An Inducible 3’-Nucleotidase/Nuclease from the Trypanosomatid Crithidia luciliae PURIFICATION

Several species of protozoan parasites of the family Trypanosomatidae have a surface membrane-associated enzyme which is capable of hydrolyzing extracellular 3'-nucleotides and nucleic acids, thereby aiding in the acquisition of nutritionally required purines and Pi from their hosts. In Crithidia luciliae, this 3'-nucleotidase/nuclease previously has been shown to be highly regulated as purine and/or Pi starvation of this trypanosomatid leads to as much as a 1000-fold increase in enzyme activity. We have purified the enzyme to apparent homogeneity from detergent extracts of purine-starved C. luciliae by heparin-agarose chromatography followed by Mono Q and Mono S fast protein liquid chromatography. The enzyme had an apparent molecular weight of 43,000 and a pI of approximately 5.8. The enzyme displayed broad pH optima, with peaks at 8.0, for both nucleotidase and nuclease activities. The pH optima shifted to lower values when the activity was assayed in the presence of sulfhydryl reagents. The enzyme was most active with 3'-AMP and poly(A) in nucleotidase and nuclease assays, respectively. As a nuclease the enzyme hydrolyzed RNA at a faster rate than single-stranded DNA with no detectable hydrolysis of double-stranded DNA. The loss of enzyme activity which occurred upon storage at acid pH was prevented by the inclusion of Zn2+ in storage buffers. The physicochemical and kinetic properties of this trypanosomatid enzyme suggest that it is similar to the class I nucleases found in fungi and in germinating seedlings of higher plants.


Several
species of protozoan parasites of the family Trypanosomatidae have a surface membrane-associated enzyme which is capable of hydrolyzing extracellular 3'-nucleotides and nucleic acids, thereby aiding in the acquisition of nutritionally required purines and Pi from their hosts. In Crithidiu luciliae, this 3'-nucleotidase/nuclease previously has been shown to be highly regulated as purine and/or Pi starvation of this trypanosomatid leads to as much as a lOOO-fold increase in enzyme activity.
We have purified the enzyme to apparent homogeneity from detergent extracts of purine-starved C. luciliae by heparin-agarose chromatography followed by Mono Q and Mono S fast protein liquid chromatography.
The enzyme had an apparent molecular weight of 43,000 and a p1 of approximately 5.8. The enzyme displayed broad pH optima, with peaks at 8.0, for both nucleotidase and nuclease activities.
The pH optima shifted to lower values when the activity was assayed in the presence of sulfhydryl reagents.
The enzyme was most active with 3'-AMP and poly (A) in nucleotidase and nuclease assays, respectively.
As a nuclease the enzyme hydrolyzed RNA at a faster rate than single-stranded DNA with no detectable hydrolysis of double-stranded DNA. The loss of enzyme activity which occurred upon storage at acid pH was prevented by the inclusion of Zn2+ in storage buffers.
The physicochemical and kinetic properties of this trypanosomatid enzyme suggest that it is similar to the class I nucleases found in fungi and in germinating seedlings of higher plants.
An enzyme, originally identified as a 3'-nucleotidase, has been localized to the surface membrane of several members of the protozoan family Trypanosomatidae.
These members include species of Leishmania (l-3) and African trypanosomes (4, 5) which are pathogenic to man, as well as the related  DTT, dithiothreitol. vealed that the enzyme was capable of hydrolyzing nucleic acids as well (7,8). It has been suggested (9) that this trypanosomatid 3'-nucleotidase/nuclease may make available purine nucleosides to these parasites which are incapable of de novo purine synthesis (10). An enzyme with similar specificity has not been identified in mammals, suggesting a difference between host and parasite which potentially may be exploited for therapy and/or diagnosis of the diseases caused by these organisms.
