Cytotoxic ribosome-inactivating lectins from plants

Biochimica et Biophysica Acta 1701 (2004) 1 – 14 www.bba-direct.com

Review

Cytotoxic ribosome-inactivating lectins from plants M.R. Hartley *, J.M. Lord Department of Biological Sciences, University of Warwick, Gibbet Hill Road, Coventry, West Midlands CV4 7AL, UK Received 20 February 2004; received in revised form 8 June 2004; accepted 16 June 2004 Available online 6 July 2004

Abstract A class of heterodimeric plant proteins consisting of a carbohydrate-binding B-chain and an enzymatic A-chain which act on ribosomes to inhibit protein synthesis are amongst the most toxic substances known. The best known example of such a toxic lectin is ricin, produced by the seeds of the castor oil plant, Ricinnus communis. For ricin to reach its substrate in the cytosol, it must be endocytosed, transported through the endomembrane system to reach the compartment from which it is translocated into the cytosol, and there avoid degradation making it possible for a few molecules to inactivate a large proportion of the ribosomes and hence kill the cell. Cell entry by ricin involves the following steps: (i) binding to cell-surface glycolipids and glycoproteins bearing h-1,4-linked galactose residues through the lectin activity of the B-chain (RTB); (ii) uptake by endocytosis and entry into early endosomes; (iii) transfer by vesicular transport to the trans-Golgi network; (iv) retrograde vesicular transport through the Golgi complex and into the endoplasmic reticulum (ER); (v) reduction of the disulfide bond connecting the Aand B-chains; (vi) a partial unfolding of the A-chain (RTA) to enable it to translocate across the ER membrane via the Sec61p translocon using the pathway normally followed by misfolded ER proteins for targeting to the ER-associated degradation (ERAD) machinery; (vi) refolding in the cytosol into a protease-resistant, enzymatically active structure; (vii) interaction with the sarcin-ricin domain (SRD) of the large ribosome subunit RNA followed by cleavage of a single N-glycosidic bond in the RNA to generate a depurinated, inactive ribosome. In addition to the highly specific action on ribosomes, ricin and related ribosome-inactivating proteins (RIPs) have a less specific action in vitro on DNA and RNA substrates releasing multiple adenine, and in some instances, guanine residues. This polynucleotide:adenosine glycosidase activity has been implicated in the general antiviral, and specifically, the anti HIV-1 activity of several single-chain RIPs which are homologous to the A-chains of the heterodimeric lectins. However, in the absence of clear cause and effect evidence in vivo, such claims should be regarded with caution. D 2004 Elsevier B.V. All rights reserved. Keywords: Ribosome-inactivating protein; Toxic lectin; RNA N-glycosidase; Retrograde transport; ER-associated degradation

1. Introduction The term lectin was introduced by Boyd and Shapleigh in 1954 [1] to describe blood group-specific proteins (agglutinins) from legumes that agglutinate erythrocytes through cross-linking reactions that involve binding to exposed sugar residues. The history of lectin research began in 1888 when Stillmark, a doctoral student at the University of Dorpat in Estonia, discovered that extracts from Ricinnus communi seeds caused the agglutination of erythrocytes, and showed that the active ingredient was a protein he termed ricin [2]. It had been known since ancient times that R.

* Corresponding author. Tel.: +44-24-765-23-521; fax: +44-24-765-23701. E-mail address: [email protected] (M.R. Hartley). 1570-9639/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.bbapap.2004.06.004

communis seeds are extremely toxic to animals, and Stillmark concluded that the heamagglutination activity of ricin was responsible for this toxicity. We now know that Stillmark’s preparations were mixtures of ricin and a related, but considerably less toxic, agglutinin termed R communis agglutinin (RCA). Early biochemical investigations on lectins were largely focused on legume lectins due to their ability to stimulate the division of lymphocytes, and it was not until 1970, when it was shown that ricin and related toxins are more toxic to Ehrlich ascites than normal cells [3], that work began on the structure, entry into cells and the mechanism of action of this group of plant lectins. Work in the 1970s by the group of Sjur Olsnes established the subunit composition of ricin and that toxicity is the result of its catalytic action on ribosomes resulting in the inhibition of protein synthesis. For this reason, ricin and related proteins with the same action on

2

M.R. Hartley, J.M. Lord / Biochimica et Biophysica Acta 1701 (2004) 1–14

ribosomes were named ribosome-inactivating proteins (RIPs). The precise nature of ricin’s activity was discovered in 1987 by Yaeta Endo and colleagues, who showed that the target in the ribosome is a universally conserved stem-loop structure in 28S rRNA that functions in elongation factor binding. At about the same time, Jon Robertus’s group solved the crystal structure of ricin and proposed a mechanism of catalysis. In recent years several other enzymatic activities have been assigned to ricin and related toxins, although it is uncertain whether these have a physiological role in intoxication. A major area of research activity for ricin, and other protein toxins, concerns their entry into cells and delivery into the cytosol. Ricin has unusual intracellular trafficking properties in which the endocytosed toxin is transported to the endoplasmic reticulum (ER) via the Golgi apparatus, and hijacks a quality control pathway that is normally used to translocate aberrant or misfolded proteins into the cytosol for degradation by the proteasome to gain entry to the cytosol. This trafficking behaviour is being exploited as a means of delivering viral or tumour-specific epitopes into the cytosol of cells of the immune system to trigger an immune response. The extreme toxicity of ricin, coupled with its relative ease of purification and widespread availability of castor oil seeds, has attracted criminals and politically motivated assassins to its use as a homicidal agent over many years. Of particular notoriety, the assassination of the Bulgarian journalist, Georgi Markov, in London in 1978 is believed to have been the result of an attack with an umbrella tip modified to contain a small metal sphere containing ricin [4]. It is also feared that an aerosolised form of ricin could be used as a weapon of mass destruction, emphasising the need for the development of an effective antidote or vaccine.

2. Distribution and classification of cytotoxic plant lectins Ricin is the best known example of a class heterodimeric plant toxins known as type 2 RIPs consisting of an enzymatic A-chain of f 30 kDa linked by a single disulfide bond to a cell-binding B chain, also of f 30 kDa, with lectin activity [5]. They are produced in seeds, tubers, bulbs, leaves and bark, depending on the species, and often as several distinct isoforms. In some plant tissues, type 2 RIPs are highly abundant; for example in mature R. communis seeds ricin accounts for about 5% of the soluble protein and elderberry (Sambucus nigra) bark produces three type 2 RIPs each representing >20% of the total protein [6]. The binding specificity of the cell binding B-chain is usually galactose and/or N-acetyl galactosamine, or less frequently N-acetyl neuramic acid [6]. Type 2 RIPs may also co-exist in their tissue of synthesis with dimeric and tetrameric agglutinins consisting of either disulfidelinked or noncovalently associated pairs or tetramers of the

basic A – B heterodimers. The agglutinins show a higher haemagglutination activity, but lower cytotoxicity than their A – B counterparts. The bacterial toxins Shiga toxin, produced by Shigella dysenteriae, and the structurally similar Shiga-like toxin (SLT), produced by enterohaemorrhagic strains of Escherichia coli, are also classified as type 2 RIPs [7]. Infection in humans causes diarrhoea leading to colitis and may progress to haemolytic uremic syndrome (HUS) as a direct result of toxin-induced kidney damage. They consist of an A-chain analogous to the Achains of plant type 2 RIPs which are noncovalently associated with a doughnut shaped pentamer of B-chains of 7.7 kDa [8]. The bacterial RIPs specifically recognise the globo series of glycolipids [9]. Some examples of type 2 RIPs are shown in Table 1. The A-chains of type 2 RIPs are structurally and functionally equivalent to single-chain, or Type 1 RIPs, which are more widely distributed in the plant kingdom. Several type 1 RIPs, for example pokeweed antiviral protein (PAP) and Mirabilis antiviral protein (MAP), are now known to be the active ingredient of a potent antiviral activity present in extracts of these plants. Although type 1 and type 2 RIPs are equally effective inhibitors of protein synthesis in cell extracts, the absence of the B-chain in type 1 means that they cannot bind to and enter cells with high efficiency, and are therefore considerably less cytotoxic [10]. Examples of type 1 RIPs are shown in Table 2. Type 2 RIPs are thought to have evolved from a gene fusion between an ancestral type 1 RIP gene, and a lectin gene [6]. A few plants, including S. nigra and Dutch iris (Iris hollandica), produce type 1 and 2 RIPs simultaneously in the same tissue [11,12]. On the basis of multiple sequence alignments, it is thought that the type 1 RIP in I. hollandica has arisen by the deletion of

Table 1 Examples of type 2 RIPs Species

Name of RIP

Organ

Sugar specificity

Ricinus communis (castor oil plant)

ricin R. communis agglutinin abrin A. precatorius agglutinin volkensin momordin viscumin

seed seed

Gal/GalNac Gal>GalNac

seed seed

Gal/GalNac Gal

root seed leaf

Gal Gal>GalNac Gal/GalNac

ebulin b

bark

NANA

SNAI SNAV (nigrin b) SNAIf shiga toxin shiga-like toxin

bark bark fruit secreted secreted

NANA GalNac>Gal NANA GB3 GB3

Abrus precatorius (rosary pea) Adenia volkensii Momordica charantia Viscum album (mistletoe) Sambucus ebulus (dwarf elder) Sambucus nigra (elderberry) Shigella dysenteriae Escherichia coli (0157:H7)

Gal, galactose; GalNac, N-acetyl galactosamine; NANA, N-acetyl neuramic acid; GB3, globotriaosyl ceremide. Compiled from Refs. [23,136].

