Chapter 1 Introduction: Aspartic and Glutamic Peptidases and Their Clans Aspartic, glutamic and metallopeptidases differ impor- enzymes from family A1 show that the molecules are bi- tantly from serine, threonine and cysteine peptidases in lobed, with the active site located between the lobes (see that the nucleophile that attacks the scissile peptide Figure 1.1). Many peptidases have bi-lobed molecules, bondisanactivatedwatermoleculeratherthanthenucleo- but in family A1, one lobe has evolved from the other by philic side chain of an amino acid. Residues that are gene duplication. Each lobe bears one of the Asp residues involved in catalysis include amino acids that act as of the catalytic dyad, and the two residues are homolo- ligands, either directly for the activated water molecule gous. Inthe modern-dayenzymes,only very limitedsimi- as in aspartic and glutamic peptidases, or by binding one larities in the amino acid sequences of the two lobes or two metal ions that in turn bind the activated water remain, but the three-dimensional structures are very sim- moleculeinthemetallopeptidases. ilar. The discovery of this situation raised the question of Aspartic peptidases are so named because Asp residues how the product of the original gene might have func- are ligands of the activated water molecule. In the great tioned as a peptidase if it consisted of only one lobe with majority of known aspartic peptidases a pair of aspartic only one aspartic. One possible answer was supplied residuesacttogethertobindandactivatethecatalyticwater when the crystal structure of a retropepsin (see molecule (Chapter 2) but in some, residues of other amino Figure 1.2) from family A2 was solved, showing a single acids replace the second Asp. In glutamic peptidases, one lobe carrying one catalytic Asp, very similar in structure ortwoGluresiduesbindandactivatethecatalyticwater. to one lobe of pepsin. The retropepsin is active only as a One notable characteristic of aspartic and glutamic homodimer, and the catalytic site forms between the peptidases is that all the enzymes so far described are molecules. The similar protein folds of the endopepti- endopeptidases, although there is no obvious reason why dases in families A1 and A2 fully support the view that this type of catalytic mechanism could not exist in an they have evolved from a common ancestor, and they are exopeptidase. grouped together in clan AA. This can be seen in Aspartic peptidases are assigned to clans AA, AC, Figure 1.3, which shows that members of the clan have AD, AE and AF. Tertiary structures solved for members similar secondary structures, and that each active site of clans AA, AD, AE and AF each show a unique protein asparticisattheendofastrand. foldunrelatedtothatofanyotherpeptidase.ForclanAC, Clan AA contains a number of additional families. there is as yet no crystal structure, but other criteria point Structures are available only for members of families to its distinctiveness. Two clans of glutamic peptidases A9 and A28, but there is conservation of the Xaa-Xaa- are recognized, GA and GB, each identified by a unique Asp-Xbb-Gly-Xbb motif around the catalytic Asp, in tertiary structure and containing just one family. Several which Xaa is a hydrophobic amino acid and Xbb is peptidase families, previously thought to be aspartic pep- either Ser or Thr (Figure 1.4) in active peptidases from tidases, are now known not to be peptidases at all, but all families. The peptidases in these families all seem lyases, catalyzing cleavage because of a property specific to have somewhat similar functions. They are monomeric to asparagine, which does not involve hydrolysis. The proteins that are assumed to require dimerization for clansandfamiliesofaspartic andglutamicpeptidasesand activity. Peptidases from families A3 and A9 act to pro- asparaginepeptidelyasesaresummarizedinTable1.1. cess viral polyproteins, whereas peptidases from family A11 process polyproteins from retrotransposons. Families A28, A32 and A33 are all retropepsin-like and peptidases Clan AA from these families are only active as homodimers. Clan AA contains the large family A1, for which the type Families A28 and A33 are found in eukaryotes, whereas example is pepsin. Crystal structures for a number of the familyA32isfoundonlyinbacteria. 