| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Laboratory of Phytopathology and Plant Protection, Katholieke Universiteit Leuven, 3001 Leuven, Belgium
1Correspondence: Laboratory of Phytopathology and Plant Protection, Katholieke Universiteit Leuven, Willem de Croylaan 42, 3001 Leuven, Belgium. E-mail: Els.VanDamme{at}agr.kuleuven.ac.be
 |
ABSTRACT |
|---|
 TOP ABSTRACT GENERAL REFLECTIONS ON RIBOSOME...STRUCTUREBIOLOGICAL AND ENZYMATIC...RIPS AS PLANT DEFENSE...CONCLUSIONSREFERENCES |
|---|
Key Words: antiviral activity • DNA glycosylase/AP lyase • polynucleotide:adenosine glycosidase
|
GENERAL REFLECTIONS ON RIBOSOME-INACTIVATING PROTEINS |
|---|
|
TOPABSTRACTGENERAL REFLECTIONS ON RIBOSOME...STRUCTUREBIOLOGICAL AND ENZYMATIC...RIPS AS PLANT DEFENSE...CONCLUSIONSREFERENCES |
|---|
. The
identification of ricin was an important milestone in biochemistry
because for the first time a well-defined biological activity was
ascribed to a plant protein. Moreover, abrin, a similar toxic protein
from the seeds of A. precatorius, also played an important
role in the early development of immunology.
Since the isolation and characterization of ricin, many structurally
and functionally related proteins have been identified in a wide
variety of plants (for review, see ref 2
). For a long time
the interest in RIPs focused on possible medical and therapeutical
applications because several of these proteins were found to be more
toxic to tumor cells than to normal cells, and hence offered a
theoretical opportunity to develop antitumor drugs that selectively
target tumor cells (3)
. Along with the efforts to develop
some RIPs into anticancer compounds, attempts were made to find out
what RIPs do and how they act. This led to the finding that RIPs are
RNA N-glycosidases that inactivate ribosomes through a site-specific
deadenylation of the large ribosomal RNA (4
, 5)
. Later it
became evident that RIPs are also capable of inactivating many
nonribosomal nucleic acid substrates (6
7
8)
and hence can
be considered polynucleotide:adenosine glycosidases (7)
.
These novel insights renewed the interest in RIPs because understanding
the enzymatic activity enhances exploitation of the unique properties
and activities of RIPs for diverse applications like immunotoxins
(9)
, abortifacients (10)
, and anti-human
immunodeficiency virus (HIV) agents (11
, 12)
. Research in
this area soon led to unexpected findings that could not easily be
explained on the basis of the ribosome-inactivating or
polynucleotide:adenosine glycosidase activity of RIPs. In combination
with the discovery of putative novel enzymatic activities of some RIPs,
this gave rise to the concept that RIPs act not only through their
classical ribosome-inactivating or polynucleotide:adenosine glycosidase
activity, but also (and in some instances, even primarily) through DNA
or RNA lyase activities. Though straightforward at first sight, the
introduction of the concept of RIPs as DNA/RNA lyases created a major
dispute because the presumed novel enzymatic activities have been and
are being contested. This review aims to provide a critical assessment
of the enzymatic activities of RIPs and the consequences thereof on
their biological activities ex planta and in planta.
Definition and classification
The term ‘ribosome-inactivating protein’ was introduced to
designate plant proteins that inactivate animal ribosomes before the
structure and enzymatic activity of the so-called RIPs were known.
After the mode of action of RIPs on ribosomes was elucidated, the term
RIP was used exclusively for these N-glycosidases. This is important
because other types of proteins that inactivate or damage ribosomes by
other mechanisms (e.g., RNases or proteases) are not considered RIPs.
For the time being, the term RIP is reserved for proteins containing an
RNA N-glycosidase domain that is structurally related to the classical
RIPs. Two major classes are distinguished: holo-RIPs and chimero-RIPs
(Fig. 1
). Holo-RIPs consist exclusively of a single RNA N-glycosidase domain.
Most holo-RIPs consist of a single, intact polypeptide of 
30 kDa and
are usually referred to as type 1 RIPs. Besides these classical type 1
RIPs, there are also a few examples of type 1 RIPs in which the
original 30 kDa polypeptide is proteolytically processed so that the
protein consists of two shorter polypeptides held together by
noncovalent interactions. These proteins have been named type 3 RIPs
(13)
, but we propose to use the term two-chain type 1 RIPs
because the term type 3 RIPs has also been introduced for RIPs with a
totally different structure. Chimero-RIPs are constructed of one or
more protomers consisting of an N-glycosidase domain linked to a
structurally different and functionally unrelated domain. Most
chimero-RIPs are so-called type 2 RIPs. The protomers of type 2 RIPs
consist of an amino-terminal domain comparable to type 1 RIPs linked by
a disulfide bridge to an unrelated carboxyl-terminal domain with
carbohydrate binding activity. Both domains are synthesized on a single
precursor that is post-translationally processed by the excision of a
linker sequence between the two domains. For historical reasons, the
amino-terminal RNA N-glycosidase domain and the carboxyl-terminal sugar
binding domain are referred to as the A- and B-chains, respectively.
Besides the classical type 2 RIPs, a 60 kDa RIP (called JIP60) has been
identified in barley (Hordeum vulgare) that consists of an
amino-terminal domain resembling type 1 RIPs linked to an unrelated
carboxyl-terminal domain with unknown function (14)
. JIP60
has been called a type 3 RIP. To avoid confusion, we reserve the ‘term
type 3 RIP’ for proteins that are structurally and evolutionary
related to JIP60.
|
|
STRUCTURE |
|---|
|
TOPABSTRACTGENERAL REFLECTIONS ON RIBOSOME...STRUCTUREBIOLOGICAL AND ENZYMATIC...RIPS AS PLANT DEFENSE...CONCLUSIONSREFERENCES |
|---|
15 type 2 RIPs have been sequenced
and/or cloned. A comparison of these sequences reveals striking
similarities between the type 1 RIPs and the A-chains of type 2 RIPs,
as well as between the B-chains of different type 2 RIPs. A close
examination indicates that the sequence similarity between the
amino-terminal and core sequences of RIPs is much higher than that of
the carboxyl-terminal sequences. It has been suggested that this
difference in degree of conservation explains why some core activities
are conserved among all RIPs whereas other activities (presumably
associated with the highly variable carboxyl-terminal sequences) are
not (15)
.
Three-dimensional structure: toward an understanding of multiple
functions of RIPs
The resolution of the crystal structure of ricin by Montfort et
al. (16)
yielded the first 3-dimensional structure of a
RIP. Ricin is a globular, glycosylated heterodimer joined by a single
disulfide bond. Approximately 50% of the A-chain consists of
-helices and ß-sheets. Three individual domains are distinguished
within this structure. A cleft at the interface of the three domains
forms the active site for RNA N-glycosidase activity. The B-chain of
ricin is a bilobal structure composed of two homologous domains. Each
domain contains one galactose binding site lying in a pocket formed in
part by a kink in the polypeptide chain by the tripeptide Asp-Val-Arg.
