Generation of pokeweed antiviral protein mutations in Saccharomyces cerevisiae: evidence that ribosome depurination is not sufficient for cytotoxicity

Abstract

Pokeweed antiviral protein (PAP) is a ribosome-inactivating protein that depurinates the highly conserved α-sarcin/ricin loop in the large rRNA. Here, using site-directed mutagenesis and systematic deletion analysis from the 5′ and the 3′ ends of the PAP cDNA, we identified the amino acids important for ribosome depurination and cytotoxicity of PAP. Truncating the first 16 amino acids of PAP eliminated its cytotoxicity and the ability to depurinate ribosomes. Ribosome depurination gradually decreased upon the sequential deletion of C-terminal amino acids and was abolished when a stop codon was introduced at Glu-244. Cytotoxicity of the C-terminal deletion mutants was lost before their ability to depurinate ribosomes. Mutations in Tyr-123 at the active site affected cytotoxicity without altering the ribosome depurination ability. Total translation was not inhibited in yeast expressing the non-toxic Tyr-123 mutants, although ribosomes were depurinated. These mutants depurinated ribosomes only during their translation and could not depurinate ribosomes in trans in a translation-independent manner. A mutation in Leu-71 in the central domain affected cytotoxicity without altering the ability to depurinate ribosomes in trans and inhibit translation. These results demonstrate that the ability to depurinate ribosomes in trans in a catalytic manner is required for the inhibition of translation, but is not sufficient for cytotoxicity.

INTRODUCTION

Pokeweed antiviral protein (PAP) is a 29 kDa ribosome-inactivating protein (RIP) isolated from the leaves of the pokeweed plant (Phytolacca americana). PAP and other RIPs such as ricin, Shiga toxin from Shigella disentaria and Shiga-like toxin from Escherichia coli catalytically remove an adenine (A4324) residue from the highly conserved sarcin/ricin loop (S/R) of the large rRNA (1–5). PAP can also remove an additional adenine and a guanine from the S/R loop (6). The depurination of the S/R loop has been reported to interfere with the elongation factor-2 catalyzed GTP hydrolysis and translocation of the peptidyl-tRNA to the P-site, resulting in the inhibition of translation (4,7). However, the role of the depurination of the S/R loop in the inhibition of translation is not fully understood. Recent results indicate that the prokaryotic initiation factor 2 (IF2), a homolog of the newly discovered eukaryotic translation initiation factor 5B (eIF5B) binds to the S/R loop (8). IF2 and eIF5B are required for joining of the ribosomal subunits during the initiation of protein synthesis (9), suggesting that the S/R loop may also be important in initiation. Therefore, analysis of how RIPs interact with the ribosome may lead to a better understanding of the role of the highly conserved S/R loop in translation.

RIPs have become important agents in agriculture and medicine mainly by virtue of their broad-spectrum antiviral activity and cytocidal properties against cancer cells (10–12). PAP and ricin have been used as the cytotoxic component of immunotoxins directed against cancer cell targets (13–16). The potent toxicity of RIPs has been exploited in biological warfare and more recently they have been used as potential bioterrorism threats (17). Understanding how RIPs interact with ribosomes and identifying the amino acids that are involved in these interactions are critical not only for protecting healthy cells from their cytotoxic effects in therapeutic applications but also for developing antidotes against their action. Substantial effort has been made to identify the amino acids involved in the chemistry of catalysis of RIPs using site-directed mutagenesis focusing on residues that are invariant among plant and bacterial RIPs. These amino acids include Glu-176 (Glu-177 in ricin), Arg-179 (Arg-180 in ricin), Tyr-72 (Tyr-80 in ricin) and Tyr-123 (Tyr-123 in ricin) in PAP (18–20). The three-dimensional X-ray structure indicates that PAP is composed of eight α-helices and a β-sheet consisting of six strands (21–23). The protein has been divided into three distinct domains: the N-terminal domain (residues 1–69), the central domain (residues 70–179) and the C-terminal domain (residues 180–262) (22,23). The highly conserved active site residues, Glu-176 and Arg-179, are located in the central domain. The amino acids Tyr-72 and Tyr-123 have been proposed to sandwich the susceptible adenine ring of rRNA into the energetically favorable conformation (20,21,24). Subsequently, the side chain of Arg-179 protonates the N-3 atom of the adenine, while Glu-176 stabilizes a positive oxocarbonium transition state (21,24). In the three-dimensional structure, there is a prominent cleft at the interface between the central and C-terminal domains, which forms the putative substrate-binding site (22). Mutagenesis of the residues of the cleft has been shown to reduce the depurination and ribosome inhibitory activity of PAP when recombinant mutant proteins were expressed in E.coli and ribosome depurination was assayed in vitro (25).

Studying various activities of PAP in vivo in plant cells is difficult due to the extreme toxicity of the protein. The yeast, Saccharomyces cerevisiae, has proven to be a powerful tool for genetic and biochemical characterization of PAP because the expression of PAP can be tightly controlled (26). Ribosomes from yeast are only 2-fold more sensitive to PAP than ribosomes from tobacco (27). PAP mutants behave similarly in yeast and in transgenic plants with respect to their toxicity and ability to depurinate ribosomes (28). When PAP cDNA was expressed in yeast from the GAL1 promoter, cell growth was inhibited (26). This positive selection system was used to isolate PAP mutants that did not inhibit cell growth. A single point mutation in the highly conserved Glu-176 (E176V) resulted in a non-toxic protein that did not depurinate ribosomes or inhibit translation in vitro (26). Point mutations near the N-terminus or a nonsense mutation at W237 produced non-toxic proteins that were able to inhibit translation by depurinating the RNA templates in vitro (6,29). Using a highly sensitive primer extension assay, we showed that these mutants did not depurinate yeast ribosomes in vivo (30). These results indicated that the ability to inhibit protein synthesis in vitro does not always correlate with the ability of the mutants to depurinate the rRNA in vivo. Our recent results provided insights into the mechanism of translation inhibition by demonstrating that PAP can inhibit translation in vitro by binding to the cap structure and depurinating the capped RNAs (6,29). These studies highlight the importance of evaluating the cytotoxicity of PAP mutants and their ability to depurinate ribosomes in vivo.

