familial cold urticaria, 1 - a s s o c i ated periodic syndro m e E-mail: [email protected] Muckle Wells syndrome and chronic in - (TRAPS); hyper IgD periodic feve r Received on June 10, 2002; accepted on fantile neurological cutaneous and ar - syndrome (HIDS); and the allelic dis- June 26, 2002. ticular syndrome). Both py rin and o rd e rs , M u ck l e - Wells syndro m e Clin Exp Rheumatol 2002; 20 (Suppl. 26): c ryo py rin contain an N-terminal do - (MWS) and familial cold urt i c a ri a S45-S53. main that encodes a -re - (FCU). A few add i t i o n a l , less we l l - lated structure, now known as the pyrin characterized fever syndromes, such as © Copyright CLINICAL AND EXPERIMEN- TAL RHEUMATOLOGY 2002. domain, or PyD.We trace the molecu - PFAPA syndrome (periodic fever apht- lar interactions mediated by these hous stomatitis and adenitis) and Key words: Hereditary fever PyDs, examine the evolution of the fa - ch ronic infantile neuro l ogical cuta- syndromes, NF- B activation, mily of molecules containing this do - neous and articular syndrome or activation, pyrin, crypyrin, Apaf-1 main, and discuss the function of PyD- CINCA (also known as neonatal onset Wlike proteins, innate immunity. containing proteins and their homo - multisystem inflammatory disease, or logues. Synthesis of the available data NOMID) might also belong in this cat- indicates that both pyrin and cryopyrin egory. Though each of these syndromes interact via their PyDs with a common has its own genetic and phenotypic adaptor , ASC. ASC itself par - peculiarities, there are also several sim- ticipates in at least three important cel - ilarities in their presentation. For exam- lular processes: apoptosis, recruitment ple, many of these syndromes involve and activation of pro-caspase-1 (with fever, urticaria, arthritis and serositis as associated processing and secretion of common clinical elements and sys- I L - 1 ) , and activation of NF- B (a temic amyloidosis as a potential com- t ra n s c ription factor invo l ved in both plication. In many cases too, inflamma- initiation and resolution of the inflam - t o ry at t a cks are accompanied by a m at o ry response). Th rough PyD:PyD robust acute phase response, and a pro- i n t e ra c t i o n s , py rin and cryo py ri n , a s minent neutrophilia ch a ra c t e ri zes the well as several related, but still unchar - inflammatory site.These features sug- acterized PyD containing proteins, ap - gest that the here d i t a ry fever syn- pear to modulate the activity of all dromes are disorders of the innate im- three of these processes, each of which mune system,an ancient arm of the im- plays a crucial role in the inflammatory mune system responsible for quick res- pathways that characterize the innate ponse to a variety of pathogenic insults immune system. to the organism. Clearly, insights into the molecular underpinnings of innate immunity, potentially available through the etiologic investigation of these dis-

S-45 REVIEW Inflammation, apoptosis and the pyrin domain / D.L. Gumucio et al. ord e rs , could have enormous cl i n i c a l ponentially, especially since that inter- significance. action required pyrin’s PyD. A yeast Among the presently known periodic two hybrid screen identified ASC (apo- fever syndromes, FMF (2, 3), MWS/- ptosis associated speck - l i ke pro t e i n FCU (4) and CINCA/NOMID (5) are containing a CARD) as a pyrin-inter- caused by mu t ations in proteins that acting protein (11). ASC was, at the contain a newly identified protein:pro- time of this finding, itself newly identi- tein interaction domain, va ri o u s ly fied (14), as a protein that forms large, called a pyrin domain (PyD) or PAAD hollow cytoplasmic aggregations called or DAPIN (6-11). Growing interest in s p e cks in apoptotic cells (Fi g. 1 ) . this 90 amino acid domain, hereafter Specks are harbingers of apoptosis, as called the PyD, stems from the fact that 100% of cells that form specks go on to its six alpha helices fo rm a deat h complete the apoptotic program (11, domain (DD)-related structure (6-11). 14). Indeed, u p - reg u l ation or ove r- The death domain is one of three relat- expression of ASC in a variety of cells ed structures seen in molecules involv- causes speck formation and apoptosis. Fig. 1. ASC specks are well-ordered cytoplas - ed in apoptosis (the other two being the Exon 1 of pyrin is both necessary and mic stru c t u re s . F l ag - t agged ASC and my c - or DED, and the sufficient for interaction of pyrin with t agged py rin constructs we re tra n s fected into caspase recruitment domain or CARD). ASC and co-localization of pyrin with HeLa cells and the proteins were recognized by double label immunofluorescence. Two cells in DDs, DEDs and CARDs are known to ASC in specks (11). Interestingly, ASC this image contain red specks (here visualized by m e d i ate pro t e i n : p rotein intera c t i o n s also has an N-terminal PyD, and its C- rhodamine staining for the myc-pyrin protein). ( 1 2 , 13). Though pro b ably evo l ve d terminal end encodes a CARD.Thus, The inset shows a higher mag n i fi c ation of a members of a single family, the indi- ASC functions as an adaptor, linking speck; the yellow color arises from the co-local- ization of ASC (green) and pyrin (red) in the vidual members interact only with their PyD containing proteins to CARD- speck. Note the hollow character of the speck. most similar re l at ives. Th at is, D D s containing ones. interact with other DDs, DEDs contact It is still unclear what significance the DEDs and CARDs touch other CARDs speck structure itself might have. How- these proteins are predicted to exist, the ( 1 2 , 13). The PyD fo rms a fo u rt h ever, it is possible that specks are a nature of such aggregates has not been gro u p , and proteins with PyDs also macroscopic reflection of the tendency probed. Moreover, evidence that ASC seem to interact with other PyDs. Inter- for ASC and pyrin (perhaps in coopera- is pro - apoptotic has been pre s e n t e d e s t i n g ly, though py rin (also called tion with additional proteins) to form (11, 14); and data indicating that pyrin m a re n o s t rin) rep resents the fo u n d i n g intermolecular aggregates. Indeed, it is can modulate the apoptotic process set protein for the PyD, it was not recog- likely that micrscopic aggregates form in motion by ASC are also available n i zed as a death domain-containing at lower concentrations of these pro- ( 1 1 ) , though there are some incon- molecule by either of the two consortia teins. Intermolecular aggregation is a gruities to deal with. First, in a variety that initially cloned it. This is because: hallmark of apoptotic proteins. In fact, of cell types, the co-expression of pyrin a) at that time, no other protein with a the activities of the effector molecules with ASC results in a dramatic increase PyD had been ch a ra c t e ri ze d, t h o u g h of apoptosis, the , are largely in the number of cells that display several could be found in the EST data- regulated by “induced proximity”. That s p e cks (ASC, but not py ri n , c a u s e s bases and b) the sequence identity is, the localized concentration, in the sp e c k form a tion when exp r essed alone). between a PyD and the related DDs, same molecular aggregate, of a number This finding would suggest that pyrin is DEDs and CARDs was too low to be of these molecules results in the local- p ro - ap o p t o t i c, t h at it enhances the re c og n i zed by simple homology ized expression of an otherwise weak apoptotic signal elicited by ASC. How- searches. However, molecular model- e ffector activity (15). This enhanced ever, paradoxically, the speck-positive ing strategies, combined with sophisti- effector activity sets in motion a cas- cells seen after ex p ression of A S C cated fold recognition algorithms, car- cade of events that shortly results in the alone die much faster than the speck- ried out simultaneously in several labo- death of the cell. The molecular aggre- positive cells seen after co-expression ratories, were able to make the structur- gate itself has been called an apopto- of pyrin with ASC. This indicates that al diagnosis (6-11). some, and the presence of these multi- pyrin enhances cell survival in the face protein aggregates has been physically of the apoptotic signal educed by ASC. The fact that the PyD looks like a pro- demonstrated, as has their central role At the time of this writing, it is still tein: protein interaction domain found in apoptosis (16, 17). u n clear whether we should place a on other molecules known to be in- Do py rin and ASC fo rm an ap o p t o- “p ro - ap o t o t i c ” or “a n t i - ap o p t o t i c ” lab e l volved in apoptosis and inflammation s o m e - l i ke stru c t u re ? It is clear that on the pyrin protein. In fact, the prob- was intri g u i n g, but when py rin wa s specks are large, well ordered molecu- lem seems further muddled (but in a found to actually interact with a pro- lar aggregates (11, 14, see Figure 1). ver y interesting way) by recent fin d i n g s apoptotic protein (11), interest rose ex- Though smaller aggregates formed by in our laboratory that the pyrin/ASC