A remarkable feature of this enzyme in C. luciliue is that the level of enzyme activity increases up to lOOO-fold when the organisms are maintained in the absence of purines and/ or Pi (6,7). The increased activity correlates with the appearance of a 43-kDa "'1 surface-labeled protein which co-migrates with enzyme activity in one-and two-dimensional electrophoresis systems.' The increase in enzyme activity is prevented by cycloheximide, an inhibitor of protein synthesis, and by actinomycin D, an inhibitor of RNA synthesis (7). However, it is not known if these metabolic inhibitors directly affect the synthesis of the enzyme protein itself or another protein which may modify the activity of a pre-existing, but inactive, enzyme protein.
This striking induction of activity in response to clearly defined, and easily manipulated, nutrient signals justifies the further study of this enzyme as a regulatory system in this family of protozoan parasites. As this induction is the only known example of such a regulated change in an enzyme or surface protein in these organisms, it may serve as a model for the differentiation events which accompany life cycle stages in these important parasites. To understand the molecular basis of the regulation at the genetic and biochemical levels it is essential to purify the enzyme to homogeneity. We report on the purification and also describe some of the physicochemical and enzymatic properties of the enzyme. The results obtained indicate that the enzyme has several properties in common with class I and other so-called singlestrand specific nucleases from germinating plants (11)(12)(13)(14)(15)(16) and fungi (17
As previously reported, purine-starved crithidia exhibited 1000.fold more enzyme activity than purine-replete organisms (6, 7). The purification was monitored by 3'-nucleotidase assay rather than nuclease assay because of the greater sensitivity and convenience of the former. In addition, we analyzed samples from the purification by activity staining gels after SDS-PAGE.
The enzyme, as demonstrated by either 3'-nucleotidase or nuclease activity staining, migrated with an apparent molecular weight of 43,000. When the purification was carried out in the absence of protease inhibitors, bands of enzyme activity which migrated faster than 43 kDa appeared following SDS-PAGE of samples of the detergent extract and subsequent stages of the purification (data not shown). We attributed these bands to proteolytic digestion of the intact enzyme.
Approximately 95% of the 3'-nucleotidase activity was associated with the membrane fraction (15,000 x g pellet) after centrifugation of whole cell lysates (Table I)  Procedures." Fractions 20-36 (denoted by the bracket) were combined to form the Q-Sepharose pool.
nuclease is an integral membrane protein.
Heparin-Agarose Chromatography-During chromatography of the detergent extract on heparin-agarose, the enzyme eluted over a broad range of the salt gradient with peak elution at 0.35 M NaCl. The specific activity of the pooled active fractions was 2-fold greater than that of the octyl-@-D-glucopyranoside extract. This step separated the enzyme from a periodic acid Schiff-staining substance which did not adsorb to the column (data not shown). This substance was presumably a lipopolysaccharide, analogous to a previously described arabinogalactan from Crithidia fasciculata (19) and to the lipophosphoglycan of Leishmania spp, (25, 26). Zon Exchange Chromatography-A portion of the heparinagarose purified material was chromatographed on Q-Sepharose; the enzyme activity eluted as a broad peak beginning at approximately 0.1 M NaCl (Fig. 1). The pooled column fractions, which were increased nearly 3-fold in specific activity and which contained 60% of the applied 3'-nucleotidase activity (Table I), were loaded on a Mono S FPLC column. This Mono S FPLC step resulted in a 5-fold increase in specific activity with a 70% yield of 3'-nucleotidase activity (Table I), SDS-PAGE of the Mono S pool revealed the presence of a prominent silver-stained band at 43 kDa which co-migrated with the enzyme activity-stained band (Fig. 4). Several silverstained bands of higher molecular weight also were observed.
