Ribosome-inactivating Proteins (Ricin and Related Proteins) || Occurrence and Taxonomical Distribution of Ribosome-inactivating Proteins Belonging to the Ricin/Shiga Toxin Superfamily

Ribosome-inactivating Proteins: Ricin and Related Proteins, First Edition. Edited by Fiorenzo Stirpe and Douglas A. Lappi.

© 2014 John Wiley & Sons, Inc. Published 2014 by John Wiley & Sons, Inc.

11

2 Occurrence and Taxonomical Distribution of Ribosome-inactivating Proteins Belonging to the Ricin/Shiga Toxin SuperfamilyChenjing Shang1, Willy J. Peumans2, and Els J. M. Van Damme1

1 Department of Molecular Biotechnology, Ghent University, Belgium2 Aalst. Belgium

Introduction

Ribosome-inactivating proteins are by definition protein synthesis inhibitors that act at the level of the ribosome.1 Though the activity of ribosomes can be affected by different types of proteins/enzymes (e.g., proteases, RNases, ribosome-binding proteins) this contribution deals exclusively with proteins that by virtue of a well-defined RNA N-glycosidase activity are capable of depurinat-ing a specific adenine in what is called the conserved α-sarcin/ricin loop of the large ribosomal RNA. The latter activity relies on the presence of a conserved structural domain (Pfam PF00161) of approximately 250 amino acid residues that occurs either as a single domain or in association with other proteins/domains. Proteins with such a domain are usually classified into the ricin/Shiga toxin superfamily because the latter two proteins are the best studied and most notorious examples of ribosome-inactivating proteins or RIPs. Though the name ricin/Shiga toxin superfamily is certainly informative, because it highlights the evolutionary relationship between two structurally different groups of toxins from two taxonomically distant phyla, it is too restrictive in the sense that there is no direct link with other types of RIPs. Therefore, it seems preferable to discuss the whole family of RIPs in terms of the presence of an RNA N-glycosidase domain structurally homologous to the A chain of both the Shiga toxin and ricin.

RIPs are classically subdivided in three main groups namely the (bacterial) Shiga and Shiga-like toxins (Stx) and the plant type 1 and type 2 RIPs. Shiga and Shiga-like toxins are built up of a cat-alytically active A subunit equivalent to a RIP domain and a B subunit, which itself is a pentamer – of five identical 89 amino acid residue polypeptides – that specifically binds to the glycolipid globo-triaosylceramide. Both the A- and B-subunits are synthesized on a single dicistronic mRNA as two separate polypeptide chains.2 Type 2 RIPs from plants are also typically described in terms of an AB structure.3–5 It should be emphasized, however, that the overall structure of, for example, ricin is completely different from that of the Shiga toxin. First, the B subunit of ricin consists of a single polypeptide chain and, second, both the A and B chains are synthesized on the very same precursor polypeptide that is post-translationally processed through the excision of a linker polypeptide be -tween the two domains. Plant type 1 RIPs are far simpler from a structural point of view since they

12 RIBOSOME-INACTIVATING PROTEINS

consist only of a RIP domain.3–5 However, this subgroup is less uniform than the type 2 RIPs and in some cases (e.g., a maize RIP b-326) enzymatic activity is acquired only after proteolytic processing of the A-domain into two smaller polypeptides. Though the mature maize RIP b-32 corresponds in fact to a proteolytically cleaved type 1 RIP, it was called a type 3 RIP to distinguish it from other type 1 RIPs known at that time.7 Unfortunately, the term type 3 RIP had already been introduced earlier for a 60 kDa jasmonate-induced barley leaf protein (JIP60) consisting of an N-terminal A domain fused to an unrelated domain with no known activity.8 The identification of JIP60 not only increases the complexity –in terms of domain structure – of the RIP family, but also argues for a clear and unambiguous classification.

At present there is some confusion about the taxonomic distribution of RIPs. The isolation and char-acterization of numerous Shiga/Shiga-like toxins and both type 1 and type 2 RIPs leave no doubt for the occurrence of the RIP domain in both bacteria and plants. Though it is – especially in review papers – often stated that RIPs are also present in fungi, algae, and even in mammalian tissues9–11 no experimental evidence has been presented yet that in any of these cases the presumed RIP activity relies on the presence of a protein with a genuine N-glycosidase domain. On the contrary, it seems pretty evident that, for example, all so-called fungal RIPs belong to the family of fungal ribotoxins, which possess a ribonuclease rather than an N-glycosidase domain.12 However, it is questionable whether our current view on the taxonomic distribution of RIPs, which is merely based on studies of isolated proteins, reflects the actual occurrence of the N-glycosidase domain in living organisms. Most RIPs identified thus far are expressed at a level allowing purification by standard biochemical tech-niques. Moreover, even though (very) low levels of RIPs can be traced through their enzymatic activity, one cannot exclude that some escape detection or, alternatively, are for practical reasons not further investigated. As a result, the distribution of the RIP domain might well be underestimated, which is very unfortunate for two main reasons. First, some RIPs with unique properties/activities cannot be exploited. Second, no detailed study of the phylogeny and molecular evolution of proteins with an N-glycosidase domain can be elaborated in the absence of a comprehensive overview of the occur-rence of the RIP domain throughout living organisms. Therefore the issue of the taxonomic distribu-tion should be readdressed to extract an overview of all proteins/genes with an N-glycosidase domain.

How to Investigate the Distribution of Proteins with an N-glycosidase Domain?

The discussion of the taxonomical distribution is in most review papers primarily based on the list of RIPs that have been isolated and (partly) characterized.3, 5 Though this approach is certainly help-ful for scientists interested in the use – in the broadest sense – of RIPs, it is obviously too restrictive for in-depth studies of the phylogeny and molecular evolution of proteins with an N-glycosidase domain. As pointed out in a recent review, preliminary screenings of plant genome and transcrip-tome databases demonstrated that an extended list of expressed RIPs as well as putative RIP genes can already be added to the existing lists of documented RIPs.5, 13 Moreover, analyses of the com-pleted plant genomes revealed the occurrence of RIP gene families, some of which encode proteins with previously unknown domain architecture. For example, the Oryza (rice) genome contains a complex set of at least 30 different genes in which a RIP domain can be identified.14 Most of these genes encode (putative) proteins consisting of a single RIP domain, but some exhibit a chimeric overall structure in which an N-terminal RIP domain precedes an unrelated C-terminal domain.