3 4 1.Introduction:AsparticandGlutamicPeptidasesandTheirClans TABLE1.1 Clansandfamiliesofasparticpeptidases.Foreachclantheresiduesdescribedforsimplicityas‘Catalytic residues’mightmoreaccuratelybedescribedas‘Ligandsofthenucleophilicwatermolecule’.Foreachfamily,the peptidasenamedisthetypeexampleusedasthefoundationforthefamilyintheMEROPSdatabase AsparticPeptides ClanAA Catalyticresidues:Asp,Asp(orHis) Fold:doubleβ-barrel(ordimerofsingleβ-barrel) Family A1A pepsinA(Homosapiens) A1B nepenthesin(Nepenthesgracilis) A2A HIV-1retropepsin(humanimmunodeficiencyvirustype1) A2B Ty3transposonpeptidase(Saccharomycescerevisiae) A2C Gypsytransposonpeptidase(Drosophilamelanogaster) A2D Osvaldoretrotransposonpeptidase(Drosophilabuzzatii) A2E retrotransposonpeptidase(Schizosaccharomycespombe) A2G retrotransposon17.6peptidase(Drosophilamelanogaster) A2H retropepsin(walleyedermalsarcomavirus) A3A cauliflowermosaicvirus-typeendopeptidase(cauliflowermosaicvirus) A3B bacilliformvirusputativeprotease(ricetungrobacilliformvirus) A9 spumapepsin(humanspumaretrovirus) A11A Copiatransposon(Drosophilamelanogaster) A11B Ty1transposonpeptidase(Saccharomycescerevisiae) A28 DNA-damageinducibleprotein1(Saccharomycescerevisiae) A32 PerPpeptidase(Caulobactercrescentus) A33 skinasparticprotease(Musmusculus) AC Catalyticresidues:Asp,Asp Fold:unknown Family A8 signalpeptidaseII(Escherichiacoli) AD Catalyticresidues:Asp,Asp Fold:unknown Family A22A presenilin1(Homosapiens) A22B impas1peptidase(Homosapiens) A24A typeIVprepilinpeptidasetypeM1(Pseudomonasaeruginosa) A24B preflagellinpeptidase(Methanococcusmaripaludis) AE Catalyticresidues:Asp(orGlu),Asp, Fold:centralbetasheetsurroundedbyhelices His(orLys) Family A25 gprpeptidase(Bacillusmegaterium) A31 HybDpeptidase(Escherichiacoli) AF Catalyticresidues:Asp,Asp,Asp,His Fold:β-cylinder Family A26 omptin(Escherichiacoli) OtherFamilies (clan‘AX’) A5 thermopsin(Sulfolobusacidocaldarius) A36 sporulationfactorSpoIIGA(Bacillussubtilis) GlutamicPeptides ClanGA Catalyticresidues:Gln,Glu Family G1 scytalidoglutamicpeptidase(Scytalidiumlignicolum) (Continued) 1.Introduction:AsparticandGlutamicPeptidasesandTheirClans 5 TABLE1.1 (Continued) ClanGB Catalyticresidues:Glu pre-neckappendageprotein(Bacillusphagephi29) Family G2 PeptideLyases ClanNA Catalyticresidues:Asp(orGlu),Asn Fold:β-sandwich Family N1 nodavirusendopeptidase(flockhousevirus) N2 tetravirusendopeptidase(Nudaureliacapensisomegavirus) N8 capsidVP0-typeself-cleavingprotein(humanpoliovirus1) ClanNB Catalyticresidues:Asn Fold: Family N6 YscUprotein(Yersiniapseudotuberculosis) ClanNC Catalyticresidues:Asn Fold: Family N7 coatprotein(mammalianorthoreovirus1) ClanND Catalyticresidues:Asn,Tyr,Glu,Arg Fold: Family N4 Tsh-associatedself-cleavingdomain(Escherichiacoli) ClanNE Catalyticresidues:Asn Fold: Family N5 self-cleavingprotein(humanpicobirnavirus) ClanPD(N) N9 intein-containingV-typeprotonATPasecatalyticsubunitA (Saccharomycescerevisiae) N10 intein-containingreplicativeDNAhelicaseprecursor(Synechocystissp. PCC6803) N11 intein-containingchloroplastATP-dependentendopeptidase (Chlamydomonaseugametos) FIGURE1.1 RichardsondiagramofthehumanpepsinA/pepstatin FIGURE1.2 Richardson diagramofthehumanimmunodeficiency complex. The image was prepared from the Brookhaven Protein Data type 1 retropepsin dimer. The image was prepared from the Bankentry(1PSO)asdescribedintheIntroduction(p.li).Catalyticresi- Brookhaven Protein Data Bank entry (5HVP) as described in the dues are shown in ball-and-stick representation: Asp32 and Asp215 in Introduction (p. li). The catalytic residue Asp25 is shown in ball-and- pinkandTyr75 in green(numberingas in thePDB entry).Pepstatinis stick representation in pink (one from each monomer; numbering as in showningreyinball-and-stickrepresentation. thePDBentry). 6 1.Introduction:AsparticandGlutamicPeptidasesandTheirClans FIGURE1.3 SecondarystructurediagramformembersofclanAA.TheimagewaspreparedasdescribedintheIntroduction(p.li). BesidesthemotifaroundthecatalyticAsp,asecondmotif isalsoconservedinfamiliesA1,A2andsomeotherfamilies ofclanAA.Thisisknownasthe‘hydrophobic-hydrophobic- Gly’ motif (Figure 1.4). In family A1, this motif is found only in the N-terminal domain. Although the function of the hydrophobic-hydrophobic-Gly motif is not understood, in family A2 it immediately precedes the invariant residue Arg123 (numbered according to human prepropepsinogen A) thatisbelievedtobeimportantfordimerization[1]. Family A1 The peptidases of family A1 are found in eukaryotes with only a handful of homologs known from marine bacteria and bacteria that are plant commensals [2], and like all aspartic peptidases they are endopeptidases. The great majorityaremostactiveatacidicpH,butafewshowactiv- FIGURE 1.4 Alignment around the active site residues and con- ityunderneutralconditions.Examplesoftheenzymesfrom served motifs for families in clan AA. Sequencesshownarefromthe mammalsincludethedigestiveenzymespepsinandchymo- typeexampleforeachfamilyorsubfamily.Numberingisaccordingtothat sin and the intracellular cathepsins D and E. Renin, which ofhumanprepepsinogenA.RegionsaroundboththeN-terminal(A1(N)) andC-terminal(A1(C))activesiteasparticsofhumanpepsinAareshown. originatesfromthekidneyandprocessesangiotensinogenin 