M.R. Hartley, J.M. Lord / Biochimica et Biophysica Acta 1701 (2004) 1–14 Table 2 Examples of type 1 RIPs Species

Name of RIP

Organ

Phytolacca americana (pokeweed)

PAP PAP II PAP-S dianthin 30 dianthin 32 saporins L1 and L2 saporins R1 – R3 saporins S5 – S7 MAP

spring leaf summer leaf seed leaf leaf leaf root seed root

Trichosanthin

root

Ebulitins a, h, and g

leaf

Pepocin

fruit

tritin-S tritin-L

seed leaf

Dianthus carryophylus (carnation) Saponaria officinalis (soapwort) Mirabilis jalapa (four o’clock) Trichosanthes kirlowii (chinese cucumber) Sambucus ebulus (dwarf elder) Cucurbita pepo (marrow) Triticum aestivum (bread wheat)

PAP, pokeweed antiviral protein; MAP, Mirabilis antiviral protein. Compiled from Refs. [11,23,136].

the B-chain of the type 2 RIP. In S. nigra, the fruit produces a GalNac-specific lectin that is a dimer of identical 32-kDa subunits corresponding to the B-chain of a type 2 RIP also made in the fruit [13]. Here, it is believed that the lectin arose through a deletion of the Achain of the type 2 RIP. Elderberry bark produces a different homodimeric lectin, S. nigra agglutinin II (SNA II), composed of 30-kDa subunits. Sequence analysis revealed that SNA II is closely related to the B-chain of a genuine type 2 RIP, SNA V (synonymous with nigrin b [27]), produced in the same tissue, except that it lacks the first eight residues. Unlike the S. nigra fruit specific homodimeric lectin, SNA II is derived from the same precursor as the genuine RIP, by different posttranslational processing events [14]. SNA II can be regarded as a toxin in its own right as its oral administration to rats has deleterious effects on the small intestine [15].

3

and project into the solvent of the active site. The third domain consists of a two-stranded antiparallel h-sheet and a helix which is anchored to the first helix in the Nterminal domain and forms part of the active site cleft and also interacts with RTB in the holotoxin. RTA contains two potential N-glycosylation sites (Asn10Phe11-Thr12 and Asn236-Gly237-Ser238), both of which are occupied [17]. Glycosylation of RTA is not required for its enzymatic activity [18]. Although elements of this three domain structure are found in other RNA-binding proteins such as RNase H from E. coli and retroviral reverse transcriptases [19], no close homologue of the Nglycosidase domain has been found in other proteins. The primary sequences of the N-glycosidase domains of RIPs show only limited sequence identity, although their tertiary structures are highly conserved, and for several RIPs, including types 1 and 2, their a-carbon backbone traces are nearly superimposable [20]. They all share the same elements of secondary structure, the main differences lying in their surface loops. A sequence alignment of the N-glycosidase domain of RIPs shows about eight invariant residues, mostly clustered in the active site, including Tyr 80, Tyr 123, Glu 177, Arg 180 and Trp 211 (RTA numbering). The N-glycosidase RIPs, including RTA, remove a specific adenine from rRNA, and crystal structures for complexes with the adenosine analog formycin monophosphate (FMP) and ApG show that the substrate purine ring is stacked between the invariant tyrosines, Arg 180 binds to N3 of the purine ring and

3. Structure of plant ribosome-inactivating proteins The crystal structure of ricin was the first to be determined in the RIP family [16]. It shows a globular, glycosylated heterodimer in which the A-chain (RTA) and B-chain (RTB) are joined by a single disulfide bond (Fig. 1). RTA consists of 267 aminoacyl residues forming eight a-helices, a six-stranded h-sheet, and a two-stranded h-sheet forming three domains. The amino-terminal 117 residues form a compact module of six h-strands and two helices. The central domain is made up of five helices of which the longest, helix E, runs through the centre of the molecule, including much of the active site. It contains the key active site residues Glu177 and Arg180, which lie on consecutive turns of the helix

Fig. 1. The three-dimensional structure of ricin. The crystal structure was determined to 2.5 A [22] (PDB code 2AAI). The a-carbon backbone of the B-chain is shown in green. The A-chain is coloured according to its structure, with a-helices in pink, h-strands in yellow and coil in grey.

4

M.R. Hartley, J.M. Lord / Biochimica et Biophysica Acta 1701 (2004) 1–14

Glu 177 interacts with the ribose [21]. Analysis of the activities of mutant forms of RTA with each of these residues substituted has confirmed their importance in catalysis (reviewed in Ref. [22]). The B-chain of ricin, related type 2 RIPs, and the elderberry lectins derived from type 2 RIP B-chains have a bilobal structure in which the two domains are homologous and arose by gene duplication [17] Each domain, in turn, consists of tripartite galactose-binding subdomains termed a, h, and g, each consisting of f 40 aminoacyl residues and believed to have arisen by two duplication/in tandem insertion events which occurred prior to its fusion with the N-glycosidase domain [17]. Each B-chain domain consist of 12 antiparallel h-strands connected by h-turns and V loops, the h-strands being arranged in a h-trefoil structure of three lobes of four h-strands [17].This h-trefoil structure is extremely ancient and widely distributed [24]. In RTB, only the 1a and 2g subdomains retain lectin activity and only the latter, high affinity site accommodates both galactopyranosides and N-acetyl galactosamines [25]. The sugar binding pockets are created by a sharp bend in the backbone formed by the sequence Asp, Val and Arg, plus a forth variable aromatic residue which provides the binding platform for the sugar. This is Trp37 in the 1a/low affinity and Tyr248 in the 2g/high affinity site [25]. Several type 2 RIPs from the genus Sambucus, including ebulin [26], nigrin b [27] and nigrin f [28], are nontoxic to cells, even though they possess active A-chains. Their lack of toxicity has been attributed to defective B-chains with reduced affinity for galactosides [29].

4. Ricin biosynthesis The different forms of ricin and the closely related agglutinins are encoded by a small multigene family composed of about eight members, at least three of which appear to be nonfunctional [30]. Ricin is synthesised in the oilstoring endosperm cells of maturing R. communis seeds where it accumulates in protein storage vacuoles [31]. Although ricin is a heterodimeric molecule, its individual RTA and RTB subunits are initially synthesised together as part of a single precursor protein [32]. Cloning the ricin precursor gene showed that it encoded a preproprotein of 576 amino acyl residues comprised of a 35-residue N-terminal extension, the first 26 residues of which are a signal sequence [33], joined to the 267 residues of RTA, a 12-residue linker peptide, and the 262 residues of RTB [34]. During ricin biosynthesis, the N-terminal signal sequence directs the precursor into the lumen of the ER where it is immediately proteolytically removed leaving proricin segregated in this cellular compartment [35]. Segregation into the ER lumen is accompanied by several cotranslational modifications of the protein. In addition to signal sequence cleavage by signal peptidase, proricin is core-glycosylated [36] and five disulfide bonds are formed within the protein [37]. Four of the disulfide bonds form

loops within the RTB sequence and the fifth forms a disulfide loop linking the C-terminal region of the RTA sequence with the N-terminal region of RTB, and ultimately becomes the interchain disulfide bond covalently joining RTAwith RTB in the ricin holotoxin. The core-glycosylated, disulfide crosslinked proricin then moves via vesicular transport from the ER, via the Golgi stack, to the vacuole [36,37]. The intracellular transport of proricin is accompanied by several largely uncharacterised posttranslational modifications. Included here are Golgi-mediated modifications to the oligosaccharide side chains which result in some sugar residues being trimmed off, and other sugar molecules being added [38]. The 12-residue peptide linking the RTA and RTB sequences carries a sequence-specific determinant for vacuolar targeting [39]. The position of the linker peptide within proricin influences the specificity of its vacuolar targeting function [40]. When proricin reaches the vacuole, acidic endoprotease(s) in this compartment remove the nine residual residues of the 35-residue N-terminal leader sequence and the 12residue linker peptide between RTA and RTB to generate the disulfide-linked heterodimeric native toxin [41]. R. communis ribosomes are themselves sensitive to the catalytic action of RTA [42], and the mechanism of ricin biosynthesis ensures the enzyme and substrate never encounter one another leaving overall protein synthesis unaffected. The RTA component of proricin is not catalytically active, becoming activated only after cleavage to release it from RTB [43]. This ensures that any proricin inadvertently misdirected into the cytosol would not inevitably depurinate R. communis ribosomes. It seems feasible that proteolytic activity in the cytosol might release catalytically active RTA from proricin, but recent work in our laboratory indicates that the nine-residue propeptide at the N-terminus of RTA is a degradation signal facilitating its rapid proteolytic destruction if it appears in the cytosol (L.M. Roberts, J.M. Lord et al., unpublished data). Biologically active heterodimeric ricin is trapped within vacuoles in the mature R. communis seeds, an intracellular compartment from which it never escapes [44,45]. When the seeds germinate, the proteins stored within the vacuole are rapidly and completely degraded by vacuolar proteases to provide a source of amino acids for proteins synthesised during early postgerminative growth until the plant becomes photosynthetically active and able to synthesise amino acids. The endosperm cells of R. communis seeds are able to synthesise and safely store ricin because the ribosomes of the cell never encounter potentially damaging or lethal quantities of RTA.

5. Ricin entry into mammalian cells 5.1. Binding to the surface of target cells In common with other catalytically active protein toxins, ricin must opportunistically bind to a component at the target cell surface in order to enter and incapacitate that cell.