Two important conclusions can be drawn from the results of the preliminary screening of the plant genomes/transcriptomes. First, the existing lists of RIPs have to be extended horizontally (more species) and vertically (more RIPs/RIP genes per species). Second, all evidence suggests that several types of

OCCURRENCE AND TAXONOMICAL DISTRIBUTION OF RIBOSOME-INACTIVATING PROTEINS 13

proteins with a novel domain architecture must be added to the present list of chimeric RIP genes (i.e., type 2 and type 3 RIPs). In addition, the identification of numerous “novel” plant RIPs by an in silico analysis raises the question of whether the same approach might reveal the possible occurrence of as yet unidentified RIPs/RIP genes outside the plant kingdom. Therefore, the issue of the taxonomic distribution of RIPs cannot be treated without a thorough exploration of the huge amount of data gen-erated by genome and transcriptome sequencing programs of both prokaryotes and eukaryotes.

Since this contribution deals exclusively with proteins/genes possessing a canonical N-glycosidase domain (equivalent to the toxin part of Stx and ricin) the availability of a complete sequence or at least sufficient sequence information is a prerequisite for any entry to be included. Accordingly, the retrieval of RIPs and RIP genes is (apart from the Stx and plant type 1 and type 2 RIPs that have been studied in detail at the protein level) based in the first place on BLAST searches. These BLAST searches were not confined to protein databases (BLASTp) but covered also all publicly accessible genome and transcriptome databases (tBLASTn searches). In a first round of searches, sequences of well-studied RIPs (e.g., ricin, pokeweed antiviral protein, Shiga toxin) were used as queries. Subsequently, sequences of newly retrieved types of RIPs/RIP genes were used in a next round of BLAST searches. The latter process was repeated until no new sequences could be identified.

As will be discussed below, a more detail screening of the databases eventually resulted in the identification of (i) numerous novel RIP genes and (ii) several previously unknown chimeric forms. Taking into consideration the existing ambiguity of the currently used classification (type 1, 2, and 3 RIPs) it seems rather inappropriate to increase the complexity of the system for the newly identi-fied chimeric forms. Therefore, a novel system is introduced based on the domain architecture of the RIPs/RIP genes. Proteins/genes consisting of a single N-glycosidase domain will be referred to as the “[A] type” and the chimeric forms as the “[AN] type” whereby N stands for the different types of (unknown) C-terminal domains.

A final remark concerns the catalytic activity of the (putative) RIPs found in different organisms. Virtually all RIPs studied thus far exhibit N-glycosidase activity, but a few lectins closely related to type 2 RIPs (e.g., a lectin from Bryonia dioica) apparently possess an A chain homolog that lacks catalytic activity (unpublished data). Though the obvious occurrence of catalytically inactive RIP domains in some plants is rather anecdotal, it is highly relevant because it demonstrates that the presence of a domain that shares a reasonable sequence similarity with a genuine RIP does not nec-essarily imply that it possesses enzymatic activity. Therefore, it cannot be excluded that some of the RIPs/RIP genes discussed in the next section lack a catalytically active domain.

Overview of the Occurrence of the N-glycosidase Domain in Living Organisms

Hitherto, genuine RIPs have been isolated exclusively from bacteria and plants. It appears, however, that (expressed) RIP genes occur also in some fungi and in a few insects. Since plant RIPs are by far the largest and most heterogeneous group, they will be discussed first. Subsequently, bacterial, fungal, and insect RIPs will be discussed in separate sections.

Plant RIPs

As already suggested in an earlier review, neither the taxonomic distribution nor heterogeneity (in terms of domain structure) are accurately reflected by the (long) list of plant RIPs that have been purified and characterized in some detail.13 However, this does not imply that the N-glycosidase domain is


ubiquitous in plants. On the contrary, the absence of a RIP domain from the first completed plant genome of Arabidopsis thaliana provided evidence for the opposite more than a decade ago. In the meantime, numerous genomes of species covering all major plant taxa have been completed, which now allows a better overview of the occurrence and overall structure of genes with an N-glycosidase domain by an in silico approach that does not depend on the availability of “protein data.”

What can be Learned from Analyses of Completed Plant Genomes?To identify RIP genes in the (nearly) completed plant genomes, extensive screenings based pri-marily on tBLASTn searches of genomic (nucleotide) sequences were set up. Searches were not confined to a BLASTp approach for the following reasons. First, several genomes are not yet anno-tated. Second, automated annotation of (putative) proteins does not identify all RIP domains. Third, the accuracy of the automated annotation is – especially for Poaceae species – insufficient (due to the presence of introns and the insertion of transposable elements).

The results of this extensive in silico screening covering a total number of 42 plant genomes are summarized in Figure 2.1, showing a phylogenetic tree of plant species with indication of the RIP gene complement. For most species, the RIP genes (if present) could readily be identified. However, for the Poaceae species the outcome of the screening is still preliminary because of the complexity of the RIP gene complement and the occurrence of (multiple) introns in some RIP genes. Apart from the identification of the genes, the in silico analysis revealed the occurrence of several as yet unknown chimeric RIPs (namely type [AD], [AC], [AX], [APM41], and [APC19]; for details see Figure 2.2) as well as at least two different lineages of the type [A] RIP. These latter two lineages are the type [AΔB] and type [AΔX] RIPs, which are derived by domain deletion events from the type [AB] and type [AX] chimeric RIPs, respectively (for details see Chapter 9). At present it is not clear how the multiple type [A] RIPs found in Poaceae species should be classified. Most probably they form a complex set of as yet unidentified type [A] RIPs and accordingly are, for practical reasons, classified here as type [Au].