1.Introduction:AsparticandGlutamicPeptidasesandTheirClans 7 FIGURE 1.7 Richardson diagram of Rhizomucor pusillus mucor- pepsin. The image was prepared from the Brookhaven Protein Data Bank entry (1MPP) as described in the Introduction (p. li). Catalytic FIGURE 1.5 Richardson diagram of pig pepsinogen A. The image residues are shown in ball-and-stick representation: Asp32 and Asp215 waspreparedfromthe BrookhavenProteinDataBankentry(3PSG)as inpink,andTyr75ingreen(numberingasinthePDBentry). described in the Introduction (p. li). The propeptide is shown in gray. Residues shown in ball-and-stick representation are: the catalytic resi- Thehomologyofthetwolobesofthemoleculeofapep- duesAsp32andAsp215(numberingasinthePDBentry);Asp11,which interactswith the propeptide, in pink; and Tyr37P (numbering the pro- sin-like peptidase is reflected by the fact that the two Asp peptide1Pto44P),whichblocksthesubstrate-bindingsite,ingreen. residues, Asp94 and Asp277 (numbering according to Alignment A1) occur in very similar sequence motifs. The core motif is commonly summarized as Asp-Thr-Gly (the ‘DTG’ motif) but a fuller description is Xaa-Xaa-Asp-Xbb- Gly-Xbb,inwhichXaaisahydrophobicresidueandXbbis either Ser or Thr. As can be seen in the sequence alignment in the MEROPS database, there are a few exceptions. Humanrenin,whichisactiveatneutralpH,hasAla280,and thisalsooccursinalmostalloftheretropepsinsfromfamily A2. It may be that this residue affects the pH optimum of the enzymes, because lack of a hydrogen bond to the usual Thr278 alters the acidity of Asp277 [4]. Histoaspartic pepti- dase from Plasmodium falciparum has the second aspartic replacedbyhistidine,andisknowntobecatalyticallyactive [5].Manyofthemammalianpregnancy-associatedglycopro- teins are non-peptidase homologs in which Asp277 is replaced byGly (Chapter 39). Acaution is needed in regard FIGURE 1.6 Richardson diagram of the human cathepsin D/pep- statincomplex.The imagewas prepared from the Brookhaven Protein totheDTGmotif:asimilarmotifisconservedinthesubtili- Data Bank entry (1LYB) as described in the Introduction (p. li). sin family, which can be described as Xaa-Xaa-Asp-Xbb- Catalyticresiduesareshowninball-and-stickrepresentation:Asp33and Gly-Xaa (in which the Asp is a member of the catalytic Asp231 in pink, and Tyr78 in green (numbering as in the PDB entry). triad) the principal difference being in the final residue of Carbohydrates are shown as CPK spheres; the carbohydrate containing the motif. Care must therefore be exercised in using the mannoseisatthetoprightofthemolecule.Pepstatinisshowningray inball-and-stickrepresentation. DTGmotiftoassignapeptidasetothepepsinfamily. There is a third residue in addition to the two aspartics theplasma,isoneoftheminorityofenzymesinthefamily that is important for catalysis in family A1. Although the thatnormallyactatneutralpH.Pepsinhomologsareknown two lobes are structurally similar, it will be noted that from plants (e.g. phytepsin, Chapter 21) from fungi (e.g. thelobe shownonthe right-handsideof Figure 1.1has an penicillopepsin, Chapter 26) and from protozoa (plasmep- extra β-hairpin loop capping the active site. This loop has sins, Chapter 15). Homologs from bacteria have now been been termed the ‘flap’, and carries residues important for identifiedfromgenomesequencingprojects[2]andthepep- specificity,notablyTyr137andThr139,whichcaninteract tidasefromShewanellaamazonensishasbeencharacterized withthesubstrateandarepartoftheS1subsite.Homologs [3]. Tertiary structures have been determined for several are known in which Tyr137 is replaced by Phe, and in the membersofthefamily(seeFigures1.1,1.5,1.6,1.7). histoasparticpeptidaseTyr137isreplacedbySer. 8 1.Introduction:AsparticandGlutamicPeptidasesandTheirClans Many fungal enzymes, including penicillopepsin, are Memapsin-1 (Chapter 13) and memapsin-2 (Chapter able to activate trypsinogen by cleavage of a LyskIle 14)aremembrane-boundendopeptidaseseachwithatrans- bond. As can be seen from the alignment in MEROPS, in membrane domain followed by a cytoplasmic domain at penicillopepsin and other fungal endopeptidases, Thr139 the C-terminus. Memapsin-2 is capable of cleaving the is replaced by Asp, which has been shown to be the membrane-bound amyloid protein precursor to release the anionicbindingsite[6]. amyloid protein that forms deposits in the brains of Most enzymes of family A1 are synthesized with sig- patientssufferingAlzheimer’sdisease,andisaβ-secretase. nal and propeptides, and enter the secretory pathway. The plant aspartic endopeptidase phytepsin (Chapter 21) Human pepsin A has a 15-residue signal peptide and a remains the only mosaic member of family A1 so far 47-residue propeptide. The other animal members of the