To identify residues that are important for ribosome depurination and cytotoxicity of PAP, site-directed mutagenesis was used to make systematic deletions from the N-and C-termini of mature PAP and to introduce point mutations. The plasmids harboring the mutated cDNAs downstream of the GAL1 promoter were transformed into yeast. The correlation between ribosome depurination, translation inhibition and cytotoxicity was examined. Here, we identify the PAP residues that are critical for ribosome depurination, inhibition of translation and cytotoxicity, and demonstrate for the first time that ribosome depurination is not sufficient for the inhibition of translation and cytotoxicity.

MATERIALS AND METHODS

Mutagenesis

Mutations of PAP cDNA were introduced using the Stratagene QuikChange™ site-directed mutagenesis kit. Mutations were introduced into a set of oligomeric primers that were used for PCR amplifications of the template plasmid with wild-type PAP in pGEM (Promega) background by Pfu DNA polymerase. After PCR amplification, the template plasmid was removed by DpnI digestion. The mutated plasmids were transformed into E.coli DH5α. Mutations were confirmed by sequencing using the T7 Sequenase version 2.0 sequencing kit (USB). The cDNAs encoding the mutants were subcloned as BamHI–HindIII fragments into the yeast expression vector YEp351 under the control of a galactose-inducible GAL1 promoter.

Yeast transformation

Yeast strain W303 (MAT α, ade2-1 trp1-1 ura3-1 leu2-3, 112 his3-11, 15 can1-1000) was transformed with YEp351, containing the mutated PAP sequences. One-half of the transformed yeast suspension was plated onto SD-Leu supplemented with 2% dextrose, and the other half was plated onto SD-Leu, containing 2% galactose. Toxicity of the PAP mutants was verified by re-plating the selected colonies onto both 2% raffinose and 2% galactose.

Growth curves

Yeast transformed with either wild-type or mutant forms of PAP were grown in the synthetic SD-Leu medium supplemented with 2% raffinose in a total volume of 100 ml until an A₆₀₀ of 0.6. Yeast cells were pelleted by centrifugation at 2000 g for 5 min, washed with SD-Leu medium and resuspended in SD-Leu medium containing 2% galactose to induce the expression of PAP or PAP mutants. At zero time (immediately following induction) and at each hour following induction, an aliquot of 1 ml was removed and A₆₀₀ was measured.

Yeast protein expression analysis

Yeast containing cDNAs encoding PAP or PAP mutants were grown as described for growth curves, in a 10 ml volume and induced with 2% galactose for 6 h. Protein expression analysis was carried out as described previously (26,31). Briefly, cells were pelleted by centrifugation at 2000 g for 5 min. Pellets were resuspended in an equal volume of cold (4°C) Buffer X (25 mM Tris–HCl, pH 7.5, 100 mM sodium vanadate, 10 mM β-glycophosphatase, 1 mM EGTA, 1 mM DTT, 1 mM PMSF and 5% glycerol) and 0.3 g of 0.5 mm diameter glass beads. The cells were vortex-mixed for 2 min and were centrifuged at 16 000 g for 5 min. Supernatant total protein was quantified by Bradford assay using BSA as a standard. Total protein (10 μg) was separated through 12% SDS–PAGE, transferred onto nitrocellulose membrane and blocked by incubation with PBST (phosphate buffered saline with 0.1% Tween-20) in 5% non-fat milk for 2 h. Proteins were probed by overnight incubation with an affinity-purified polyclonal antibody to PAP (1:5000) in PBST–5% milk and secondary goat anti-rabbit immunoglobulin G (IgG) conjugated to horseradish peroxidase (1:5000) in PBST–5% milk for 1.5 h. Mutant PAP proteins were visualized by chemiluminescence using a Renaissance kit (NEN, DuPont). To confirm equal loading of total protein, blots were stripped with 8 M guanidine hydrochloride and re-probed with a polyclonal G6PD antibody (1:5000; Chemicon, Temecula, CA) and horseradish peroxidase-conjugated secondary donkey anti-goat IgG (1:5000).

rRNA depurination assay

Yeast cells (100 ml) grown as described for growth curves were harvested following a 6-h induction of PAP and PAP mutants and were used to isolate ribosomes as described previously (31). To determine whether PAP mutants depurinated the S/R loop when expressed in vivo, primer extension analysis was performed essentially as described in Hudak et al. (30). Purified ribosomal RNAs (1 μg) were incubated with a 5′ ³²P-end-labeled oligonucleotide primer (5′-GGCGTTCAGCCATAATCC-3′) complementary to the 3′ end of yeast 25S rRNA. The total reaction volume was 15 μl, to which 5 μl of formamide buffer was added to stop the extension without the precipitation of RNA and resulting cDNA. An aliquot of this reaction (4 μl) was separated on a 6% polyacrylamide/7 M urea gel and visualized by autoradiography. To determine the position of rRNA depurination, a sequencing ladder of DNA corresponding to the yeast 25S rRNA was separated on the same gel (6,30).

To determine whether depurination occurs in cis or trans, yeast cells expressing wild-type PAP or PAP mutants were grown overnight in SD-Leu containing 2% raffinose to an A₆₀₀ of 0.6. The cells were then diluted to A₆₀₀ of 0.3 and resuspended in SD-Leu containing 2% galactose to induce PAP expression. After 1 h of galactose induction (Gal1), aliquots were removed for protein (1 ml) and RNA (9 ml) analysis. Dextrose (20%) was added to a final concentration of 2%, and 10 mg/ml cycloheximide was added to a final concentration of 0.1 mg/ml. At 2, 4 and 6 h of growth after the initial induction, a 9 ml aliquot was removed for RNA analysis and an 1 ml aliquot was removed for protein analysis. A parallel culture was induced on galactose for 6 h and was sampled in the same way. Protein expression was analyzed by immunoblot analysis. The extent of depurination was quantified using a dual primer extension assay as described previously (32).

In vivo [³⁵S]methionine incorporation

Yeast cells were grown to an A₆₀₀ of 0.6 in SD-Leu, -Met, 2% raffinose. The cells were then resuspended at an A₆₀₀ of 0.3 in 2% galactose for 4 h in order to induce either wild-type PAP or mutant PAP expression. At time zero, [³⁵S]methionine was added to cells growing on galactose. At various timepoints, 800 μl of yeast cells were removed for growth measurements, and additional aliquots of 800 μl were assayed for methionine incorporation in triplicate as described by Parikh et al. (32). Briefly, the yeast was added to 200 μl of 100% trichloroacetic acid (TCA), incubated for 10 min on ice, followed by 20 min at 70°C. The precipitate was then filtered through 24 mm glass microfiber filters (VWR), washed with ice-cold 5% TCA followed by ice-cold 95% ethanol. Filters were dried for several hours and incorporation was quantified in a scintillation counter. The CPM was normalized to the A₆₀₀ reading. Rates of translation were determined from these results and tabulated as CPM per A₆₀₀ per minute. The rates were determined for each construct in a minimum of two independent experiments compared with the rates of the vector control, NT616, which contains the same vector as wild-type PAP with the luciferase gene.