Several important conclusions can be drawn from Figure 2.1. First, RIP genes are apparently absent from 24 out of 42 completed genomes. Even within the group of flowering plants, more than half of all species investigated (20 out of the 38) lack RIP gene(s). Second, in some families (e.g., Euphorbiaceae and Poaceae) RIP genes are found in all sequenced genomes whereas in others (e.g., Rosaceae) RIP genes occur in some genomes (Malus domestica and Prunus persica) but are absent from others (Fragaria vesca). Third, there are striking differences between the RIP gene complement of the dif-ferent species, ranging from a single gene to a complex set of genes encoding both type [A] and chimeric RIPs. Extended RIP gene families are apparently common in Poaceae species, but there are striking interspecific differences with respect to both the gene number and the domain architecture.

Transcriptome Analyses Yield a More Complete Overview of the Occurrence of Expressed RIP Genes in PlantsThough indicative, the summary presented in Figure 2.1 might not be very representative because it covers only three species from which one or more RIPs have been purified and characterized (namely maize (Zea mays), water melon (Cucumis sativus), and castor bean (Ricinus communis)). Most RIPs have been isolated, indeed, from species that are not particularly interesting for whole genome sequencing. Therefore in silico searches for RIP genes were also extended to all other plant genome and transcriptome sequences.

A general overview of the documented occurrence of (expressed) RIP genes within the major taxa of green plants is presented in Figures 2.3 and 2.4. Apart from a single representative of the Gnetophyta (Gnetum gnemon), RIP sequences were exclusively found in flowering plants (angio-sperms or Magnoliophyta). To illustrate the overall occurrence and phylogenetic distribution within

Solanum lycopersicum: No RIP gene

Asterids

Euro

sid

s II

Eurosids I

Flo

werin

g p

lants

Vascula

r pla

nts

Land p

lants

Gre

en p

lants

Lequmes

Monocots

Grasses

Rosid

s

Solanum tuberosum: No RIP gene

Mimulus guttatus: No RIP gene

Vitis vinifera: No RIP gene

Eucalyptus grandis: No RIP gene

Citrus clementina: No RIP gene

Citrus sinensis: No RIP gene

Carica papaya: No RIP gene

Arabidopsis lyrata: No RIP gene

Arabidopsis thaliana: No RIP gene

Capsella rubella: No RIP gene

Thelungiella halophila: No RIP gene

Brassica rapa: No RIP gene

Linum usitatissimum: No RIP gene

Jatropha curcas: 7 [Ac]

Manihot esculenta: 4 [Ac]

Ricinus communis: 8 [AB]; 7 [Ac]

Populus trichocarpa: 1 [AX]

Cannabis sativa: 7 [Ac]; 2 [AΔx]; 1 [AX]

Prunus persica: 4 [Ac]

Malus domestica: 3 [Ac]; 1 [AB]

Fragaria vesca: No RIP gene

Cajanus cajan: No RIP gene

Glycine max: No RIP gene

Medicago truncatula: No RIP gene

Lotus japonicus: No RIP gene

Aquilegia coerulea: 1 [AB]

Cucumis sativus: 2 [AB]; 3 [ΔAB]

Phoenix dactylifera: 1 [AΔx]; 1 [ΔAX]

Brachypodium distachyon*: 15 [Au];1 [AC];1 [AP]

Oryza sativa*: >30 [Au]; 3 [AC]; 1 [AP]

Zea mays*: 7 [Au]; 2 [AB]; 1 [AC]; 1 [AD]

Sorghum bicolor*: 14 [Au]; 1 [AB]

Setaria italica*: >5 [Au]; 2 [AC]

Amborella trichopoda: No RIP gene

Selaginella moellendorffii: No RIP gene

Physcomitrella patens: No RIP gene

Chlamydomonas reinhardii: No RIP gene

Volvox carteri: No RIP gene

Citrullus lanatus: 2 [Ac]

Gossypium raimondii: 1 [AB]

Theobroma cacao: 2 [AΔx], 3[AX]; 1[ΔAX]

Figure 2.1 Schematic overview of the presence/absence of RIP genes in the currently completed plant genomes. The dendro-

gram reflects only the overall phylogeny of the species listed. The presence or absence of RIP genes is indicated. *denotes

preliminary results.


Parent type:type [AB]

[AB] lineage

[AX] lineage

[AC] lineage

[AD] lineage

[APM41] lineage

[APc19] lineage

Offspring type:type [AΔB]

Offspring type:type [AΔx]

Offspring type:type [AΔc]?

Offspring type:type [AΔD]?

Parent type:type [AC]

?

?

Parent type:type [AD]

Signal

peptide

EP=expressed

proteinX domain C domain D domain

RIP domain

Classical type 2 RIP

Vacuolar form (common)

Cytoplasmic form (common)

Classical type 1 RIP

Malus domestica EP

Cannabis sativa EP

Fagus sylvatica EP

Muscari armeniacum

type 1 RIP

Hordeum vulgare JIP 60

Zea mays EP

Oryza sativa EP

Triticum aestivum EP

Possibly some poaceae type [Au] RIP

Possibly some poaceae type [Au] RIP

Cytoplasmic form (rare)

Vacuolar form (rare)

Peptidase C19

domain

Peptidase M41

domain

Parent type:type [AX]

Figure 2.2 Schematic overview of the domain architectures identified in plant RIP genes.