described,containinga104-residueinsertintheC-terminal family have propeptides homologous to that of pepsino- lobe.Thisishomologoustopartofthemammaliansaposin gen A, but the propeptides of plant and fungal members, precursor, which contains four saposin peptides. Once though similar in size, show little relationship to each proteolytically excised from their precursor, saposins are other or to the pepsinogen A propeptide. The mechanism lysosomal proteins that act as activators for β-galactosyl- of formation of the catalytically active pepsin from its ceramidase and β-galactosidase. The phytepsin insert does pepsinogen precursor is described in Chapter 2. Some notcorrespondtoasinglesaposindomain,butratherover- pepsin homologs are membrane proteins, for example the lapstheC-terminalthirdofonedomainandtheN-terminal plasmepsin precursors from the malarial parasite two-thirds of the next. Not only cysteine residues but also Plasmodium falciparum are type II membrane proteins, potential N-linked glycosylation sites are conserved in this which are then internalized with the substrate hemoglobin novelpartofthesequence. into the food vacuole where they are activated by being Aphylogenetictreefor familyA1showsthatthemost released from the membrane (see Chapter 15). Yapsin 1 divergent sequences are those of nucellin and related pro- (Chapter 36) is membrane-linked through a glycosylpho- teinsfromplants.Consequently,familyA1isdividedinto sphatidinylinositol anchor. Nucellin and other homologs two subfamilies, with subfamily A containing predomi- from Arabidopsis appear to be intracellular proteins (see nantly animal and fungal peptidases, whereas subfamily Chapter 22). Bacterial homologs lack N-terminal signal B contains peptidases from plants. Nucellin (not yet peptidesandpropeptides[2]. shown to be a peptidase) is found in the nucellus of the The gene structures of vertebrate pepsin homologs germinating seed. The large set of pepsin homologs in show striking conservation of intron/exon junctions and Arabidopsis thaliana, in contrast to other eukaryotes, has intron phases, reflecting gene duplication events that have been identified as a lineage-specific protein cluster [7]. A occurred subsequent tothe introductionofintrons intothe similar cluster of plant-specific peptidases is known in ancestral gene. Genes of pepsin homologs from some family S8. Memapsins and plasmepsins (Chapters 15 and plants and fungi show a much more diverse pattern of 16) also are derived from very ancient divergences and introns and exons, presumably because the divergence of are integral membrane proteins. It is not possible to pro- animals,plantsandfungiprecededtheacquisitionofintrons. vide a time scale for the tree, because the mammalian Intronsareabsentfromthegenes offamilyA1 members in gastric enzymes (e.g. gastricsin, pepsin A) are evolving Drosophilamelanogaster.Thismayexplainwhyfewpepsin fasterthanthevacuolarandlysosomalenzymes(e.g.plas- homologs are multidomain proteins, because such proteins mepsin, cathepsin D). (A similar situation is observed in may be theresult of exon-shuffling which would only have peptidase family C1 where the plant vacuolar peptidase been possible when introns were present in the ancestral aleurain is unusually closely related to lysosomal cathep- genes. sin H.) Although examples of the family are known from There is a striking conservation of disulfide bonds some single-celled organisms, such as species of in the secreted peptidases of family A1 in mammals; Plasmodium, Eimeria and Dictyostelium, there is no pep- they have three disulfide bonds, the first and third of sin homolog in the completely sequenced genome of the which are also present in many of the fungal homologs. pathogenic microsporidian Encephalitozoon cuniculi or in The disulfide bonds are different in the different lobes, thezooflagellates LeishmaniaorTrypanosoma.Nohomo- implying that the bonds were introduced after the gene loghasbeenidentifiedinanyarchaean,andthehomologs duplication that gave rise to the two domains. Cathepsin thatare knownfrom bacteria have avery limiteddistribu- D (Chapter 8; Figure 1.6) has an additional disulfide tion, including the genera Shewanella and Marinomonas. bond. Cathepsin E (Chapter 6) is unusual in that the pro- It is not clear whether homologs in single-celled organ- enzyme is a disulfide-linked dimer, and the active isms and bacteria have resulted from ancient horizontal enzyme can be either a dimer or a monomer. Although gene transfers or if a pepsin homolog was present in the cysteine residues are conserved between memapsins-1 ancestor of all living things and was subsequently lost in and-2,theyaredifferentlyplacedtothoseinpepsin. Archaeaandsomeprotozoanphyla. 