In the present study, four pulchellin isoforms termed P I, P II, P III and P IV were isolated by affinity, ion exchange and chromatofocusing chromatographies, and investigated with respec
Trang 1inactivating pulchellin isoforms from Abrus pulchellus
seeds
Priscila V Castilho1,2, Leandro S Goto2, Lynne M Roberts3and Ana Paula U Arau´jo1,2
1 Programa de Po´s-graduac¸a˜o em Gene´tica e Evoluc¸a˜o, Universidade Federal de Sa˜o Carlos, Brazil
2 Instituto de Fı´sica de Sa˜o Carlos, Universidade de Sa˜o Paulo, Sa˜o Carlos, Brazil
3 Department of Biological Sciences, University of Warwick, Coventry, UK
Ribosome-inactivating proteins (RIPs; rRNA
N-glyco-sidases; EC 3.2.2.22) are found predominantly in
plants but they may also occur in fungi and bacteria
[1] Collectively, unless mutated, they are all
rRNA-specific N-glycosidases capable of selectively cleaving a
glycosidic bond to release an adenine within the
uni-versally conserved sarcin⁄ ricin loop of the large rRNA
in 60S ribosomal subunits [2] This modification
pre-vents the binding of elongation factors and thereby
irreversibly inhibits protein synthesis in eukaryotic
cells Despite this common activity, RIPs can vary in
their physical properties and cellular effects [3]
Currently, RIPs are divided into three groups
Type 1 RIPs comprise a single catalytically active
sub-unit of 25 ± 30 kDa, whereas type 3 RIPs consist of
an amino-terminal domain, resembling type 1 RIPs,
linked to a carboxyl-terminal domain with unknown function [4] In contrast, type 2 RIPs contain at least one ribosome inactivating A-chain and a correspond-ing number of carbohydrate-bindcorrespond-ing B-chains, with the latter generally showing a preference for b1-4 linked galactosides [3] It follows that, although type 1 and type 2 RIPs are active against ribosomes in vitro, only the type 2 proteins are cytotoxic due to the presence of
a B-chain that mediates surface binding and entry of holotoxin into susceptible cells From studies of the biosynthesis of type 2 RIPs in their producing tissues,
it is apparent that both polypeptides are made in cor-rect stoichiometry by being derived from a single pre-cursor through the excision of a intervening peptide sequence [5] The two polypeptides remain covalently joined, however, by a disulfide bridge between cysteine
Keywords
Abrus pulchellus; characterization; cloning;
isoforms; ribosome-inactivating protein
Correspondence
A P U Arau´jo, Grupo de Biofı´sica
Molecular, IFSC, PO Box 369,
Zip 13560-970, Sa˜o Carlos, Brazil
Fax: +55 16 33715381
Tel: +55 16 33739834
E-mail: anapaula@ifsc.sc.usp.br
(Received 14 November 2007, revised 11
December 2007, accepted 20 December
2007)
doi:10.1111/j.1742-4658.2008.06258.x
Abrus pulchellusseeds contain at least seven closely related and highly toxic type 2 ribosome-inactivating pulchellins, each consisting of a toxic A-chain linked to a sugar binding B-chain In the present study, four pulchellin isoforms (termed P I, P II, P III and P IV) were isolated by affinity, ion exchange and chromatofocusing chromatographies, and investigated with respect to toxicity and sugar binding specificity Half maximal inhibitory concentration and median lethal dose values indicate that P I and P II have similar toxicities and that both are more toxic to cultured HeLa cells and mice than P III and P IV Interestingly, the secondary structural character-istics and sugar binding properties of the respective pairs of isoforms corre-late well with the two toxicity levels, in that P I⁄ P II and P III ⁄ P IV form two specific subgroups From the deduced amino acids sequences of the four isoforms, it is clear that the highest similarity within each subgroup is found to occur within domain 2 of the B-chains, suggesting that the disparity in toxicity levels might be attributed to subtle differences in B-chain-mediated cell surface interactions that precede and determine toxin uptake pathways
Abbreviations
GalNAc, N-acetylgalactosamine; IC 50 , half maximal inhibitory concentration; LD 50 , median lethal dose; P I, pulchellin isoform I; P II, pulchellin isoform II; P III, pulchellin isoform III; P IV, pulchellin isoform IV; RIP, ribosome-inactivating protein.