S-46 REVIEW

and other still uncl a s s i fied diseases m a rked by fever and infl a m m at i o n (22). There is also evidence that, both in human populations, and during pri- mate evolution, pyrin has been subject to positive Darwinian selection (23- 27). Interestingly, it appears that certain of the FMF-causing mu t ations have been selected for, indicating that under some env i ronmental circ u m s t a n c e s , a mutant pyrin allele actually confers a benefit to the host. One proposed sce- nario is that a heightened state of in- flammatory alert (e.g., increased levels of acute phase response proteins such as CRP and SAA), which is apparently bestowed even on heterozygotes carry- ing a mutant pyrin gene, results in more certain clearance of dangerous patho- gens (26, 27). Cryopyrin, recently shown by Hoffman et al. to be the protein product of the causative gene (CAIS1) for MWS and FCU (4), also appears to be centrally l i n ked to a my riad of infl a m m at o ry syndromes. Cryopyrin is so named for its N-terminal stru c t u ral re l at i o n s h i p with py rin (both proteins contain a PyD) and for its role in cold-stimulated urt i c a r ia. Just as we we re scrat ch i n g our heads as to how mutations in a sin- gle protein can cause cold-induced rashes in some individuals (FCU) and s e n s o ri n e u ral hearing loss in others (MWS), Feldmann et al. linked muta- tions in cryopyrin to another inflamma- Fig. 2. Structural comparison of pyrin domain-containing proteins. As much as possible, and where t o ry syndro m e, NOMID/CINCA (5). known, motifs are shown as ribbon diagrams. These proteins contain many common This syndrome has its own peculiari- motifs, including the pyrin domain (10), nucleotide binding site (NBS)(62), Leucine Rich Repeat (LRR) (63), and caspase recruitment domain (CARD) (64). Some of these proteins have domains ties, including migratory skin rashes, unique to them as well, such as the WD40 repeats of Apaf-l (65), which have been implicated in apop- ch ronic meningitis and ab n o rm a l i t i e s tosome formation (16,17),the Toll/Interleukin 1 receptor (TIR) domain (66) of RPS2 from arabadop- in cartilage growth. sis,and the leucine zipper motif (67) of the tobacco N protein. The protein most unlike the others in this So what does cryopyrin do ?The first structural family is pyrin itself, which, in addition to a pyrin domain, also contains a B-box type (68), a coiled coil (69), and a B30.2 domain (70). hints came from analysis of its struc- ture. Figure 2 summarizes the structure of the proteins that are germane to this interaction may be different in different pyrin (18), followed by pyrin-induced d i s c u s s i o n , and Table I serves as a cells, and both labels could potentially modulation of ASC function. Exactly q u i ck - re fe rence guide for their func- fit (data to be presented). h ow the py rin mu t ations affect this tional significance and confusing no- Thus far, we have established connec- pathway is a major detail that remains menclature. At the same time that cryo- tions between clinical infl a m m at i o n to be filled in,but evidence that pyrin is pyrin was first identified as the MWS/- and the protein product of the FMF a central component of the body’s in- FCU protein, Manji et al. reported the locus (pyrin), between pyrin and ASC, fl a m m at o ry response continues to characterization of PYPAF1, or pyrin- and between the ASC/pyrin interaction build.There is now evidence that muta- containing Apaf1-like protein (28). The and apoptosis. At this point, we can po- tions in pyrin might also predispose an P Y PAF fa m i ly designation re fe rs to stulate that at least part of the mecha- individual to other inflammatory condi- Apaf-1-like proteins that contain a PyD nism of the inflammatory response in- tions, including Bechet’s disease (19, at the N-terminus; PYPAF1 is cryopy- volves cytokine-induced activation of 20), inflammatory bowel disease (21), rin. The cryopyrin structure was inter-

S-47 REVIEW Inflammation, apoptosis and the pyrin domain / D.L. Gumucio et al.