1.Introduction:AsparticandGlutamicPeptidasesandTheirClans 9 Other Families in Clan AA the serine or threonine that occurs in most members of familyA1,butisalanineinrenin.Bothreninandretropep- Family A2 contains polyprotein-processing enzymes from sins act at neutral pH, and Ala35 may be important for a positive-strand RNA viruses and retrotransposons. Viral neutral pH optimum [4]. A subset of retropepsins from RNA encodes a number ofpolyproteins,and inthe human oncoviruses (e.g. Mason(cid:1)Pfizer virus retropepsin) and immunodeficiency virus the genes are known as gag, pol avian retroviruses (e.g. Rous sarcoma virus retropepsin) andenv.Thegagandenvgenesencodestructuralproteins, are larger proteins with N-terminal extensions, some of whereas the pol gene encodes three enzymes that are nec- whicharehomologoustoEscherichiacolidUTPase. essaryfor: (a) transcribing the viral RNA toDNA (reverse Peptidases of family A2 are found in retrotransposons transcriptase); (b) incorporating the DNA into the host as well as retroviruses. Uncontrolled multiplication of genome (ribonuclease H); and (c) processing the polypro- retrotransposons throughout a genome could damage it, tein (retropepsin). In some of the viruses (e.g. Rous sar- so it is not surprising that mechanisms exist whereby ret- coma virus) the retropepsin is part of the Gag polyprotein, rotransposons are inactivated. It has been estimated that whereas in others (e.g. Mason(cid:1)Pfizer monkey virus) the a large proportion of a multicellular eukaryote genome retropepsin is encoded by a separate RNA gene known as consists of inactivated retrotransposon sequences. vprt. In many viruses, the gag and pol genes are tran- Inactivation can be by mutation of the catalytic Asp of scribed contiguously because of a ribosomal frameshift, the protease, or by a more general scrambling of the ret- and a Gag-Pol polyprotein is synthesized by the infected rotransposon sequence. As a result, there are a number of cells.Theretropepsinisrequiredforprocessingofallthree non-peptidasehomologsinfamilyA2. polyproteins, although initial stages of Env polyprotein An evolutionary tree for family A2 shows considerable processing are performed by cellular enzymes. In retro- divergence in sequence, probably because of the high rate transposons (e.g. yeast Ty3 transposon endopeptidase [8]) of mutation in RNA viruses (which lack an efficient error- the endopeptidase is part of a pol polyprotein that also correcting mechanism) [10]. The most divergent branches contains a reverse transcriptase and an endonuclease. The of the tree are for the retrotransposon peptidases, which retrotransposon endopeptidases in family A2 were at one implies that retroviruses are derived from retrotransposons. timeassignedtofamiliesA10andA13. Family A2 is divided into several subfamilies because of Tertiary structures of several endopeptidases from thedeepdivergencesamongsttheproteinsequences. family A2 have now been determined, including the ret- FamilyA3containsapolyprotein-processingendopepti- ropepsins from human immunodeficiency viruses type 1 dase from pararetroviruses; these are double-stranded DNA and type 2, macaque immunodeficiency virus, equine viruses that infect plants. The genome contains an open infectious anemia virus, feline immunodeficiency virus, readingframe(ORFV)thatisanalogoustothepolgeneof Rous avian sarcoma virus and avian myeloblastosis-asso- retroviruses. ORF V encodes a polyprotein containing a ciated virus. The structures are similar, and show a single reverse transcriptase homologous to that of retroviruses, domain bearing a single catalytic aspartic. The fold of and an endopeptidase at the N-terminus. Mutagenesis con- the domain resembles that of the N-terminal domain of firmed that Asp45 (numbering as in the family A3 align- pepsin A, so the retropepsin monomer also carries the ment in MEROPS) is catalytic in cauliflower mosaic virus specificity site on the ‘flap’, and the catalytically active endopeptidase (Chapter 60) and weak inhibition with pep- retropepsin dimer therefore has two ‘flaps’. The structure statin prevented polyprotein processing [11]. A sequence of the human immunodeficiency virus type 1 retropepsin that had been thought to be a transposon known as Gypsy is shown in Figure 1.2. The presence of the two flaps in from Drosophila is now believed to be a retrovirus [12]; it a retropepsin dimer is a major difference from the has a domain homologous to the pararetrovirus endopepti- eukaryotic endopeptidases of family A1. In a family A1 dase,buttheaspartichasbeenreplaced. member such as human pepsin A, the flap constitutes the Family A9 includes the polyprotein-processing asymmetric aspect of the structure, and interactions enzyme from spumaretroviruses. The endopeptidase is between residues on the flap