The crystal structures of at least seven type 1 and three type 2 RIPs
have been reported. In general, the tertiary structures of the
different RIPs are well conserved, as demonstrated by the fact that the
-carbon traces of most RIPs are virtually superimposable. There are,
however, some major differences, especially in the carboxyl-terminal
region and surface loop structures. These differences are believed to
account for the differences in activity and substrate specificity of
the different RIPs.
Due to the discovery of novel putative enzymatic activities of RIPs, intensive efforts are under way to find structural evidence of their multiple functions. However, in the absence of a definitive proof for enzymatic activities other than RNA N-glycosidase and polynucleotide:adenosine glycosidase activity, this issue remains controversial.
|
BIOLOGICAL AND ENZYMATIC ACTIVITIES |
|---|
|
TOPABSTRACTGENERAL REFLECTIONS ON RIBOSOME...STRUCTUREBIOLOGICAL AND ENZYMATIC...RIPS AS PLANT DEFENSE...CONCLUSIONSREFERENCES |
|---|
, 18)
. Though the B-chains of different type 2 RIPs
share high sequence similarity and virtually identical 3-dimensional
structures, there are pronounced differences in sugar binding
specificity. These differences in lectin activity and specificity are
important because the toxicity and cytotoxicity of type 2 RIPs is
(partly) determined by the binding of the B-chain to a sugar-containing
receptor on the cell surface. Due to the extreme toxicity of ricin and
abrin, type 2 RIPs are usually associated with highly toxic proteins
(2)
. However, type 2 RIPs show marked differences in
(cyto)toxicity. Ricin, for example, causes 50% cell death at
concentrations below 1 ng/ml whereas some elderberry type 2 RIPs show
no effect at 1 mg/ml (19)
.
Type 1 RIPs were discovered in 1925 when Duggar and Armstrong
(20)
observed that the so-called Phytolacca
americana antiviral protein (PAP) inhibits the transmission of
tobacco mosaic virus (TMV) in plants. However, only in 1978 was PAP
recognized as an inhibitor of protein synthesis (21)
.
Many, but certainly not all, type 1 RIPs are antiviral proteins. Unlike
some type 2 RIPs, type 1 RIPs are not cytotoxic and do not behave as
toxins because they are not able to cross the cell membrane on their
own. Some specialized animal cells, however, can import type 1 RIPs by
endocytosis and subsequently become sensitive to the RIP activity. For
example, the abortifacient effect of trichosanthin is ascribed to the
active uptake of the RIPs by trophoblasts (22)
.
The molecular basis for the biological activities remained unclear
until the inhibitory activity of type 1 and type 2 RIPs on protein
synthesis was discovered. Once it turned out that the so-called
single-chain protein synthesis inhibitors share a substantial sequence
similarity with the A-chain of ricin, the first functional link between
type 1 and type 2 RIPs became obvious and the search for a common
working mechanism started. This search soon revealed that ricin, abrin,
and PAP inhibit cell-free protein synthesis by irreversibly
inactivating the ribosomes in such a way that the function of
elongation factors EF-1 and EF-2 is blocked (2
, 23)
.
Though the target of RIPs was identified, the question remained as to
how RIPs inactivate ribosomes.
Enzymatic activities
It is now generally accepted that all RIPs are enzymes and that
some have multiple enzymatic activities. This section gives a brief
review of the classical enzymatic activities of RIPs and addresses the
actual dispute about the presumed novel enzymatic activities.
Classical enzymatic activities
Site-specific RNA N-glycosidase activity toward ribosomes
Endo and co-workers (4)
demonstrated for the first
time that RIPs are enzymes. Ricin recognizes a highly conserved region
in the large 28S rRNA and cleaves a specific N-C glycosidic bond
between an adenine and the nucleotide on the RNA whereby the adenine
residue is removed (Fig. 2
). Due to the removal of this adenine, the deadenylated (or abasic) site
becomes unstable and a ß elimination reaction can occur after the RNA
is treated with acidic aniline, whereby the 3'-end of the rRNA is
cleaved and can be detected by electrophoresis. For the most often used
substrate—rat liver ribosome—this specific site is
A4324 in 28S rRNA. This site is usually depicted
as being present in a single-stranded loop, called the sarcin/ricin
loop. It is located in domain VII some 400 nucleotides from the 3' end
of the rRNA. Subsequent work revealed that this particular
site-specific RNA N-glycosidase activity is a common property of all
identified type 1 and type 2 RIPs.
|
Although all RIPs exhibit RNA N-glycosidase activity toward ribosomes,
there are marked differences in substrate specificity. For example,
ricin is highly active toward mammalian and yeast ribosomes but poorly
active or even inactive on plant and Escherichia coli
ribosomes (2)
. In contrast, PAP depurinates ribosomes from
plants, bacteria, yeasts, and lower and higher animals. Most type 1
RIPs have a rather broad specificity whereas type 2 RIPs have a
preference for animal ribosomes. Both RIPs and ribosomes contribute to
the apparent substrate specificity. Since the rRNA target structure is
universally conserved, differences in sensitivity between ribosomes
most likely reside within the ribosomal proteins, which may either
allow or prevent access of the RIPs to the sarcin/ricin loop. Vater et
al. (24)
identified rat liver ribosomal proteins L9 and
L10e as the binding target of the ricin A-chain, whereas yeast
ribosomal protein L3 was identified as the binding factor of PAP
(25)
. The specific interaction between PAP and L3 probably
explains the broad-spectrum activity of PAP toward ribosomes from
species of different taxonomic groups because L3 is highly conserved in
ribosomes. Differences in activity and ribosome substrate specificity
are also due to differences in the structure of different RIPs. This
was demonstrated by an approach in which PAP-ricin A-chain protein
hybrids were created and examined for activity on rabbit reticulocyte
and E. coli ribosomes. According to the results of these
experiments, the amino-terminal half of the hybrid proteins determines
the substrate specificity. Structurally dissimilar surface polypeptide
loops apparently do not play a role (26)
.
An important question with regard to the substrate specificity is whether RIPs act on conspecific ribosomes. Although it has been proposed that RIPs do not affect ribosomes of the plant in which they are expressed, recent studies have unambiguously demonstrated that RIPs are fully capable of deadenylating and inactivating conspecific ribosomes.
Site-specific RNA N-glycosidase activity toward ribosomal RNA
Although the rRNA in native ribosomes is the preferred substrate
for RIPs, naked (deproteinized) rRNA and a synthetic 35-residue
oligoribonucleotide that mimics the sarcin/ricin loop, can also serve
as the substrate for RIP activity (27)
. Most probably all
RIPs are capable of deadenylating the same target adenine from naked
rRNA as from native ribosomes. However, many and perhaps all RIPs
deadenylate naked rRNA at multiple sites. Moreover, some RIPs are
capable of deadenylating naked rRNA from nonsubstrate ribosomes. For
example, ricin A-chain is able to act on naked E. coli 23S
rRNA, but not the intact E. coli ribosomes. Two important
conclusions can be drawn from the apparent differences in activity and
substrate specificity of RIPs toward ribosomes and naked rRNA. First,
given the huge difference in activity on native ribosomes and naked
rRNA, it seems unlikely that rRNA can be considered a physiologically
relevant substrate. Second, the apparent loss of specificity of some
RIPs toward ribosomes from different origin confirms the role of the
ribosome in the determination of the substrate specificity of the RIPs.