Amborellales

Magnoliids

Commelinids

Fabids

Rosids

Asterids

Lamiids

Campanulids

Malvids

Monocots

Eudicots

Core eudicots

An

gio

sperm

s

Nymphaeales

AustrobaileyalesPiperalesCanellalesMagnoliales [AB]Chloranthales

CommelinalesZingiberalesPoales [Au], [AB], [AC], [AD], [AP]Arecales [AΔX], [AB]

DasypogonaceaeAsparagales [AΔB], [AΔX]LilialesPandanalesDioscoreales

PetrosavialesAlismatalesAcorales

CeratophyllalesRanunculales [AΔB], [AB]SabiaceaeProtealesBuxalesTrochodendrales

GunneralesCucurbitales [AΔB], [AB], [ΔAB]

Fagales [AΔX]

Rosales [AΔB], [AΔX], [AB]

Malpighiales [AΔB], [AB], [AX]

Malvales [AΔX], [AB], [AX]BrassicalesHuertealesSapindales [AB]PicramnialesCrossosomatalesMyrtales

GeranialesVitalesSaxifragalesDilleniaceaeBerberidopsidalesSantalales [AB]Caryophyllales [AΔB], [ΔX]Cornales [AB]Ericales [AΔB], [AB]Garryales

GentianalesLamiales [AΔB]SolanalesBoraginaceaeAquifolialesEscallonialesAsterales [AB]Dipsacales [AB], [ΔAB]ParacryphialesApiales [AΔX], [AB], [AX]

Bruniales

Zygophyllales

Fabales [AB], [AX]

CelastralesOxalidales

Figure 2.3 Schematic overview of the documented occurrence of (expressed) RIP genes within the major taxa of Angiosperms.

The dendrogram (based on APG III) reflects only the overall phylogeny of the taxa listed.


the latter taxon, a summary of the results of the transcriptome screenings is superimposed on the phylogenetic tree of the Angiosperm Phylogeny Group III system (APG III, 2009), which is the most recent version of a modern, primarily molecular-based, system of plant taxonomy. Further details about each RIP-positive group are given in Tables 2.1 and 2.2, listing approximately 100 non-grass and over 20 grass species. Though illustrative, the results of the present overview have to be interpreted with care because they are highly biased by the availability of sequence information. No conclusion can be drawn, indeed, neither with respect to the possible absence of RIP genes from a particular species, family, or higher order taxon nor for what concerns the apparent frequent occurrence in some taxonomic groups (e.g., Caryophyllaceae, Cucurbitaceae). What can be con-cluded with certainty is that RIP genes (i) occur over a wide taxonomic range of flowering plants and (ii) are particularly prominent (or possibly even ubiquitous) in Poaceae. Moreover, it appears that the genome of most grasses possesses a fairly complex family of RIP genes, some of which might be confined to the Poaceae.15

Eudicots (+)(3)

Monocots (+)(3)

Magnoliids (+)(2)

Ceratophyllales (–)

Chloranthales (–)

Austrobaileyales (–)

Nymphaeales (–)

Amborellales (–)

Coniferophyta (–)

Gnetophyta (–)

Ginkgophyta (+)(1)

Cycadophyta (–)

Polypodiopsida (–)

Equisetopsida (–)

Marratiopsida (–)

Ophioglossopsida (–)

Psilotopsida (–)

Lycopodiophyta (–)

Bryophyta (–)

Marchantiophyta (–)

Anthocerophyta (–)

Chlorophyta (–)

Figure 2.4 Schematic overview of the documented occurrence of (expressed) RIP genes within the major taxa of Viridiplantae.

The dendrogram reflects only the overall phylogeny of the taxa listed (based on Palmer et al.).15 (1) Refers to Gnetum gnemon,

containing two [AB] RIPs. (2) refers to Cinnamomum camphora, containing three [AB] RIPs. (3) For details on the occurrence of

RIP genes in Eudicots and Monocots, we refer to Figure 2.3.


Bacterial RIPs – Shiga Toxin Group

Genuine Shiga and Shiga-Like ToxinsThe best known and, until recently the only identified, group of bacterial RIPs are the so-called Shiga and Shiga-like toxins. The Shiga toxin itself is produced and secreted by Shigella dysenteriae. Nearly identical toxins were also found in some Escherichia coli strains (more specifically in the Shiga toxigenic group of E. coli (STEC)). All members of the Shiga toxin family have the same

Table 2.1 Detailed overview of the documented taxonomic distribution of the different types of RIPs within the Magnoliids and

Monocot divisions of the Angiosperms (flowering plants). Taxa are ordered according to the dendrogram shown in Figure 2.3.

MAGNOLIIDS

Laurales Lauraceae Cinnamomum camphora 3 [AB]

MONOCOTSCommelinids Arecaceae Elaeis guineensis* [AB]

Arecales Phoenix dactylifera [AΔX], [ΔAX]

CommelinidsPoales

Bromeliaceae Ananas comosus [Au]

Poaceae B#

E

P

Phyllostachys edulis [Au], [AB]

Oryza sativa* >30 [Au], 3 [AC], [AP]

Agrostis capillaries [Au]

Avena barbata [Au], [AC]

Brachypodium distachyon* 15 [Au], [AC], [AP]

Festuca arundinacea [Au]

Festuca pratensis [Au]

Aegilops speltoides [Au]

Hordeum vulgare >20 [Au], [AC]

Leymus cinereus [Au]

Pseudoroegneria spicata Multiple [Au]

Secale cereale Multiple [Au]

Triticum aestivum >20 [Au], [AB], [AP]

P#

A

C

C

A

D

Saccharum officinarum [Au], [AB]

Asparagales

Sorghum bicolor* 14 [Au], [AB]

Zea mays* 7 [Au], 2 [AB], [AC], AD]

Panicum virgatum [AB], [Au], [ΔAC]

Setaria italica* >5 [Au], 2 [AC]

Agavaceae Yucca filamentosa 2 [AΔX]

Asparagaceae Asparagus officinalis 8 [AΔX]

Hyacinthaceae Charybdis maritime [AΔX]

Drimiopsis kirkii [AΔX]

Hyacinthus orientalis [AΔX]

Muscari armeniacum [AΔX]

Iridaceae Iris brevicaulis 3 [AΔB]

Iris fulva 2 [AΔB]

Iris hollandica 3 [AΔB], 2 [AB]

Ruscaceae Ophiogon japonicus [AΔX]

Polygonatum multiflorum 2 [AB]

*Completed genomes

#BEP clade (Bambusoideae, Ehrhartoideae and Pooideae); PACCAD clade (Panicoideae, Aristidoideae, Centothecoideae,

Chloridoideae, Arundinoideae, Danthonioideae)


Table 2.2 Detailed overview of the documented taxonomic distribution of the different types of RIPs within the Eudicotyledons

of the Angiosperms (flowering plants). Taxa are ordered according to the dendrogram shown in Figure 2.3.