Trang 2residues at the C-terminal of the A-chain and the
N-terminal of the B-chain
Overall, relatively few type 2 RIPs are known [6]
Ricin and abrin (from Ricinus communis and
Abrus precatorius seeds, respectively) were recognized
more than a century ago, with others (e.g mistletoe
lectin I [viscumin] from Viscum album [7], modeccin [8]
and volkensin [9] from the roots of Adenia digitata and
Adenia volkensii, respectively, and pulchellin from
Abrus pulchellus seeds [10,11]) being discovered within
the last 30 years The greatest number of RIPs have
been found in the Caryophyllaceae, Sambucaceae,
Cucurbitaceae, Euphorbiaceae, Phytolaccaceae and
Poaceae [1] Although many are potentially useful in
agriculture and medicine because of their antiviral
properties [12] and cell killing characteristics (e.g in
‘immunotoxins’) [13], the complete distribution map,
mode of cell entry⁄ action and the role(s) of RIPs in
nature remain only partly understood
Plants commonly produce several RIP isoforms
encoded by multigene families that could possess
adap-tations related to their specific role in plant tissues [6]
Therefore, widening our knowledge of the occurrence,
structural properties and biological functions of RIPs
will contribute to an understanding of their role(s)
in vivo Abrus pulchellus tenuiflorus
(Leguminosae-Papi-lionoideae) seeds contain a highly toxic type 2 RIP
named pulchellin It exhibits specificity for galactose
and galactose-containing structures, can agglutinate
human and rabbit erythrocytes, and kills mice and the
microcrustacean Artemia salina at very low
concentra-tions [10] Similar to the RIP in A precatorius seeds
[14], this toxic activity is presented by a mixture of
clo-sely related isoforms In the present study, four
pulchel-lin isoforms were isolated, and their amino acids
sequences deduced by cDNA cloning and verified by
MS Half maximal inhibitory concentration (IC50) and
median lethal dose (LD50) values from HeLa cells and
mice divided them into two subgroups: the more toxic
forms (P I and P II) and the less toxic forms (P III and
P IV) In similar pairwise combinations, their
interac-tion with specific sugars was also shown to differ From
a comparison of deduced amino acid sequences within
each subgroup, it is striking that the members of each
show closest identity in domain 2 of the B-chain The
potential implications of this are discussed
Results
Nomenclature of the toxic pulchellin lectins
The abbreviation P is followed by the Roman
numer-als I, II, III and IV and refers to each pulchellin
isoform (P I, P II, P II and P IV) The A-chain of P II was formerly cloned and named recombinant pulchel-lin A-chain [15] The heterologous expression and refolding of a recombinant pulchellin binding chain was previously reported [16], although this recombi-nant pulchellin binding chain does not correspond to any of the four B-chains presented here
Purification of four pulchellin isoforms from
A pulchellus seeds Using a combination of affinity, ion exchange and chromatofocusing chromatography, four pulchellin isoforms were isolated from A pulchellus seeds The protein eluting from an affinity column with lactose suggested protein homogeneity However, isoeletric focusing revealed multiple bands (data not shown), indicating the presence of related isoforms in the affin-ity-purified preparation The very distinct differences
in isoelectric points suggested that the ion exchange chromatography could be used for the separation of the various isoforms Using an anion exchanger, four peaks were resolved (Fig 1A) and proteins were iso-lated Denaturating gels revealed pulchellins of approx-imately 62 kDa, which, upon reduction, showed a pattern of two bands of approximately 28 and 34 kDa, related to A and B-chains respectively (Fig 1B) The slight differences in the migration pattern of the A-chains is possibly attributable to glycosylation differences Lanes 5 and 9 (Fig 1B) relate to the peak indicated by an asterisk (Fig 1A) and showed hetero-geneity, which was further confirmed in LC-MS⁄ MS assays Although samples from the asterisked peak in Fig 1A displayed hemagglutination and toxicity toward mice (data not shown), additional efforts to cleanly isolate the isoform were not successful and further characterization was abandoned A chromato-focusing step was included to separate the P III and
P IV isoforms from the eluate P III⁄ P IV (Fig 1C) Isoelectric focusing gave pI of 5.8, 5.7, 5.5 and 5.2 for the four isoforms respectively
Secondary structure of the pulchellin isoforms and melting temperature
CD-spectral analyses were performed as described in the Experimental procedures As can be seen from Fig 2, the far-UV CD spectra of the pulchellin iso-forms suggest only subtle differences in the content of secondary structure, which was confirmed by the spec-tral deconvolution using cdpro software Thermal sta-bility was also monitored by CD, following changes
in each spectrum with increasing temperature The
Trang 3predicted content of secondary structure and melting temperatures found for the four isoforms are given in Table 1
Sequence comparison of pulchellins from
A pulchellus seeds Using RT-PCR and a primer set, full length cDNA clones were prepared and sequenced with primer walk-ers as detailed in the Experimental procedures Accord-ing to the extent of similarity, seven different sequences were found amongst all analyzed clones Nanoelectro-spray LC-MS⁄ MS was also carried out on the individ-ual proteins to demonstrate correspondence between each cloned sequence and a native pulchellin isolated from mature seeds To generate samples for mass anal-ysis, intact proteins were isolated by chromatographic
0
50
100
150
200
250
300
*
P I
Elution volume (mL)
Absorbance in 280 nm (a.u.) 0
20 40 60 80 100
66
45
30
20
kDa
B-chain A-chain
0
20
40
60
80
100
0 20 40 60 80 100
P IV
P III % of buffer B
Elution volume (mL)
C
B
A
Fig 1 Mono Q elution profile of pulchellin isoforms (A) The four
isoforms were eluted with a linear gradient of 0–20% 1 M NaCl
in 20 m M Tris–HCl, pH 8, for 45 min (dashed line) at a flow rate
of 1 mLÆmin)1 The peaks referring to each isoform are indicated
by arrows The asterisk indicates the peak containing a mixture
of other isoforms (B) Gel visualization of proteins eluted Lane 1,
molecular weight markers Numbers on the left indicate the Mr
values of the standards in thousands Lanes 2–9, 5 lg of peak
P I, P II, P III ⁄ P IV and a mixture of other isoforms (*),
respec-tively, in the presence (lanes 2–5) or absence (lanes 6–9) of
2-mercaptoethanol (C) Elution profile of P III and P IV
Chromato-focusing chromatography of the P III ⁄ P IV peak previously
iso-lated from the Mono Q column Samples were dialyzed against
10 m M sodium phosphate buffer, pH 7.0 The column was
simi-larly equilibrated and P III and P IV were separated by a linear
gradient (dashed line) of 10 m M sodium phosphate buffer, pH 5.8
from 0–100% for 20 min, holding for 5 min in 100% buffer B.