Polynucleotide:adenosine glycosidase activity
RIPs were originally thought to act exclusively on ribosomes or
rRNA. Attempts to identify other substrates were hampered by the lack
of a sensitive method to detect possible reaction products other than
the Endo fragments. A breakthrough in methodology was the development
of a high-performance liquid chromatography (HPLC) fluorescence based
method to quantify the amount of free adenine released from various
substrates by RIPs (28)
. This novel method not only
offered a direct method to measure deadenylation of ribosomes, but also
led to the discovery of unexpected activities of RIPs on other
substrates. Several RIPs were found to act on ribosomal RNA at multiple
sites (29)
. Moreover, it was demonstrated that saporin-L1,
a type 1 RIP from the leaves of Saponaria officinalis, is
capable of removing multiple adenine residues from various nucleic acid
substrates such as herring sperm DNA, poly(A), tRNA, and even TMV RNA
(30
, 31)
. Subsequent testing of 52 RIPs further revealed
that all RIPs tested act on herring sperm DNA and poly(A), and roughly
one-third of the RIPs tested also act on TMV RNA (7)
.
In the first experiments, herring sperm DNA was used as a substrate to
determine the polynucleotide:adenosine glycosidase activity of RIPs.
Later, the range of the substrates was extended to some mammalian
nuclear and mitochondrial DNAs (32)
. No detailed studies
have been made of the polynucleotide:adenosine glycosidase activity of
RIPs on conspecific nucleic acids. Therefore, it is still unclear which
kind of nucleic acids can be modified in vivo. RIPs also strongly
differ from each other as to their polynucleotide:adenosine glycosidase
activity and substrate specificity. So the question is whether all RIPs
have similar polynucleotide:adenosine glycosidase reactivity
toward nucleic acid substrates and require identical or different
conditions (such as pH, temperature, and cofactor requirements) for
optimal activity.
An issue that deserves some special attention is the deadenylation of
viral RNA by RIPs. Many RIPs are potent inhibitors of animal and/or
plant viruses. The mode of action for the antiviral activity of RIPs is
not understood, but there is good evidence that this activity does not
rely solely on the inactivation of ribosomes. As an alternative
mechanism, a direct interaction with viral RNA or DNA has been
proposed. If so, the key question is whether RIPs can use viral nucleic
acids as a substrate. As already mentioned above, about one-third of
the 52 RIPs tested deadenylate TMV RNA (7)
. More recent
studies demonstrated that the pokeweed antiviral proteins PAP-I,
PAP-II, and PAP-III cause a concentration-dependent deadenylation of
genomic HIV-1 RNA, TMV RNA, and bacteriophage MS2 RNA
(33)
. In contrast, the A-chain of ricin did not release
detectable quantities of adenine from the same viral RNAs. These
findings leave no doubt that some RIPs can use viral RNA as substrates,
but also indicate that an active RNA N-glycosidase site is not
sufficient for the recognition and deadenylation of viral RNA.
The obvious capacity of RIPs to deadenylate different polynucleotides
implies that they can be considered polynucleotide:adenosine
glycosidases. Since the polynucleotide:adenosine glycosidase activity
is much broader than ribosome-inactivating activity (which is only a
special case of polynucleotide:adenosine glycosidase activity), Stirpe
and co-workers proposed to replace the term ribosome-inactivating
protein with polynucleotide:adenosine glycosidase (7)
.
There has been no consensus to implement this novel term.
Depurination of capped mRNA: a novel and specific type of
polynucleotide:adenosine glycosidase activity
In a recent paper, Hudak et al. (8)
proposed a novel
mechanism for the inhibition of translation by PAP that is based on a
specific depurination of capped mRNA. Using wild-type PAP and three
different PAP mutants that do not depurinate tobacco or rabbit
ribosomes, the authors show that PAP inhibits the in vitro translation
of brome mosaic virus and potato virus X RNAs without depurinating
ribosomes. PAP and some of its mutants inhibit the translation of
capped (but not uncapped) luciferase transcripts, indicating that these
RIPs can distinguish between capped and uncapped mRNAs. Treatment of
the capped mRNAs with PAP and PAP mutants in the presence of the cap
analog m7pppG prevented their translational inactivation, suggesting
that these RIPs recognize the cap structure on the mRNAs. The direct
effect of wild-type PAP on brome mosaic virus RNA and capped and
uncapped luciferase transcripts was further investigated by checking
the possible depurination of these RNAs. Analysis of the PAP-treated
RNAs revealed that the capped but not the uncapped RNAs were degraded
after treatment with acidic aniline, and hence were depurinated in
vitro. Based on these results, it was concluded that PAP may inhibit
translation by binding to the cap structure and depurinating the RNA
and that depurination of capped viral RNA may be the primary mechanism
for the antiviral activity of PAP.
Although the results presented in this paper are important because they
describe for the first time a possible mechanism to explain the
antiviral activity of PAP, some questions remain. One problem concerns
the activities of PAP and its mutants. Besides wild-type PAP, the
following mutants were used: 1) PAPx, an active site mutant
(E176V); 2) PAPn, a mutant with a substitution (G75D) in the
amino-terminal sequence; and 3) PAPc, a mutant lacking the
carboxyl-terminal 25 amino acid residues. According to the authors,
only wild-type PAP depurinates tobacco and rabbit reticulocyte
ribosomes. This contradicts earlier reports because PAPc was originally
described as an enzymically active, carboxyl-terminal deletion mutant
of PAP with an intact active site and ribosome-inactivating activity
(34)
. A second problem concerns the results of the
depurination activity of wild-type PAP on mRNA. At high concentrations,
PAP apparently exhibited RNase activity (because brome mosaic virus RNA
and capped luciferase transcripts were degraded without treatment with
acidic aniline). A third problem concerns the statement of the authors
that wild-type PAP, and PAPn as well as PAPc (which both are presumably
devoid of ribosome-inactivating activity) have a direct effect on
capped mRNA that results in a significant inhibition of message
translation. In the case of PAP, evidence is presented that this
particular effect can be ascribed to the depurinating activity of PAP.
However, since no such data are shown for PAPn and PAPc, it is unclear
whether these mutants can also depurinate capped mRNA. This is an
important issue because the authors state that "PAPn and PAPc
maintain the ability to depurinate viral RNAs and mRNA’ and assume
that because neither PAPn nor PAPc depurinates ribosomes, the mutations
alter their association with ribosomes (8)
.
Polynucleotide: guanosine glycosidase activity
According to some reports, RIPs are also capable of removing
guanine residues from both eukaryotic and prokaryotic rRNA. By 1987,
Endo and co-workers (4)
had reported that treatment of rat
ribosomes with ricin caused the removal of G4323
from the sarcin/ricin loop. A similar activity has been reported for
PAP. Using a quantitative HPLC technique, Rajamohan et al.