Table I. The players: PyD-containing proteins and proteins structurally related to PyD proteins.

Protein Other names Function Reference

Pyrin Marenostrin Product of MEFV locus; modulates ASC-mediated apoptosis 2, 3, 11

ASC TMS 1/PYCARD Adaptor that binds pyrin, cryopyrin, DEFCAP, PYPAF7; 7, 11, 14, 49, 50 activates NF- B at high concentrations; activates pro-caspase-1; forms”specks” in cells; pro-apoptotic

Cryopyrin PYPAF1/NALP3 Responsible for FCU, MWS, NOMID/CINCA. 4, 5, 28, 50 Binds ASC, activating NF- B and caspase-1

PYPAF7 Binds ASC, activating NF- B and caspase-1 50

DEFCAP CARD7/NAC/ Enhances Apaf-1-mediated caspase 9 activation; pro-apoptotic; 7, 56, 57 NALP1 binds ASC

N (tobacco) Intracellular plant R protein 60

Rps2 (Arab) Intracellular plant R protein 61

Apaf- 1 CED4-related; binds cytochrome c and ATP, activates caspase 9 34, 35 by induced proximity

Nodl CARD4 Intracellular LPS receptor; activates caspase 9; activates NF- B; 31, 32, 37 mediates JNK and NF- B activation by S. flexneri

Nod2 Intracellular LPS receptor; activates caspase 9; activates NF- B; 38, 46-48 mutations associated with Crohn’s disease and Blau syndrome

POP1 ASC2 Uncharacterized; predicted inhibitor of ASC function (see Figure 3) 50

ICEBERG Binds pro-caspase-1, inhibits activation and caspase-1-dependent 53, 54 IL-1 secretion

Pseudo-ICE COP Binds pro-caspase-1, inhibits activation and casp1-dependent IL-1 54, 55 secretion; interacts with RICK and activates NF- B.

Note that other, less well-characterized PyD-containing proteins are not listed here. These include the NOD-related proteins PYPAF2 (NBS1/NALP2), PYPAF3,PYPAF4/NALP4,PYPAFS, PYPAF6,PYPAF8/MATER, NAIP and CIITA discussed by others (29,50). In addition, proteins not in the NOD/Apaf family, but containing a pyrin domain (though more divergent) include:AIM2 (absent in melanoma), several interferon inducible genes such as IFI16 and MDNA,and the viral proteins M013L from myxoma virus and gp013L from rabbit fibroma virus (9). The latter two are PyD-only proteins; it remains to be seen whether these viral proteins can inhibit ASC-mediated caspase-1 activation. esting because of its similarity to Apaf- the NBS and LRR motifs are also nals to the activation of the execution 1 (discussed further below), but more found in other proteins besides cryopy- pathway (33-36). The C-terminal WD- compelling was its similarity to the rin and the NODs. Before getting back 40 motifs of Apaf-1 bind cytochrome c NOD proteins (members of the same to the function of cryopyrin, a tangen- released from mitochondria. This caus- Apaf-1-like structural family, but con- tial diversion to provide a brief over- es a structural alteration in the Apaf-1 taining a CARD rather than a PyD at view of some of these other proteins molecule that allows its homo-oligo- the N-term i nus). Two stru c t u ral do- will be instructive from both a structur- m e ri z ation v i a the region containing mains are shared by cryopyrin and the al and a functional point of view. the NBS and the recruitment of caspase NOD proteins (Fig. 2). First, there is a Nod1 (which also goes by the name 9 v i a its N-terminal CARD domain. domain that binds ATP and mediates CARD4, ref. 31) was originally identi- The oligomerization results in the con- homo-oligomerization. We will refer to fied on the basis of its structural homol- centration of several caspase 9 mole- it as the NBS or nu cleotide binding ogy to Apaf-1 (31, 32). Apaf-1 (apop- cules in the molecular aggregate, re- site, but it has also been called a NOD, totic protease-activating factor-1) is a sulting in proximity-induced activation nucleotide-binding oligomerization do- central component of the apoptosis ma- of caspase 9 (36). The NOD proteins main (29), or a NACHT (30). The se- chine; it was originally discovered in function similarly (32, 37, 38). That is, cond shared domain is composed of a the worm, C. elegans, and named Ced- the C-terminal LRRs act as a ligand series of leucine-rich repeats (LRRs); 4 (33). Apaf-1/Ced-4 is a transducer of binding domain (the ligand itself is dis- LRRs function as protein:protein inter- the mitochondrial or intrinsic pathway cussed below); ligand binding induces action motifs. As shown in Figure 2, of apoptosis, connecting cell death sig- a probable conformational change that