RESULTS

Mutations within the N-terminal domain of PAP

PAP is synthesized as a 33 kDa precursor in pokeweed plants, with a 22 amino acid signal sequence and a 29 amino acid C-terminal extension (26). The 29 kDa mature protein is secreted into the cell wall matrix via the exocytotic vesicular pathway in pokeweed (33). To identify the residues that are important for cytotoxicity within the N-terminal domain of PAP, site-directed mutagenesis was used to make systematic deletions. N-terminal deletions were made by introducing a Met codon and deleting the residues upstream of the Met by introducing a BglII site.

Cytotoxicity of each mutant was analyzed by its ability to grow on plates containing galactose. As shown in Table 1, the deletion of the N-terminal signal peptide and 14 amino acids from the N-terminus of mature PAP in S14M did not alleviate the cytotoxicity observed when synthesis of wild-type PAP was induced in yeast. The level of rRNA depurination in each mutant after 6 h of induction in yeast was analyzed by primer extension in Figure 1A and was expressed as a percentage of the level of depurination in wild-type PAP in Table 1. As shown in Table 1, the level of depurination in S14M was 70% of wild-type PAP. Similarly, the substitution of Lys-15 with Met in K15M produced a protein that was cytotoxic and depurinated ribosomes to 79% of wild-type levels. These results indicated that the deletion of the N-terminal signal sequence and the first 15 amino acids from the N-terminus of mature PAP did not have a major effect on cytotoxicity or the ability to depurinate ribosomes. In contrast, changing Tyr-16 to Met in Y16M generated a deletion mutant that was not cytotoxic and depurinated ribosomes only to 6% of the wild-type PAP level. Deletion of the residues downstream of Tyr-16, such as T18M, resulted in a non-toxic protein that again depurinated ribosomes to 7% relative to the wild-type PAP. As the deletion of the first 15 residues did not affect cytotoxicity or the ability to depurinate ribosomes, tyrosine at position 16 may be critical for these properties. To address this, we constructed point mutations of the tyrosine to either Ala (Y16A) or Phe (Y16F). Point mutations at Tyr-16 created proteins that were cytotoxic and depurinated ribosomes, indicating that Tyr-16 is not entirely responsible for these characteristics (Table 1 and Figure 1A). Similarly, point mutations at Lys-15 (K15A) and Ser-14 (S14A) created proteins that were cytotoxic and depurinated ribosomes (Table 1 and Figure 1A). The doubling time of cultures grown in galactose-containing liquid medium was calculated from growth curves and was used to confirm plate growth results. The doubling time of Y16M was 7.1 h, similar to the doubling time of E176V, the active site mutant at 6.4 h, rather than wild-type PAP with a doubling time of 10.4 h. In contrast, doubling times of the cytotoxic mutants S14A, K15A and Y16A were 10.2, 10.2 and 9.5 h, respectively, which were similar to the doubling time of wild-type PAP.

Table 1. Effect of mutations on the cytotoxicity of PAP, its ability to depurinate ribosomes and to inhibit translation.

Expression vector	Mutation	Cytotoxicity	Depurination (% of wild-type)	Doubling time (h)	Translation (% vector control)
N-terminal domain mutants
NT418	S14M	Yes	Yes (70)	7.8	51.5
NT413	K15M	Yes	Yes (79)	10	32.6
NT542	Y16M	No	No (6)	7.1	82.1
NT549	T18M	No	No (7)	7.3	77.8
NT525	S14A	Yes	Yes (106)	10.2	35.3
NT360	K15A	Yes	Yes (83)	10.2	37.9
NT558	Y16A	Yes	Yes (97)	9.5	32.3
NT548	Y16F	Yes	Yes (103)	11	37.3
Central domain mutants
NT502	N69A	Yes	Yes (103)	10.5	22.5
NT501	N70A	No	Yes (96)	7	71.7
NT538	L71R	No	Yes (105)	8.5	33.2
NT241	Y72A	No	Yes (28)	6.5	88.9
NT532	V73E	No	Yes (90)	7.5	73.1
NT533	M74R	No	Yes (103)	8.5	22.4
NT255	G75D	No	No (0)	6.4	100
NT534	Y76A	No	Yes (99)	10	32.8
NT503	D92A	Yes	Yes (101)	10	25
NT242	Y123A	No	Yes (61)	7.1	88.8
NT483	Y123F	Yes	Yes (98)	11.4	27.5
NT485	Y123I	No	Yes (81)	6.9	56.5
NT224	E176V	No	No (0)	6.4	75.2
C-terminal domain mutants
NT246	W237*	No	No (0)	6	70.7
NT509	L240*	No	No (0)	6.5	82.5
NT552	R241*	No	No (0)	6.2	81.6
NT510	V242*	No	No (5)	7	84.8
NT333	E244*	No	No (3)	6.4	90.1
NT486	A250*	No	Yes (43)	6.8	71.5
NT347	L251*	No	Yes (35)	6	94.5
NT420	L252*	No	Yes (39)	6	96.3
NT456	N253*	Yes	Yes (64)	7	68.6
NT443	Y254*	Yes	Yes (100)	7.8	41.1
NT233	T262*	Yes	Yes (96)	10.5	38.9
NT457	L252K	Yes	Yes (94)	8	32.8
NT232	C259A	Yes	Yes (97)	7.5	36.9
Control
NT188	WT	Yes	Yes (100)	10.4	30.3
NT616	VC	No	No (0)	6	100

^*Denotes introduction of a stop codon at the indicated position.

Figure 1.

Analysis of ribosome depurination in yeast. Yeast expressing wild-type PAP (WT), the vector without PAP (VC), the N-terminal (A), C-terminal (B) or the central domain (C) mutants was grown for 6 h to induce PAP expression, and rRNA isolated from each culture was incubated with an end-labeled primer complimentary to the 3′ end of the yeast 25S rRNA. The resulting fragments representing the primer extension products that have stopped prematurely at the depurination site are indicated with an arrow. To determine the position of depurination in the rRNA, a sequencing ladder of DNA corresponding to the yeast 25S rRNA was separated on the same gel.