M.R. Hartley, J.M. Lord / Biochimica et Biophysica Acta 1701 (2004) 1–14

RTB binds to h-1,4-linked galactose residues. Since such residues are widely present on mammalian cell surface glycoproteins and glycolipids, most cell types readily bind significant quantities of ricin. HeLa cells, for example, contain 3  107 ricin binding sites per cell [46], although not all of these sites will be involved in ricin uptake. The relative significance of glycoproteins versus glycolipids as receptors for ricin has not been clarified, although a role for glycosphingolipids as receptors has not been ruled out [47]. Native ricin is also able to bind to and enter certain cells by virtue of its own oligosaccharide side chains. A limited number of cell types express at their surface mannose receptors to which ricin is able to bind, in the presence of concentrations of free galactose that prevent ricin binding to surface galactosides, via its high mannose glycans. Mannose receptor-mediated uptake of ricin has been described in both macrophages [48] and rat liver endothelial cells [49]. 5.2. Endocytic uptake of ricin Ricin enters mammalian cells by endocytosis. Because ricin binds to a variety of cell surface components, it is able to utilise different endocytic mechanisms to enter cells. The binding and internalisation of ricin by Vero cells (African Green Monkey kidney cells) was first observed to use clathrin-dependent endocytosis [50]. However, it soon became apparent that clathrin-independent mechanisms were also involved in toxin uptake. For example, Hep-2 cells, in which clathrin-coated pit formation was blocked by hypotonic shock and potassium ion depletion, remained sensitive to ricin [51]. In addition, in HeLa cells, overexpression of a trans-dominant negative mutant of the GTPase dynamin, which is required for clathrin-mediated endocytosis [52], failed to protect cells against ricin, although it did protect against diphtheria toxin which is known to enter cells in a clathrin-dependent manner [53]. A different study, however, has implicated a role for dynamin in ricin entry [54], although this role might be in intracellular trafficking rather than endocytic uptake. 5.3. Transport of ricin from the cell surface to the Golgi complex Ricin that enters cells endocytically is initially delivered to early endosomes, from where most of it appears to be either recycled back to the cell surface or delivered via late endosomes to lysosomes where it is proteolytically degraded. Raising the pH of endosomes/lysosomes, which protects against protein toxins such as diphtheria toxin that utilise a low pH-dependent mechanism to translocate into the cytosol from endosomes [55], does not protect against ricin. Treating cells with brefeldin A (BFA) or the ionophore monensin, compounds that influence Golgi morphology and function, confers resistance to ricin but not to diphtheria toxin. Resistance to ricin is also observed when endosome-to-Golgi transport is blocked by low tempera-

5

ture [56]. These findings imply that reaching or progressing beyond the Golgi complex is required for ricin intoxication, a conclusion consistent with the demonstration that endocytosed ricin can be visualised within the perinuclear Golgi complex, in particular the trans-Golgi network (TGN) [57]. Biochemical evidence has confirmed that endocytosed ricin reaches the Golgi complex. A tyrosine sulfation site was engineered into recombinant RTA, which was reassociated with RTB to form holotoxin. Internalisation of unlabelled ricin into cells in the presence of [35S]-sulfate led to the radiolabelling of a proportion of the toxin, demonstrating that it had reached the Golgi complex, which is the intracellular location of tyrosyl sulfotransferase [58]. Ricin does not appear to pass through late endosomes as it moves from early endosomes to the TGN. Early to late endosome transport requires the low pH-dependent formation of carrier vesicles [59]. However, ricin transport to the Golgi is not affected by treatments that perturb the low pH of endosomes [56,60]. Furthermore, overexpression of a dominant negative mutant of Rab9, a small GTPase that regulates vesicular transport between late endosomes and the TGN [61,62], failed to protect cells from ricin intoxication [63] and did not prevent sulfation of internalised ricin [64]. Collectively, these data indicate that ricin may be transported to the Golgi complex by a pathway that is independent of late endosomes (Fig. 2). Existence of a direct pathway from early/recycling endosomes to the Golgi has been demonstrated using the ricin-related bacterial toxin, Shiga toxin produced by S. dysenteriae [65 – 67]. Recently, specific proteins such as Rab6V and cognate SNAREs have been implicated in the direct early endosome-to-TGN transport step utilised by Shiga toxin [68]. The Rab9-independent transport of ricin suggests that this protein might likewise exploit the direct route from early endosomes to the Golgi, although this remains to be experimentally confirmed. It is also possible that ricin might use more than one pathway from the endosomal compartments to the Golgi complex due to the multiplicity of its binding partners. 5.4. Transport of ricin from the Golgi complex to the ER Although internalised ricin has been visualised in the Golgi, it has not been visualised in the ER in spite of repeated efforts to directly identify the protein in this compartment. Treatment of cells with brefeldin A (BFA), a drug that causes the Golgi stack to collapse and redistribute to the ER [69], protects against ricin intoxication in most cell types [70,71]. BFA treatment does not prevent the fusion of ricin-containing vesicles with the TGN, indicating that RTA does not translocate into the cytosol from this compartment. BFA treatment did not protect against ricin in MDCK cells, and EM studies of such cells revealed that the toxin did not alter the morphology of the Golgi complex [71]. Ilimaquinone, a drug that causes Golgi fragmentation in certain cell types, also protects such cells against ricin [72]. Furthermore, transient

6

M.R. Hartley, J.M. Lord / Biochimica et Biophysica Acta 1701 (2004) 1–14

Fig. 2. Ricin trafficking in mammalian cells. Ricin binds to the cell surface through the lectin activity of its B chain (red circles). Some surface bound toxin enters the cell by endocytosis. After initially reaching early endosomes, the bulk of the endocytosed toxin either recycles back to the cell surface or proceeds via late endosomes to lysosomes where it is degraded. The productive route for intoxication occurs for a small proportion of the toxin that undergoes vesicular transport from early endosomes to the trans-Golgi network (TGN) and, via the Golgi complex, to the lumen of the ER. Here the ricin subunits reductively separate, allowing the catalytically active A chain (yellow circles) to retro-translocate into the cytosol.

expression of mutant GTPases that regulate vesicular transport in the early secretory pathway partially protected cells against ricin [63], whereas the addition of an ER retrieval sequence (-KDEL) to the C-terminus of RTA increased the cytotoxicity of free RTA [73] and reconstituted holotoxin [74,75]. This suggests that the RTA-KDEL interacts with Golgi-to-ER cycling KDEL receptors. RTA does not naturally have a KDEL or related signal, but toxin might ‘piggyback’ interacting with something that recycles between the Golgi and the ER because it does possess a retrieval signal, such as the ER molecular chaperone calreticulin [76]. The most direct evidence for Golgi-to-ER transport of ricin, however, has been shown by the BFA-sensitive uptake and subsequent core-glycosylation of a non-glycosylated toxin [58]. Posttranslational core-glycosylation of an endocytosed protein provides compelling evidence for the retrograde transport of ricin from the cell surface to the ER where the oligosaccharyl transferase responsible for the core-glycosylation is exclusively located. The presence of this coreglycosylated RTA in a cytosolic fraction further indicated that membrane translocation takes place only after delivery of the toxin to the ER has been accomplished [58]. How is the retrograde transport of ricin from the Golgi to the ER achieved? The classic Golgi-to-ER transport step involves movement via coatamer protein 1 (COP-1)-coated vesicles [77]. A second, largely uncharacterised, Golgi-to-ER transport step is COP-1-independent. This step mediates the transport of some cycling Golgi glycosylation enzymes and possibly lipids, and is controlled by Rab6 [78]. Both Golgito-ER pathways are utilised by protein toxins. For example, cholera toxin and Pseudomonas exotoxin A exploit KDEL receptors that are cycled within COP-1-coated vesicles [79 – 82]. In contrast, Shiga toxin uses the COP-1-independent route [78,83].

5.5. Membrane translocation of RTA Why does ricin need to traffic to the ER to exert a toxic effect? It is likely that ricin’s complex intracellular journey must be undertaken to reach a cellular destination where the membrane translocation of RTA can occur. Some bacterial toxins form pores in biological membranes, either directly or upon exposure to low pH. Ricin, however, does not form pores in membranes. One appealing possibility is that ricin must be endocytosed to a compartment that contains in its membrane pre-existing protein-conducting channels through which the toxin can exit into the cytosol. In the endomembrane system the only organelle that is known to meet this criteria is the ER, whose membrane contains numerous protein translocases (known as Sec61 translocons) [84] and peptide transporters [84]. While peptide transporterdeficient cells have been shown to be as sensitive as to ricin as parental peptide transporter-positive cells [85], evidence is accumulating that protein toxin exploits the Sec61 protein translocons to cross the ER membrane [75,86 – 88]. It is possible that RTA is expelled from the ER by the quality control system that operates there [89]. Newly synthesised proteins are scrutinised by molecular chaperones in the ER lumen. Aberrant proteins that are unable to assume their biologically active conformation, or orphan subunits that have failed to become incorporated into the appropriate multimeric complexes, are detected at this stage and eliminated by proteolytic degradation. At first it was believed that proteolytic degradation occurred within the ER itself [90]. More recently, it has been recognised that such proteins are exported from the ER [91] and are proteolytically degraded in the cytosol. This aspect of quality control is known as ER-associate protein degradation (ERAD) [92,93]. Cytosolic degradation is carried out by protea-

M.R. Hartley, J.M. Lord / Biochimica et Biophysica Acta 1701 (2004) 1–14

somes, a step that is usually facilitated by ubiquitylation of the protein being eliminated. How are the toxin A chains that retrotranslocate from the ER perceived as ERAD substrates by the cell? Evidence has been presented consistent with cholera toxin A fragment and RTA unfolding to some extent in the ER lumen. In the case of cholera A, this may be caused by an interaction with the ER resident molecular chaperone protein disulfide isomerase [94], while in the case of RTA unfolding may be a consequence of an interaction with the lipid membrane itself [95]. In this way, the unfolded toxin A chains would be ERAD candidates leading to their export into the cytosol. This is a high-risk strategy for the unfolded toxin since the normal fate of ERAD substrates is rapid and complete proteasomal degradation. Clearly, since the toxins are effective enzymes in the target cell cytosol, at least a proportion must avoid degradation and refold into its native conformation. Interestingly, all the toxin A chains that are believed to retrotranslocate from the ER have an uncharacteristically low lysine content, and it has been proposed that this represents an evolutionary strategy to allow them to subvert ERAD [96]. Experimental support for this proposal has been obtained for both RTA [97] and cholera A [98], where the introduction of additional lysines into the proteins reduces their toxicity by increasing the extent of ubiquitin-mediated proteasomal degradation. Apparently the few lysine residues that occur in the A chains of ER-translocating toxins are tolerated because they are not sites for ubiquitin attachment. Little is known about the way that unfolded toxin that escapes proteasomal degradation refolds into its native conformation although evidence has been presented that RTA might use its target substrate, the ribosome itself, as a refolding chaperone [99].