S-48 Inflammation, apoptosis and the pyrin domain / D.L. Gumucio et al. REVIEW a l l ows homo-oligo m e ri z ation thro u g h the pathogen-containing plant cell. tein that clearly functions as an adaptor. the NBS region and re c ruitment of Recognizing the remarkable structural L i ke py ri n , c ryo py ri n / P Y PAF1 inter- another protein via the CARD domain. similarities between the NOD proteins acts with ASC via PyD:PyD contact In the case of Nodl and Nod2, this re- and the plant R proteins, Inohara et al. and is recruited to ASC specks (28). cruited protein is RICK (RIP2/CAR- predicted that the LRRs of Nod1 and Whether cryo py rin also modulat e s D I A K ) , a CARD-containing kinase Nod2 would bind bacterial pro d u c t s ASC mediated apoptosis as pyrin does that itself binds to and activates IKK / such as LPS.Their predictions proved has not yet been tested.The interaction NEMO, the regulatory subunit of the I- correct, as they were able to demon- between cryopyrin and ASC requires B (IKK) complex (39). Homo-ol- strate the intracellular binding of LPS the PyD of both proteins; thus the re- igomerization of Nodl or Nod2 results by Nodl and show that different bacter- cruitment of ASC to cryopyrin essen- in the activation of RICK, and thence ial molecules had differing abilities to tially supplies cryopyrin with a CARD the IKK complex by induced pro- activate these two NOD proteins (37). domain, which is then free to recruit ximity; the result is NF- B activation. This finding suggests that NODs could still another effector protein to the The activation of NF- B plays a central function in a manner similar to the R complex. This recruited protein might role in cellular responses to stress, in- proteins. That is, they may be involved be RICK or a functional homolog, flammatory signals and pathogens (40, in intracellular pat h ogen re c og n i t i o n , since cryopyrin appears to enhance the 41). Both Nodl and Nod2 also display functioning as re c ep t o rs and signal otherwise weak ability of ASC to acti- pro-apoptotic tendencies, and activate transducers for the innate immune sys- vate NF- B, and this activation pro- the caspase 9 pathway, but the molecu- tem. In fa c t , G i ra rdin et al. re c e n t ly ceeds through the IKK complex (28). lar details of this (e.g., do the Nods d e m o n s t rated that Nod1 mediates an Furthermore, deletion of the LRR do- activate caspase 9 and NF- B in res- intracellular host response to invasive main from cryopyrin increases the abil- ponse to different triggers ? Is there a Shigella flexneri (45), resulting in the ity of the truncated protein to activate t e m p o ral aspect to be considere d ? ) a c t ivation of both NF- B and the NF- B. This suggests that for cryopy- have still to be worked out. In this con- stress-induced kinase, JNK (c-Jun N- rin, as with Apaf-1 and the NODs, the text, it is interesting that the NF- B t e rminal kinase). Dominant negat ive C-terminal region might comprise a li- pathway seems to be critical both for fo rms of Nod1 (ve rsions lacking the gand binding domain that constitutes a the initiation of the inflammatory res- CARD domain or consisting of only molecular “switch”. If so, identification p o n s e, and for its resolution (42). the LRR domain) inhibit both of these of the ligand becomes an import a n t While the initiation phase includes NF- activities. These data, along with those goal with potential clinical signifi- K B - s t i mu l ated induction of pro - i n- of Inohara et al. reveal a novel intracel- cance. flammatory cytokines and suppression lular response system in which Nod1 There are a few more puzzle pieces that of apoptosis, the NF- B-mediated re- functions to sense bacterial LPS and can be inserted (though these by no solution phase is characterized by the initiate an inflammatory response. means complete the picture). First, re- expression of anti-inflammatory mole- If Nods are intracellular sensors that call that mutations in both pyrin and cules and the induction of apoptosis. can reg u l ate the innate immune re s- cryopyrin cause fever and inflamma- Another set of proteins with structural ponse, then we might expect that muta- tion, and note that both proteins inter- s i m i l a rity to cryo py ri n , N O D s , a n d tions in these proteins might be in- act with ASC. This turns out to be an Apaf-1 is the intracellular group of vo l ved in human infl a m m at o ry syn- important connection, since two differ- plant cytosolic disease resistance (R) dromes. This prediction too, has been ent laboratories have now demonstrat- proteins. In the R proteins, the N termi- b o rne out. Mutations in Nod2 (also ed that the CARD domain of ASC in- nal domain is not a CARD or PyD, but called CARD15) are associated both teracts directly with the CARD domain a different kind of protein recruitment with Cro h n ’s disease, and with Blau of pro- (49, 50). Caspase-1 domain (re s e m bling either a leucine Syndrome (46-48). It is interesting that (also called interleukin-1 converting zipper or To l l / i n t e rleukin-1 re c ep t o r, the Crohn’s disease mutations are thus enzyme or ICE) is the protease respon- TIR). The R proteins constitute a very far seen in the LRR domain, s o m e sible for the processing and secretion of large family of proteins that mediate re- causing a tru n c ation of that domain IL-1 and IL-18. It also processes the sistance to a variety of pathogens, in- (46, 47), while the Blau syndrome mu- precursor of IL-6 and participates in the cluding bacteria, viruses and fungi (43, t ations lie in the NBS (48). Furt h e r secretion of IL-4, IL- and TNF- (51, 44). The C-terminal LRR domain of work will be needed to determine how 52). Caspase-1 -/- mice exhibit a major each type of R protein is capable of re- and why these different mutations in defect in IL-1,3 processing, and dimin- cognizing and binding to a pathogen- the same molecule encode different in- ished secretion of these other cy t o- specific molecule called an Avirulence flammatory phenotypes. kines; in addition, they show minor de- factor (Av r ) , t h e reby constituting a Coming back to cryopyrin, recall that ficits in some apoptotic pathways. “one R protein-one pathogen” recogni- its structure resembles a NOD protein, The interaction between ASC and pro- tion system. The consequence of this except that it has an N-terminal PyD caspase-1 potentiates pro-caspase-1 ac- Avr:LRR interaction is the ap o p t o t i c rather than a CARD. However, as it tivity by induced proximity and thereby death (the hypersensitive response) of happens, cryopyrin binds ASC, a pro- potently induces the processing and se-

S-49 REVIEW Inflammation, apoptosis and the pyrin domain / D.L. Gumucio et al.