Immunoblot analysis shown in Figure 2A indicated that protein from N-terminal mutants was detected following 6 h of induction in yeast, though the expression patterns varied. Wild-type PAP expressed in yeast is present in two forms, the mature protein at 29 kDa, which co-migrates with the purified protein from plants and a higher molecular weight form (33 kDa), which co-migrates with the precursor form of PAP (26). As shown in Figure 2A, protein expressed in yeast harboring the S14A and the Y16F mutants showed the two forms previously observed with wild-type PAP (26,31,32). The N-terminal deletion mutants, S14M, Y16M and T18M migrated slower on SDS–PAGE than the mature form of PAP (Std). As the signal sequence and the N-terminal amino acids are deleted in these mutants, their slower migration on SDS–PAGE suggests that they are not processed at their C-termini. The N-terminal deletion mutants were expressed at lower levels than the point mutants, possibly due to the destabilization of the protein. Higher levels of protein accumulated in Y16M and T18M, which were not toxic and did not depurinate ribosomes than in S14M, which was toxic and depurinated ribosomes, indicating that the lack of ribosome depurination and cytotoxicity observed in yeast expressing Y16M and T18M is not due to lower level of protein expression.

Figure 2.

Immunoblot analysis of PAP mutants. Total protein (10 μg) from yeast expressing the N-terminal (A), C-terminal (B and C) and central domain (D and E) mutants was separated on 12% SDS–PAGE. Proteins were transferred onto nitrocellulose membrane and probed with polyclonal PAP antiserum (1:5000). Purified PAP (9 ng) from pokeweed leaves was used as a standard (Std). The immunoreactive bands corresponding to mature and precursor forms of PAP are indicated by arrows. The blots were subsequently stripped and re-probed with polyclonal antibodies against glucose 6-phosphate dehydrogenase (G6PD) (1:1000) as loading controls.

The ability of PAP mutants to inhibit total translation in vivo in yeast was analyzed by [³⁵S]Met incorporation as described previously (32). Total translation was measured at 6 h after induction in yeast expressing wild-type PAP or PAP mutants and was expressed as a percentage of total translation in cells harboring the vector (VC) in Table 1. As observed previously, the PAP expression in yeast reduced the total translation to 30.3% of the vector control levels (32). In contrast, the total translation was at 82.1% of the vector control level in Y16M and at 77.8% of the vector control level in T18M, which was similar to the level of translation (75.2%) observed with the inactive PAP, E176V (Table 1). These results indicated that unlike wild-type PAP, which inhibited translation, Y16M did not inhibit total translation in yeast, demonstrating that the first 16 amino acids of PAP are critical for ribosome depurination, translation inhibition and cytotoxicity.

Mutations within the C-terminal domain of PAP

Sequential deletions were made from the 3′-terminus of PAP by introducing stop codons. Analysis of the C-terminal deletion mutants indicated that a reduction in cytotoxicity was positively correlated with a decline in ribosome depurination. The first mutation, T262^*, which deleted the C-terminal extension and the final amino acid of the mature protein was cytotoxic and depurinated ribosomes to 96% of wild-type PAP (Figure 1B and Table 1). This result was expected, given that the mature form of PAP expressed in yeast is cytotoxic and enzymatically active. As shown in Figure 2B, this mutant produced a protein that co-migrated with the mature form of PAP, providing evidence that the upper form observed on SDS–PAGE is processed at its N-terminus, but not at its C-terminus (Figure 1A). Sequential deletion of amino acids upstream of Thr-262 until Leu-252 produced truncated proteins that were cytotoxic and depurinated ribosomes (Table 1). However, increased deletion size resulted in decreased level of depurination relative to wild-type PAP (Figure 1B). For example, Y254^* depurinated ribosomes to wild-type levels, whereas N253^* depurinated ribosomes to 64% of wild-type PAP. Cytotoxicity was lost when L252 and the remaining C-terminal amino acids were deleted (Table 1). The L252^* was capable of depurinating ribosomes, albeit at lower levels than wild-type PAP (39%). Similarly, L251^* was not cytotoxic and depurinated ribosomes to only 35% of the wild-type levels. The lack of toxicity of both L252^* and L251^* may be due to their lower levels of depurination relative to wild-type PAP, suggesting that a threshold level exists at which yeast cells will tolerate some degree of ribosome depurination without a reduction in overall growth. Depurination activity of PAP gradually decreased upon sequential deletion of C-terminal amino acids and was abolished when a stop codon was introduced at E244 (Figure 1B). Sequential deletion analysis from the C-terminus of PAP also indicated that cytotoxicity was lost prior to the ability to depurinate ribosomes. The substitution of L252 for Lys (L252K) did not alter cytotoxicity or the depurination ability of PAP (Table 1), indicating that L252 alone is not responsible for cytotoxicity. Rather, our results suggest that L252 and residues downstream are important determinants of cytotoxicity. This observation is supported by the sequential increase in both toxicity and depurination observed between L252^* and Y254^* (39 and 64%, respectively).

PAP contains two disulfide bonds, one between residues Cys-34 and Cys-259 and the other between Cys-85 and Cys-106 (21). To determine whether the disulfide bond involving the amino acid within the C-terminus (Cys-259) was required for cytotoxicity or depurination activity, C259A was expressed in yeast. This protein maintained its cytotoxicity and depurinated ribosomes to 97% relative to wild-type PAP (Figure 1B and Table 1). Therefore, maintenance of this disulfide bond is not essential for ribosome depurination or cytotoxicity.

Immunoblot analysis shown in Figure 2B indicated that the C-terminal deletion mutants produced proteins that had similar mobility as the mature PAP. As the size of the C-terminal deletion increased, proteins decreased in size relative to the mature PAP (Figure 2C). This decrease in mobility was expected given that mutants such as L240^* and R241^* were missing 23 and 22 amino acids from their C-terminus, respectively. The C-terminal deletion mutants migrated on SDS–PAGE consistent with their size, indicating that they are processed at their N-termini. The deletion mutants, such as A250^*, L251^* and L252^* that were not cytotoxic and caused a low level of depurination, expressed more protein than the mutants like Y254^* and T262^* that were cytotoxic and depurinated ribosomes at similar levels as wild-type PAP. These results demonstrate that the lack of cytotoxicity and the lower level of depurination observed with the C-terminal deletion mutants are not due to the lower levels of protein expression.