EUDICOTYLEDONS

RANUNCULALES Ranunculaceae Actaea racemosa 3 [AΔB]

Adonis aestivalis [AB]

Aquilegia coerulea* [AB]

Eranthis hyemalis [AB]

Menispermaceae Cissampelos mucronata [AB]

C

O

R

E

E

U

D

I

C

O

T

Y

L

E

D

O

N

S

R

O

S

I

D

S

F

A

B

I

D

S

Cucurbitales Cucurbitaceae Bryonia dioica [AΔB], [AB]

Citrullus lanatus* 2 [AΔB], 2 [AB]pseudo

Cucumis figarei [AΔB]

Cucumis melo* 3 [AB], 3 [ΔAB]

Cucumis sativus* 2 [AB], 3 [ΔAB]

Gynostemma pentaphyllum 4 [AΔB]

Luffa cylindrica 2 [AΔB]

Momordica balsemina [AΔB]

Momordica charantia 2 [AΔB] [AB]

Siraitia grosvenorii ≥3 [AΔB], [ΔAB] or [AB]?

Trichosanthes cordata 2 [AB]

Trichosanthes cucumerina [AΔB]

Trichosanthes dioica [AB]

Trichosanthes kirilowii 4 [AΔB], [AB]

Fagales Fagaceae Fagus sylvatica 8 [AΔX]

Rosales Cannabaceae Cannabis sativa* 7 [AΔB], 2 [AΔX], [AX]

Humulus lupulus [AΔX]

Rosaceae Malus domestica* 3 [AΔB], [ΔAB]

Prunus armeniaca ≥1 [AΔB]

Prunus mume ≥ 2[AΔB]

Prunus persica* 4 [AΔB]

Fabales Fabaceae Abrus precatorius 5 [AB]

Abrus pulchellus 4 [AB]

Acacia mangium [AX]

Malpigiales Euphorbiaceae Euphorbia esula 9 [AΔB], 5 [ΔAB]

Euphorbia characias [ΔAB]

Euphorbia serrata 2 [AΔB]

Gelonium multiflorum [AΔB]

Jatropha curcas* 7 [AΔB]

Manihot esculenta* 4 [AΔB]

Ricinus communis* 7 [AΔB], 8 [AB]

Vernicia fordi [AΔB]

Passifloraceae Adenia volkensii [AB]

Saliceae Populus trichocarpa* [AΔX], 3 [ΔAB]

M

A

L

V

I

D

S

Malvales Malvaceae Gossypium hirsutum [AB]

Gossypium raimondii* [AB]

Theobroma cacao* 2 [AΔX] (or truncated [AX]), 3

[AX], [ΔAX]

Sapindales Sapindaceae Paullinia cupana [AB]


A

S

T

E

R

I

D

S

Santalales Olaceae Ximenia americana [AB]

Santalaceae Viscum album 10 [AB]

Viscum articulatum [AB]

Caryophyllales Aizoaceae Mesembryanthemum crystallinum

5 [AΔB]

Amaranthaceae Amaranthus tricolor 2 [AΔB]

Amaranthus viridis 4 [AΔB]

Atriplex patens [AΔB]

Beta vulgaris 6 [AΔB], [AΔX]

Celosia cristata [AΔB]

Chenopodium album [AΔB]

Chenopodium quinoa [AΔB]

Spinacia oleracea 2 [AΔB]

Caryophyllaceae Dianthus caryophyllus [AΔB]

Dianthis chinensis 3 [AΔB]

Saponaria officinalis 4 [AΔB]

Silene latifolia 5 [AΔB]

Stellaria media [AΔB]

Nyctaginaceae Bougainvillea x buttiana 3 [AΔB]

Bougainvillea spectabilis 3 [AΔB]

Mirabilis jalapa [AΔB]

Mirabilis expansa [AΔB]

Phytolaccaceae Phytolacca americana 6 [AΔB]

Phytolacca acinosa [AΔB]

Phytolacca heterotepala [AΔB]

Phytolacca insularis 2 [AΔB]

Phytolacca octandra [AΔB]

Cornales Hydrangaceae Hydrangea macrophylla [AB]

Ericales Actinidiaceae Actinidia deliciosa [AΔB], [AB]

Theaceae Camellia sinensis [AΔB], 2 or 3 [AB], 3 [ΔAX]

Ebenaceae Diospyros kaki [AΔB] or [AB]?

Polemoniaceae Ipomopsis aggregata [AB]

L

A

M

I

I

D

S

Lamiaceae Clerodendron inerme [AΔB]

Clerodendron aculeatum [AΔB]

C

A

M

P

A

N

U

L

I

D

S

Asterales Asteraceae Artemisia annua [AB]

Centaurea maculosa [AB]

Centaurea solstitialis [AB]

Chrysanthemum x morifolium [AB]

Helianthus tuberosus [AB]

Parthenium argentatum [AB]

Dipsacales Adoxaceae Sambucus nigra Multiple [AB] and [ΔAB]

Sambucus ebulus Multiple [AB] and [ΔAB]

Sambucus sieboldiana Multiple [AB] and [ΔAB]

Apiales Araliaceae Panax ginseng [AB], [AX], [ΔAX]

Apiaceae Bupleurum chinense Most probably [AΔB]