Fig. 4. Bootstrapped consensus cladogram of pyrin domains from pyrin, ASC, POPI and the seven other PYPAF proteins identified to date. The pyrin domains of ASC (accession number N M _ 0 1 3 2 5 8 ) , POP1 (accession nu m b e r Fig. 3. Summary of molecular interactions and signaling pathways in which pyrin and cryopyrin funo- A F 4 5 4 6 6 9 ) , P Y PAF8 (accession nu m b e r tion. Both pyrin and cryopyrin interact with ASC and stimulate one or more of three ASC-mediated AY 0 5 4 9 8 6 ) , P Y PAF2 (accession nu m b e r- activities: apoptosis, NF- B activation (possibly through RICK) and caspase-l-activation, leading to A F 3 1 0 1 0 6 ) , P Y PAF3 (accession nu m b e r IL-1 processing and secretion. Caspase-1 activation can also trigger apoptosis, but it is not clear that A F 4 6 4 7 6 5 ) , P Y PAF5 (accession nu m b e r this is the path followed in this case. ASC itself, however, induces apoptosis, and this pathway is mod- X M _ 1 1 3 7 0 0 ) , P Y PAF4 (accession nu m b e r ulated by pyrin. NF- B activation is believed to have two, temporally separate activities: early induc- X M _ 0 8 5 9 7 2 ) , P Y PAF7 (accession nu m b e r tion of the pro-inflammatory response and later resolution of the inflammatory response.These two AY 0 9 5 1 4 6 ) , c ryo py rin (accession nu m b e r activities rely on different effector genes,and have opposite effects on apoptosis. It is likely that other A F 4 1 0 4 7 7 ) , DEFCAP (accession nu m b e r PYPAF proteins also figure in the upstream portion of these pathways. See text for details. NM_013258) and py rin (accession nu m b e r AF018080) were aligned using ClustalX. A max- i mum pars i m o ny phy l ogenetic tree was con- c retion of IL-1 , a major py roge n i c the NOD-like and PYPA F - l i ke pro- structed with PAUP version 4.0blO. A bootstrap and pro-inflammatory cytokine. Srini- teins. Experiments in several laborato- a n a lysis was perfo rmed with 1000 rep l i c at e s , using a full heuristic search and including groups vasula et al. suggest that the interaction ries had shown that expression of just compatible with 50% majority-rule consensus. of ASC and caspase-1 creates a signal- the interacting domains (e. g. , t h e Significant support for related sub-groups within ing complex,the , in res- CARD or PyD) of these proteins has a the fa m i ly is only found for the PYPA F 2 /- ponse to pro - i n fl a m m at o ry stimu l i dominant negat ive effect on effe c t o r PYPAF3 pair and for ASC/POP1. Some trees also supported a grouping for PYPAF1/crypopy- (49). PYPAF7 and its cousin, cryopyrin function. Ap p a re n t ly, these ex p e ri- rin and PYPAF7, but the bootstrap value for this (PYPAF1), can each impart a synergis- ments had already been done by the grouping, 57, was too low to be considered sig- tic activation of pro-caspase-1 and a re- genome itself. Two proteins, pseudo- nificant. sultant increase in IL-1 secretion (49). ICE (which has also been named COP, It remains to be tested whether other or CARD-only protein) and ICEBERG t ree from a bootstrapped maximu m PyD containing proteins known to (53-55), have been identified that con- parsimony analysis. The starburst di- interact with ASC (pyrin, DEFCAP and sist essentially of only a CARD domain stribution of branches suggests that an- the remaining PYPAFs) also regulate t h at is highly homologous to the cient gene duplications exploded the the activity of this putative inflamma- CARD domain of pro-caspase-1 (92% number of these proteins,and that since some. If so, then a more complete (if and 53%, respectively). Either of these that time, few additional gene duplica- still rather fuzzy) pat h way begins to p roteins can prevent the caspase-1 tions have taken place. However, two emerge: an inflammatory trigger sig- dependent processing and secretion of clusterings are well supported by the nals activation of PyD-containing pro- IL-1 . However, the two are function- b o o t s t rap analy s i s : P Y PAF2 and teins, recruitment of ASC, activation of a l ly diffe re n t , since pseudo-ICE can PYPAF3 as well as ASC and POP.The caspase-1 and processing and secretion activate NF- B and can enhance TNF- l atter pair invites some further func- of IL-1 . In addition, for some PyD induced NF-kB activation, but ICE- tional speculation. Recall that ASC ap- proteins (e.g., cryopyrin), ASC mediat- BERG cannot (54). pears to act as an adaptor for several of ed activation of NF- B is enhanced, Interestingly, the database also contains the PyD-containing pro t e i n s , t h o u g h adding to the pro-inflammatory signal, a PyD-only protein (POP), noted previ- not all have been tested for ASC bind- while for others (e. g. , py ri n ) , A S C - ously by Wang et al. (50). Since POP ing. Since the analysis shows that of all mediated apoptosis is altered (Fig. 3). has not yet been characterized, we exa- the PyD proteins, POP is most closely As might be predicted, these processes mined its relatedness to the PyDs of related to AS C, and since for all of the of apoptosis and inflammation require py ri n , ASC and seve ral of the other death domain-related structures includ- tight control. Thus, it is perhaps not PyD-containing proteins of the PYPAF ing the PyD, similarity breeds familiar- surprising that there are inhibitors of group. Figure 4 shows the consensus ity (i.e., interaction), POP might be ex-