6. The action of RIPs on ribosomes Early work on the effects of the A-chains of ricin, abrin and moddecin on the inhibition of protein synthesis by cellfree systems showed that the ribosomal 60S subunit alone was affected and that the turnover number (Kcat) of these Achains on ribosomes was in the order of 1500 min 1 [5]. The fact that ribosome inactivation occurred in simple buffers suggested that there are no cofactor requirements. The nature of the enzymatic modification to ribosomes was discovered by Endo and Tsurugi [100] to be the cleavage of the Nglycosidic bond between an adenine residue and ribose in 28S rRNA, releasing free adenine and leaving the rRNA depurinated at a single site—A4324 in the case of rat 28S rRNA. This lies near the middle of the most highly conserved sequence present in large subunit rRNA (5VAGUACGAGAGGA-3V where the underlined A is removed by RTA) that lies in domain VII of 25S/28S rRNA, approximately 400 nucleotides from the 3V end, or at A2990 in

7

helix 90 in domain VI of E. coli 23S rRNA [101]. This sequence forms part of a universally conserved stem-loop structure in rRNA that is part of the recognition/binding site for both the eukaryotic/prokaryotic elongation factor 1 (eEF1/EF-Tu) and the eukaryotic/prokaryotic elongation factor 2 (eEF-2/EF-G) complexes [102]. It is also the site of action of the structurally unrelated endoribonuclease a-sarcin, produced by Aspergillus giganteus [103], and is named the sarcin/ricin domain (SRD). The catastrophic effects of depurination by RTA or cleavage by a-sarcin on ribosome function, together with the finding that single point mutations introduced into in the E. coli SRD at several positions show a dominant lethal phenotype [104], demonstrate the functional importance of the SRD. It is noteworthy that two different groups of ribotoxins, exemplified by ricin and asarcin, have evolved independently to attack the ribosome at its ‘‘Achilles heel’’. It has recently been shown that mammalian ribosomes treated with both of these toxins undergo conformational changes in 28S rRNA in domains II and V, which are far removed from SLD in domain VII, as well as in the latter [105]. It is likely that the action of the toxins alters the dynamic flexibility of the ribosome, especially that necessary for the transition between the pre- and posttranslocational states of the elongation cycle. Indeed, the SRD structure, as described below, is only marginally stable under physiological conditions, and it has been suggested that alternations in SRD structure may be responsible for driving protein synthesis in a directional fashion [106]. The importance of the SRD in ribosome function, and the finding that this is the only site in the f 7000 nucleotides of the rRNAs of the native ribosome that is a substrate for RIP action, has attracted the attentions structural molecular biologists, and it stands as one of the most thoroughly investigated RNA structures. The first structural models of the SRD based on phylogenetic considerations depicted it as a 7-base-pair stem and a 15-nucleotide loop containing the near universally conserved 12-nucleotide sequence shown above [107]. However, a solution NMR structure of a synthetic oligoribonucleotide (29-mer) that mimics the mammalian SRD revealed a compact, but irregular helix containing a bulged G cross strand A stack and a GAGAcontaining tetraloop [108]. This structure has been confirmed by X-ray crystal structures for the mammalian SRD [109], the E. coli SRD [110], and has essentially the same features as the structure in the native Haloarcula marismortui ribosome [111], showing that its association with ribosomal proteins does not significantly affect its fold. The crystal structure of part of the rat SRD is shown in Fig. 3. In all of these models, A4324 (A2660) (rat numbering; E. coli numbering in parentheses), the base removed by RIPs, is stacked with G4325 (G2661) and A4326 (A2662) and the Watson –Crick faces of these bases in the minor groove are exposed and available for interaction with ligands. Although attempts have been made to obtain a structure for the SRD docked to RTA, the instability of the complex has precluded this [112]. However, it can be inferred from the structure of

8

M.R. Hartley, J.M. Lord / Biochimica et Biophysica Acta 1701 (2004) 1–14

Fig. 3. Structure of part of the sarcin/ricin domain in rat 28S rRNA. The GAGA tetraloop contains the non-canonical G4323 – A4326 base pair and is shut off by the C4322 – G4327 base pair. The N-glycosidic bond cleaved by RIPs is indicated by the arrow (PDB code 430D, [108]).

RTA bound to the substrate analog FMP, in which the FMP is stacked between Tyr80 and Tyr123, that A4324 flips out of the stack with G4325 and A4326 to form a similar sandwich structure in the active site (Fig. 4). The demonstration that RTA is able to depurinate naked 28S rRNA with an identical specificity to that in native ribosomes shows that ribosomal proteins per se are not required for the reaction [113]. However, the kinetic parameters for the depurination of ribosomes and naked 28S rRNA reveal that whilst the Km values are similar (2.6 and 5.8 AM, respectively), the turnover numbers (Kcat) differ by a factor of c. 105 (1777 and 0.02 min 1 respectively) [113]. This suggests that RTA has a similar affinity for both substrates, but that the presence of ribosomal proteins is required for efficient catalysis, possibly by inducing a conformation in the SRD that is favourable for catalysis. However, ribosomal proteins can also block RTA action, as evidenced by the finding that E. coli naked 23S rRNA is susceptible to depurination at A2660 whereas the native ribosome is totally refractory. [101]. The most direct evidence for the requirement for a ribosomal protein for RIP activity has come from work on the effects of the expression of PAP in yeast. The expression of PAP is lethal to wild-type yeast as a result of ribosome depurination, whereas cells expressing the mak8-1 allele of ribosomal protein L3, which differs from the wild type at two aminoacyl residues, survive PAP expression and their ribosomes are

not depurinated in vivo [114]. Although PAP interacts with both free wild-type L3 and the mak8-1 proteins in vitro, as demonstrated by co-immunoprecipitation with monoclonal L3 antibody, PAP does not interact with ribosomes containing the mak8-1 protein, suggesting that this mutation alters its conformation in the ribosome. The involvement of L3 in RIP interaction is supported by the finding that L3, together with L6, footprint to the SRD in E. coli 23S rRNA [115]. The nature of the structural features in the SRD required for recognition and catalysis by RIPs (the identity elements) has been deduced from studies using a 35-mer synthetic oligoribonucleotide substrate that mimics the SRD [107]. RTA is able to depurinate the ‘‘wild-type’’ substrate with the same specificity as the native ribosome, but is unable act on certain ‘‘mutant’’ substrates. From these studies, it was possible to deduce that the minimum structure required for RTA action is a GAGA motif flanked on either side by two bases capable of forming Watson –Crick base pairs, i.e., a GAGA tetraloop. Although the site of action of most of the N-glycosidase RIPs studied to date on native ribosomes is the adenine equivalent to A2660 (E. coli numbering), the RNA identity elements vary to some extent from RIP to RIP. For example, the single-chain RIP PAP does not require the tetraloop structure, and the second guanine in the target site sequence GAGA can be replaced by adenine [116]. PAP action on ribosomes also differs from that of RTA and the majority of other RIPs in that it removes both G2259 and A2660 (E. coli numbering) in the SRD of rabbit reticulocyte, tobacco and yeast ribosomes [117]. This relaxation of the identity elements and the site of glycosidic bond cleavage for PAP could explain its ability to release multiple

Fig. 4. Location of the AMP analogue FMP in the active site of ricin Achain. The critical active site residues Arg180, Tyr123 and Tyr80 are coloured with carbon atoms in grey, oxygen atoms in red, nitrogen atoms in blue and FMP in green (PDB code 1FMP [20]).

M.R. Hartley, J.M. Lord / Biochimica et Biophysica Acta 1701 (2004) 1–14

adenines and guanines from naked RNA substrates, as described below.

7. Non-ribosomal substrates for RIPs The assertion that the ribosome is the only site of action of the catalytic A-chains of toxic lectins and the single-chain RIPs has been challenged in recent years. Proposed substrates for action include nucleic acids other than rRNA in native ribosomes and non-nucleic acids. 7.1. Polynucleotide:adenosine glycosidase There have been numerous reports that some, and perhaps all, RIPs possess an activity that results in multiple, non-sequence-dependent, deadenylations in both RNA and DNA substrates through an activity termed polynucleotide:adenosine glycosidase (PAGase). Barbieri et al. [118] assayed 15 different RIPs (including several RIP isoforms) for release of adenine from viral genomic RNAs of MS 2, TMV and AMCV under conditions optimised for specific depurination on rat ribosomes (pH 7.8, 50 mM NH4+, 10 mM Mg2 +). Thirteen of the RIPs (all type 1) showed some activity on at least one of the substrates, whereas reduced ricin holotoxin and recombinant SaporinS6 (a type 1 RIP isoform from Sapponaria officinalis seeds) showed none. The same study was extended to examine the action of a total of 52 RIPs on poly(A), naked rRNA and DNA at pH 4.0, the optimum for poynucleotide:adenosine glycosidase activity. All RIPs released multiple adenines from DNA, whereas their activity on the RNA substrates varied widely. In contrast to its specific depurination activity on ribosomes, unreduced ricin holotoxin was more active on DNA than the reduced form, possibly due to differences in steric hindrance with the different substrates. On the basis of this work, the authors proposed that RIPs should be named polynucleotide:adenosine glycosidases in preference to RNA N-glycosidases. However, it is difficult to reconcile these findings, and any claims as to their possible physiological significance, with those of the groups of Endo, Wool and ourselves, showing that RTA and several type 1 RIPs retain a similar specificity for naked rRNA, synthetic oligoribonucleotides that mimic the SRD and the SRD in native ribosomes. To our knowledge, there are no reports in which the kinetic parameters of the two activities have been compared for the same RIP preparation under conditions which approximate to physiological, making it difficult to asses their relative contributions to cytotoxicity. The most frequently cited evidence for a role for the PAGase activity of RIPs comes from the finding that RIPs such as PAP, which exhibit PAGase activity on RNA substrates (including viral RNAs), possess a potent antiviral activity towards infected cells, whereas RTA, which lacks PAGase activity on RNA, also lacks antiviral activity. For