Flow rate = 1 mLÆmin)1 The peaks relating to each isoform are
indicated by arrows.
200 210 220 230 240 250 –4
–3 –2 –1 0 1 2 3
–1 ·cm
Wavelength (nm) Fig 2 Circular dichroism spectra of P I, P II, P III and P IV CD spectra of P I (solid), P II (dash), P III (dot) and P IV (dash dot) were measured in the far-UV range (195–250 nm) in 1 mm path length quartz cuvettes and recorded as an average of 16 scans CD spec-tra were measured in protein solution of 0.125 mgÆmL)1 (Tris
20 m M , pH 8, 10 m M NaCl added).
Table 1 Secondary structure content (expressed as %) and melt-ing temperatures found for P I, P II, P III and P IV Secondary struc-ture values were obtained by the spectral deconvolution using
1 The melting temperatures were calculated based on CD thermal scans (at 232 nm) of the proteins.
Secondary structure
Melting temperatures (C)
Trang 4methods previously described and subjected to tryptic
digestion The acquired masses were then compared
with those deduced from peptide sequences encoded by
the seven cDNA clones The deduced precursor
pro-teins (prepropulchellins) from the cDNA clones, which
were found to correspond with the isolated mature P I,
P II, P III and P IV isoforms, are shown in Fig 3
Stretches of sequence that matched the calculated
masses obtained by LC-MS⁄ MS are underlined For example, our results showed that the first ten residues
of the mature proteins (EDPIKFTTEG) were the same for P I, P III and P IV, but P II differs in that it con-tains an additional arginine as the third residue and has glutamine instead of lysine as the sixth residue: ED-RPIEFTTE LC-MS⁄ MS analysis of a tryptic digest of mature P II revealed a peptide mass compatible with a
Fig 3 Deduced amino acid sequences from the cDNA clones of P I, P II, P III and P IV aligned to abrin-a (pdb 1ABR), ricin (pdb 2AAI) and mistletoe lectin I (pdb 1CE7) The peptides selected by the PROTEINLYNX 2.0 software are underlined As a databank, the program used the seven amino acids sequences deduced from seven pulchellin cDNA clones that contained the immature precursors The signal peptides were predicted based on the program SIGNAL P The amino acids numbers were based on the mature proteins Conserved amino acids are highlighted in gray, conserved residues only amongst pulchellin isoforms are shown in bold Residues involved in the active site cleft, pre-dicted by homology to abrin and ricin, are indicated by an asterisk Glycosylation sites have a black background and residues forming the two carbohydrate-binding sites, first (Æ) and second (:), predicted by homology to mistletoe, abrin and ricin [17], are boxed Ten cysteines that form one interchain and four intrachain disulfide bonds are marked by (^) Dashes denote gaps introduced to obtain maximal homology.
Trang 5fragment containing these changes (underlined at the
N-terminus of P II)
Overall, the four isoforms precursors have 562 (P I),
563 (P II) and 561 (P III and P IV) aminoacyl
resi-dues The protein outside this family to which they
showed the highest amino acid identity was abrin, at
approximately 94% Besides abrin, the pulchellin
iso-forms were also compared with ricin and mistletoe
lec-tin I (Fig 3), with which they showed approximately
47% amino acid identity to both sequences
The respective A-chains contain 251 (P II) and 250
(P I, P III and P IV) amino acids P I and P IV
A-chains have two N-glycosylation sites, whereas P II
and P III only one In the four pulchellin A-chains,
the residues involved in the active site cleft are the same as in abrin and ricin A-chains This suggests that the catalytic reaction is exactly the same The sugar binding pulchellin B-chains are 264 (P I and P II) or
263 (P III and P IV) amino acids in length and contain two N-glycosylation sites Soler et al [17] defined two homologous carbohydrate binding sites that were shared in mistletoe lectin I, ricin-d and abrin-a B-chains Based on these previously published observa-tions, we predict residues comprising the two sugar binding pockets in the pulchellins (Fig 3)
In order to compare the similarity of the A- and B-chains of the four isoforms, a pairwise alignment was performed and the values of identity expressed in
Fig 3 (Continued).