*Completed genomes

Table 2.2 (Continued )


canonical AB5 molecular structure. Basically, an A subunit (equivalent to a RIP domain) with catalytic activity forms through non-covalent interactions a complex with a so-called B subunit, which itself is a pentamer of five identical smaller polypeptides (Figure 2.5) that interact with specific glycolipids like globotriaosylceramide (Gb3).2 Though the term AB5, as well as the under-lying overall structure of the Shiga toxins, might be reminiscent of the canonical AB structure of ricin and other type 2 RIPs from plants, there are three fundamental differences. First, there is no evolutionary relationship between the B chain of ricin and the polypeptide constituting the B sub-unit of the Shiga toxin. Second, the AB5 structure of the Shiga toxin results from non-covalent interactions between the A and B subunits, whereas in ricin the A and B chains are covalently linked

Genes with a Shiga-toxin A domain

Genes with a Shiga-toxin A domain only

Genes with a domain distantly to the Shiga-toxin A domain

Genes with an A domain resembling plant type A RIP

RIP [As]-U: dicistronicNon-secreted

Non-secreted

RIP [A]

RIP [AM]Monocistronic

Signal

peptide

RIP [As]-P-W: tricistronic

Genuine Shiga-toxin genes (discistronic)

RIP [As]-Bs

Shiga and

shiga-like toxins

Protein homolog

of Stx A subunit

Enterobacter cloacaeCitrobacter freundiiEscherichia coliShigella sonneiShigella dysenteriae

Rickettsiella grylli

Burkholderia ambifaria

Flavobacterium columnare

Streptomyces coelicolor

Streptomyces lividans TK24Streptomyces scabieiStreptomyces somaliensis

Streptomyces lysosuperificusStreptomyces sp. Mg1

Streptomyces secreted

type A RIP

Streptomyces non-secreted

type A RIP

Micromonospora sp. ATCC 39149Chimeric protein

Shiga toxin A or related domain

Domain resembling plant RIP

Unknown domains

Part of Pasteurella ToxA domain

Brukholderia sp. CCGE1002

RIP [As]

Figure 2.5 Schematic overview of the domain architectures identified in bacterial RIP genes.


through an interchain disulfide bond. Third, the A and B chains of ricin are synthesized on a single precursor polypeptide that undergoes a complex post-translational processing, whereas the A and B chains of the Shiga toxins are synthesized as two separate polypeptides on a single discistronic mRNA.

Genuine Shiga and Shiga-like toxins and/or their genes have been identified in five different bac-terial species, all of which are classified in the family Enterobacteriaceae (Gammaproteobacteria; Table 2.3). The genes encoding these toxins are presumably part of the genome of lambdoid prophages.

Protein Homologs of Stx A SubunitA (monocistronic) gene was identified in the genome of Rickettsiella grylli that resembles the gen-uine Stx genes but lacks the second cistron encoding the B polypeptide (Figure 2.5). Since no other gene occurs in the genome with a sequence encoding a B-polypeptide homolog, one can reasonably expect that Rickettsiella grylli does not produce a Shiga-like toxin. In principle Rickettsiella grylli can express a protein consisting only of an A chain. Alternatively, an as yet unidentified (toxic?) protein composed of an A chain homolog and another non-covalently bound subunit might be synthesized.

A similar gene occurs in the Betaproteobacterium Burkholderia sp. CCGE1002, but in this species the RIP gene is located on plasmid pBC201. In this case also the nature and structure of the presumed RIP remain unclear.

Table 2.3 Overview of documented domain architectures of RIP genes.*

PLANTS

[AΔB]** Type [A] derived from type [AB] RIP (most classical type 1 RIPs)

[AΔx] Type [A] derived from type [AX] RIP (some classical type 1 RIPs; e.g. muscarin)

[Au] Type [A] from undefined origin (e.g. maize RIP b-32; described as type 3 RIP)

[AB] Chimer with C-terminal ricin-B domain (type 2 RIPs)

[AC] Chimer with C-terminal (unknown) C domain (JIP 60; type 3 RIP)

[AD] Chimer with C-terminal (unknown) D domain

[AX] Chimer with C-terminal (unknown) X domain

[APM41] Chimer with C-terminal peptidase M41domain

[APC19] Chimer with C-terminal peptidase C19 domain

FUNGI[AΔxf] Fungal homolog of plant [AΔx] RIP

[AXf] Fungal homolog of chimeric plant [AX] RIP

INSECTS[AI] Chimer with C-terminal (unknown) I domain

BACTERIA[A] Bacterial homolog of plant type [A] RIP (Streptomyces coelicolor RIPsc)

[As] Type [A] equivalent to Shiga toxin A-subunit

[As]-[Bs]° Shiga and Shiga-like toxins

[As]-[U]° Chimer with unknown C-terminal domain

[As]-[P]-[W]° Chimer with PMT-C3 and unknown C-terminal domain

[AM] Chimer with C-terminal (unknown) M domain

*For a schematic representation see Figures 2.2, 2.5, 2.6, and 2.7

**Domain (s) between square brackets are synthesized on a single mRNA

°Polycistronic gene


Distantly Related Non-secreted Protein Homolog of Stx A Subunit

Proteins/genes distantly related to the Stx A subunit have been identified in two bacterial species. A Burkholderia ambifaria gene encodes a protein that clearly corresponds to a RIP domain and is most probably synthesized on a dicistronic mRNA. The latter fact might indicate that Burkholderia ambifaria synthesizes a novel type of toxin consisting of an A domain and an unknown non- covalently bound subunit(s) (Figure 2.5). The genome of Flavobacterium columnare (Bacteroidetes; Flavobacteriales) also possesses a gene with a RIP domain (Figure 2.5). This RIP domain is synthe-sized most probably on a tri-cistronic mRNA. Interestingly, the protein translated from the second cistron corresponds to the C-terminal domain of Pasteurella multocida ToxA (a bacterial protein toxin that modulates G proteins). This domain belongs to a family of transglutaminase-like proteins, with active site Cys–His–Asp catalytic triads.16

Unlike the proteins belonging to the Stx group described above, the Burkholderia ambifaria and Flavobacterium columnare RIPs are synthesized without a signal peptide and accordingly might be retained within the cytoplasm of the bacterial cells.