9

example, Rajamohan et al. [119] showed that both RTA and three PAP isoforms have comparable in vitro protein synthesis inhibitory activities, but only the latter showed PAGase activity on several viral RNAs, including HIV-I, and caused inhibition of HIV-I replication in peripheral blood mononuclear cells. The authors hypothesise that the PAGase activity of PAP is responsible for its antiviral activity, and that the highly conserved active site structure shared by RTA and PAP that is responsible for rRNA depurination is not sufficient for recognition and depurination of viral RNA. Tumer et al. [120] showed that an intact active site of PAP is necessary for antiviral activity, toxicity and in vivo rRNA depurination. However, a deletion mutant lacking the C-terminal 25 aminoacyl residues retained its antiviral activity, but not its toxicity or in vivo rRNA depurination activity when expressed in transgenic tobacco plants, suggesting that the antiviral activity can be dissociated from toxicity. It has been suggested [120] that the Cterminal domain of PAP is involved in membrane translocation, so it is unclear why it is not required for antiviral activity. In addition to PAGase activity, PAP also possesses a deguanylation activity, and releases approximately equimolar amounts of guanine and adenine from HIV-I RNA as well as other RNA substrates [121]. Modelling studies showed that guanine is able to fit into the active site pocket of PAP very much like adenine [121]. This may also have relevance to the observation that PAP depurinates capped, but not uncapped, RNAs in vitro [117], and binds to the m7Gppp cap structure of luciferase mRNA [122]. PAP does not remove or depurinate the cap structure, but the finding that the PAP-treated luciferase mRNA gives rise to discrete primer extension products terminated at positions corresponding to particular A and G residues throughout the mRNA suggests that PAP acts in cis after binding to the cap. The mechanism for such an activity is completely unknown. The sequence context of the depurination sites does not reveal any conserved features, and the majority of sites cannot form a tetraloop structure resembling the RIP target site in rRNA. Studies on PAP expression in yeast have shown that it specifically destabilises its own mRNA by a mechanism that requires an intact active site, but can be separated from depurination of the SRD [123]. It seems unlikely that PAP mRNA could be targeted through a cap binding mechanism, as it would not provide specificity. The diverse nature of these observations makes it difficult to incorporate them into an all-embracing hypothesis for the physiological role of PAP and other single chain RIPs. If the broad spectrum antiviral action works through specifically targeting the viral RNA, rather than a cell suicide mechanism, the RIP must recognise some feature present in all viral RNAs, but absent from all host RNAs. Such a substrate would have to be destabilised by the RIP more efficiently than the action on the SRD of the hosts’ ribosomes in order to avoid toxicity. It is therefore important that the observations for

10

M.R. Hartley, J.M. Lord / Biochimica et Biophysica Acta 1701 (2004) 1–14

HIV-I RNA depurination by PAP from in vitro studies are followed up by in vivo studies in which viral RNAs and host control mRNAs and rRNA from PAP-treated cells are assayed for depurination by the sensitive primer extension technique. Although much of the recent work on the antiviral activity of RIPs has tended to diminish the role of RIPs in causing a suicide response in infected cells [124], there is no doubt that the SRD-specific ribosome depurination is an important causative event in the cytotoxic action of RIPs, as it can be detected soon after entry of the toxin, and correlates closely with inhibition of protein synthesis [125]. 7.2. DNA lyase Over the years, there have been numerous reports that RIPs cleave both single stranded and double stranded DNA (reviewed in Ref. [24]). One suggestion to account for this activity, from work on MAP30, is that RIPs act as both DNA glycosidases and apurinic site lyases [126]. For MAP30, it was proposed from structural studies that apurinic sites in DNA, generated by the N-glycosidase activity, bind to a conserved tryptophan residue (Trp190) located on the protein surface in the vicinity of the N-glycosidase active site cleft, which brings the apurinic site in close contact with a conserved lysine side chain (Lys195). DNA backbone cleavage is postulated to occur through the nucleophilic attack by the lysine amino group on the C1V ribose of the apurinic site. However, it is clear that the DNAse and RNAse activities ascribed to some RIPs are due to contamination [127,128]. It has been claimed that genomic DNA fragmentation activity mediated by the type 1 RIP saporin-6 from Sapporina officinalis in human lymphoma cells is dependent on a catalytic activity that is distinct from its RNA N-glycosidase activity [129]. However, several studies (reviewed by Nielsen and Boston [130]) have shown that RIP-treated cells undergo apoptosis, one characteristic of which is nuclear DNA fragmentation. The precise relationship between RIP-mediated inhibition of protein synthesis and apoptosis is unclear, and RTB alone, which is not a protein synthesis inhibitor, is known to cause apoptosis in human myeloid leukemia cells, albeit less efficiently than ricin holotoxin [131]. It is interesting to note that several lectins unrelated to RIPs, including phytohaemagglutinin, concanavalin A and wheat germ agglutinin, also induce apoptosis [132,133]. However, Brigotti et al. [134] have recently shown that ricin and Shiga toxin cause damage to nuclear DNA in human endothelial cells, which occurs concomitantly with the inhibition of protein synthesis, and is thus not a secondary effect resulting from inhibition of protein synthesis or apoptosis. 7.3. Action on non-nucleic acid substrates Ricin and other RIPs have been reported to possess a number of activities on non-nucleic acid substrates, in-

cluding superoxide dismutase, chitinase, phosphatase and phospholipase, though it is uncertain whether any of these is physiologically relevant to their toxic action [24]. Recently it has been reported that ricin holotoxin, but not RTA and RTB separately, possesses a lipase activity that resides in the subunit interface, comprising the lipase catalytic triad residues postulated to be Ser211 and His40 from RTA and Asp94 from RTB on the basis of their positions relative to those in a reference lipase active site [135] These residues are conserved in the toxic type 2 RIP abrin, but not in the barely toxic type 2 RIPs ebulin 1 and mistletoe lectin 1. Mutation of Ser211 to Ala resulted in the loss of lipase activity of the reassociated holotoxin, but did not affect uptake and intracellular routing by mouse lymphocytes, or RNA N-glycosidase activity. Working on the assumption that inhibition of protein synthesis observed following treatment of the lymphocytes with the wild-type and mutant toxins is directly proportional to the RTA translocation rate [136], it was calculated that the Ser211Ala mutant RTA translocated at a rate of only 64% of that of wild-type RTA, thus implying a role for the lipase activity in the translocation of RTA. However, this interpretation is only valid if the stability of the translocated mutant and wildtype A chains is identical.

8. Conclusions For ricin and related type 2 RIPs to mediate their potent toxic effects on mammalian cells, they must negotiate the endomembrane system and translocate across an internal membrane to reach the cytosol and inactivate ribosomes. To achieve this, ricin interacts with a number of intramolecular carriers and transport intermediates, subverts the ER-associated degradation pathway, and refolds in the cytosol, where it resists degradation. Unravelling the molecular details of these processes is currently a major research activity of several laboratories. In addition, the intracellular trafficking properties of ricin are being exploited therapeutically. Disarmed versions are being developed as carriers to deliver certain tumour and virus epitopes to elicit an immune response. The evidence reviewed here suggests that the toxic action of ricin is the result of the inhibition of protein synthesis brought about by the depurination of the sarcin/ricin domain of 28S rRNA, although the relationship between this event and apoptosis is unclear. All RIPs possess a less specific polynucleotide:adenosine glycosidase (PAGase) activity on DNA that is independent of sequence context, and most single-chain RIPs (but not ricin A-chain) show PAGase activity on most RNA substrates, including viral RNAs. The single-chain RIP pokeweed antiviral protein also removes multiple guanines from capped RNA substrates. Although these nonspecific deadenylation and deguanylation activities have been implicated in both the cytotoxic and antiviral activities of RIPs, they have not been

M.R. Hartley, J.M. Lord / Biochimica et Biophysica Acta 1701 (2004) 1–14

confirmed to occur in vivo and their contribution is currently unknown. The recent report [134] that RIPs can damage DNA independently of their protein synthesis inhibitory activity is potentially significant, and requires confirmation.

Acknowledgements Work in our laboratory at Warwick is funded by the UK Biotechnology and Biological Sciences Research Council and the Wellcome Trust. We thank the current and past members of the laboratory for their contribution to the research describe in part here.

References [1] W.C. Boyd, E. Shapleigh, Diagnosis of subgroups of blood A and AB by the use of plant agglutinins, J. Lab. Clin. Med. 44 (1954) 235 – 237. [2] H. Stillmark, Uber ricin, eines gifiges ferment aus den samen von Ricinnus communis L. und anderen Euphorbiacen, Inaugural dissertation, University of Dorpat, Estonia, (1888). [3] J.-Y. Lin, K.-Y. Tserng, C.-C. Chen, L.-T. Lin, T.-C. Tung, Abrin and ricin: new antitumor substances, Nature 227 (1970) 292 – 293. [4] B. Knight, Ricin-a potent homicidal poison, Br. Med. J. 278 (1979) 350 – 351. [5] S. Olsnes, A. Phil, in: P. Cohen, S. van Heyringen (Eds.), Toxic Lectins and Related Proteins in Molecular Action of Toxins and Viruses, Elsevier, Amsterdam, 1982, pp. 51 – 105. [6] E.J.M. Van Damme, W.J. Peumans, A. Barre, P. Rouge, Plant Lectins: a composite of several distinct families of structurally and evolutionary related proteins with diverse biological roles, Crit. Rev. Plant Sci. 17 (1998) 576 – 692. [7] A.D. O’Brien, R.K. Holmes, Shiga and shiga-like toxins, Microbiol. Rev. 51 (1987) 206 – 220. [8] P.E. Stein, A. Boodhoo, G.J. Tyrrell, J.L. Brunton, R.J. Read, Crystal structure of the cell-binding B oligomer of verotoxin-1 from E. coli, Nature 355 (1992) 748 – 750. [9] C.A. Lingwood, Role of verotoxin receptors in pathogenesis, Trends Microbiol. 4 (1996) 147 – 152. [10] L. Barbieri, M.G. Battelli, F. Stirpe, Ribosome-inactivating proteins from plants, Biochim. Biophys. Acta 1154 (1993) 237 – 282. [11] F.M. de Benito, L. Citores, R. Iglesias, J.M. Ferreras, F. Soriano, J. Arias, E. Mendez, T. Girbes, Ebulitins: a new family of type 1 ribosome-inactivating proteins (rRNA N-glycosidases) from leaves of Sambucus ebulus L. that co-exist with the type 2 ribosome-inactivating protein ebulin 1, FEBS Lett. 360 (1995) 299 – 302. [12] E.J.M. Van Damme, A. Barre, L. Barbieri, P. Valbonensi, P. Rouge, F. Van Leuven, F. Stirpe, W.J. Peumans, Type I ribosome-inactivating proteins are the most abundant proteins in iris (Iris hollandica var. Professor Blaauw) bulbs: characterization and molecular cloning, Biochem. J. 324 (1997) 963 – 970. [13] E.J.M. Van Damme, S. Roy, A. Barre, P. Rouge, F. Van Leuven, W.J. Peumans, The major elderberry (Sambucus nigra) fruit protein is a lectin derived from a truncated type 2 ribosome-inactivating protein, Plant J. 12 (1997) 1251 – 1260. [14] E.J.M. Van Damme, A. Barre, P. Rouge, F. Van Leuven, W.J. Peumans, Characterization and molecular cloning of SNAV (nigrin b), a GalNAc-specific type 2 ribosome-inactivating protein from the bark of elderberry (Sambucus nigra), Eur. J. Biochem. 237 (1996) 505 – 513.