Trang 6percentage are shown in Table 2 The B-chain domains
were predicted by comparison with ricin domains (pdb
2aai) Domains are defined by a repeating pattern of
disulfide-bonded loops in each half of the polypeptide,
analogous to those described for the ricin B-chain,
which suggested that this lectin arose as a product of
gene duplication [18] Domain 1 comprises residues
251–387 (P I, P III and P IV) or 252––388 (P II), and
domain 2 comprises 388–514 (P I), 389–515 (P II) or
388–513 (P III and P IV)
In vivo toxicity
The addition of pulchellin isoforms to cultures of
HeLa cells resulted in high inhibition of protein
syn-thesis (Fig 4) The IC50 values showed that P I and
P II have similar toxicity [21.7 ngÆmL)1(0.375 nm) and
22.7 ngÆmL)1(0.391 nm), respectively] and are approxi-mately five-fold more toxic then the others [101.9 ngÆmL)1 (1.76 nm) for P III and 98.4 ngÆmL)1 (1.7 nm) for P IV] LD50experiments also showed vari-ability in the toxicity to mice, with the most potent toxin being P II (15 lgÆkg)1), followed by P I (25 lgÆkg)1), P IV (60 lgÆkg)1) and P III (70 lgÆkg)1) These results indicate that the pulchellin isoforms are highly toxic, but not as much as mistletoe lectin (LD50 5–10 lgÆkg)1) [19], ricin (IC50 0.001 nm and LD50 2.6 lgÆkg)1) [20] and abrin (IC50 0.0037 nm and LD500.56 lgÆkg)1) [21]
Agglutination and carbohydrate-binding of the B-chains
It has been observed on several occasions that different type 2 RIPs from a single plant differ from each other with respect to their agglutination activity and⁄ or spec-ificity [22,23] To check whether this also holds true for pulchellin isoforms, the agglutination properties and carbohydrate-binding affinity were studied in some detail The pulchellin isoforms were examined for their hemagglutination potential using blood of three spe-cies: human (types A+, B+and O+), rabbit and horse where they showed blood group specificity and distinct hemagglutination activity P I and P II promoted hem-agglutination of human erythrocytes at 22.5 and 27.5 ngÆmL)1, respectively Although P I showed only activity towards human erythrocytes, P II was able
to agglutinate rabbit (27.5 ngÆmL)1) and horse (41.7 ngÆmL)1) erythrocytes P III and P IV aggluti-nated only rabbit blood (18.5 ngÆmL)1 and 12.3 ngÆmL)1, respectively)
To determine their carbohydrate binding specificity,
a series of hemagglutination inhibition assays were carried out using 14 sugars of three classes Whereas agglutination was inhibited by galactose and its deriva-tives [such as N-acetylgalactosamine (GalNAc), methyl-a-d-galactopyranoside], it was evident that, at doses up to 100 mm, glucose, mannose, a-methylman-noside, fucose, maltose, xylose and saccharose did not inhibit agglutination (Table 3)
All four pulchellins were shown to interact with ga-lactosides, although the minimum sugar concentration that promoted inhibition of hemagglutination varied The failure to bind glucose, mannose, a-methylmanno-side, fucose, maltose, xylose and saccharose shows that
an axial hydroxyl group at C4 is not only an impor-tant binding group for the lectin, but also that a reversed configuration at this position might prevent sugar recognition P I and P II were able to inhibit hemagglutination in the presence of GalNAc whereas
Table 2 Identity of pulchellin isoforms in a pairwise alignments.
Values are expressed as a percentage (%) The B-chain domains
were defined by comparison with ricin B-chain (PDB: 2aai).
Domain 1 comprises residues 251–387 (P I, P III and P IV) or
252–388 (P II), and domain 2 comprises residues 388–514 (P I),
389–515 (P II) or 388–513 (P III and P IV).
A-chain
B-chain
5
15
25
35
45
55
65
75
85
95
Fig 4 Inhibition of protein synthesis in HeLa cells Each isoform
was diluted serially in DMEM ⁄ fetal bovine serum and added to
HeLa cells at the concentrations shown The incorporation of
[ 35 S]methionine into new cellular proteins was subsequently
deter-mined as described in the Experimental procedures Each value is
the mean for triplicate samples h, P I; , P II; s, P III; d, P IV.