(33)
demonstrated that recombinant PAP released a guanine
residue from E. coli rRNA. Similarly, it has been shown by a
highly sensitive primer extension assay that wild-type PAP removes
G4323 from rabbit reticulocyte, tobacco, and
yeast ribosomes (8)
. Modeling studies confirmed that a
guanine base fits into the active site pocket of PAP without disturbing
the geometry and can be sandwiched between the side groups of
Y72 and Y123 in the same
manner as an adenine base. No evidence has been presented for a
possible deguanylating activity of other RIPs, which implies that such
activity cannot be extrapolated to all RIPs. Moreover, according to the
results of recent experiments with highly purified gelonin, momordin,
PAP-S, and saporin-S6, none of these RIPs is able to remove bases other
than adenine (35)
. Since a highly sensitive HPLC method
was used in these experiments to detect the bases released from DNA and
rRNA and the sensitivity limits of the assay were able to detect the
presence of bases at less than 1/100th of the concentration of the
released adenine, the question of the possible deguanylating activity
of RIPs has to be readdressed.
Novel enzymatic activities
During the last decade, putative evidence has accumulated that
some RIPs possess enzymatic activities other than RNA N-glycosidase and
polynucleotide:adenosine glycosidase activity (summarized in
Table 1
). Most of the novel enzymatic activities are related to a presumed
RNase or DNase activity. Other enzymatic activities reported for
individual RIPs include phosphatase activity on lipids
(36)
, phosphatase activity on nucleotides
(37)
, chitinase activity (38)
, and superoxide
dismutase activity (39)
. Most if not all of these novel
enzymatic activities are still controversial because it cannot be
excluded that the observed activities are due to contaminants. Several
independent reports leave no doubt that presumed pure RIP preparations
indeed contain contaminating nucleases that can be separated from the
RIPs by appropriate techniques (35
, 40
, 41)
.
|
RNase activity
Several papers report a presumed RNase activity of RIPs. For
example, Mock et al. (42)
demonstrated that preparations
of - and especially ß-momorcharin (corresponding to momordin I and
momordin II, respectively) possess a poly(U)-specific RNase activity.
However, it has been shown recently that RIP and RNase activity can be
separated by red Sepharose chromatography of an impure momordin II
preparation, yielding momordin II, which is essentially free of RNase
activity (41)
. The possible RNase activity of RIPs is
still of interest. Recently, Hudak et al. (8)
observed
that PAP degraded capped luciferase transcripts and concluded that, at
high concentrations, PAP will degrade RNA. To verify whether the PAP
preparation was free of contaminating nucleases, the authors performed
an RNA activity gel assay. Since all RNase activity comigrated in a
single band with PAP, it was concluded that the observed degradation of
RNA was not due to contaminating nucleases. However, this conclusion
may be too optimistic, because it cannot be ruled out that a genuine
RNase comigrated with PAP on SDS-PAGE. Most plant RNases have a
molecular mass of 25–30 kDa, and hence can comigrate with
PAP. Moreover, plant leaves usually contain high levels of a mixture of
different RNases. It is questionable, therefore, whether a purification
scheme that includes only ammonium sulfate precipitation, anion
exchange chromatography, and cation exchange chromatography
(8)
can yield an RNase-free RIP preparation. This concern
is based not only on our own experience, but is supported by the recent
finding that nuclease activities of commercial preparations of gelonin,
momordin I, PAP-S, and saporin-S6 can be quantitatively removed by dye
chromatography on red Sepharose (35)
. Until RNase activity
can be demonstrated in essentially pure preparations, there is no
evidence that RIPs exhibit RNase activity.
DNase activity
Numerous efforts have been made during the last decade to
demonstrate that RIPs possess DNase activity. Several RIPs were
reported to cleave and linearize in vitro single-stranded M13 phage DNA
(43)
as well as supercoiled DNA (6
, 44
45
46)
.
In general, the observed DNase activity occurs only at high RIP
concentrations. Accordingly, these RIPs are considered to act as
inefficient nucleases. It is also believed that in contrast to the RNA
N-glycosidase activity, recognition of the supercoiled or
single-stranded DNA as a substrate for nuclease activity is not site
specific, but may be determined by the 3-dimensional structure of the
DNA (46)
. At first little attention was given to these
novel activities of RIPs because the observed DNase activities were
weak and conceptual problems arose to explain these unexpected
activities. Later, the issue of the presumed DNase activity of RIPs
became controversial because serious concerns were formulated about
contaminating nucleases in the RIP preparations. For example,
contaminating DNase and RNase have been detected in and removed from
apparently pure ricin and momordin II, respectively (40
, 41)
. Similarly, indirect evidence was presented that a PAP
isoform from leaves and several other natural type 1 RIPs are
contaminated with nucleases (40)
. Even though the possible
contamination of RIP preparations with nucleases remains a major
concern, intensive efforts are under way to assess the DNase activity
of RIPs and unravel the molecular mechanism of the presumed nuclease
activity.
Experiments with dianthin 30, saporin-S6, and gelonin indicated that
these type 1 RIPs exhibit nuclease activity toward single-stranded DNA
(43
, 47)
. Moreover, the polypeptide responsible for the
zinc-activated degradation of linear single-stranded DNA by native and
(bacterial) recombinant gelonin was identified as gelonin by
zymography, which strengthened the idea that this RIP acts as a DNase.
In a subsequent paper, Nicolas et al. (48)
presented
evidence that commercial preparations of gelonin, ricin, and PAP damage
single-stranded DNA (pUC18) by the removal of a protein-specific set of
adenines, followed by cleavage at the resulting abasic sites.
Oligonucleotides (28-mers) were deadenylated but not cleaved. Using a
borohydride
trapping,2
assay, it was concluded that the reaction of gelonin
proceeds via the enzyme-DNA imino intermediate that is
characteristic of DNA glycosylase/apurinic/apyrimidinic (AP) lyases.
Accordingly, they concluded that RIPs meet the established criteria to
be classified as DNA glycosylase/AP lyases. No explanation was provided
for the apparent lack of AP lyase activity on the oligonucleotides
except to say that this was the first report of the absence of AP lyase
activity of a native DNA glycosylase/AP lyase on an unmodified
oligodeoxyribonucleotide. In a later paper, the same authors provided
more details on the interaction of gelonin with single- and
double-stranded oligonucleotides and revisited their previous
conclusion that gelonin has an associated AP lyase activity
(49)
. At neutral pH and in the presence of zinc and
ß-mercaptoethanol, gelonin catalyzed the removal of adenine from
single-stranded DNA and, to a lesser extent, from normal base pairs and
mismatches in double-stranded DNA. Deadenylation of single- and
double-stranded oligonucleotides containing multiple adenines generated
unstable products with several abasic sites and resulted in strand
breakage and duplex melting, respectively. However, in contrast to
single-stranded DNA, the oligonucleotides were not cleaved. A
reinvestigation of the AP lyase activity of gelonin using the
borohydride trapping assay led to a different conclusion. The fact that
the gelonin-oligodeoxyribonucleotide complex formed only slowly after
base removal suggested that it was a result of a fortuitous encounter
between a lysine residue and the abasic site. Accordingly, it was
concluded that gelonin is a monofunctional glycosylase that forms a
covalent complex with its substrate and NOT a DNA glycosylase/AP lyase
(49)
. Though straightforward, this novel conclusion did
not explain the previously described DNA glycosylase/AP lyase activity
of gelonin.