Analysis of total translation by [³⁵S]Met incorporation indicated that the total translation in L251^* and L252^* was at 94.5 and 96.3% of the vector control levels, respectively (Table 1). These results indicated that the total translation was not inhibited in yeast expressing L251^* and L252^*, which depurinated ribosomes to 35 and 39% of wild-type levels, respectively. The doubling time of L251^* and L252^* at 6 h was similar to the inactive PAP, E176V, at 6.4 h, demonstrating that these mutants did not inhibit growth. In contrast, total translation was inhibited in Y254^*, T262^*, L252K and C259A, which depurinated ribosomes at similar levels as wild-type PAP (Table 1 and Figure 1B).

Mutations within the central domain of PAP

Based on the X-ray structure of PAP, Tyr-72 and Tyr-123 in the central domain are proposed to be involved in binding the target adenine base (21). As shown in Table 1 and Figure 1C, substitution of Tyr-72 for Ala (Y72A) resulted in a protein that was not cytotoxic and depurinated ribosomes at much lower levels than the wild-type PAP. The relative degree of depurination in Y72A was 28% of wild-type PAP, which confirms earlier reports that this amino acid is important for depurination of the rRNA (21). Replacement of Tyr-123 with Ala (Y123A) again produced a mutant form of PAP, which was not cytotoxic, but exhibited a higher level of ribosome depurination (61% of wild-type levels) than the Tyr-72 mutant (Figure 1C and Table 1). Substitution of Y123 with Ile (Y123I) resulted in a mutant protein with a growth rate comparable to the active site mutant, E176V (6.9 and 6.4 h respectively). Ribosomes from cells expressing Y123I were depurinated 81% relative to wild-type PAP, without cytotoxicity, providing evidence that cytotoxicity may not be entirely due to ribosome depurination (Figure 1C and Table 1). Replacing Tyr-123 with Phe (Y123F) generated a toxic protein with a doubling time of 11.4 h. The Y123F depurinated ribosomes to 98% of the levels observed with the wild-type protein (Figure 1C and Table 1).

Immunoblot analysis shown in Figure 2D confirmed that both forms of PAP were expressed in Y72A, Y123A, Y123F and Y123I, as observed with the active site mutant, E176V. Several mutants, such as Y72A and Y123A, which caused a lower level of depurination, were expressed at higher levels than mutants like D92A and Y123F, which caused a similar level of depurination as wild-type PAP (Figure 2D). The observation was that Y123I was not cytotoxic despite 81% ribosome depurination relative to wild-type PAP, whereas Y123F was cytotoxic and depurinated ribosomes, indicated that the maintenance of the phenolic ring structure was likely important for the role of Y123 in cytotoxicity, but not in rRNA depurination.

To determine whether the lack of cytotoxicity observed with Y123A and Y123I mutants is due to their inability to inhibit translation, we analyzed total translation in cells expressing these mutants by [³⁵S]Met incorporation. As shown in Table 1, unlike wild-type PAP, which reduces translation to 30.3%, Y123A did not inhibit total translation. Total translation in yeast expressing Y123A was at 88.8% of the vector control levels at 6 h after induction. Similarly, total translation in yeast expressing Y123I was at 56.5% of the vector control levels. In contrast, total translation was reduced to 27.5% in cells expressing Y123F. These results indicated that ribosome depurination is not sufficient for the inhibition of total translation. The phenolic ring of Tyr-123 is not critical for ribosome depurination, but is critical for the inhibition of translation and cytotoxicity, suggesting that Tyr-123 does not play a major role in ribosome depurination, but may be involved in another interaction that is required for the inhibition of translation and cytotoxicity.

The sequences in the central domain of PAP reveal similarity to a ribonucleoprotein (RNP) motif present on many RNA-binding proteins (34–35). The RNP motif is characterized by conserved RNP2 and RNP1 sequences separated from each other by about 40 amino acid residues (34). Some RNA-binding proteins lack consensus RNP1 and RNP2 sequences, but have similar structurally significant residues. The sequences that resemble the RNP2 motif lie between amino acids 71 and 76 in PAP. A previously characterized mutant in this motif, G75D, did not bind ribosomes efficiently, demonstrating the involvement of Gly-75 in ribosome binding (30,36). As shown in Figure 1C and Table 1, this mutant did not depurinate ribosomes and was not cytotoxic. Point mutations in the other residues of the RNP motif, including Asn-70 (N70A), Leu-71 (L71R), Val-73 (V73E), Met-74 (M74R) and Tyr-76 (Y76A) caused ribosome depurination to similar levels as wild-type PAP, but eliminated cytotoxicity (Figure 1C and Table 1). When viability was assayed by plating different dilutions of cells on glucose plates after galactose induction in liquid media, the expression of wild-type PAP reduced the viability of cells by almost 3 logs by 10 h of post-induction. In contrast, N70A, L71R, V73E, M74R and Y76A displayed minimal loss of viability by 10 h of post-induction (data not shown). These results demonstrated that the putative RNP domain of PAP contains amino acids that are essential for depurination of rRNA (Tyr-72), binding to ribosomes (Gly-75) or for another interaction required for cytotoxicity. Immunoblot analysis shown in Figure 2E indicated that both forms of PAP are expressed in N69A, N70A and M74R. In contrast, predominantly the upper form of PAP is detected in V73E, G75D and Y76A, suggesting that the C-terminal processing of the PAP precursor is affected in these mutants.

To determine whether ribosome depurination in the central domain mutants leads to the inhibition of translation, we examined total translation in yeast by [³⁵S] incorporation. As shown in Table 1, total translation was not inhibited in N70A, although ribosomes were depurinated at wild-type levels. Total translation at 6 h post-induction was at 72% of the vector control levels, similar to the active site mutant, E176V, at 75% of the vector control levels (Table 1). Growth of this mutant was also not inhibited (Table 1), and its doubling time at 7.0 h was similar to the doubling time of the active site mutant E176V at 6.4 h. In contrast, translation was inhibited in yeast harboring D92A, which, like N70A, depurinated ribosomes at similar levels as wild-type PAP. The D92A inhibited growth of yeast cells with a doubling time of 10 h and inhibited total translation to the same extent as wild-type PAP (Table 1). Analysis of total translation in other mutants in the central domain indicated that V73E, which depurinated ribosomes at 90% of the wild-type levels, did not inhibit translation. Translation was at 73% of the vector control levels in this mutant, similar to the level of translation observed in the active site mutant, E176V (Table 1). In contrast, as reported previously, translation was inhibited in L71R (32), as well as in M74R and in Y76A (Table 1). The doubling times of these mutants were similar to wild-type PAP, consistent with the translation inhibition (Table 1).