Trang 7P III and P IV did not The hemagglutination
inhibi-tion caused by methyl- a -d-galactopyranoside suggests
that the -OH on C2, C3 and C4, which have the same
configuration as those in galactose and lactose, are
responsible for the strong interaction with the
iso-forms Interestingly, P II was the only isoform with
affinity for rhamnose As a result, P II lacked the
galactose and⁄ or N-acetyl galactosamine specificity
that is a characteristic feature of the archetypal type 2
RIP (with few exceptions)
The most striking difference in sugar binding
prefer-ence was observed with GalNAc (Table 3) We
there-fore performed cytotoxicity assays in which the
various pulchellins were pre-incubated or not with free
GalNAc to determine whether this sugar can prevent
surface binding of toxin in a manner that might
indi-cate a possible basis for the distinctive subgroup
potencies (Fig 4) For P I and P II, we observed
improved levels of cellular protein synthesis as the
concentration of pre-mixed GalNAc was increased
(Fig 5) The reason for the different plateaus seen with
P I (where the rescue of protein synthesis reaches a
maximum of approximately 50%) and P II (where
res-cue of protein synthesis reaches a maximum of
approx-imately 85%) is not known, but the variation may
reflect suboptimal binding of GalNAc in one or both
binding pockets of P I However, in striking contrast
with the rescue of protein synthesis observed for P I
and P II, protection against P III and P IV was
mar-ginal, even when toxin was pre-treated with 100 mm
GalNAc Taken together with the inability of P III
and P IV to inhibit hemagglutination in vitro in the
presence of this sugar, these data suggest that the
bind-ing and uptake of these two isoforms does not require
receptors containing GalNAc The pulchellin isoforms
P I and P II are clearly different since, in the presence
of this sugar, both hemagglutination and cytotoxicity are inhibited
Discussion
The present study reports the isolation and initial char-acterization of four pulchellin type 2 RIPs and their encoding cDNA sequences Seven cDNAs were com-pletely sequenced and four were correlated with the iso-forms isolated from mature seeds Since the pulchellin isoforms contain both A-chain and B-chain sequences connected in sequence (Fig 3), they are clearly made
as precursors This is compatible with other type 2 RIPs The precursors contain a very similar 34 residue N-terminal pre-sequence, and a short intervening linker peptide joining the A- and B-chains, that must be removed during protein maturation upon their biosyn-thesis The pre-sequence resembles a true endoplasmic signal peptide to direct the proteins into the secretory pathway The additional N-terminal sequence may function in a manner akin to the N-terminal propeptide found in preproricin [24] It is most likely cleaved after
an Asn residue once the protein is deposited in vacu-oles The intervening linker peptides are also extremely similar and, by analogy to that of preproricin, may well contain a vacuolar targeting signal [25]
Alignment of the immature polypeptide sequences (Fig 3) shows that some residues are conserved only amongst the pulchellin isoforms (Fig 3) Although
Table 3 Carbohydrate-binding specifity of P I, P II, P III and P IV.
In the first well, 100 lL of each sugar at 100 m M was placed and
50 lL was taken and serially two-fold diluted in wells containing
50 lL of NaCl ⁄ P i Then, 50 lL of each isoform solution
(112 lgÆmL)1) was added to the wells Following incubation, 50 lL
of a 1% erythrocyte solution was added Numbers indicate the
minimal concentration that inhibits agglutination.
Sugar
Minimum concentration for inhibition (m M )
Methyl-a- D -galactopyranoside 1.56 3.12 100 25
0 10 20 30 40 50 60 70 80 90
[N -acetyl B-D-galactosamine] (mM)
100
Fig 5 Competition of pulchellin entry by N-acetyl- D -galactosamine.
A single dose of toxin (200 ngÆmL -1 P I and P II, or 800 ngÆmL)1
P III and P IV), previously shown capable of inhibiting 90% protein synthesis within 4 h, was used in all preincubations Each toxin was mixed with increasing concentrations of GalNAc in DMEM ⁄ FCS for 30 min at 37 C The mixtures were added to cells for 4 h and remaining protein synthesis determined as detailed
in the Experimental procedures Each value is the mean for tripli-cate samples h, P I; , P II; s, P III; d, P IV.