It has also been suggested that the presumed DNA damaging activity of
gelonin may be responsible for the elimination of the parasite 6 kb
extrachromosomal mitochondrial DNA of Plasmodium
falciparum-infected erythrocytes (48
, 50)
. This
peculiar activity was believed to be independent of the
ribosome-inactivating activity of gelonin because it was not abolished
after boiling the protein (50)
. However, the possible
involvement of a contaminant could not be excluded. On the contrary,
the lack of cytotoxicity of a bacterial recombinant gelonin may suggest
the presence of contaminants in the gelonin purified from plant
material.
Another RIP that has received attention for its possible nuclease
activities is PAP. Wang and Tumer (51)
demonstrated that
PAP cleaves double-stranded supercoiled DNA using the same active site
required to depurinate rRNA and pG-1 DNA. Their rationale was that
wild-type PAP isolated from leaves cleaved double-stranded supercoiled
DNA, whereas the recombinant active site mutant PAPx isolated from
leaves of transgenic tobacco did not affect the same DNA. Though the
results seem convincing, there are two caveats. First, cleavage of the
double-stranded plasmid DNA occurs only at high concentrations of PAP.
Linearized DNA is not formed at an enzyme/substrate ratio <10. Second,
the arguments to prove the purity of the protein preparations are not
convincing. The fact that amino-terminal sequencing yielded a single
signal does not preclude the presence of small quantities of
contaminating proteins and does not take into account the possible
occurrence of blocked proteins. Moreover, silver staining of the
purified proteins after SDS-PAGE revealed the presence of some extra
bands. It is also important to note that wild-type PAP was a commercial
preparation from pokeweed leaves whereas PAPx was isolated from
transgenic tobacco leaves. Even if the same purification procedure is
followed, the content and identity of contaminating proteins may be
different. The PAPx preparation (from tobacco) may have been free of
contaminating DNase or DNA glycosylase/AP lyase activity whereas that
of PAP was not. A standard commercial preparation of PAP from seeds
(PAP-S) was also contaminated with nuclease(s), which could be removed
by chromatography on red Sepharose (35)
.
The authors presented evidence that the nicked DNA generated by PAP
contained free 3'-hydroxyl termini that could be labeled. This finding
implies that PAP does not cleave the DNA by a ß elimination reaction
like genuine DNA glycosylase/AP lyases, but rather removes adenines as
a monofunctional glycosylase and cleaves the DNA like an endonuclease.
Accordingly, PAP does not behave as a DNA glycosylase/AP lyase as
concluded by Nicolas et al. (48)
on the basis of the
results of a borohydride trapping assay. Since the solution structure
of MAP30 indicates that the putative DNase activity of RIPs is based on
a DNA glycosylase/AP lyase mechanism, the results obtained with PAP are
unexpected and may well be due to contaminating nuclease(s).
A critical assessment of the results obtained with ricin, gelonin, and
PAP along with the recent finding that a relatively simple
chromatographic technique allows removal of nuclease activity from RIP
preparations refuels the debate about the presumed nuclease activity of
RIPs. Structural studies of MAP30, one of the type 1 RIPs from seeds of
the bitter melon (Momordica charantia), have been used to
support the idea that RIPs are DNA glycosylase/AP lyases
(52)
. MAP30 is intensively studied because of its unique
biological activities. This RIP not only exhibits a potent antitumor
activity against human cancer cell lines, but also markedly inhibits
HIV-1 infection in lymphocytes and monocytes and suppresses viral
replication in HIV-infected cells. Since MAP30 has no adverse effects
on normal cells, its antitumor and antiviral properties make it a
candidate for therapeutical applications. Earlier observations
indicated that mechanisms unrelated to the ribosome-inactivating
activity contribute to the antitumor and antiviral activity, which
suggests that MAP30 exhibits multiple activities. Recently, Wang et al.
(52)
determined the solution structure of recombinant
MAP30 (expressed in E. coli) by NMR and reported structural
insights into the presumed multiple functions of this RIP. The RNA
N-glycosidase site of MAP30 closely resembles that of other RIPs.
Residues Y70, Y109,
E158, and R161 are involved
in the RNA N-glycosidase activity and are all located in a deep pocket
that specifically accommodates an extrahelical adenine base. The pocket
itself lies in the middle of a groove that runs along one side of the
polypeptide. Binding experiments with the HIV-1 long terminal repeat
indicated that specific residues in the groove interact with this DNA.
These residues are either required for ribosome inactivation activity
or surround the active site. Modified (i.e., abasic) as well as
unmodified HIV long terminal repeat and MAP30 formed imino
intermediates that could be trapped by reduction with borohydride. The
formation of these imino intermediates was interpreted as evidence that
MAP30 acts as a DNA glycosylase/AP lyase: in a first step, an adenine
base is removed from the DNA substrate and in the second step a
covalently linked DNA/MAP30 intermediate is formed that is trapped by
the reduction of borohydride. In addition, it was suggested that the
DNA glycosylase/AP lyase activity of MAP30 makes the HIV long terminal
repeat unsuitable as a substrate for the HIV integrase as well as the
DNA gyrase and that the DNA glycosylase/AP lyase may contribute to the
antiviral and antitumor activities of MAP30. The determination of the
solution structure also yielded important information about the active
sites for depurination and AP lyase activities. According to the data,
DNA depurination and AP lyase activities are carried out at distinct
but contiguous subsites. The structures of MAP30 and other RIPs suggest
that the RNA N-glycosylase site indeed cannot be an AP lyase site
because there is no nearby amino group that can act as a nucleophile.
However, it is proposed that after depurination, the abasic DNA site
(which is expected to have a high specific binding affinity for
aromatic side chains) binds to the side chain of a strictly conserved
tryptophan residue (W190 in MAP30) located on the
protein surface adjacent to the RNA N-glycosylase site. As a result of
this binding, the abasic site comes close to a lysine side chain
(K195 in MAP30), and the side chain amino group
of this lysine can function as a nucleophile that attacks the C1' of
the ribose of the abasic site. Although the proposed mechanism of DNA
glycosylase/AP lyase activity seems attractive, there are a few
caveats. First, it remains to be demonstrated that MAP30 possesses DNA
glycosylase/AP lyase activity. The recent finding that standard
preparations of another type 1 RIP from the bitter melon (called
momordin II) contain contaminating nucleases, which account for the
previously observed DNA glycosylase/AP lyase activity, cannot but
emphasize the need for a thorough purity check. Second, the proposed
mechanism is partly based on the assumption that gelonin and PAP also
possess DNA glycosylase/AP lyase activity. However, as discussed above,
there is still no definitive proof for such activity. Third, as
demonstrated by Nicolas et al. (49)
, the formation of the
gelonin–oligodeoxyribonucleotide complex proceeds slowly after base
removal and probably is the result of a fortuitous encounter between a
lysine residue and the abasic site. The same may hold true for the
complex formation between MAP30 and the HIV long terminal repeat.
The results described by Wang et al. (52)
have also been
addressed by Putnam and Tainer (53)
, who fully support the
view that MAP30 is an RNA and DNA glycosylase as well as a DNA/AP
lyase. However, these authors did not mention that no direct evidence
was provided by Wang and collaborators for the presumed DNA/AP lyase
activity and that some of their most important conclusions were based
on an extrapolation from a study with ricin, gelonin and PAP. In
addition, the comments by Putnam and Tainer (53)
were
formulated before Nicolas et al. (49)
published their
novel data on the enzymatic activity of gelonin and questioned the
accuracy of the borohydride trapping assay.