Adsorption of the enzyme to the Q-Sepharose (anion exchange) column decreased at pH values lower than 6.5 and adsorption to the Mono S column (cation exchange) decreased at pH values greater than 5.0 (data not shown). These observations were consistent with the determination by two-dimensional electrophoresis that the p1 of the enzyme was approximately 5.8 (see below). In order to obtain better resolution than that which was possible on the large capacity Q-Sepharose column, a portion of the heparin-agarose purified material was subjected to Mono Q FPLC. The resulting column profile (Fig. 2) revealed two peaks of enzyme activity. These peaks were pooled separately, designated pool 1 and pool 2, and used for Mono S FPLC. Activity stained gel patterns revealed no differences between the enzyme collected in pool 1 and pool 2. The combined pools contained 77% of the 3'-nucleotidase activity loaded onto the column, and exhibited a 5-fold increase in specific activity (260 units/mg) .
A column profile for the Mono S FPLC chromatography of the Mono Q FPLC pool 2 is shown in Fig. 3. Virtually identical results were obtained for pool 1 (data not shown). The active fractions from pools 1 and 2 were combined and are referred to as the Mono Q-Mono S pool. The gel pattern of the Mono Q-Mono S pool after SDS-PAGE revealed a single silverstained band at 43 kDa which co-migrated with enzyme activity (Fig. 4).
Purification Summary-The yields and specific activities of the enzyme during each step of the purification are shown in Table I. Each step of the purification also was analyzed by SDS-PAGE followed by staining for protein and for 3'-nucleotidase activity (Fig. 4). Identical results were obtained when parallel gels were stained for nuclease activity (data not shown).
The crithidial 3.nucleotidase/nuclease was purified 490fold for 3'-nucleotidase activity and 2780-fold for nuclease activity. The reduction, upon purification, in the ratio of 3'. nucleotidase to nuclease activities was reproducible in many different preparations.
One explanation for the reduction in ratios is that an enzyme capable of hydrolyzing 3'-nucleotides, but which does not possess nuclease activity, was removed during purification to a much greater extent than the 3'nucleotidase/nuclease. However, such an enzyme would have to be regulated by purine starvation to a similar extent as the 3.nucleotidase/nuclease. Also, as the column fractions were pooled on the basis of 3'.nucleotidase activity rather than nuclease activity, the selective loss of this postulated additional 3.nucleotidase was unlikely. A more likelv explanation is that a cellular component(s), which inhibited nuclease activity to a greater extent than 3'-nucleotidase activity, was removed during purification.
For example, such a component could inhibit access of large polynucleotide substrate molecules, but not nucleotides, to the 3'-nucleotidase/nuclease during enzyme assays. The observation that the total amount of nuclease activity actually increased during the first few purification steps (Table I) lends support to this explanation. The purified enzyme and crude octyl-$-D-glucopyranoside extracts of membranes were assayed for two other enzyme activities associated with trypanosomatid surface membranes, acid phosphatase, and <5-nucleotidase (1, 3, 27). Neither of these actiGties were detected in Mono S-purified material, but both activities were present in the octyl-/j-II-glucopyranoside extract (data not shown). was used in the first dimension to resolve proteins over a broad pH range (pH 4.7-8.5). Following electrophoresis in the second dimension, the gels were silver-stained for protein or stained for nuclease or 3.nucleotidase activity (Fig. 5). The gel staining patterns revealed co-migration of protein with nuclease and 3'.nucleotidase activities at a p1 of 5.8 and an Mr of 43,000. Similar gels, using isoelectric focusing (pH range 4.5-6.5) in the first dimension to obtain better resolution around pH ,5.8, supported the conclusion that the same enzyme is responsible for both activities (data not shown).