11

[15] A. Pusztai, S.W.B. Ewen, G. Grant, W.J. Peumans, E.J.M. Van Damme, L. Rubio, Digestion 46 (1990) 308 – 316. [16] W. Montfort, J.E. Villafranca, A.F. Monzingo, S. Ernst, B. Katzin, E. Rutenber, N.H. Xuong, R. Hamlin, J.D. Robertus, The threedimensional structure of ricin at 2.8 A, J. Biol. Chem. 262 (1987) 5398 – 5403. [17] E. Rutenber, J.D. Robertus, Structure of ricin B-chain at 2.5 A, Proteins Struct. Funct. Genet. 10 (1991) 260 – 269. [18] M. O’Hare, L.M. Roberts, P.E. Thorpe, G.J. Watson, B. Prior, J.M. Lord, Expression of ricin A-chain in Escherichia coli, FEBS Lett. 216 (1987) 73 – 78. [19] M.P. Ready, B.J. Katzin, J.D. Robertus, Ribosome inhibiting proteins, retrovoiral reverse transcriptases and RNAse H share common structural elements, Proteins 3 (1988) 53 – 59. [20] A.F. Monzingo, E.J. Collins, S.R. Ernst, J.D. Irvin, J.D. Robertus, The 2.5A structure of pokeweed antiviral protein, J. Mol. Biol. 233 (1993) 705 – 715. [21] A.F. Monzingo, J.D. Robertus, X-ray analysis of substrate analogs in the ricin-A chain active site, J. Mol. Biol. 227 (1992) 1136 – 1145. [22] J.M. Lord, L.M. Roberts, J.D. Robertus, Ricin: structure, mode of action and some current applications, FASEB J. 8 (1994) 201 – 208. [23] J.E. Villafranca, J.D. Robertus, Ricin B chain is the product of gene duplication, J. Biol. Chem. 256 (1981) 554 – 556. [24] E.L.S. Van Damme, Q. Hao, Y. Chen, A. Barre, F. Vandenbussche, S. Desmyter, P. Rouge, W.J. Peumans, Ribosome-inactivating proteins: a family of plant proteins that do more that inactivate ribosomes, Crit. Rev. Plant Sci. 20 (2001) 395 – 465. [25] T. Hatakeyama, N. Yamasaki, G. Funatsu, Identification of the tryptophan residue located at the low-affinity saccharide binding site of ricin D, J. Biochem. 100 (1986) 781 – 788. [26] L. Citores, F.M. De Benito, R. Iglesias, J.M. Ferreras, P. Argueso, P. Jimenez, A. Testera, E. Camafetia, E. Mendez, T. Girbes, Characterization of a new non-toxic two-chain ribosome-inactivating protein from rhizomes of dwarf elder (Sambucus ebulus L.), Cell. Mol. Biol. 43 (1997) 485 – 499. [27] T. Girbes, L. Citores, J.M. Ferreras, M.A. Rojo, R. Iglesias, R. Munoz, F.J. Arias, M. Calonge, J.R. Garcia, E. Mendez, Isolation and partial characterization of nigrin b, a non-toxic novel ribosomeinactivating protein from the bark of Sambucus nigra L, Plant Mol. Biol. 22 (1993) 1181 – 1186. [28] L. Citores, F.M. De Benito, R. Iglesias, J.M. Ferreras, P. Jimenez, P. Argueso, G. Farias, E. Mendez, T. Girbes, Isolation and characterization of a new, non-toxic two-chain ribosome-inactivating protein from fruits of elder (Sambucus nigra L.), J. Exp. Bot. 47 (1996) 1577 – 1585. [29] J.M. Pascal, P.J. Day, A.F. Monzingo, S.R. Ernst, J.D. Robertus, R. Iglesias, Y. Perez, J.M. Ferreras, L. Citores, T. Girbes, .8A crystal structure of a non-toxic type II ribosome-inactivating protein, ebulin 1, Proteins 43 (2001) 319 – 326. [30] J.W. Tregear, L.M. Roberts, The lectin gene family of Ricinus communis: cloning of a functional gene and three lectin pseudogenes, Plant Mol. Biol. 18 (1992) 515 – 525. [31] J.M. Lord, L.M. Roberts, J.D. Robertus, Ricin: structure, mode of action, and some current applications, FASEB J. 8 (1994) 201 – 208. [32] A.G. Butterworth, J.M. Lord, Ricin and Ricinus communis agglutinin subunits are all derived from a single sized polypeptide precursor, Eur. J. Biochem. 137 (1983) 57 – 65. [33] J.-B. Ferrini, M. Martin, M.-P. Taupiac, B. Beaumelle, Expression of functional ricin B chain using the baculovirus system, Eur. J. Biochem. 233 (1995) 772 – 777. [34] F.I. Lamb, L.M. Roberts, J.M. Lord, Nucleotide sequence of cloned cDNA coding for preproricin, Eur. J. Biochem. 146 (1995) 265 – 270. [35] L.M. Roberts, J.M. Lord, The synthesis of Ricinus communis agglutinin. Co-translational and post-translational modifications of agglutinin polypeptides, Eur. J. Biochem. 119 (1981) 31 – 41. [36] J.M. Lord, Synthesis and intracellular transport of lectin and storage

12

[37]

[38]

[39]

[40]

[41] [42]

[43]

[44]

[45]

[46]

[47] [48]

[49]

[50]

[51]

[52]

[53]

[54]

[55] [56]

[57]

M.R. Hartley, J.M. Lord / Biochimica et Biophysica Acta 1701 (2004) 1–14 protein precursors in endosperm from castor bean, Eur. J. Biochem. 146 (1985) 403 – 409. J.M. Lord, Precursors of ricin and Ricinus communis agglutinin. Glycosylation and processing during intracellular transport, Eur. J. Biochem. 146 (1985) 411 – 416. J.M. Lord, S.M. Harley, Ricinus communis agglutinin B chain contains a fucosylated oligosaccharide side chain not present on ricin B chain, FEBS Lett. 189 (1985) 72 – 76. L. Frigerio, N.A. Jolliffe, A. Di Cola, D. Hernandez Felipe, N. Paris, J.-M. Neuhaus, J.M. Lord, A. Ceriotti, L.M. Roberts, The internal propeptide of the ricin precursor carries a sequence-specific determinant for vacuolar sorting, Plant Physiol. 126 (2001) 167 – 175. N.A. Jolliffe, A. Ceriotti, L. Frigerio, L.M. Roberts, The position of the proricin vacuolar targeting signal is functionally important, Plant Mol. Biol. 51 (2003) 631 – 641. S.M. Harley, J.M. Lord, In vitro endoproteolytic cleavage of castor bean lectin precursors, Plant Sci. 41 (1985) 111 – 116. S.M. Harley, H. Beevers, Ricin inhibition of in vitro protein synthesis by plant ribosomes, Proc. Natl. Acad. Sci. U. S. A. 79 (1982) 5935 – 5938. P.T. Richardson, M. Westby, L.M. Roberts, J.H. Gould, A. Colman, J.M. Lord, Recombinant proricin binds galactose but does not depurinate 28S ribosomal RNA, FEBS Lett. 255 (1989) 15 – 20. R.J. Youle, A.H.C. Huang, Protein bodies from the endosperm of castor bean. Subfractionation, protein components, lectins, and changes during germination, Plant Physiol. 58 (1976) 703 – 709. R.E. Tully, H. Beevers, Protein bodies of castor bean endosperm. Isolation, fractionation and characterization of protein components, Plant Physiol. 58 (1976) 710 – 716. K. Sandvig, S. Olsnes, A. Pihl, Kinetics of binding of the toxic lectins ricin and abrin to surface receptors of human cells, J. Biol. Chem. 251 (1976) 3977 – 3984. B. Spiklsberg, G. van Meer, K. Sandvig, Role of lipids in the retrograde pathway of ricin intoxication, Traffic 4 (2003) 544 – 552. B.M. Simmons, P.D. Stahl, J.H. Russell, Mannose receptor-mediated uptake of ricin toxin and ricin A chain by macrophages. Multiple intracellular pathways for A chain translocation, J. Biol. Chem. 261 (1986) 7912 – 7920. S. Magnusson, R. Kjeken, T. Berg, Characterization of two distinct pathways of endocytosis of ricin by rat liver endothelial cells, Exp. Cell Res. 205 (1993) 118 – 125. B. van Deurs, O.W. Petersen, A. Sundan, S. Olsnes, K. Sandvig, Receptor-mediated endocytosis of a ricin-colloidal gold conjugate in Vero cells: intracellular routing to vacuolar and tubulo-vesicular portions of the endosomal system, Exp. Cell Res. 159 (1985) 287 – 304. M. Moya, A. Dautry-Varsat, B. Goud, D. Louvard, P. Boquet, Inhibition of coated pit formation in Hep2 cells blocks the cytotoxicity of diphtheria toxin but not that of ricin, J. Cell Biol. 101 (1985) 548 – 559. J.E. Hinshaw, S.L. Schmid, Dynamin self assembles into rings suggesting a model for coated vesicle budding, Nature 374 (1995) 190 – 192. J.C. Simpson, D.C. Smith, L.M. Roberts, J.M. Lord, Expression of mutant dynamin protects cells against diphtheria toxin but not against ricin, Exp. Cell Res. 239 (1998) 293 – 300. A. Llorente, A. Rapak, S.L. Schmid, B. van Deurs, K. Sandvig, Expression of mutant dynamin inhibits toxicity and transport of ricin to the Golgi apparatus, J. Cell Biol. 140 (1998) 553 – 563. K. Sandvig, S. Olsnes, Diphtheria toxin entry into cells is facilitated by low pH, J. Cell Biol. 87 (1980) 828 – 832. K. Sandvig, B. van Deurs, Endocytosis, intracellular transport, and cytotoxic action of Shiga toxin and ricin, Physiol. Rev. 76 (1996) 949 – 966. B. van Deurs, K. Sandvig, O.W. Petersen, S. Olsnes, K. Simons, G. Griffiths, Routing of internalised ricin and ricin conjugates to the Golgi complex, J. Cell Biol. 102 (1986) 37 – 47.