Trang 8three isoforms (P I, P II and P IV) have the nine
con-served cysteines in the B-chains, P III has only eight of
these and lacks Cys506, indicating that it must lack
one of the usual four intra-chain disulfide bridges
The primary sequences of the catalytic A-chains
were found to be only slightly different (Fig 3) but
not in the pairwise manner indicated from cytotoxities
(Fig 4) Indeed, virtually all of the changes within the
pulchellin A-chains revealed pairwise identity of P II⁄
P III and P I⁄ P IV (Table 2) However, these
differ-ences lie outside the residues that are known, from
other ribosome inactivating proteins, to determine the
major folds and the catalytic site (Asn71, Tyr73,
Tyr112, Arg123, Gln159, Glu163, Arg166, Glu194,
Asn195, Trp197, P I numbering) [26] Indeed, these
residues are retained in positions corresponding exactly
to those in the A-chains of ricin and abrin [26]
Over-all, it is therefore unlikely that the A-chains differ
sig-nificantly in catalytic activity
The toxicity values found for the pulchellin isolectins
divided them into two subgroups, the more toxic forms
(P I and P II) and the less toxic forms (P III and
P IV) It was suggested that the presence⁄ absence of a
carbohydrate chain close to the RNA-binding sites
could influence the different toxicities found for
mistle-toe lectins I and III [27] P I and P IV A-chain have
two N-glycosylation sites, whereas P II and P III have
only one In this sense, and in contrast to mistletoe
lectins, no correspondence between glycosylated⁄
nonglycosylated A-chain and their biological activities
could be found
The amino acids residues most likely involved in the
two B-chain sugar binding sites also vary, although
only very slightly (Fig 3) Although the first putative
sugar binding site is the same in P II and P III, it
dif-fers by a single residue (Trp instead of Tyr) in both P I
and P IV This may be analogous to ricin B-chain in
which Trp and Tyr side chains have been reported to
provide a flat binding surface for galactose, although
they do not make more specific interactions with the
sugar [28] If similar to the present study, then this
sub-tle difference between sugar binding site 1 in P II and
P III may have no functional consequence in relation
to carbohydrate binding From the putative C-terminal
sugar binding site (identical in P I and P II but
differ-ing by a sdiffer-ingle residue (Trp instead of Tyr) in both
P III and P IV), the same logic may also apply In the
present study, the absence of any marked difference
between the actual sugar binding residues suggests that
the simplest explanation for the different
haemaggluti-nation and cytotoxicities between the two subgroups
(Table 3, Figs 4 and 5), is that flanking residues may
be critical in preventing a P III⁄ P IV interaction with
GalNAc This hypothesis was also raised for the mistletoe lectin I [29] The pairwise alignment of the isoforms reveals that, although the highest primary sequence similarity of each subgroup is found in the C-terminal half of the B-chains (domain 2; Table 2), and that there is only a single conserved aromatic sub-stitution in the residues that make up the putative sugar binding pockets, there is some interesting varia-tion in the flanking regions around the second sugar binding pocket that could influence the P III⁄ P IV specificity and binding properties Of particular interest
is the substitution G488R presented by P III⁄ P IV
In summary, our data describe a preliminary charac-terization of a family of pulchellins and reveal a num-ber of clear differences in B-chain behaviour We speculate that variations within domain 2 (C-terminal half) of these lectins may be relevant for the different patterns of cell surface binding that are likely to influ-ence receptor clustering, entry of these toxins into cells and ultimately their toxicities Further studies aim to investigate the proposed structure–function relation-ships experimentally
Experimental procedures
Abrus pulchellus seeds were obtained from a plant culti-vated in the garden of our laboratory, in Sa˜o Carlos-SP, Brazil Escherichia coli DH5-a (Promega, Madison, WI, USA) was used for plasmid amplification Oligonucleotide synthesis was produced by IDT, Inc (Coralville, IA, USA) Restriction endonucleases and DNA ladders were obtained from Promega Immobilized d-galactose was purchased from Pierce (Rockford, IL, USA) Mono Q 5⁄ 50 and Mono
P 5⁄ 50 were purchased from GE Healthcare (GE Health-care, Little Chalfont, UK) Sugars were purchased from Sigma (St Louis, MO, USA) All other chemicals used were
of analytical grade
Isolation of pulchellin isoforms
Dehulled seeds of A pulchellus were ground in a mixer and the fine flour obtained was suspended (1 : 10, w⁄ v) in the extraction buffer (20 mm Tris–HCl, pH 8, containing
150 mm NaCl) for 3 h at 4C and centrifuged at 12 000 g for 20 min at 4C The supernatant was loaded onto an immobilized d-galactose column, previously equilibrated with the same buffer The unbound material was eluted from the column with the extraction buffer, whereas the adsorbed proteins were obtained in a single peak after elu-tion with a soluelu-tion containing 150 mm NaCl and 100 mm lactose The fractions containing the pulchellin were dia-lyzed against 5% acetic acid in order to remove the lactose