S-50 Inflammation, apoptosis and the pyrin domain / D.L. Gumucio et al. REVIEW pected to bind ASC readily, thus in- gens, it is clear that at least one inciting We have examined here only a subset hibiting ASC’s interaction with other t ri gger is simply temperat u re, and it of the PyD-containing proteins related PyD motifs. If this prediction is borne seems unlikely that the attack-produc- to pyrin and cryopyrin. Several more out, POP’s future value may lie in its ing ligand (if there is one) for cryopy- lurk in the database. However, it is al- therapeutic potential for the treatment rin, in FCU at least, is a bacterial or ready possible to appreciate the amaz- of those infl a m m at o ry diseases in v i ral product. [NOMID/CINCA may ing interconnectedness of the mole- wh i ch ASC may play a role (FMF, differ in this regard.] Thus, there may cular interactions that encircle the cri- FCU/MWS, CINCA/NOMID, and po- be an intracellular, stress induced mole- tical processes: inflammation, apopto- tentially others). cule to which the cryopyrin LRR can sis, NF- B activation, caspase-1 acti- It is interesting that the intera c t i o n s respond.Then there is the question of vation and cytokine secretion. Th e s e between PyD proteins in general seem the mutations, all of which are thus far interactions describe a remarkable set to be weak, or perhaps unstable. For o b s e rved in and around NAC H T /- of stru c t u ral and functional para l l e l s example, using GST pull-down assays, NBS/NOD domain (the domain between pathogen recognition systems S ri n ivasula et al. found no ev i d e n c e responsible for oligomerization). Inte- and components of the innate immune that ASC forms stable complexes with re s t i n g ly, this is true for FCU and p at h way inside of mammalian cells. pyrin, cryopyrin, DEFCAP or PYPAF3 MWS as well as NOMID, despite the The large number of remaining ques- (49). However, other laboratories, us- phenotypic diffe rences in these thre e tions still to be answered, however, is ing yeast two hybrid screens, immu- syndromes. Could it be that one type of sure to keep this field “hot” for some noprecipitation, immunolocalization or mutant molecule in the oligomer group years to come. the mammalian two - hy b rid system, leads to failure to reach a critical mass find evidence for intera c t i o n s , a l b e i t for induced proximity activation, while References 1. GALON J , AKSENTIJEVICH I , McDERMOTT F, lower affinity ones (7, 11, 50). Perhaps another mu t ation causes homo-oligo- O’SHEA J, KASTNER DL: TNFRSF1A muta- these weak interactions are important. merization to occur too readily ? Do tions and autoinflammatory syndromes. Curr CARD-CARD domain intera c t i o n s some mutations alter the conformation Opin Immunol 2000; 12: 479-86. may be of high affinity because most of cryopyrin so as to change either the 2.INTERNATIONAL FMF CONSORTIUM: Ancient missense mutations in a new member of the are critical for the kind of velcro-like upstream sensing or downstream effec- RoRet gene family are likely to cause familial stable recruitment needed for induced tor functions? More data is needed. Mediterranean fever. Cell 1997; 90: 797-807. p roximity activation. On the other And what of pyrin? In the region out- 3.FR E N C H FMF CO N S O RT I U M: A candidat e h a n d, PyD-PyD interactions may be side of the PyD, it is clearly the most gene for familial Mediterranean feve r. Nat Genet 1997; 17:25-31. primarily of a regulatory nature; thus, divergent of the PyD containing pro- 4. HOFFMAN HM, MUELLER JL, B ROIDE DH, by necessity, easy to perturb to allow teins. Most of its mutations fall within WANDERER AA, KOLODNER RD: Mutation of for a sensitive on/off switch. A propen- the C-terminal rfp or B30.2 domain a new gene encoding a putative pyrin-like pro- sity to act as a regulatory switch could (58), suggesting that this domain is cri- tein causes familial cold autoinfl a m m at o ry syndrome and Muckle-Wells syndrome. Nat also ultimat e ly underlie the fact that tical for the function of the pro t e i n . Genet 2001; 29: 301-5. mutations in these proteins cause dis- However, the mouse and rat genomic 5. F E L D M A N N J, PRIEUR A M , QUA RTIER P e t eases that present in a “periodic” fash- sequences both contain a fra m e s h i f t al.: Chronic infantile neurological cutaneous and articular syndrome is caused by mutations ion. A potential modulat o ry role has that results in the expression of a trun- in CMS1, a gene highly expressed in polymor- already been suggested for pyrin (in re- cated form of pyrin that lacks the rfp phonuclear cells and chondrocytes. Am J Hum gulating ASC induced apoptosis, ref. domain entirely (59)! How is it that this Genet 2002; May 24 (epub ahead of print). 1 1 ) , c ryo py rin (in activating NF- B d o m a i n , so seemingly critical to the 6. BE RT I N J, Di ST E FA N O P S : The PYRIN do- main:A novel motif found in apoptosis and in- and caspase-1-dependent IL-1 function of pyrin, as judged by the ef- flammation proteins. Cell Death Differ 2000; p rocessing re f. 28, 5 0 ) , P Y PAF7 (in fects of the mutations that it harbors, is 7:1273-4. regulating NF- B activation and cas- missing altogether in rodents? Extrapo- 7. MARTINON F, HOFMANN K, TSCHOPP J: The p a s e - 1 - d ependent IL-1 p ro c e s s i n g, lating from the function of other PyD- pyrin domain:A possible member of the death d o m a i n - fold fa m i ly implicated in ap o p t o s i s ref. 50), and DEFCAP (in modulating containing proteins, it is tempting to and inflammation. Curr Biol 2001; 11: R118- Ap a f - 1 - d ependent ap o p t o s i s , re f. 56, think that this apparent paradox could 20. 57). be reconciled if the C-terminal rfp do- 8. STAUB E, DAHL E, ROSENTHAL A: The DAPIN Though some shape is beginning to main acts as an intramolecular regula- family:A novel domain links apoptotic and in- t e r fe ron response proteins. Trends Bioch e m emerge from the connection of these tory domain, functionally (but not stru- Sci 2001; 26: 83-5. molecular dots, t h e re are still many cturally) equivalent to Apaf-1’s WD40 9. PAW L OWSKI K, PIO F, CHU Z, REED JC, outstanding questions. For ex a m p l e, motifs or the LRRs of the NOD-like GODZIK A: PAAD - A new as- proteins. Would this mean that the rfp sociated with apoptosis, cancer and autoim- what are the triggers for the periodic in- mune diseases. Trends Biochem Sci 2001; 26: flammatory attacks suffered by indivi- domain binds a ligand that then con- 85-7. duals with FMF, MWS or FCS ? trols the on/off switch for pyrin func- 10. FA I R B ROT H E R W J, GORDON NC, H U M K E Though cryopyrin seems to look and tion? If so, what is the ligand? And is EW et al.:The PYRIN domain: a member of the death domain-fold superfamily. Protein function like a NOD protein, and might apoptosis the function (the only func- Sci 2001; 10: 1911-8. be expected to be able to sense patho- tion?) that is controlled? 11. RICHARDS N, SCHANER P, DIAZ A et al.:

S-51 REVIEW Inflammation, apoptosis and the pyrin domain / D.L. Gumucio et al.