and the substrate are quite partoftheGagpolyproteininthehumanfoamyvirus,but specific, most notably involving Tyr137 (human prepepsi- part of the Pol polyprotein in simian foamy virus [13]. nogen numbering) and residues P1(cid:1)P20 of the substrate. The catalytic aspartic has been identified by mutagenesis The two flaps of the retropepsin dimer interact not only [14]; as can be seen from Figure 1.4, it is contained in a with the substrate but also with each other. From the ter- motif that differs from that of pepsin A in that residue 92 tiary structure of a retropepsin-inhibitor complex it is is not hydrophobic. There is also no hydrophobic-hydro- known that each flap forms a short β-sheet, one with phobic-Glymotif. each half of the inhibitor, P4(cid:1)P1 and P10(cid:1)P30 [9]. Family A15 includes the polyprotein-processing endo- ThealignmentforfamilyA2inMEROPSshowsthatin peptidases from rice tungro and sugarcane bacilliform the sequence around the catalytic aspartic, Ala96 replaces viruses. These are plant pararetroviruses from the 10 1.Introduction:AsparticandGlutamicPeptidasesandTheirClans badnavirus group. The endopeptidase has been shown to stalk assembly for the sessile stage [21]. A single active beessentialforprocessing,andthecatalyticAsphasbeen site Asp has been identified within an ATG motif, and the identifiedbymutagenesis[15]. peptidase is assumed to be active as a homodimer. PerP is Retrotransposons of fungi, plants and animals contain a multidomain, transmembrane protein with the peptidase endopeptidases related to retropepsin. Often an endopepti- unitintheC-terminaldomain. dase has been identified only by the presence of an Asp- Thr-Gly motif, or something similar. In the Pao retrotran- Clan AF sposon from Bombyx mori an Asp-Gly-Ser motif has been proposed [16], though it could equally well be Asp-Asp- Clan AF is the third clan of aspartic peptidases for which Gly. In some cases, the hydrophobic-hydrophobic-Gly information about protein fold is available. Clan AF con- motif is also present (see Figure 1.4). In a few cases an tains omptin (Chapter 68) in family A26. Omptin is a endopeptidase has been biochemically characterized, usu- membrane-bound protein that cleaves between paired allybymutagenesisoftheproposedcatalyticAspresulting basic amino acids. The tertiary structure indicates that the inabolitionof polyproteinprocessing.Among thoseendo- omptin molecule takes the form of a cylinder inserted peptidases that have been so characterized is the into the bacterial outer membrane with four catalytic resi- Drosophilatransposoncopia(familyA11)[17]. dues at the outer surface forming Asp-Asp and Asp-His The relationships between the putative retrotransposon pairs. This uniquefold isshown inFigure1.8. Thefamily endopeptidases are often difficult to determine, because is known only from proteobacteria. Before its crystal the sequences are very diverse presumably because the structure had been determined, omptin was thought to be rate of mutation is very high. It has been possible to a serine peptidase and had been assigned to the now detect a significant relationship between a retrotransposon deleted peptidase family S18. A homolog from Yersinia endopeptidase from soya bean and peptidases in family pestis has coagulant and fibrinolytic activities, and may A11andsomesequencesarenowincludedinfamilyA2. possiblybeinvolvedinplaguetransmission(Chapter69). None of the proteins included in family A28 have been shown to be peptidases, but the tertiary structure of the DNA damage inducible peptidase DDI1 from Clan AC Saccharomyces cerevisiae is similar to that of retropep- Clan AC contains family A8, for which the type example sin [18]. is signal peptidase II (Chapter 62). This enzyme is Family A33 includes the skin aspartic peptidase SASPase (Chapter 57) which was first isolated from human epidermis and may be involved in development of normal skin. Knockout of the SASPase gene leads to a finely wrinkled skin surface [19] and unprocessed SASpaseisfound inthestratumcorneumofpsoriaticepi- dermis, whereas only the processed form occurs in the upper layers of normal skin [20]. Activity against the oxi- dized insulin B-chain has been shown, and the peptidase autoactivates, which can be inhibited with the retropepsin inhibitor Indinavir. The active site Asp has been identi- fied by mutagenesis [21]. Although no tertiary structure has been solved, family A33 is included in clan AA becauseofthemotifaroundtheactivesiteAsp. Family A32 includes retropepsin-like sequences from bacteria. In Caulobacter crescentus, the PerP peptidase is important for processing the PodJ protein. Caulobacter alternate between two phenotypes: a sessile phase and a swarmer phase. In the sessile phase a stalk forms, whereas in the swarmer phase a flagellum and pili develop. The PodJ protein is important for recruiting the correct compo- nentstoassembletheorganellestoonepoleofthedividing bacterium. The PerP peptidase removes a C-terminal pep- FIGURE 1.8 Richardson diagram of omptin. The image was pre- paredfromtheBrookhavenProteinDataBankentry(1I78)asdescribed tide producing a truncated form of PodJ. The full-length intheIntroduction(p.li).Thecatalyticresiduesareshowninball-and- PodJ is important in the swarmer phase for pili assembly, L stickrepresentation:Asp83,Asp85and,Asp210inpink,andHis212in whereas the shortened form PodJS recruits components for darkblue(onemonomerisshown;numberingasinthePDBentry). 