In summary, one can conclude that all RIPs possess
ribosome-inactivating and polynucleotide:adenosine glycosidase
activity. Other enzymatic activities cannot be excluded, but there has
been no unambiguous evidence for any of the proposed activities. Even
the presumed DNase or AP-lyase activity, which is believed to play a
critical role in some biological activities of RIPs, is based on
circumstantial evidence because the purity of the preparations has not
been properly assessed. It is surprising that the issue of possible
contamination of RIP preparations was not taken seriously even after
Day et al. (40)
clearly demonstrated that the DNase
activity attributed to PAP and ricin is due to contamination. As a
result, the elaborate efforts to prove that some RIPs have DNase or DNA
glycosylase/AP lyase activity and to elucidate the mechanisms of these
activities may have been trivial, because the recent results reported
by Barbieri et al. (35)
leave no doubt that at least
momordin, gelonin, PAP, and saporin are completely devoid of nuclease
or DNA glycosylase/AP lyase activity. Considering the pros and cons,
the debate about the presumed DNA glycosylase/AP lyase activity of
gelonin, PAP, and MAP30 can only be resolved by repeating the
experiments with RIP preparations purified according to the method
described by Barbieri et al. (35)
or an equivalent
technique.
The apparent lack of DNase and DNA glycosylase/AP lyase activity of
RIPs raises important questions about the biological activities
described for these proteins. For example, gelonin and MAP30 have
potent antiviral and antitumor activities that are believed to be
independent of their ribosome inactivation activity. In the case of the
anti-HIV-1 activity, the observed inhibitory activity has been ascribed
to the inhibition by MAP30 and gelonin of the three specific reactions
catalyzed by the viral integrase (54)
. It is not clear
whether any enzymatic activity is required for the RIPs to inhibit the
viral integrase. Studies with synthetic polypeptides showed that a 33
amino acid stretch corresponding to
K10-K42 of gelonin is the
shortest peptide necessary and sufficient for HIV-1 inhibition, DNA and
RNA binding, and ribosome inactivation (44)
. Although this
peptide can in principle inhibit HIV-1 and bind RNA and DNA, it is
unlikely that it can inactivate ribosomes because there is no active
RNA N-glycosidase site. Probably the highly basic peptide (IEP=10.58)
inhibits the in vitro translation just by binding to the mRNA. To
corroborate the presumed ribosome-inactivating activity of the peptide,
the possible deadenylation of the ribosomes should have been checked.
Moreover, contradictory results were obtained with proteolytic
fragments generated from MAP30 and gelonin. Fragments of gelonin
spanning G1-E195 and
G1-E219 had no
ribosome-inactivating activity, indicating that the removal of a
relatively short carboxyl-terminal fragment abolishes the RNA
N-glycosidase activity (55)
. However, the same fragments
inhibited HIV-integrase and converted supercoiled HIV-long terminal
repeat into relaxed and linear forms as well as native gelonin, which
led the authors to conclude that the antiviral and antitumor activities
of MAP30 and gelonin are independent of ribosome-inactivating activity.
The apparent DNase activity of the native RIPs and the proteolytic
fragments derived thereof may be indicative of the presence of
contaminating DNase(s). Even so, this does not explain the inhibition
of the HIV-integrase. Perhaps the inhibition of the viral integrase
just relies on the specific binding of the RIPs (or RIP fragments) to
the HIV long terminal repeat end sequences and does not require any
enzymatic activity. It should be emphasized, however, that until the
same experiments have been done with nuclease-free RIP preparations, no
decisive conclusions can be drawn as to whether the observed effects
are due to the RIPs or to contaminants.
The same reasoning applies to the elimination of the 6-kb
extrachromosomal DNA of Plasmodium falciparum, which does
not depend on the ribosome-inactivating activity of gelonin
(50)
. Until the same results have been obtained with
nuclease-free gelonin, it cannot be excluded that the observed
elimination of the parasite extrachromosomal DNA is due to a
contaminant.
Enzymatic regulators of RIPs
RIPs exhibit RNA N-glycosidase activity at a very low
concentration. Although the given IC50 value for
an individual RIP varies between laboratories due to differences in
plant sources or activity assay systems, it is generally accepted that
the minimal concentration required for type 1 RIPs to depurinate animal
ribosomes is at pM concentrations, whereas that for plant or
microbial ribosomes is at nM concentrations. Type 2 and 3 RIPs are
generally less active than type 1 RIPs. It is therefore helpful to know
which kinds of molecules can regulate RIP activity and how the
enzymatic activities of RIPs are regulated.
Some but not all type 1 RIPs have a marked requirement for ATP and
supernatant factors for activity (56
, 57)
. Both
requirements are directed toward the ribosome: preincubation of the
ribosome with the cofactors increases the rate of subsequent
inactivation by the RIP. Although ATP hydrolysis occurs,
phosphorylation of ribosomal proteins is not required. Different
cofactors have been identified for different RIPs (58)
.
The cofactor requirements of the type 1 RIP gelonin (from
Gelonium multiflorum) were studied in some detail. One of
the supernatant factors for gelonin was identified as
tRNAtrp (59
, 60)
. Only avian
(chicken) and mammalian (beef, rat, and rabbit)
tRNAsTrp are active, whereas yeast
tRNATrp is completely devoid of stimulating
activity. Construction of chimeric tRNATrp
transcripts identified several nucleotides in bovine
tRNATrp as recognition elements for
gelonin-stimulating activity (61)
. The reason why only
some RIPs show cofactor requirements is unknown.
Inhibitors of RIP activity were also identified. Adenine is a product
of the reaction catalyzed by RIPs both as an RNA N-glycosidase and as a
polynucleotide:adenosine glycosidase. It behaves as an uncompetitive
inhibitor for RIPs by binding to the enzyme–substrate complex
(62)
. A computer-assisted search identified pteroic acid
as a potent ricin activity inhibitor. Kinetic assays confirmed this
conclusion. The crystal structure of the complex revealed that the
pterin ring displaces Y80 and binds in the
adenine pocket, making specific hydrogen bonds to active site residues.
The benzoate moiety of pteroic acid binds on the opposite side of
Y80, making van der Waals contact with the
tyrosine ring and forming a hydrogen bond with
N78 (63)
.
Recently developed in vitro genetics techniques applied to select
nucleic acid ligands that bind to RIPs identified a group of novel RNA
aptamers as RIP activity inhibitors (64
, 65)
. Several RNA
aptamers were found preferentially bound to pepocin (a type 1 RIP from
Cucurbita pepo) and ricin A-chain. These aptamers show
strong inhibition activity on the RNA N-glycosidase activity of both
RIPs. A conserved hairpin motif, which is different from the sequence
of the sarcin/ricin loop domain in 28S rRNA was identified in the
aptamer sequences. It was suggested that RIPs (at least in case of
pepocin) may contain two RNA recognition sites: one site could bind
efficiently to a portion of the RNA substrate, and the other is the
catalytic site that hydrolyzes the adenine residue in rRNA
(65)
.
|
RIPS AS PLANT DEFENSE PROTEINS |
|---|
|
TOPABSTRACTGENERAL REFLECTIONS ON RIBOSOME...STRUCTUREBIOLOGICAL AND ENZYMATIC...RIPS AS PLANT DEFENSE...CONCLUSIONSREFERENCES |
|---|
RIPs occur in many, but certainly not all, plant species. Conclusive evidence for the lack of RIPs has been obtained only for Arabidopsis thaliana, as this plant does not express detectable amounts of RIPs or contain a sequence encoding a putative RIP in its whole genome. This implies that RIPs are not ubiquitous and do not play a universal role in the growth, development, or defense of plants. However, the question remains as to why some plants produce RIPs.