To isolate the isoforms, the samples were dialyzed against
20 mm Tris–HCl, pH 8, containing 10 mm NaCl (buffer A)
Trang 9and 1 mL of sample containing approximately 1 mg was
loaded onto a Mono Q 5⁄ 50 column previously
equili-brated with the same buffer Elution was performed with a
linear gradient of buffer B (20 mm Tris–HCl, pH 8,
con-taining 1 m NaCl) from 0% to 20% for 40 min followed by
20–100% in 5 min The corresponding peaks of P I and
P I) were collected, dialyzed and freezed The peaks related
to the P III and P IV were dialyzed against 10 mm sodium
phosphate, pH 7 (buffer A from Mono P) and submitted to
a second chromatographic step in a Mono P 5⁄ 50
chro-matofocusing column previously equilibrated with the same
buffer One milliliter samples containing P III and P IV
(approximately 0.5 mg) were isolated by an elution gradient
of 0–100% of buffer B (10 mm sodium phosphate, pH 5.7)
for 20 min, holding for 5 min in 100% buffer B The
corre-sponding peaks of P III and P IV were collected, dialyzed
and freezed In the two last chromatographic steps (ion
exchange and chromatofocusing), the flow rate was
main-tained at 1 mLÆmin)1, the protein level was monitored at
280 nm and the pressure was maintained under 5.5 MPa
SDS⁄ PAGE was used to monitor the isolation as well as
the estimation of the apparent molecular weights and
struc-tural properties of the pulchellin isoforms
Isoeletric focusing
Isoelectric focusing of the proteins was carried out on Phast
System (Pharmacia, Uppsala, Sweden) Samples
reconstitu-ted in MilliQ water were applied to Phast Gel IEF, pH 3–9,
and run according to the standard program Gels were
stained with Comassie brilliant blue The range of pI values
of each protein was estimated by using standard markers
CD measurements
CD experiments were performed on Jasco J-715
Spectro-polarimeter (Jasco Inc., Tokyo, Japan) equipped with a
thermoelectrically controlled cell holder CD spectra of the
four isolated isoforms were measured in the far-UV range
(195–250 nm) in 1 mm path length quartz cuvettes,
recorded as the average of 16 scans CD spectra were
measured in 0.125 mgÆmL)1 of protein solution (20 mm
Tris, pH 8, 10 mm NaCl added) Analyses of the protein
CD spectra for determining the secondary structure
fractions were performed using cdpro software, comprising
the three programs: selcon3, cdsstr and contin [30]
CD thermal scans were used to examine the melting
tem-perature of the proteins Spectra were measured at 5C
intervals in the temperature range 20–100C with an
aver-age time of 3 s, an equilibration time of 3 min, and a band
width of 1 nm The CD signal at 232 nm was recorded as a
function of temperature, h232 (T) The wavelength 232 nm
was chosen because of the maximal difference between the
denatured and the native protein spectra observed at this
wavelength
cDNA cloning and amino acid sequence dedution
of the isoforms from A pulchellus
Total RNA was isolated from immature A pulchellus seeds previously frozen in liquid nitrogen, using the RNAeasy Plant Mini Kit (Qiagen, Hilden, Germany) Total RNA was quantified at 260 nm (Hitachi U-2000 spectrophotometer; Hitachi, Vienna, Austria) RT-PCR (Super Script Choice System for DNA Synthesis, Gibco BRL., Paisley, UK) was performed in two steps In the first step, for cDNA single strand synthesis, 600 ng of RNA, 0.5 lg of oligo(dT) primer and 10 mm of dNTPs were incubated for 5 min at 65C Subsequently, 4 lL of the first strand buffer 5 X and 2 lL
of dithiothreitol (0.1 m) was added and the reaction was incubated for 2 min at 42C Finally 1 lL of Superscript II was added and the reaction was incubated for an additional
1 h at 65C After the cDNA synthesis, the reaction was precipitated with ethanol [31] In the second RT-PCR step,
in order to isolate and amplify the cDNAs of the pulchellin isoforms, the whole amount of the cDNA obtained in the reaction described above was used Several primer designs, based on the N-terminal amino acid sequence of the iso-forms and on the DNA sequence of pulchellin A- [15] and B- [16] chains, were tested These included: pair 1: primer
AC-3¢) and primer anti-sense Oligo (dT)12–18 (Invitrogen, Carlsbad, CA, USA) and pair 2: primer sense Nterm (5¢-ATG GAC AAA ACT TTG AAR CTA CTG ATT TTA TG-3¢) and anti-sense Cterm (5¢-TTA AAA CAA AGT AAG CCA TAT TTG RTT NGG YTT-3¢) The reaction mixtures [75 mm of MgSO4, 100 pmol of each primer,
10 mm of dNTPs (Promega), 5 lL of buffer HiFi 10 X (Invitrogen), 2 U of Taq Platinum (Invitrogen), and MilliQ water to a final volume of 50 lL] were submitted to PCR The conditions were initial denaturation of 2 min at 94C followed by 40 cycles of denaturation (94C for 30 s), annealing (50C for 30 s) and extension (68 C for 2 min) and a final extension of 68C for 7 min
The amplified products were resolved in agarose gels and the DNA was eluted using Perfectprep Gel Cleanup kit (Eppendorf, Westbury, NY, USA) For the ligation mix-ture, 25 ng of each amplified product were ligated to 50 ng
of TOPO-TA (pCR 2.1) (Invitrogen) following the manu-facturer’s protocol E coli DH5- a cells were transformed [31], plasmids were isolated and the positive clones were screened by EcoRI digestion
DNA sequencing