Interaction between pyrin and the apoptotic ponse and evolution of familial Mediter- cifically by Cf-2. Proc Natl Acad Sci USA speck protein (ASC) modulates ASC-induced ranean fever. Lancet 1999; 353: 1415. 2000; 97: 8807-14. apoptosis. J Biol Chem 2001; 276: 39320-9. 28. MANJI GA, WANG L, GEDDES BJ et al.: 45. GIRARDIN SE, TOURNEBIZE R,MAVRIS M et 12. AR AV I N D L , D I X I T V M , KO O N I N E V: Th e P Y PA F 1 , a PYRIN-containing Ap a f 1 - l i ke al.: CARD4/Nod1 mediates NF-kappaB and domains of death: Evolution of the apoptosis protein that assembles with ASC and regu- JNK activation by invasive Shigella flexneri. machinery. Trends Biochem Sci 1999; 24: 47- lates activation of NF-kappa B. J Biol Chem EMBO Rep 2001; 2: 736-42. 53. 2002; 277: 11570-5. 46. OGURA Y, BONEN DK, INOHARA N et al.: A 13. HOFMANNK: The modular nature of apoptot- 29. INOHARA N, OGURA Y, NUNEZ G: Nods: A frameshift mutation in NOD2 associated with ic signaling proteins. Cell Mol Life Sci 1999; family of cytosolic proteins that regulate the s u s c eptibility to Cro h n ’s disease. N at u re 55: 1113-28. host response to pathogens. Curr Opin Micro - 2001; 411: 603-6. 14. MASUMOTO J, TANIGUCHI S,AYUKAWA K et biol 2002; 5: 76-80. 47. H A M P E J, C U T H B E RT A , C ROUCHER P J e t al.: ASC, a novel 22-kDa protein, aggregates 30. KOONIN EV, ARAVIND L:The NACHT fam- al.:Association between insertion mutation in d u ring apoptosis of human pro mye l o cy t i c i ly - A new group of predicted NTPa s e s NOD2 gene and Crohn’s disease in German l e u kemia HL-60 cells. J Biol Chem 1 9 9 9 ; implicated in apoptosis and MHC transcrip- and British populations. Lancet 2001; 357: 274: 33835-8. tion activation. Trends Biol Sci 2000; 25:223- 1925-8. 15. SALVESEN GS, DIXIT VM: Caspase activation: 4. 48. MICELI-RICHARD C, LESAGE S,RYBOJAD M The induced-proximity model. P roc Nat l 31. BERTIN J, NIR WJ, FISCHER CM et al.: Human et al.: CARD15 mutations in Blau syndrome. Acad Sci USA: 1999; 96: 10964-7. CARD4 protein is a novel CED-4/Apaf-1 cell Nat Genet 2001; 29:19-20. 16. SHI Y: : The cellular engine for death family member that activates NF-kap- 49. SRINIVASULA SM, POYET J-L, RAZMARA M, the activation of caspase-9. Structure (Camb) paB. J Biol Chem 1999; 274: 12955-8. DATTA , P, ZHANG Z, ALNEMRI E S:Th e 2002; 10: 285-8. 32. INOHARA N,KOSEKI T, DEL PESO L et al.: PYRIN CARD protein ASC is an activating 17. ACEHAN D, JIANG X, MORGAN DG, HEUSER Nod1, an Apaf-1-like activator of caspase-9 a d aptor for caspase-1. J Biol Chem 2 0 0 2 ; JE, WANG X, AKEY CW:Three-dimensional and nu clear fa c t o r- k ap p a B. J Biol Chem (epub ahead of print). structure of the apoptosome: implications for 1999; 274: 14560-7. 50. WANG L, MANJI G, GRENIER JM et al. : assembly, procaspase-9 binding, and activa- 33. METZSTEIN MM, STANFIELD GM, HORVITZ PYPAF7: A novel PYRIN-containing Apaf1- tion. Mol Cell 2002; 9: 423-32. HR: Genetics of programmed cell death in C. like protein that regulates activation of NF-kB 18. CENTOLA M, WOOD G, FRUCHT DM et al.: e l ega n s:Pa s t , p resent and future. Tre n d s and caspase- 1-dependent cytokine process- The gene for familial Mediterranean fever, Genet 14: 410-16. ing. J Biol Chem 2002; May 17 (epub ahead MEFV, is expressed in early leukocyte devel- 34. ZOU H, HENZEL W J, LIU X, LUTSCHG A , of print). opment and is regulated in response to in- WANG X: Apaf-1, a human protein homolo- 51. KUIDA K,ZHENG TS, NA S et al.:Altered cy- flammatory mediators. Blood 2000; 95:3223- gous to C. elegans CED-4, participates in cy- tokine export and apoptosis in mice deficient 31. tochrome c-dependent activation of caspase- in interleukin-lb converting enzyme. Science 19. LIVNEH A, AKSENTIJEVICH I, LANGEVITZ P 3. Cell 1997; 90: 405-13. 1995; 267: 2000-3. et al.: A single mutated MEFV allele in Israeli 35. LI P, NIJHAWAN D,BUDIHARDJO I, SRINIVA- 52. LI P, ALLEN H, BANERJEE S, SESHADRI T: p atients suffe ring from familial Mediter- SULA SM,AHMAD M,ALNEMRI ES,WANG X: Characterization of mice deficient in inter- ranean fever and Behcet’s disease (FMF-BD). Cytochrome c and dATP-dependent forma- leukin-1 beta conve rting enzyme. J Cell Eur J Hum Genet 2001; 9: 191-6. tion of Apaf-1/caspase-9 complex initiates an Biochem 1997; 64: 27-32. 20. TOUITOU I, MAGNE X, MOLINARI N et al.: apoptotic protease cascade. C e l l 1997; 91: 53. HUMKE EW, SHRIVER SK, STA R OVA S N I K MEFV mutations in Behçet’s disease. Hum 479-89. M A , FA I R B ROTHER W J, D I X I T V M : I C E- Mutat 2000; 16: 271-2. 36. SRINIVASULA SM, AHMAD M, FERNANDES- BERG: A novel inhibitor of interleukin-1beta 21. CATTAN D, NOTARNICOLA C, MOLINARI N, ALNEMRI T, ALNEMRI E S:Autoactivation of generation. Cell 2000; 103: 99-111. TOUITOU I: Inflammatory bowel disease in procaspase-9 by Apaf-1-mediated oligomer- 54. DRUILHE A,SRINIVASULA SM,RAZMARA M, non-Ashkenazi Jews with familial Mediter- ization. Mol Cell 1998; 1: 949-57. AHMAD M, ALNEMRI ES: Regulation of IL- ranean fever. Lancet 2000; 355: 378-9. 37. INOHARA N, OGURA Y, CHEN FF, MUTO A, 1beta ge n e ration by Pseudo-ICE and ICE- 22. BOOTH DR, LACHMANN HJ, GILLMORE JD, NUNEZ G: Human Nod1 confers responsive- B E R G, t wo dominant negat ive caspase re- BOOTH SE, HAWKINS PN: Prevalence and si- ness to bacterial lipopolysaccharides. J Biol cruitment domain proteins. Cell Death Differ gnificance of the familial Mediterranean fever Chem 2001; 276: 2551-4. 2001; 8: 649-57. gene mu t ation encoding py rin Q148. Q J M 38. OGURA Y, INOHARA N, BENITO A, CHEN FF, 55. LEE SH, STEHLIK C,REEDJC: Cop,a caspase 2001; 94: 527-31. YAMAOKA S, NUNEZ G: Nod2,a Nodl/Apaf-1 re c ruitment domain-containing protein and 23. SCHANER P, RICHARDS N,WADHWA A, AK- family member that is restricted to monocytes inhibitor of caspase-1 activation processing. J SENTIJEVICH I, KASTNER D, TUCKER P, GU- and activates NF-kappaB. J Biol Chem 2001; Biol Chem 2001; 276: 34495-500. MUCIO D: Episodic evolution of pyrin in pri- 276: 4812-8. 56. HLAING T, GUO RF, DILLEYKA et al.: Mole- m at e s : human mu t ations re c ap i t u l ate an- 39. INOHARA N,KOSEKI T, LIN J et al.: An in- cular cloning and characterization of DEF- cestral amino acid states. Nat Genet 2001; 27: duced proximity model for NF-kappa B acti- CAP-L and -S, two isoforms of a novel mem- 318-21. vation in the Nodl/RICK and RIP signaling ber of the mammalian Ced-4 family of apop- 24. AKSENTIJEVICH I, TOROSYAN Y, SAMUELS J pathways. J Biol Chem 2000; 275: 27823-31. tosis proteins. J Biol Chem 2001; 276: 9230- et al.: Mutation and haplotype studies of fa- 40. TAK PP, FIRESTEIN GS: NF- B:A key role in 8. milial Mediterranean fever reveal new ances- inflammatory diseases. J Clin Invest 2001; 57. CHU ZL,PIO F, XIE Z et al.:A novel enhancer tral relationships and evidence for a high car- 107: 7-11. of the Ap a fl apoptosome invo l ved in cy- rier frequency with reducedpenetrance in the 41. LI X, STARK GR: NF B-dependent signaling tochrome c-dependent caspase activation and A s h kenazi Jewish population. Am J Hum pathways. Exp Hematol 2002; 30: 285-96. apoptosis. J Biol Chem 2001; 276: 9239-45. Genet 1999; 64: 949-62. 42. LAWRENCE T, GILROY DW, COLVILLE-NASH 58. TOUITOU I: The spectrum of familial Medi- 25. STOFFMAN N, MAGAL N, SHOHATT et al.: PR,WILLOUGHBY DA:Possible new role for terranean fever (FMF) mutations. Eur J Hum Higher than expected carrier rates for familial NF B in the resolution of inflammation. Nat Genet 2001; 9: 473-83. Mediterranean fever in various Jewish ethnic Med 2001; 7: 1291-7. 59. CHAE JJ, CENTOLA M, AKSENTIJEVICH I et groups. Eur J Hum Genet 2000; 8: 307-10. 43. FEYS BJ,PARKER JE: Interplay of signaling al.: Isolation, genomic organization, and ex- 26. BOOTH DR, GILLMORE JD, LACHMANN HJ pathways in plant disease resistance. Trends pression analysis of the mouse and rat ho- et al. :The genetic basis of autosomal domi- Genet 2000; 16: 449-55. mologs of MEFV, the gene for familial medi- nant familial Mediterranean fever. QJM 2000; 44. DIXON MS, GOLSTEIN C, THOMAS CM, VAN t e rranean feve r. Mamm Genome 2000; 11: 93: 217-21. DER BIEZEN EA,JONES JD: Genetic complex- 428-35. 27. TUNCA M, KIRKALI G, SOYTURK M,AKAR S, ity of pathogen perception by plants: The ex- 60. WHITHAM S, DINESH-KUMAR SP, CHOI D, PEPYS MB, HAWKINS PN: Acute phase res- ample of Rcr3, a tomato gene required spe- HEHL R, CORR C, BAKER B: The product of