Several lines of evidence support the idea that RIPs play a role in
plant defense. A distinction should be made between the different types
of RIPs, because only type 2 RIPs can gain access to the cytoplasm of
intact cells through a receptor-lectin-mediated uptake process
(2
, 9)
. In principle, type 2 RIPs are potential toxins for
all organisms if they can gain access to the cell surface and bind to
suitable glycan receptors, but in practice their action spectrum is
restricted to higher and lower animals because bacteria and fungi are
protected by an impenetrable cell wall. Highly toxic type 2 RIPs like
ricin and abrin are believed to protect the seeds they inhabit against
plant-eating organisms. Though the oral toxicity of most other type 2
RIPs is much lower than that of ricin, the accumulation of large
quantities of these low-toxicity proteins (e.g., in vegetative storage
tissues) probably offers some protection against phytophagous
invertebrates and/or herbivorous animals. It has been suggested that
abundant type 2 RIPs are storage proteins that can be used as aspecific
defense proteins if the plant is challenged by a predating organism
(66)
.
Type 1 and type 3 RIPs are not cytotoxic and exhibit no documented oral
toxicity toward higher animals or invertebrates. This implies that
these RIPs cannot fulfill the same defensive role against plant
predating organisms as type 2 RIPs. According to some reports, type 1
RIPs have a direct effect on plant pathogenic fungi. However, the
observed antifungal activity was very weak and may be questioned
because possible contamination of the RIPs with genuine antifungal
proteins was not corroborated. In contrast, the antiviral activity of
type 1 RIPs is well documented and attempts are under way to exploit
RIP genes to protect transgenic plants against viral infection
(67)
.
RIPs as antiviral proteins
Antiviral activity in vitro
It has been known since the 1920s that crude extracts of pokeweed
leaves inhibit plant virus infection when the extracts and virus are
mixed and rubbed onto the surface of a test plant (20
, 21)
. Later, the active protein component was isolated and named
as pokeweed antiviral protein. Subsequently, purified RIPs from many
sources have been shown to potently inhibit the infection of test
plants with diverse plant viruses. All RIPs with proven antiviral
activity toward plant viruses are type 1 RIPs except the type 2 RIP
from Eranthis hyemalis (68)
.
Although there is no doubt about the antiviral activity of RIPs, the underlying mechanism has not been elucidated. Basically, three explanations can be put forward. First, RIPs act directly on the virus particles or viral nucleic acids by means of their polynucleotide:adenosine glycosidase activity. Second, RIPs selectively gain entrance to the cytosol of infected cells and destroy the protein synthesis machinery so that the virus cannot replicate and infect neighboring cells. In this so-called ‘local suicide’ model, access to the cytosol becomes possible when the integrity of the plasmalemma is breached by a virus vector such as an aphid or when cells are damaged during mechanical inoculation. Third, RIPs act indirectly through an activation of the plant’s defense system.
A direct effect on intact virus particles is unlikely because the
nucleic acids are physically protected from polynucleotide:adenosine
glycosidase activity of the RIPs. However, as soon as the virus
particles are disassembled in the cell, the viral RNA or DNA can be
attacked and deadenylated by the RIPs. Moreover, PAP can inhibit the
translation of capped viral RNAs without depurinating ribosomes
(8)
. It is questionable, however, whether damage or
translation inhibition of viral RNA is of any relevance in vivo: if the
cell is damaged during infection, the loss of cellular integrity almost
inevitably results in a massive degradation of proteins and nucleic
acids by endogenous lyases liberated from lysosomes; in the event that
virus particles succeed in penetrating the cytoplasm without disturbing
the cellular integrity, there will be no RIP molecules around to damage
the RNA. The same reasoning holds for the relevance of the inactivation
of ribosomes by RIPs in the local suicide hypothesis. This model also
implies that viral infection is accompanied by severe cell damage.
Since protein synthesis will stop almost instantaneously when the cells
lose their structural integrity, depurination of the ribosomes by RIPs
could hardly have any direct effect. The possible antiviral action of
RIPs through a direct effect on viruses/viral nucleic acids or through
a local suicide of cells is also contradicted by the observation that
RIP-containing plants are not resistant against viruses. For example,
even pokeweed plants, which express high levels (up to 5% of the total
protein) of a potent antiviral protein in their leaves, are susceptible
to plant viruses. It can be concluded, therefore, that the antiviral
activity of RIPs probably relies on an indirect mechanism that
activates the defense system of the plant. Evidence for such an
indirect mechanism has not yet been obtained from plants that normally
express RIPs, but only from transgenic tobacco expressing recombinant
RIPs.
Antiviral activity of RIPs in transgenic plants
Experiments with transgenic plants provide convincing evidence for
the in planta antiviral activity of RIPs. Expression of PAP conferred
the tobacco plant resistance against infection by mechanically and
aphid-transmitted viruses including potato virus X, potato virus Y, and
cucumber mosaic virus (69)
. Similarly, transgenic tobacco
expressing trichosanthin was found to be completely resistant against
infection by turnip mosaic virus (70)
, and potato plants
constitutively expressing Phytolacca insularis antiviral
protein exhibited resistance against potato virus X, potato virus Y,
and potato leafroll virus (71)
. Unfortunately, all these
RIPs caused a severe phenotype because of their toxicity for the
transgenic plants. Therefore, attempts were made to use antiviral RIPs
with a lower toxicity toward host plants like a low-toxic isoform of
PAP from leaves (PAP II) (72)
and a mutant of PAP (in casu
PAPc) that exhibits full antiviral activity in planta but is not
cytotoxic. Due to the separation of antiviral activity and
ribosome-inactivating activity of PAPc, virus-resistant tobacco lines
with a normal phenotype could be obtained (34
, 73)
.