1.Introduction:AsparticandGlutamicPeptidasesandTheirClans 11 responsible for the removal of the signal peptide from the precursors with specialized leader peptides at the N-ter- N-terminus of the precursor of murein lipoprotein, one of mini that are removed by the type IV prepilin leader pep- the most abundant bacterial cell wall proteins. The cyste- tidase. These leader peptides are 6 to 8 residues long and ine residue that is to become the new N-terminus of the rich in charged amino acids, quite unlike the leader pep- lipoprotein must be modified by substituting the sulfur tides removed by signal peptidase I (Chapter 774). All with a CH (OOCR )CH(OOCR )CH -group before signal mature type IV pilins have a methylated N-terminal Phe 2 1 2 2 peptidase II will cleave it. Signal peptidase II was first residue. The type IV prepilin leader peptidase is located suggested to be an aspartic peptidase when it was found in the inner membrane, and cleavage and methylation of to be inhibited, albeit weakly, by pepstatin [22]. A more the pilin precursors appears to occur on the external face effective inhibitor is globomycin [23]. The enzyme is of the membranes. The type IV prepilin leader peptidase membrane-bound, and has four membrane-spanning cleaves GlykPhe bonds, and is also responsible for the domains and two larger, periplasmic regions. Studies on methylation of the new N-terminal phenylalanine. Prior signal peptidase II from Bacillus subtilis have shown that to the identification of Asp149 and Asp213 as active site two aspartic residues are essential for activity (Asp114 residues [26], members of the family had been thought and Asp141 in the alignment for family A8 in MEROPS) to be cysteine peptidases. The active site aspartic residues [24]; these are located at the ends of transmembrane occur in the motifs Xaa-Xaa-Asp-Xaa-Xbb-Xcc-Xcc- domains,atthemembrane(cid:1)periplasminterface. Xcc-Xaa-Pro and Xaa-Gly-Xcc-Gly-Asp-Xaa-Lys-Xaa- Clan AC is somewhat hypothetical in that crystal Xaa-Xaa (where Xaa is hydrophobic, Xbb is charged and structureshavenotyetbeenobtained. Xccisanyaminoacid). Some archaea possess a flagellum. The flagellin pro- Clan AD tein is synthesized as a precursor with a positively charged leader peptide; the leader peptide is removed Families A24 and A22 contain membrane-inserted endo- bythepreflagellinpeptidasebeforetheproteinisincorpo- peptidases with eight transmembrane domains, and the rated into the filament [27]. The preflagellin peptidase catalytic sites either within themembrane orvery close to are included in subfamily B. The tertiary structure of the it.Theirgroupingtogetherinclan ADissupportedby the preflagellin peptidase from Methanococcus maripaludis observation that putative catalytic Asp residues occur in has been solved [28]. The structure shows a homodimer, similar sequence motifs: Gly-Tyr-Gly-Asp-Phe in family and each monomer consists of a bundle of five helices A24 and Gly-Leu-Gly-Asp-Phe in family A22 [25] (see arranged around a short, central helix bearing one of Figure1.9). the two active site aspartics. The other aspartic is some Family A24 is divided into two subfamilies. distance away and the protein must undergo a conforma- Subfamily A contains the bacterial type IV prepilin pepti- tional change to bring the aspartics into close contact dase (Chapter 63) and subfamily B contains the archaean (seeFigure1.10). preflagellin peptidase (Chapter 64). Pili are hair-like The second family in the clan, family A22, is divided structures that occur on the surfaces of some bacterial intotwosubfamilies.SubfamilyAcontainstheproteinpre- cells, and each is assembled from one or more protein senilin (forms 1 and 2) that is the putative γ-secretase of subunits known as pilins. Type IV pili are found in Alzheimer’sdisease(Chapters65and66).Amyloidprecur- Gram-negative pathogens, including Neisseria gonor- sor protein is a neuronal type I transmembrane receptor- rhoeae and Pseudomonas aeruginosa, and are thought to like protein. Proteolytic cleavages by the β- and γ-secre- be responsible for