S-52 Inflammation, apoptosis and the pyrin domain / D.L. Gumucio et al. REVIEW

the tobacco mosaic virus resistance gene N: al.: (Leucine rich repeat structure) MMDB 2000; 408: 111-5. similarity to toll and the interleukin-1 recep- Id: 17799; PDB Id: 1H6T. 67. PHELAN P, GORFE AA, JELESAROV I, MARTI tor. Cell 1994; 78:1101-15. 64. SHI Y: (CARD structure) MMDB Id: 13221 D N,WA RWICKER J, BOSSHARD H R: S a l t 61. MINDRINOS M, K ATAGIRI F, YU GL, AU S- PDB Id: 2YGS. bridges destabilize a leucine zipper designed UBEL FM: The A. thaliana disease resistance 65. CARR S , PENFOLD CN ,BAMFORD V, JAMES for maximized ion pairing between helices. gene RPS2 encodes a protein containing a R, HEMMINGS AM: The structure of TolB, an Biochemistry 2002; 41: 2998-3008. nucleotide-binding site and leucine-rich re- essential component of the tol-dep e n d e n t 68. BORDEN KLB, FREEMONT PS: (Zinc finger peats. Cell 1994; 78: 1089-99. translocation system, and its protein-protein structure) MMDB Id: 5201 PDB Id: 1FRE. 62. JAROSZEWSKI L, RYCHLEWSKI L, REED JC, interaction with the translocation domain of 69. DAY CL, ALBER T: Crystal structure of the GODZIK A: ATP-activated oligomerization as colicin E9. Structure Fold Des 2000; 8: 57- a m i n o - t e rminal coiled-coil domain of the a mechanism for apoptosis regulation: fold 66. APC tumor suppressor. J Mol Biol 2000; 301: and mechanism prediction for CED-4. Pro - 66. X U Y, TAO X, SHEN B, HORNG T et al..: 147-56. teins 2000; 39: 197-203. Structural basis for signal transduction by the 70. SETO MH, LIU HL, ZAJCHOWSKI DA, WHIT- 63. SCHUBERTW-D, GOBEL G, DIEPHOLZ M et To l l / i n t e rleukin-1 re c eptor domains. Nat u r e LOW M: Protein fold analysis of the B30.2-

S-53