Antiviral activity of RIPs in transgenic plants: working mechanism
Experiments with transgenic plants expressing PAP and PAP mutants
were able to dissect to some extent the mechanism of the in planta
antiviral activity of RIPs (8
, 34
, 69
, 73
74
75)
. Wild-type
PAP confers viral resistance to transgenic tobacco, but is cytotoxic
and depurinates tobacco ribosomes in vivo. PAPx, an active site mutant
(E176V) of PAP that possesses no RNA N-glycosidase activity, is not
cytotoxic and does not depurinate tobacco ribosomes in vivo, but also
does not protect the transgenic plants against viruses. PAPc, a
carboxyl-terminal deletion mutant with an intact RNA N-glycosidase
site, exhibits the same antiviral activity as wild-type PAP in planta
but is devoid of cytotoxicity and does not depurinate the tobacco
ribosomes in vivo. These observations led the authors to conclude that
1) an intact RNA N-glycosidase site of PAP is required for
antiviral activity, toxicity, and in vivo depurination of the host
ribosomes, 2) an intact active site is not sufficient for
all these activities, and 3) an intact carboxyl terminus is
required for toxicity and in vivo depurination of the host ribosomes
but not for the antiviral activity. In addition, the authors suggested
that the carboxyl terminus of PAP may not contain crucial viral
recognition determinants because an intact carboxyl terminus is not
crucial for antiviral activity (73
, 75)
. It is
questionable, however, whether the results obtained with PAPc allow one
to draw the conclusion that antiviral activity and toxicity can be
simply separated. An alternative explanation could be that the tobacco
cells succeed better in correctly targeting PAPc than wild-type PAP
(for example, because its protein trafficking machinery may be confused
by the carboxyl terminus of PAP). If so, the observed lack of toxicity
of PAPc mutants is not due to its incapability to depurinate ribosomes,
but rather to the lack of access to the ribosomes in vivo.
A next step in the elucidation of the antiviral working mechanism was
the finding that PAPv (a PAP mutant with two amino acid substitutions,
L20R and Y49H, showing antiviral activity and toxicity) causes a
constitutive expression of the acidic form of PR-1 (class II
pathogenesis related protein PR-1). This acidic PR-1 is normally
induced in tobacco by salicylic acid (SA). However, upon expression of
PAP the synthesis of PR-1 occurred independently of SA accumulation.
Reciprocal grafting experiments further revealed that wild-type scions
grafted on transgenic root stocks and wild-type root stocks grafted
with transgenic scions acquired resistance against TMV and potato virus
X (74)
. The wild-type root stocks and scions did not
express PAPv, nor was there any increase in the expression level of
PR-1 or salicylic acid content. Based on these results, it was
concluded that PAP generates through its enzymatic activity a soluble
signal that can be translocated to the upper and lower parts of the
grafted plants and induces resistance to virus infection by a mechanism
that is not linked to accumulation of salicylic acid or induction of PR
1 (34
, 74)
.
Expression of PAP induced not only the synthesis of acidic PR-1, but
also that of acidic PR-2 and the basic isoform of PR-3 (class I PR-3)
without a significant increase in the concentration of salicylic acid
(73)
. The simultaneous induction of class I and class II
PR proteins by the ectopically expressed PAP is important because both
classes normally are induced by different pathways. In tobacco, PR-1
and class II isoforms of PR-2 and PR-3 are induced by salicylic acid
and play a role in systemic acquired resistance, whereas class I
isoforms of pathogenesis-related proteins are induced by ethylene or by
stress. Since transgenic tobacco producing PAP constitutively expresses
both basic and acidic isoforms of PR proteins in the absence of an
increase of salicylic acid, it was suggested that PAP acts through the
activation of a downstream regulatory signal that is common for both
pathways. Irrespective of the exact mode of action, the constitutive
expression of PR proteins increased the survival of transgenic tobacco
seedlings in soil inoculated with the phytopathogenic fungus
Rhizoctonia solani (73)
.
Markedly different results were obtained with transgenic
tobacco-expressing PAPn, a mutant of PAP with the single amino acid
substitution G75D (75)
. PAPn does not depurinate rabbit
reticulocyte lysate or tobacco ribosomes in vitro (8)
. The
mutant is not toxic for tobacco but confers transgenic tobacco plants
resistance to infection with viruses and R. solani. PAPn,
does not depurinate ribosomes in vivo and in contrast to PAP does not
efficiently bind to tobacco ribosomes. Though the latter observation
prompted the authors to conclude that G75 in the
amino-terminal domain is critical for binding of PAP to the ribosomes,
the observed reduced binding may rely simply on the introduction of a
negative charge in the polypeptide chain. Moreover, the issue of the
binding of PAP and its mutants to the tobacco ribosomes is of little
relevance because it is unlikely that the RIPs and ribosomes can bind
to each other, as they are supposedly located in different cellular
compartments. Similar to PAP and PAPc, PAPn-expressing tobacco plants
do not show an elevated level of salicylic acid. However, in contrast
to PAP and PAPc, the expression of PAPn does not induce the expression
of acidic PR proteins but only that of the basic isoform of PR-3. Based
on these findings, it was suggested that translation inhibition
activity and ribosome binding of PAP through the amino-terminal domain
are sufficient to induce the expression of acidic isoforms of PR
proteins in transgenic tobacco. The fact that PAPn confers transgenic
tobacco resistance against viruses and R. solani indicates
that the acidic PR proteins are not involved in the resistance
mechanism, which agrees well with the results of the above-described
grafting experiments. Apart from PR proteins, the authors also checked
the transcriptional activation of the wound inducible genes encoding a
wound-inducible protein kinase (WIPK) and the tobacco proteinase
inhibitor PI-II in transgenic tobacco expressing PAPn and PAPv. Both
PAPv and PAPn activated WIPK, a marker for SA-independent pathogen
response, and PI-II, which is a typical marker of wound response.
According to these data, PAPn and PAPv activate an SA-independent
signal transduction pathway reminiscent of the wounding response in
transgenic tobacco. It is not clear whether the same applies to
wild-type PAP, and PAPc and PAPx because no data were shown for these
forms. It is not entirely clear what activities are required to elicit
this signal transduction pathway. Since it is believed that PAPn does
not depurinate tobacco ribosomes neither in vitro nor in vivo it is
suggested that the translation inhibitory activity of PAPn is
sufficient to trigger the signal transduction. According to the authors
the in vivo depurination of capped mRNA in transgenic tobacco cells by
the transgene PAPn may be the onset to a SA-independent
stress-associated signal transduction that eventually leads to
resistance against viruses and fungi. Though this hypothesis is
certainly attractive, it remains to be demonstrated that the presumed
depurination of capped mRNAs occurs in vivo.
In summary, one can conclude that the antiviral mechanism of PAP in transgenic tobacco relies on an SA-independent signal transduction pathway reminiscent of the wounding response. Though there are indications that an active RNA N-glycosylase site is an absolute requirement for antiviral activity, no further conclusions can be drawn as yet about other structural requirements. In addition, there is no experimental evidence that the results obtained with PAP and its mutants can be extrapolated to other type 1 RIPs.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTGENERAL REFLECTIONS ON RIBOSOME...STRUCTUREBIOLOGICAL AND ENZYMATIC...RIPS AS PLANT DEFENSE...CONCLUSIONSREFERENCES |
|---|
RIPs probably play a role in plant defense. Type 2 RIPs are defense proteins directly targeted against plant eating organisms, whereas type 1 RIPs are involved in the defense against viruses and perhaps microorganisms. The protective activity of both type 1 and type 2 RIPs definitely depends on their enzymatic activity. However, in contrast to type 2 RIPs, type 1 RIPs act indirectly through an activation of the plant’s defense system(s).
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTGENERAL REFLECTIONS ON RIBOSOME...STRUCTUREBIOLOGICAL AND ENZYMATIC...RIPS AS PLANT DEFENSE...CONCLUSIONSREFERENCES |
|---|