[58] A. Rapak, P.O. Falnes, S. Olsnes, Retrograde transport of mutant ricin to the endoplasmic reticulum with subsequent translocation to cytosol, Proc. Natl. Acad. Sci. U. S. A. 94 (1997) 783 – 788. [59] M.J. Clague, S. Urbe, F. Aniento, J. Gruenberg, Vacuolar ATPase activity is required for endosomal carrier vehicle formation, J. Biol. Chem. 269 (1994) 21 – 24. [60] E.L. Melby, K. Prydz, S. Olsnes, K. Sandvig, Effect of monensin on ricin and fluid phase transport in polarized MDCK cells, J. Cell. Biochem. 46 (1991) 251 – 260. [61] D. Lombardi, T. Soldati, M.A. Riederer, J. Lin, S. Pfeffer, Rab9 functions in the transport between late endosomes and the transGolgi network, EMBO J. 12 (1993) 677 – 682. [62] M.A. Riederer, T. Soldati, A.D. Shapiro, Y. Goda, M. Zerial, S. Pfeffer, Lysosome biogenesis requires Rab9 function and receptor recycling from late endosomes to the trans-Golgi network, J. Cell Biol. 125 (1994) 573 – 582. [63] J.C. Simpson, C. Dascher, L.M. Roberts, J.M. Lord, W.E. Balch, Ricin cytotoxicity is sensitive to recycling between the endoplasmic reticulum and the Golgi complex, J. Biol. Chem. 270 (1995) 20078 – 20083. [64] T.G. Iversen, G. Skretting, A. Llorente, P. Nicoziani, B. van Deurs, K. Sandvig, Endosome to Golgi transport of ricin is independent of clathrin and the Rab9- and Rab11-GTPases, Mol. Biol. Cell 12 (2001) 2099 – 2107. [65] B.J. Nichols, J. Lippincott-Schwarz, Endocytosis without clathrin coats, Trends Cell Biol. 11 (2001) 406 – 412. [66] F. Mallard, C. Antony, D. Tenza, J. Salamero, B. Goud, L. Johannes, Direct pathway from early/recycling endosomes to the Golgi apparatus revealed through the study of Shiga toxin B-fragment transport, J. Cell Biol. 143 (1998) 973 – 990. [67] F. Mallard, B.L. Tang, T. Galli, D. Tenza, A. Saint-Pol, X. Yue, C. Antony, W. Hong, B. Goud, L. Johannes, Early/recycling endosomes-to-TGN transport involves two SNARE complexes and a Rab6 isoform, J. Cell Biol. 156 (2002) 653 – 664. [68] L. Johannes, The epithelial cell cytoskeleton and intracellular trafficking: shiga toxin B-subunit system, retrograde transport, intracellular vectorization, and more, Am. J. Physiol.: Gasterointest. Liver Physiol. 283 (2002) G1 – G7. [69] J. Lippincott-Schwarz, L.C. Yuan, J.S. Bonifacino, R.D. Klausner, Rapid redistribution of Golgi proteins into the ER in cells treated with brefeldin A: evidence for membrane cycling from Golgi to ER, Cell 56 (1989) 801 – 813. [70] T. Yoshida, C. Chen, M. Zhang, H.C. Wu, Disruption of the Golgi apparatus by brefeldin A inhibits the cytotoxicity of ricin, modeccin and Pseudomonas toxin, Exp. Cell Res. 192 (1991) 389 – 395. [71] K. Sandvig, K. Prydz, S.H. Hansen, B. van Deurs, Ricin transport in brefeldin A-treated cells: correlation between Golgi structure and toxic effect, J. Cell Biol. 115 (1991) 971 – 981. [72] M.P. Namibar, H.C. Wu, Ilimaquinone inhibits the cytotoxicities of ricin, diphtheria toxin and other protein toxins in Vero cells, Exp. Cell Res. 219 (1995) 671 – 678. [73] R. Wales, L.M. Roberts, J.M. Lord, Addition of an endoplasmic reticulum retrieval sequence to ricin A chain significantly increases its cytotoxicity to mammalian cells, J. Biol. Chem. 268 (1993) 23986 – 23990. [74] R. Wales, J.A. Chaddock, L.M. Roberts, J.M. Lord, Addition of an ER retention signal to the ricin A chain increases the cytotoxicity of the holotoxin, Exp. Cell Res. 203 (1992) 1 – 4. [75] J. Wesche, A. Rapak, S. Olsnes, Dependence of ricin toxicity on translocation of the toxin A chain from the endoplasmic reticulum to the cytosol, J. Biol. Chem. 274 (1999) 34443 – 34449. [76] P.J. Day, S.R. Owens, J. Wesche, S. Olsnes, L.M. Roberts, J.M. Lord, An interaction between ricin and calreticulin that may have implications for toxin trafficking, J. Biol. Chem. 276 (2001) 7202 – 7208. [77] P. Cosson, F. Letourneur, Coatamer interaction with di-lysine endoplasmic reticulum retention motifs, Science 263 (1994) 1629 – 1631.

M.R. Hartley, J.M. Lord / Biochimica et Biophysica Acta 1701 (2004) 1–14 [78] A. Girod, B. Storrie, J.C. Simpson, L. Johannes, B. Goud, L.M. Roberts, J.M. Lord, T. Nilsson, R. Pepperkok, Evidence for a COP-I-independent transport route from the Golgi complex to the endoplasmic reticulum, Nat. Cell Biol. 1 (1999) 423 – 430. [79] V. Chaudhary, Y. Jinno, D. FitzGerald, I. Pastan, Pseudomonas exotoxin contains a specific sequence at the carboxyl terminus that is required for cytotoxicity, Proc. Natl. Acad. Sci. U. S. A. 87 (1990) 308 – 312. [80] W.I. Lencer, C. Constable, M.G. Jobling, H.M. Webb, S.R. Ruson, J.L. Madara, T.R. Hirst, R.K. Holmes, Targeting of cholera toxin and Escherichia coli heat-labile toxin in polarized epithelia; role of COOH-terminal KDEL, J. Cell Biol. 131 (1995) 951 – 962. [81] I. Majoul, P.I.H. Bastiaens, H.-D. So¨ling, Transport of an external Lys-Asp-Glu-Leu (KDEL) protein from the plasma membrane to the endoplasmic reticulum: studies with cholera toxin in Vero cells, J. Cell Biol. 133 (1996) 777 – 789. [82] M.E. Jackson, J.C. Simpson, A. Girod, R. Pepperkok, L.M. Roberts, J.M. Lord, The KDEL retrieval system is exploited by Pseudomonas exotoxin A, but not by Shiga-like toxin-1, during retrograde transport from the Golgi complex to the endoplasmic reticulum, J. Cell. Sci. 112 (1999) 467 – 475. [83] J. White, L. Johannes, F. Mallard, A. Girod, S. Grill, S. Reinsch, P. Keller, B. Tzschaschel, A. Echard, B. Goud, E.H. Stelzer, Rab6 coordinates a novel Golgi to ER retrograde transport pathway in live cells, J. Cell Biol. 147 (1999) 743 – 759. [84] A.E. Johnson, M.A. van Maes, The translocon: a dynamic gateway at the ER membrane, Annu. Rev. Cell Dev. Biol. 15 (1999) 799 – 842. [85] G. Russ, F. Esquivel, J.W. Yewdell, P. Cresswell, T. Spies, J.R. Bennink, Assembly, intracellular location and nucleotide binding properties of the human peptide transporters TAP1 and TAP2 expressed by recombinant vaccinia viruses, J. Biol. Chem. 270 (1995) 21312 – 21318. [86] E.J.H.J. Wiertz, D. Tortorella, M. Bogyo, J. Yu, W. Mothes, T.R. Jones, T.A. Rapoport, H.L. Ploegh, Sec61-mediated transfer of a membrane protein from the endoplasmic reticulum to the proteasome for destruction, Nature 384 (1996) 432 – 438. [87] R.K. Plemper, S. Bo¨hmler, J. Bordallo, T. Sommer, D.H. Wolf, Mutant analysis links the translocon and BiP to retrograde protein transport for ER degradation, Nature 388 (1997) 891 – 895. [88] J.C. Simpson, L.M. Roberts, K. Ro¨misch, J. Davey, D.H. Wolf, J.M. Lord, Ricin A chain utilises the endoplasmic reticulum-associated protein degradation pathway to enter the cytosol of yeast, FEBS Lett. 459 (1999) 80 – 84. [89] L. Ellgaard, M. Molinari, A. Helenius, Setting the standards: quality control in the secretory pathway, Science 286 (1999) 1882 – 1888. [90] R.D. Klausner, R. Sitia, Protein degradation in the endoplasmic reticulum, Cell 62 (1990) 611 – 614. [91] B. Tsai, T.A. Rapoport, Retro-translocation of proteins from the endoplasmic reticulum to the cytosol, Nat. Rev., Mol. Cell Biol. 3 (2002) 246 – 255. [92] J.L. Brodsky, A.A. McCracken, ER protein quality control and proteasome-mediated protein degradation, Semin. Cell Dev. Biol. 10 (1999) 507 – 513. [93] R.K. Plemper, D.H. Wolf, Retrograde protein translocation: ERADication of secretory proteins in health and disease, Trends Biochem. Sci. 24 (1999) 266 – 270. [94] B. Tsai, C. Rodighero, W.I. Lencer, T.A. Rapoport, Protein disulfide isomerase acts as a redox-dependent chaperone to unfold cholera toxin, Cell 104 (2001) 937 – 948. [95] P.J. Day, T.J.T. Pinheiro, L.M. Roberts, J.M. Lord, Binding of ricin A chain to negatively charged phospholipid vesicles leads to protein structural changes and destabilizes the lipid bilayer, Biochemistry 41 (2002) 2836 – 2843. [96] B. Hazes, R.J. Read, Accumulating evidence suggests that several A-B toxins subvert the endoplasmic reticulum-associated protein degradation pathway to enter target cells, Biochemistry 36 (1997) 11051 – 11054.