Although Y123A depurinated ribosomes, it is possible that the timing of depurination is altered in this mutant relative to the wild-type PAP. To examine the extent of depurination that occurred during the 10-h induction, we used a previously described dual primer extension assay (32), which employs two different primers. One primer that hybridizes downstream of the depurination site (Dep.) was used to examine the extent of depurination. Another primer that hybridizes upstream of the depurination site (rRNA) was used to quantify the total amount of 25S rRNA (Figure 3A). The ratio of the depurination fragment compared with the control fragment allowed for accurate quantification of the extent of depurination (Figure 3B) (32). As shown in Figure 3A and B, the maximal level of depurination was observed 4 h after induction of wild-type PAP expression, and the extent of depurination decreased after 4 h in yeast expressing wild-type PAP. In contrast, depurination increased slowly up to 10 h during induction of Y123A expression (Figure 3A and B). These results indicated that Y123A depurinated ribosomes more slowly than wild-type PAP, suggesting that the depurination that is observed in Y123A might be occurring only during its translation.

Figure 3.

Analysis of ribosome depurination in yeast expressing PAP and Y123A. Yeast expressing wild-type PAP or Y123A were resuspended in SD-Leu plus galactose to induce PAP expression up to 10 h. (A) At indicated times after induction, total RNA was isolated and analyzed by primer extension analysis using two different end-labeled primers, the depurination primer (Dep.) used to measure the extent of depurination and the 25S rRNA (rRNA) primer used to measure the amount of 25S rRNA present at the indicated times (in hours). Primers were also extended separately as marked in the last panel. (B) The extent of depurination was quantified by calculating the ratio of the depurination fragment to the 25S rRNA fragment.

We have previously shown that PAP depurinates ribosomes in trans in a manner that is independent of translation (32). To determine whether Y123 mutants can depurinate ribosomes in trans, we induced PAP expression by growing yeast for 1 h on galactose containing medium followed by inhibition of PAP gene transcription and translation by shifting cells to medium containing glucose and cycloheximide. Aliquots were taken after 1 or 6 h of induction on galactose (Gal1 and Gal6), or after 6 h of growth on glucose plus cycloheximide (Glu6) and analyzed by immunoblot analysis. As shown in Figure 4, translation of wild-type PAP, Y123A, Y123F or Y123I was inhibited when cells were grown on glucose plus cycloheximide for 6 h (Glu6). In contrast, wild-type PAP and the Y123 mutants accumulated when cells were grown for 6 h on galactose (Gal6). Primer extension analysis shown in Figure 5A was used to analyze the extent of depurination that occurred when transcription and translation were both inhibited (Glu + CHX) and the results were quantified in Figure 5B. A low level depurination was observed with wild-type PAP after 1 h of induction on galactose (1). Depurination reached maximal levels when transcription and translation were inhibited for another hour on glucose plus cycloheximide (2 Glu + CHX) and increased up to 6 h (6 Glu + CHX) (Figure 5A and B). These results indicate that ribosome depurination occurs predominantly in trans in cells expressing wild-type PAP. In contrast, in cells expressing Y123A, very little depurination was observed when cells were shifted to media containing glucose plus cycloheximide (2 Glu + CHX). The extent of depurination that occurred in trans did not increase during 6 h of growth on glucose plus cycloheximide (6 Glu + CHX). Although similar levels of PAP protein were expressed in cells expressing wild-type PAP or Y123A (Figure 4), very little depurination occurred in trans when transcription and translation were inhibited in cells expressing Y123A (Figure 5A and B). This mutant depurinated ribosomes at a similar level as wild-type PAP only after it was grown for 6 h on galactose (Gal6). These results demonstrated that Y123A depurinated ribosomes only during its translation and could not depurinate ribosomes in trans. Furthermore, as total translation was not inhibited in yeast expressing Y123A (Table 1), ribosome depurination that occurred up to 6 h in these cells did not inhibit total translation.

Figure 4.

Immunoblot analysis of the Tyr-123 mutants. Yeast cells expressing wild-type PAP, Y123A, Y123F or Y123I were resuspended in SD-Leu plus galactose to induce PAP expression. After 1 h of galactose induction (Gal1), glucose and cycloheximide were added. After 6 h of growth on glucose plus cycloheximide (Glu6) or 6 h of growth on galactose (Gal6), aliquots were removed and subjected to immunoblot analysis using polyclonal PAP antiserum (1:5000). Purified PAP (9 ng) from pokeweed leaves was used as a standard. The immunoreactive bands corresponding to mature and precursor forms of PAP are indicated by arrows. The blot was subsequently stripped and re-probed with anti-G6PD (1:1000) antibodies as loading controls.

Figure 5.

Analysis of ribosome depurination in trans during expression of the Tyr-123 mutants. Yeast cells expressing wild-type PAP, Y123A, Y123F or Y123I were resuspended in SD-Leu plus galactose to induce PAP expression. After 1 h of galactose induction (1), glucose and cycloheximide were added. After 2, 4 and 6 h of growth on glucose plus cycloheximide (Glu + CHX) aliquots were removed and subjected to primer extension analysis (A). A parallel culture was induced on galactose for 6 h (6) and analyzed by primer extension analysis. (B) The extent of depurination was quantified by calculating the ratio of the depurination fragment (Dep.) to the 25S rRNA fragment (rRNA) and plotted.

These experiments were repeated with Y123F and Y123I. The primer extension analysis shown in Figure 5A indicated that very little depurination occurred in trans in yeast expressing Y123I when transcription and translation were inhibited for 6 h on glucose plus cycloheximide (6 Glu + CHX). Ribosomes were depurinated in Y123I only after 6 h of growth on galactose in the absence of cycloheximide (Gal6), indicating that Y123I depurinated ribosomes only during its translation. In contrast, ribosomes were depurinated when transcription and translation were inhibited in yeast expressing Y123F (6 Glu + CHX), indicating that Y123F depurinated ribosomes in trans. These results demonstrate that ribosome depurination is not always accompanied by inhibition of translation. Rather, depurination in trans is required for the inhibition of translation.