pH 0ptLma-The effect of pH on the 3'-nucleotidase and nuclease activities of the Mono S-purified crithidial3'.nucleotidase/nuclease, when assayed under standard conditions, is shown in Fig. GA. Similar results were obtained using a combination of 0.1 M NaAc (pH 4.5-6.5), MES (pH LO-7.0), and HEPES (pH 6.5-9.0) instead of Tris maleate to buffer the reactions. The results indicated broad pH optima, around pH 8.0, for both activities. When the crithidial enzyme was assayed in the presence of 1 mM DTT (Fig. 6B) to approximate the conditions under which plant class I nuclease activities were assayed in some reports (11,13), a shift in the pH optima to lower values occurred. The shift in pH optima of the crithidial enzyme, from 8.0 in the absence of DTT, to 6.0-6.5 in the presence of DTT, may be explained by our finding that 1 mM DTT inhibited 85% of the enzyme activity when assayed at pH 8.0, but only 20% of the activity when assayed at pH 6.0 when compared to control assays without DTT (Fig. 7). Similar results were obtained using cysteine instead of DTT. A, purified enzyme was assayed for 3'-nucleotidase and nuclease activities as described under "Experimental Procedures" except 0.1 M Tris maleate at the indicated pH values was included as buffer in the reaction mixtures. The 3'nucleotidase assays contained approxima~ly 3.3 ng of protein/ml and the nuclease assays contained approximately 170 ng of protein/ ml. The maximum 3'nucleotidase activity was 1860 units/mg, and the maximum nucIease activity was 495 units/mg. 3, assays were performed as described in A but included 1 mM DTT in the reaction mixtures. The maximum 3'-nucleotkiase activity was 950 units/mg, and the maximum nuclease activity was 260 units/mg. The symbols used are: 0, 3'nucleotidase activity; 0, nuclease activity. FK. 7. Differential sensitivity of the crithidial 3'-nucleotidase/nuclease to DTT at pH 6.0 and 8.0. Crithidial3'-nucleotidase/nucIease was assayed for 3'nucleotidase activity at pH 6.0 (0) or pH 8.0 (0) with DTT included in the reaction mixtures at the indicated concentrations. Control values (445 units/mg at pH 6.0 and 760 units/mg at pH 8.0) were obtained by assay without DTT at the indicated pH. enzyme at various pH values will be considered in more detail below. The results were not infmenced by the purity of the enzyme preparations (detergent extracts to homogeneous enzyme) used in repetitions of these experiments (data not shown).
Substrute Speciticity- Table  II shows the substrate specificities of the purified nuclease when assayed at pH 8.0 for 3'-nucleotidase and nuclease activities. The results revealed a preference by the enzyme for 3'-ribonucleotide substrates with no activity on either deoxyribonucleotides or cyclic nucleotides. The Km of the enzyme for 3'-AMP was approximately 0.4 mM; purified enzyme exhibited a Vmax of 1470 units/mg. Pi was not released from nucleic acid substrates as indicated by lack of nucleotidase activity with poly(A) as substrate. As a nuclease, the enzyme preferentially hydrolyzed poly(A) andpoly(U).
The crithidial nuclease did not hydrolyze Procedures." Reaction mixtures contained 5 mM of the indicated substrates (except poly(A), at 10 mg/ml) and approximately 20 ng of protein/ml. * Purified 3'-nucleotidase/nuclease was assayed for nuclease activity at pH 8.0 as described. Reaction mixtures contained 10 mg/ml of the indicated substrates and approximately 1.2 fig of protein/ml. double-stranded DN& however, the enzyme was capable of hydrolyzing single-stranded, denatured DNA, although to a lesser extent than RNA.
End Products of Digestion-As indicated above, Pi was not detected following the incubation of the purified enzyme with poly(A); this result was consistent with a predicted mechanism of hydrolysis which yielded nucleotides that possess terminal 5'-phosphate and 3'-hydroxyl groups as products. Further evidence that the products of poly(A)hydrolysis by the 3'-nucleotidase/nuclease were 5'-nucleotides was provided by the observation that the addition of purified snake venom 5'-nucleotidase, which specifically releases Pi from 5'nucleotides (28), resulted in the release of Pi from the products of 3'~nucleotidase/nuclease digestion of poly(A) (data not shown).