13

[97] E.D. Deeks, J.P. Cook, P.J. Day, D.C. Smith, L.M. Roberts, J.M. Lord, The low lysine content of ricin a chain reduces the risk of proteolytic degradation after translocation from the endoplasmic reticulum to the cytosol, Biochemistry 41 (2002) 3405 – 3413. [98] C. Rodighero, B. Tsai, T.A. Rapoport, W.I. Lencer, Role of ubiquitination in retro-translocation of cholera toxin and escape of cytosolic degradation, EMBO Rep. 3 (2002) 1222 – 1227. [99] R.H. Argent, A.W. Parrott, P.J. Day, P.G. Stokley, L.M. Roberts, J.M. Lord, S.E. Radford, Ribosome-mediated folding of partially unfolded ricin A chain, J. Biol. Chem. 275 (2000) 9263 – 9269. [100] Y. Endo, K. Tsurugi, RNA N-glycosidase of ricin A chain. Mechanism of action of the toxic lectin ricin on eukaryotic ribosomes, J. Biol. Chem. 262 (1987) 8128 – 8130. [101] M.R. Hartley, G. Legname, R.W. Osborn, Z. Chen, J.M. Lord, Single-chain ribosome-inactivating proteins from plants depurinate Escherichia coli 23S ribosomal RNA, FEBS Lett. 290 (1991) 65 – 68. [102] D. Moazed, J.M. Robertson, H.F. Noller, Interaction of elongation factors EF-G and EF-Tu with a conserved loop in 23S rRNA, Nature 334 (1998) 362 – 364. [103] Y. Endo, I.G. Wool, The site of action of a-sarcin on eukaryotic ribosomes. The sequence in the a-sarcin cleavage site in 28S ribosomal ribonucleic acid, J. Biol. Chem. 257 (1982) 9054 – 9060. [104] A. Marchant, M.R. Hartley, Mutational studies on the a-sarcin loop of Escherichia coli 23S ribosomal RNA, Eur. J. Biochem. 226 (1994) 141 – 147. [105] S.L. Larsson, M.S. Sloma, O. Nygard, Conformational changes in the structure of domains II and V of 28S rRNA in ribosomes treated with the translational inhibitors ricin or a-sarcin, Biochim. Biophys. Acta 1577 (2002) 53 – 62. [106] I.G. Wool, A. Gluck, Y. Endo, Ribotoxin recognition of ribosomal RNA and a proposal for the mechanism of translocation, Trends Biochem. Sci. 17 (1992) 266 – 269. [107] I.G. Wool, C.C. Correll, Y.-L. Chan, Structure and function of the sarcin-ricin domain, in: R.A. Garrett, S.R. Douthwaite, A. Liljas, A.T. Marteson, P.B. Moore, H.F. Noller (Eds.), The Ribosome: Structure, Function, Antibiotics and Cellular Interactions, ASM Press, Washington, DC, 2000. [108] A.A. Szewczak, P.B. Moore, The sarcin/ricin loop, a modular RNA, J. Mol. Biol. 247 (1995) 81 – 98. [109] C.C. Correll, A. Munishkin, Y.-L. Chan, Z. Ren, I.G. Wool, T.A. Steitz, Crystal structure or the ribosomal RNA domain essential for binding elongation factors, Proc. Natl. Acad. Sci. U. S. A. 95 (1998) 13436 – 13441. [110] C.C. Correll, I.G. Wool, A. Munishkin, The two faces of the Escherichia coli 23S rRNA sarcin/ricin domain: the structure at 1.11 A resolution, J. Mol. Biol. 292 (1999) 275 – 287. [111] N. Ban, P. Nissen, J. Hansen, P.B. Moore, T.A. Steitz, The complete atomic structure of the large ribosomal subunit at 2.4A resolution, Science 289 (2000) 905 – 920. [112] L. Holmberg, O. Nygard, Depurination A4526 in 28S rRNA by the ribosome-inactivating proteins from barley and ricin results in different ribosome conformations, J. Mol. Biol. 259 (1966) 81 – 94. [113] Y. Endo, K. Tsurugi, The RNA N-glycosidase of ricin A-chain: the characteristics of the enzymatic activity of ricin A-chain with ribosomes and rRNA, J. Biol. Chem. 263 (1988) 8735 – 8739. [114] K.A. Hudak, J.D. Dinman, N.E. Tumer, Pokeweed antiviral protein accesses ribosomes by binding to L3, J. Biol. Chem. 274 (1999) 3859 – 3864. [115] T. Uchiumi, N. Sato, A. Wada, A. Hachimori, Interaction of the sarcin/ricin domain of 23S ribosomal RNA with proteins L3 and L6, J. Biol. Chem. 274 (1999) 681 – 686. [116] A. Marchant, M.R. Hartley, The action of pokeweed antiviral protein and ricin A-chain on mutants in the a-sarcin loop of Escherichia coli 23S ribosomal RNA, J. Mol. Biol. 254 (1995) 848 – 855. [117] K.A. Hudak, P. Wang, N.E. Tumer, A novel mechanism for inhibi-

14

[118]

[119]

[120]

[121]

[122]

[123]

[124]

[125]

[126]

M.R. Hartley, J.M. Lord / Biochimica et Biophysica Acta 1701 (2004) 1–14 tion of translation by pokeweed antiviral protein: depurination of the capped RNA template, RNA 6 (2000) 369 – 380. L. Barbieri, P. Valbonesi, E. Bonora, P. Gorini, A. Bolognesi, F. Stirpe, Polynucleotide:adenosine glycosidase activity of ribosomeinactivating proteins: effects on DNA, RNA and poly (A), Nucleic Acids Res. 25 (1997) 518 – 522. F. Rajamohan, T.K. Venkatachalam, J.D. Irvin, F.M. Uckun, Pokeweed antiviral protein isoforms PAP-1, PAP-II and PAP-III depurinate RNA of human immunodeficiency virus (HIV)-1, Biochem. Biophys. Res. Commun. 260 (1999) 453 – 458. N.E. Tumer, D.-J. Hwang, M. Bonness, C-terminal deletion mutant of pokeweed antiviral protein inhibits viral infection but does not depurinate host ribosomes, Proc. Natl. Acad. Sci. U. S. A. 94 (1997) 3866 – 3871. F. Rajamohan, I.V. Kurinov, T.K. Venkatachalam, F.M. Uckun, Deguanylation of human immunodeficiency virus (HIV-1) RNA by recombinant pokeweed antiviral protein, Biochem. Biophys. Res. Commun. 263 (1999) 419 – 424. K.A. Hudak, J.D. Bauman, N.E. Tumer, Pokeweed antiviral protein binds to the cap structure of eukaryotic mRNA and depurinates the mRNA downstream of the cap, RNA 8 (2002) 1148 – 1159. B.J. Parikh, C. Coetzer, N.E. Tumer, Pokeweed antiviral protein regulates its own stability by a mechanism that requires depurination but can be separated from depurination of the a-sarcic/ricin loop of rRNA, J. Biol. Chem. 277 (2002) 41414 – 41428. M.S. Bonness, M.P. Ready, J.D. Irvin, T.J. Mabry, Pokeweed antiviral protein inactivates pokeweed ribosomes: implications for the antiviral mechanism, Plant J. 5 (1994) 173 – 183. J.H. Gould, M.R. Hartley, P.C. Welsh, D.K. Hoshizaki, A. Frankel, L.M. Roberts, J.M. Lord, Alteration of an amino acid outside the active site of ricin A chain reduces its toxicity towards yeast ribosomes, Mol. Gen. Genet. 230 (1991) 81 – 90. Y.-X. Wang, N. Neamati, J. Jacob, S.J. Stahl, J.D. Kaufman, P.L. Huang, P.L. Huang, H.E. Winslow, Y. Pommier, P.T. Wingfield, S. Lee-Huang, A. Bax, D.A. Torchia, Solution structure of anti-HIV-1

[127]

[128]

[129]

[130]

[131]

[132]

[133]

[134]

[135] [136]

and anti-tumor protein MAP30: structural insights into its multiple functions, Cell 99 (1999) 433 – 442. P.J. Day, J.M. Lord, L.M. Roberts, The deoxyribonuclease activity attributed to ribosome-inactivating proteins is due to contamination, Eur. J. Biochem. 258 (1998) 540 – 545. L. Barbieri, P. Valbonensi, F. Righi, G. Zuccheri, F. Monti, P. Gorini, B. Samori, F. Stirpe, Polynucleotide:adenosine glycosidase is the sole activity of ribosome-inactivating proteins on DNA, J. Biochem. 128 (2000) 883 – 889. S. Bagga, D. Seth, K. Batra, The cytotoxic activity of ribosomeinactivating protein saporin-6 is attributed to its rRNA N-glycosidase and internucleosamal DNA fragmentation activities, J. Biol. Chem. 278 (2002) 4813 – 4820. K. Nielsen, R.S. Boston, Ribosome-inactivating proteins: a plant perspective, Annu. Rev. Plant Physiol. Plant Mol. Biol. 52 (1999) 785 – 816. N. Hasegawa, Y. Kimura, T. Oda, N. Komatsu, T. Muramatsu, Isolated ricin B chain-mediated apoptosis in U937 cells, Biosci. Biotechnol. Biochem. 64 (2000) 1422 – 1429. P. Stanley, Surface carbohydrate alteration of mutant mammalian cells selected for resistance to plant lectins, in: W.J. Lennarz (Ed.), The Biochemistry of Glycoproteins and Proteoglycans, Plenum Press, New York, 1981, pp. 161 – 190. M. Kim, M.V. Rao, D.J. Tweady, M. Prakash, U. Galili, E. Gorelik, Lectin-induced apoptosis of tumour cells, Glycobiology 3 (1993) 447 – 453. M. Brigotti, R. Alfieri, P. Sestili, M. Bonelli, P.G. Petronini, A. Guidarelli, L. Barbieri, F. Stirpe, S. Sperti, Damage to nuclear DNA induced by Shiga toxin I and ricin in human endothelial cells, FASEB J. 16 (2002) 365 – 372. J. Morlon-Guyot, M. Helmy, S. Lombard-Frasca, D. Pignol, G. Pieroni, B. Beaumelle, J. Biol. Chem. 278 (2003) 17006 – 17011. T.H. Hudson, D.M. Neville, Temporal separation of protein toxin translocation from processing events, J. Biol. Chem. 262 (1987) 16484 – 16494.