Since PAP mutants, such as L71R, inhibited translation, but were not cytotoxic, we examined ribosome depurination in L71R to determine whether it depurinates ribosomes in trans. The expression of L71R and PAP was induced by growing yeast on galactose for 1 h followed by shifting cells to medium containing glucose and cycloheximide. Aliquots were taken after 1 or 6 h of induction on galactose (Gal1 and Gal6), or after 6 h of growth on glucose plus cycloheximide (Glu6) and analyzed by immunoblot analysis. As shown in Figure 6A, translation of L71R was inhibited when cells were grown on glucose plus cycloheximide for 6 h (Glu6). In contrast, L71R accumulated when cells were grown for 6 h on galactose (Gal6). Primer extension analysis indicated a low level of depurination in L71R, as with PAP when transcription and translation were inhibited after 1 h of induction on galactose (Figure 6B and C). Depurination increased when transcription and translation were inhibited for another hour on glucose plus cycloheximide (2 Glu + CHX) and continued to increase up to 6 h (6 Glu + CHX) (Figure 6B and C). These results demonstrate that ribosome depurination occurs predominantly in trans in cells expressing L71R, providing further evidence that ribosome depurination in trans is required for the inhibition of translation.

Figure 6.

Analysis of ribosome depurination in trans during expression of the L71R mutant. Yeast cells expressing wild-type PAP or L71R were resuspended in SD-Leu plus galactose to induce PAP expression. After 1 h of galactose induction (1), glucose and cycloheximide were added. After 2, 4 and 6 h of growth on glucose plus cycloheximide (Glu + CHX), aliquots were removed and subjected to immunoblot analysis (A) or primer extension analysis (B). A parallel culture was induced on galactose for 6 h (6) and analyzed by immunoblot and primer extension analysis. The extent of depurination was quantified by calculating the ratio of the depurination fragment (Dep.) to the 25S rRNA fragment (rRNA) and plotted (C).

DISCUSSION

N-terminal deletion analysis of PAP indicated that deletion of the first 16 amino acids of PAP, including the highly conserved tyrosine (Tyr-16), eliminated its cytotoxicity and the ability to depurinate ribosomes in vivo. Tyr-16 of PAP corresponds to Tyr-21 of the ricin A-chain (RTA), which is one of the nine highly conserved amino acids found in similar plant and bacterial toxins that have the same mode of action as PAP (37). A previous study that deleted the first 20 amino acids from the N-terminus of RTA or amino acids 21–23 reported that the mutant proteins synthesized using an E.coli S30 extract were inactive against ribosomes from rabbit reticulocyte lysate (38). However, the same mutant proteins synthesized in reticulocyte lysate were active on the ribosomes of the lysate (39). The apparent differences in activity were attributed to differences in folding of the proteins in E.coli versus the reticulocyte lysate system (40). The results obtained with the deletion of the first 20 amino acids of RTA in reticulocyte lysate are consistent with our results with the K15M mutant of PAP, which is active on yeast ribosomes.

Mutation analysis of Tyr-16 indicated that Tyr-16 by itself is not entirely responsible for ribosome depurination and cytotoxicity. Amino acids upstream of Tyr-16 also contribute. The N-terminal amino acids of PAP may be critical for proper processing of the C-terminus, because when they are deleted, the resulting protein does not appear to be processed at its C-terminus as the wild-type PAP (Figure 2A). Similarly, the deletion of the N-terminal amino acids of ricin affected its processing at the 3′ end (40). Folding of nascent ricin chains occurs as peptidyl-tRNA on ribosomes, and N-terminal deletions affect termination and release of the protein from the ribosome (40). Our results suggest that folding of the nascent PAP protein may also occur on the ribosome. This is consistent with the gradual increase in ribosome depurination observed with increasing length of the C-terminus (Figure 1B) and with ribosome depurination that occurs in a translation-dependent manner in mutants like Y123A and Y123I, which do not depurinate ribosomes in trans (Figure 5).

Analysis of the C-terminal mutants indicated a positive correlation between sequential deletion of amino acids from the C-terminus and reduced depurination of ribosomes. For example, deletions between T262^* and E244^* showed a level of depurination that declined from 100 to 3% of the wild-type protein. Therefore, the ability to depurinate ribosomes was gradually lost with increased truncation of PAP from the C-terminus. The C-terminal deletions also showed that cytotoxicity was lost before the ability of PAP to depurinate ribosomes. Specifically, cytotoxicity was lost only after the truncation of the protein at Leu-252. Depurination ability, however, did not cease until a stop codon was introduced at Glu-244. The Leu-252 alone was not responsible for cytotoxicity, the residues downstream also contributed.

Structural analysis of another Type I RIP, trichosanthin (TCS), indicated that the corresponding leucine in TCS, Leu-240, is hydrogen bonded to Pro-35 (41). Therefore, Leu-240 in TCS is critical for maintaining the interaction between the N- and C-terminal domains (41). Pro-35 of TCS corresponds to the conserved Pro-38 in PAP. By analogy to TCS, the C-terminal amino acids may stabilize the structure of PAP by hydrogen bonding to N-terminal residues. Deletion of the last five amino acids or the penultimate five from the C-terminus of the ricin A-chain resulted in proteins that maintained their ribosome depurination activity. When aligned with PAP, these amino acids of ricin are downstream of the Leu-252 (38). Therefore, the deletion mutants of ricin at this region of the C-terminus are similar to deletion mutants of PAP because removal of these amino acids did not result in decreased cytotoxicity or ribosome depurination.

The loss of cytotoxicity before ribosome depurination with progressive deletion of the C-terminal amino acids of PAP suggests that these amino acids play an additional important role in cytotoxicity besides proper folding of the active site. PAP is localized in the ER in yeast (unpublished data). As the ribosomal target of PAP is intracellular, the cytotoxic action of PAP requires its translocation across the ER membranes into the cytosol. RIPs, such as ricin and Shiga toxin are internalized by endocytosis. Subsequently, they undergo retrograde transport by means of the Golgi complex to reach the ER lumen from where they escape to exert their cytotoxic effect in the cytoplasm (42–46). The mechanism by which PAP accesses its intracellular substrate is unknown, but clearly both its synthesis and passage back into the cytosol would require a transmembrane movement. It has been shown that the efficient internalization of surface and transmembrane proteins requires a signal sequence in the cytoplasmic tail of these proteins (47–51). Often this signal contains a tyrosine or di-leucine motif important for internalization and/or lysosomal targeting (47–51). Therefore, the C-terminal di-leucine motif (Leu-251 and Leu-252) of PAP may mediate its transport across membranes and hence its subcellular localization and cytotoxicity. Consistent with this observation, we have previously shown that W237^* PAP, which is missing 25 amino acids from its C-terminus, binds more tightly to ribosomes than wild-type PAP (30). The missing 25 amino acids contain the di-leucine motif that appears to be required for dissociation from ribosomes, since the mutant protein remains bound to the ribosomes (30). Even though the W237^* mutant binds tightly to ribosomes, it does not depurinate them, suggesting that the deleted portion of the protein also alters its enzymatic properties.