RIPs in Actinobacteria

Streptomyces Coelicolor type [A] RIPThe production and biochemical characterization of the gene product SCO7092 (RIPsc) from the Gram-positive soil bacterium Streptomyces coelicolor expressed in E. coli demonstrated for the first time the occurrence of a bacterial RIP other than the classical Stx. Moreover, sequence comparisons revealed that RIPsc was more closely related evolutionarily to some plant type [A] RIPs than to the Stx A subunit (Figure 2.5).17 Besides S. coelicolor a nearly identical protein occurs in S. lividans, whereas similar proteins (46% and 41% sequence similarity, respectively) are found in S. scabei and S. somaliensis. All these RIPs are synthesized with a signal peptide and hence are likely to be secreted.

Other Streptomyces RIPsRIP sequences were retrieved in two additional Streptomyces species. The genome of Streptomyces lysosuperificus possesses a single gene encoding a protein with a RIP domain. Two similar genes located on different loci occur in the genome of Streptomyces sp. Mg1. All three (putative) RIPs are synthesized without a signal peptide and hence might be retained in the cytoplasm (Figure 2.5).

Within the Actinobacteria a single RIP gene was identified outside the genus Streptomyces. Micromonospora sp. ATCC 39149 (Actinomycetales; Micromonosporaceae) possesses a gene encoding a chimeric RIP comprising an N-terminal A domain fused to a C-terminal domain of approximately 250 amino acid residues that hitherto has not been found in any known protein (Figure 2.5). This novel chi-meric protein will further be referred to as a type [AM] RIP (A fused to unknown C-terminal domain).

Fungal RIPs

Previously Described Fungal Ribosome-Inactivating Proteins: No Evidence for the Presence of an N-glycosidase DomainSeveral reports claim the isolation and characterization of RIPs from fungi.9–11 However, until now no evidence has been presented that any of these presumed RIPs comprises a domain that is struc-turally and evolutionarily related to the A chains of either bacterial Stx or plant RIPs.


One group of fungal proteins/genes that is often referred to as RIPs belongs to the family of fungal ribotoxins. These ribotoxins are a family of extracellular ribonucleases that inactivate ribo-somes by specifically cleaving a single phosphodiester bond located at the universally conserved sarcin/ricin loop of the large rRNA.12 Though ribotoxins inactivate ribosomes, both their catalytic activity and structure fundamentally differ from those of the RIPs with an N-glycosidase domain, and hence fall beyond the scope of this review.

Besides ribotoxin homologs, a number of partly characterized proteins were described as fungal RIPs. Most of these proteins were isolated from mushrooms (e.g., Calvatia caelata, Flammulina velutipes, Hypsizigus marmoreus, Lyophyllum shimeji, and Pleurotus tuber-regium). Though the authors ascribe a protein synthesis inhibitory activity to all these proteins, they did not provide sufficient sequence information for formal identification. Evidently, this obvious lack of sequence information – in an era where scores of fungal genomes have been completely sequenced – can no longer be justified, and accordingly none of these “novel “ or “new” fungal RIPs can for the time being be considered members of the N-glycosidase family, and hence will not be further discussed here.

In Silico Analyses Provide Evidence for the Presence of Genes with an N-glycosidase Domain in a Few FungiScreening of genome and transcriptome databases using some recently identified plant RIP genes as a query revealed the occurrence of two different types of genes with an N-glycosidase domain. A first type referred to as [AΔXf], that occurs in the genome of the nematode trapping fungus Arthrobotrys oligospora (Ascomycota) encodes a protein that consists of a single RIP domain and most probably can be considered a homolog of plant type [AΔX] RIP (Figure 2.6). No homologs/orthologs of the A. oligospora RIP gene could be retrieved from any other sequenced fungal genome or transcriptome. The second type of fungal RIP gene encodes a chimeric protein that, based on its sequence, can be considered a fungal homolog of the plant [AX] chimers and accord-ingly will be referred to as [AXf]. Type [AXf] genes were identified in eight Ascomycota species, all of which belong to the Sordariomycetes (Table 2.4). Pseudogenes lacking an ORF due to the occurrence of frame shifts and/or stop codons were retrieved in two additional species of the same group. Two important conclusions can be drawn from Table 2.4. First, RIP genes are confined to a few fungal taxonomic groups. Second, the absence of a Magnaporthe poae [AXf] homolog from the genomes of M. grisea and M. oryzeae illustrates that even within the very same genus the [AXf] gene is not retained.

RIP [Axf]

RIP [AXf]

[Axf] domain [Xf] domain

Cytoplasmic

Arthrobotrys oligospora

Epichloe sp.

Cordyceps sp.

Neotyphodium gansuenseMagnaporthe oryzeae

Vacuolar

Figure 2.6 Schematic overview of the domain architectures identified in fungal RIP genes.


Insect RIPs

Hitherto no RIP has been isolated from any animal species. However, in silico analyses of genome and transcriptome databases revealed that two mosquito species express genes encoding chimeric proteins comprising an N-terminal RIP domain fused to an unidentified C-terminal domain of approximately 200 amino acid residues. To distinguish this novel type of chimer from previously described RIPs it will further be referred to as the [AI] type, built up of an A domain and an “Insect specific” C-terminal domain (Figure 2.7). A single type [AI] RIP gene occurs in the genome of Culex quinquefasciatus (southern house mosquito), whereas three paralogs are found in Aedes aegypti (yellow fever mosquito). The presence of perfectly matching (partial) EST sequences indicates that all four insect RIP genes are expressed.

Table 2.4 Summary of the occurrence and domain architecture of RIP genes in Bacteria.