Plasmids were sequenced [32] using an ABI Prism 377 auto-mated DNA sequencer (Perkin Elmer, Waltham, MA, USA) following the manufacturer’s protocol The primers used for sequencing each whole cDNA sequence were M13 Forward and Reverse (Invitrogen), and six primer walkers: Asense (5¢-CTA GGG TTA CAG GCC TTG AC-3¢), Bsense ⁄ Xho
Trang 10(5¢- CCG CTC GAG TTA AAA CAA ATG AAG-3¢),
CTC-3¢) Each isoform whole sequence was submitted to a
BLAST script databank search [33], which returned the
highest sequence identity to preproabrin The predicted
protein sequence was aligned using clustal w software
(http://www.ebi.ac.uk/clustalw/) in the default set up to
prep-roabrin, proricin and mistletoe lectin I for identity analysis
The nucleotide sequences of the four isoforms were
(EU008735, EU008736, EU008737 and EU008738, for P I,
P II, P III and P IV respectively)
Amino acid sequence analysis
Samples of each pulchellin isoform were submitted to
SDS⁄ PAGE and electroblotted on a poly(vinylidene
difluo-ride) membrane Polypeptides were excised from the blots
and the N-terminal region was sequenced on an Applied
Biosystems model 477A protein sequencer interfaced with
(Applied Biosystems, Weiterstadt, Germany) The standard
Edman degradation procedure was used [34]
LC-MS⁄ MS analysis of tryptic peptides
Pulchellin isoforms (P I, P II, P III and P IV) (100 lg) were
desalted and dried in a SpeedVac SPD12P concentrator
(Thermo Savant, Holbrook, NY, USA) The samples were
solved in 25 lL of 50% (v⁄ v) acetonitrile and 50 mm
NH4HCO3; subsequently 5 lL of 45 mm dithiothreitol were
added to each sample After incubation for 1 h at 56C,
5 lL of 100 mm iodoacetamide were added followed by 2 h
of incubation in the dark at room temperature After
five-fold dilution with 100 mm NH4HCO3, samples were treated
with 2 U of trypsin (sequencing grade, modified, Promega)
LC-MS⁄ MS analyses were performed in a Q-TOF ultima
API mass spectrometer (Micromass, Manchester, UK)
cou-pled to a capillary liquid chromatography system (CapLC,
Waters, Milford, MA, USA) A nanoflow ESI source was
used with a lockspray source for lockmass measurement
during all chromatographic runs The digested protein was
desalted online using a Waters Opti-Pack C18 trap column
The mixture of trapped peptides was then separated by
elu-tion with a water⁄ acetonitrile ⁄ formic acid gradient through
a Nanoease C18 (75 lm inner diameter) capillary column
The column was washed with 90% A solution (0.1% formic
acid) and 10% B solution (90% acetonitrile with 0.1%
formic acid) for 20 min Peptides were eluted by a 60 min
linear gradient from 10–50% B solution holding for 40 min
in 50% B Data were acquired in a data-dependent mode,
and multiplycharged peptide ions (+2 and +3) were automatically mass selected and dissociated in MS⁄ MS experiments Typical LC and ESI conditions were: flow of
200 nLÆmin)1, nanoflow capillary voltage of 3 kV, block temperature of 100C and a cone voltage of 100 V The MS⁄ MS spectra were processed using proteinlynx 2.0 software (Waters) Search parameters used the fixed cys-teine carbamidomethylation and the variable methionine oxidation as modifications The PKL file generated was used to perform a database search using the deduced pep-tide sequences provided by the sequences previously cloned
Cytotoxicity assays
HeLa cells were maintained in DMEM⁄ fetal bovine serum (10%) Cells were seeded at 1.5· 104⁄ well in a 96-well tis-sue culture plate, allowed to grow overnight and incubated for 4 h with 100 mL DMEM⁄ fetal bovine serum containing graded concentrations of pulchellin isoforms Subsequently, cells were washed twice with NaCl⁄ Pi and incubated in NaCl⁄ Pi containing 10 lCiÆmL)1 [35S] methionine for
30 min Labelled proteins were precipitated with three washes in 5% (w⁄ v) trichloroacetic acid and the amount of radiolabel incorporated was determined after the addition
of 100 mL⁄ well of scintillation fluid, by scintillation count-ing in a Micro-Beta 1450 Trilux counter (Perkin Elmer, Waltham, MA, USA) For each value, the level of protein synthesis was taken as a percentage of toxin-free control cells, and the mean from four replicate samples was calcu-lated Where appropriate, toxins were pre-incubated with increasing concentrations of N-acetyl-d-galactosamine
Toxicity to mice
The toxic activity of the pulchellin isoforms was determined
by simple intraperitoneal injection in female Swiss mice All animal procedures were performed in accordance with the
US National Research Council’s guidelines for care and use
of laboratory animals and this work was approved by the Animal Experimentation Ethics Commission of the Federal University of Sa˜o Carlos The protein samples were dilluted
in buffer 20 mm Tris–HCl, pH 8, containing 10 mm NaCl Groups of six animals were used in each different dose and the group received the same proportion of toxin⁄ body mass The mice had free access to food and water and tests were carried out over 48 h The results were evaluated as the median lethal dose Each assay had an animal control that received only the dilution buffer described above
Hemagglutination and hemagglutination-inhibition assays
Hemagglutination assays were carried out using normal human (A+, B+and O+), horse and rabbit erythrocytes in