The central domain of PAP includes the active site residues, Tyr-72, Tyr-123, Glu-176 and Arg-179 that are proposed to directly participate in the catalytic depurination of the rRNA (21). The X-ray crystal structures of substrate analogs bound to ricin A-chain suggested that the adenine ring of the substrate stacks between the side chains of the comparable tyrosines in ricin, Tyr-80 and Tyr-123 (52). However, Tyr-80 is positioned in parallel with the adenine ring, whereas Tyr-123 is not (52). A similar situation also exists in PAP (21). Mutation analysis indicated that alterations of Tyr-80 were more disruptive for catalysis than alterations of Tyr-123 in ricin A-chain (53). However, tyrosine mutants of the ricin A-chain were expressed at 100–1000-fold lower than the wild-type in E.coli, which precluded their kinetic analysis (20,54).

Our results indicate that alanine substitution at Tyr-72 significantly reduced ribosome depurination by PAP, while substitution of an alanine at Tyr-123 did not cause a similar reduction in the depurination of ribosomes. Substitution of an isoleucine in place of Tyr-123 allowed depurination at similar levels as wild-type PAP, but eliminated cytotoxicity. In contrast, substitution of a Phe in place of Tyr-123 resulted in a cytotoxic protein that depurinated ribosomes. These results demonstrate that Tyr-72 of PAP plays a more critical role in the catalytic depurination of ribosomes than Tyr-123.

Analysis of depurination in Y123A indicated that ribosome depurination occurred gradually and more slowly in this mutant compared to the wild-type PAP. These results suggested that Y123A depurinates ribosomes only during its translation in a stoichiometric manner. To determine whether Tyr-123 mutants can depurinate ribosomes in trans in a catalytic manner, we inhibited transcription and translation of the Tyr-123 mutants and quantified the extent of depurination that occurred under these conditions. Unlike wild-type PAP, which depurinated ribosomes in trans, the Y123A and the Y123I mutants caused significantly reduced levels of trans depurination. In contrast, the Y123F mutant was able to depurinate ribosomes in trans. As similar levels of mutant proteins were produced in yeast expressing the Tyr-123 mutants, the differences in trans depurination observed are not due to the differences in the level of expression. Furthermore, since mutations in the same residue had very different effects on depurination and cytotoxicity, it is unlikely that altered subcellular localization of the mutant proteins is responsible for the observed effects. These results indicate that unlike Tyr-72, which is essential for depurination of the ribosome, Tyr-123 is not as critical for depurination, but appears to be more critical for recognition of the ribosome substrate. This is consistent with the X-ray crystal structure of the protein where the side chain of Tyr-123 is partially exposed on the surface, while the side chain of Tyr-72 is mostly buried (21). As total translation is not inhibited in cells that express Y123A or Y123I even though the ribosomes are depurinated, ribosome depurination that occurs during translation is not accompanied by the inhibition of protein synthesis. The ability of PAP to recognize ribosomes and to depurinate them in trans is required for the inhibition of total translation.

It is generally believed that the substrate specificity of RIPs is determined by residues that are distant from the active site and are in domains that recognize ribosomal proteins (54). However, except for Gly-75 in PAP (30), residues critical for ribosome recognition have not been characterized for RIPs. One interesting difference between PAP and ricin is that PAP can attack bacterial ribosomes, whereas ricin A-chain cannot (55). Although Tyr-123 is conserved between PAP and ricin, residues at positions 120–125 show considerable sequence and structural variation between PAP and ricin (21). Our results indicate that Tyr-123 is not critical for ribosome depurination, but may be more critical for determining the substrate specificity of PAP versus ricin.

The L71R mutation in the putative RNP2 motif in the central domain did not affect ribosome depurination, but affected cytotoxicity. In contrast to Y123A, L71R inhibited translation and depurinated ribosomes in trans. The lack of cytotoxicity of L71R indicates that ribosome depurination in trans is not enough for cytotoxicity. We have previously shown that PAP targets its own mRNA in addition to rRNA in vivo (32). Subsequent analysis indicated that specific cellular mRNAs are affected in cells expressing PAP (unpublished data). These studies complemented the in vitro findings that PAP can bind to the cap structure and depurinate certain mRNAs downstream of the cap (6,29). Although L71R depurinated ribosomes in trans and inhibited translation, it did not destabilize its own mRNA, indicating that Leu-71 is not critical for translation inhibition, but is required for mRNA destabilization (32). These results suggest that cytotoxicity of PAP may not be only due to ribosome depurination, but activity of PAP on mRNA may also be required for cytotoxicity.

Mutations in Asn-69 or Asp-92 did not affect the cytotoxicity or the depurination ability of PAP (Table 1). A previous study showed that simultaneous alanine substitutions at residues N69N70, F90N91D92, as well as R122Y123 substantially reduced the depurination and ribosome inhibitory activity of PAP when mutant proteins were produced in E.coli and depurination was assayed in vitro (25). Although ribosome depurination was reduced in the R122Y123 mutant, ribosome binding was not altered (56). Our study focused exclusively on characterizing individual mutations in vivo and showed that point mutations at Asn-69 or Asp-92 did not affect ribosome depurination, translation inhibition or cytotoxicity of PAP. However, single mutations in Asn-69 and Asp-92 were also shown to affect depurination and ribosome binding in vitro (56). As previously observed with ricin (40), the differences observed in depurination activity may be due to differences in the proper folding and the activity of the mutant proteins expressed in E.coli versus yeast. The mutant proteins may not be folding properly in E.coli. In contrast, yeast cells might contain factors that facilitate folding of these proteins. Furthermore, as the mutations were introduced into the precursor form of PAP in the previous study (25), they may have different effects on folding of the precursor protein.

In summary, we have identified PAP residues that are critical for cytotoxicity, but not for depurination of ribosomes, indicating that cytotoxicity is not dependent entirely on ribosome depurination. Here, we demonstrate that ribosome depurination in trans is required for inhibition of translation and provide evidence that ribosome depurination in trans is not sufficient for cytotoxicity. Future experiments will characterize the interaction of the PAP mutants with ribosomes and mRNA to further elucidate the role of these residues in cytotoxicity.