P

R

O

T

E

O

B

A

C

T

E

R

A

Gamma-

proteobacteria

Enterobacteriales;

Enterobacteriaceae

Enterobacter cloacaeCitrobacter freundiiEscherichia coliShigella sonneiShigella dysenteriae

[As][Bs]*°

[As][Bs]

2 [As][Bs]

[As][Bs]

[As][Bs]

Legionellales;

Coxiellaceae

Rickettsiella grylli [As]

Beta-

proteobacteria

Burkholderiales;

Burkholderiaceae

Burkholderia ambifaria MC40-6

Burkholderia sp CCGE1002

[As][U]

[As]

A

C

T

I

N

O

B

A

C

T

E

R

I

A

Actinobacteridae;

Actinomycetales

Micromonosporineae;

Micromonosporaceae

Micromonospora sp. ATCC 39149 [AM]

Streptomycineae;

Streptomycetaceae

Streptomyces coelicolor A3 [A]

Streptomyces lividans TK24 [A]

Streptomyces lysosuperificus [A]

Streptomyces scabiei [A]

Streptomyces somaliensis [A]

Streptomyces sp. Mg1 [A]

Bactero-idetes Flavobacteriia;

Flavobacteriales

Flavobacteriaceae Flavobacterium columnare [As][P] [W]

*Domain(s) between square brackets correspond to a single cistron. A sequence of two or three single domains between square

brackets means that they correspond to a single polycistronic gene.

°[As] and [Bs]: A and B domains identical or related to the A and B domains, respectively, of the Shiga toxins

[U]; [W]: undefined domains

RIP [AI] Culex quinquefasciatus

aedes aegypti

Figure 2.7 Schematic overview of the domain architectures identified in insect RIP genes.


Conclusions

A renewed screening of the databases allowed the updating of the overall taxonomic distribution of the RIP domain and the domain architecture of the corresponding genes. In silico analyses revealed that the RIP domain is more widespread than was previously inferred from the classical biochemical and molecular approach, and has to be extended to fungi and insects. In addition, the identification of several novel chimeric forms implies that the heterogeneity of RIP genes in terms of domain architecture is no longer covered by the classical Shiga toxins and type 1, type 2, and type 3 plant RIPs, and argues for a novel classification system. Therefore one has to take into account that, as can be expected from the preliminary analyses of the RIP gene families in grass species, the present picture is still incomplete.

Acknowledgments

This research was supported by the Research Council of Ghent University and the Fund for Scientific Research (FWO-Vlaanderen, Brussels, Belgium). Chenjing Shang acknowledges the receipt of a CSC Grant from the Chinese Government.

References

1. Stirpe F, Battelli MG. Ribosome-inactivating proteins: progress and problems. Cell Mol Life Sci. 2006;63:1850–1866.

2. Johannes L, Römer W. Shiga toxins–from cell biology to biomedical applications. Nat Rev Microbiol. 2010;8:105–116.

3. Barbieri L, Battelli MG, Stirpe F. Ribosome-inactivating proteins from plants. Biochim Biophys Acta. 1993;1154:237–282.

4. Nielsen K, Boston RS. Ribosome-inactivating proteins: A plant perspective. Ann Rev Plant Physiolg. 2001;52:785–816.

5. Van Damme EJM, Hao Q, Chen Y, et al. Ribosome-inactivating proteins: A family of plant proteins that do more than inac-

tivate ribosomes. Crit Rev Plant Sci. 2001;20:395–465.

6. Walsh TA, Morgan AE, Hey TD. Characterization and molecular cloning of a proenzyme form of a ribosome-inactivating

protein from maize. Novel mechanism of proenzyme activation by proteolytic removal of a 2.8-kilodalton internal peptide

segment. J Biol Chem. 1991;266:23422–23427.

7. Mehta AD, Boston RS. Ribosome-inactivating proteins. In: Bailey-Serres J, Gallie DR, (eds.) A Look Beyond Transcription:

Mechanisms Determining mRNA Stability and Translation in Plants. Rockville, MD: American Society of Plant Physiologists,

1998:145–152.

8. Chaudhry B, Müller-Uri F, Cameron-Mills V, et al. The barley 60 kDa jasmonate-induced protein (JIP60) is a novel

ribosome-inactivating protein. Plant J. 1994;6:815–824.

9. Girbés T, Ferreras JM, Arias FJ, Stirpe F. Description, distribution, activity and phylogenetic relationship of ribosome-

inactivating proteins in plants, fungi and bacteria. Mini Reviews in Medicinal Chemistry 2004;4:461–476.

10. Reyes AG, Anné J, Meija A. Ribosome-inactivating proteins with an emphasis on bacterial RIPs and their potential medical

applications. Future Microbiol. 2012;7:705–717.

11. Stirpe F. Ribosome-inactivating proteins. Toxicon. 2004;44:371–383.

12. Lacadena J, Alvarez-García E, Carreras-Sangrà N, et al. Fungal ribotoxins: molecular dissection of a family of natural killers.

FEMS Microbiol Rev. 2007;31:212–237.

13. Peumans WJ, Van Damme EJM. Evolution of plant ribosome-inactivating proteins. In: Lord JM, Hartley MR, (eds.) Toxic Plant Proteins Plant Cell Monographs Vol. 18. Berlin: Springer, 2010:1–26.

14. Jiang SY, Ramamoorthy R, Bhalla R, et al. Genome-wide survey of the RIP domain family in Oryza sativa and their expres-

sion profiles under various abiotic and biotic stresses. Plant Mol Biol. 2008;67:603–614.

15. Palmer JD, Soltis DE, Chase MW. The plant tree of life: an overview and some points of view. Am J Bot. 2004;91:1437–1445.

16. Wilson BA, Ho M. Recent insights into Pasteurella multocida toxin and other G-protein-modulating bacterial toxins. Future Microbiol. 2010;5:1185–1201.

17. Reyes AG, Geukens N, Gutschoven P, et al. Streptomyces coelicolor encodes a type I ribosome-inactivating protein. Microbiol.

2010;156:3021–3030.

Documents

Ribosome-inactivating Proteins (Ricin and Related Proteins) || Occurrence and Taxonomical Distribution of Ribosome-inactivating Proteins Belonging to the Ricin/Shiga Toxin Superfamily