adhesion of the organism to the surface tases towards the C-terminus can release a soluble form of host epithelial cells, and thus implicated in infection and also produce the amyloidogenic peptide of about 42 and virulence. The pilins are secreted to the periplasm as aminoacidresidues.Severallinesofevidenceindicate that presenilin is, or is intimately associated with, the γ-secre- tase. Thus, hereditary Alzheimer’s disease is commonly correlatedwithmutationsinthepresenilingenes,andengi- neered mutation of conserved aspartic residues in preseni- lin-1 produce Alzheimer’s-like abnormalities [29]. The γ-secretase is inhibited by pepstatin [30], and two aspartic residues, Asp257 and Asp385, have been implicated in the activity of presenilin. These occur in transmembrane domains in sequence motifs similar to that of the type IV FIGURE1.9 Alignmentaroundtheputativeactivesiteresiduesfor prepilin peptidases in family A24 (see Figure 1.9) [25]. families in clan AD. Sequences shown are from the type example for Presenilin homologs are known from plants as well as ani- each family or subfamily. Numbering is according to that of PseudomonasaeruginosatypeIVprepilinpeptidase. mals,sotheyprobablyhavediversefunctions. 12 1.Introduction:AsparticandGlutamicPeptidasesandTheirClans FIGURE 1.10 Richardson diagram of the preflagellin peptidase FIGURE 1.11 Richardson diagram of the HybD peptidase. The from Methanococcus maripaludis. The image was prepared from the image was prepared from the Protein Data Bank entry (1CFZ) as ProteinDataBankentry(3S0X)asdescribedintheIntroduction(p.li). describedintheIntroduction(p.li).Onemonomerofthehomotetramer One monomer of the homodimer is shown. The active site residues is shown.The catalytic cadmiumionis shown as a silver CPKsphere. Asp518 and Asp579 are shown in ball-and-stick representation in pink Thenickelligandsare shownin ball-and-stickrepresentation:Glu16in (numberingasinthePDBentry). darkblue,Asp62inpinkandHis163inpurple. Subfamily B contains the recently identified mamma- by aspartic. The presence of nickel is important for sub- lian signal peptide peptidase that is responsible for strate recognition by the hydrogenase activator [35]. A C- degradingthesignalpeptidesliberatedbysignalpeptidase terminal consensus sequence, Asp-Pro-Cys-Xaa-Xaa-Cys- I in the endoplasmic reticulum. Signal peptide peptidase Xaa-Xaa-(Arg/His)1, is present in the substrate and after is predicted to have seven transmembrane domains, and the cleavage, the cysteine residues in this consensus two aspartic residues are conserved throughout the sequence act as nickel ligands. The hydrogenase activator family and are predicted to be active site residues. The C- is not affected by typical inhibitors of serine or metallo- terminal halves of signal peptide peptidase and presenilin peptidases. E. coli possesses two other homologs, the 1 are homologous, and the active site aspartic residues HyaD protein, which may process the large subunit of occur in similar motifs [31]. Atleastfive human homologs hydrogenase 1 at the C-terminus, and the HycI protein, are known [32]. Homologs are present in many eukaryotes which processes the large subunit of hydrogenase 3. The also,butnonehasyetbeendetectedinabacterium. genes are part of the hyc operon that also controls expres- sionofthelargesubunitgenehycE[36].Thehydrogenase processingendopeptidasesformfamilyA31. Clan AE The second family in the clan, family A25, contains There are two families in clan AE, and the relationship the GPR endopeptidase from Bacillus megaterium between them was discovered by Pei & Grishin [33]. The (Chapter 71). This endopeptidase is involved in sporula- type structure (see Figure 1.11) is from the HybD endo- tion in bacteria, and is active in the degradation of small, peptidaseofE.coli[34],whichprocessestheprecursorof acid-soluble spore proteins that are not bound to DNA the large subunit of hydrogenase 2 at the C-terminus [37,38] in the final stages of spore formation. (Chapter 70). Hydrogenases are nickel-containing, multi- Endopeptidase GPR is also self-processing, releasing a subunit enzymes, and the processing enzyme was previ- 16(cid:1)residuepropeptideduringstageII[37],andtheN-ter- ously believed to be a nickel enzyme, binding the metal minal leucine before small acid-soluble spore proteins with a glutamic, an aspartic and a histidine residue. It is become substrates. Even though the tertiary structure for now thought that these three residues form the catalytic the precursorhadbeen solved(see Figure 1.12)itwas not triad of an aspartic peptidase [35]. In several sequences clear to what catalytic type the enzyme belonged [39]. from other bacteria and archaea, the glutamic is replaced The enzyme was not inhibited by any of the usual