Crude detergent extracts of crithidial membranes, in contrast to purified 3'-nucleotidase/nuclease, released significant amounts of Pi from comparable levels of poly(A) when assayed in a similar manner. The results suggested that another enzyme(s) present in the o~tyl-~-D-glu~opyranoside extracts, presumably including the aforementioned crithidial5'-nucleotidase, was responsible for the released Pi+ Inhibitor S'ensitiuity-The crithidial 3'-nucleotidase/nuclease was not inhibited by commonly used phosphatase inhibitors including tartrate, molybdate, and fhroride ions when these inhibitors were included in standard reaction mixtures at a concentration of 1 mM. Under the conditions of such experiments, the activity was partially sensitive to 1 mM EDTA (40% inhibition), although incubation of the 3'-nucleotidase/nuclease with 1 mM EDTA prior to enzyme assays completely abolished enzyme activity. The activity loss caused by EDTA could be recovered by preincubating the enzyme before the assay for 30 min in 0.3 mM ZnSO+ provided the ZnSOd was diluted to subinhibitory concentrations during the assay. The enzyme was extremely sensitive to Zn2+ and sulfhy-dry1 reagents in these assays at pH 8.0, as 1 mM ZnS& 1 mM cysteine, or 1 mM DTT inhibited over 90% of the enzyme activity, The enzyme activity was inhibited to a much lesser extent when the assays were carried out at pH 6.0 as mentioned above for thiol reagents and in Fig. 7, and indeed Zn2+ protected against activity losses when the enzyme was stored at acid pH (see below).

Stabilization of Enzyme Activity by Zn'+ and Restoration of
Enzyme Activity by Cysteine-Initially during our efforts to purify the enzyme, activity was lost at the Mono S FPLC stage of purification at which time the Q-Sepharose-or Mono Q FPLC-purified enzyme was placed into Buffer D at pH 5.0, but without ZnSO+ Chromatography at low pH at this step was necessary as adsorption of the enzyme to the Mono S column was reduced at pH values greater than 5.0. In an effort to maintain enzyme activity of the Q-Sepharose and Mono Q-purified enzyme at pH 5.0, we evaluated the effects of various reagents and identified the protective effects of ZnSOd (Table III). Reducing agents, Mp, and the substrate 3'-AMP did not protect against activity loss at acid pH. In the absence of added Zn*+, almost 90% of the enzyme activity was lost upon incubation of the enzyme for 24 h at pH 5.0. In contrast, less than 10% of the activity was lost during storage at pH 7.0. The addition of Zn*+, at concentrations up to 1 mM, protected against the loss at pH 5.0. Paradoxically, when these concentrations of ZnSOh were present in the reaction mixtures during enzyme assays under standard conditions, the activity was strongly inhibited (see above). In the experiment shown in Table III, ZnSO* was diluted sufficiently during the assays to prevent such inhibition.
In contrast to Zn'+, neither cysteine nor DTT were able to protect against the loss of activity at acid pH. However, crithidial 3'-nucleotidase/nuclease activity which was lost during incubation at pH 5.0, and to a lesser extent at pH 7.0, was almost fully restored by incubation with 1 mM cysteine for 30 min at 4 'C immediately before the enzyme was diluted and assayed. As with Zn'+, the final concentrations of cysteine in the assay mixtures were subinhibitory.
ZnSOd, when used in place of cysteine in similar experiments, did not restore enzyme activity (data not shown). Final concentrations of cysteine in the reaction mixtures were 300-fold less than the indicated concentrations and were subinhibitory in every case. Activities were compared to that of the enzyme incubated at pH 7.0, without cysteine.

DISCUSSION
We have purified a strongly regulated enzyme, originally described as a 3'-nucleotidase (6), from the trypanosomatid protozoan C. luciliae. We have shown that the enzyme also possesses nuclease activity based on the observations that the 3'-nucleotidase and nuclease activities co-purify and co-migrate in both one-and two-dimensional electrophoresis. The properties of the enzyme described in this report are similar to those of a number of so-called single-strand specific nucleases (29), especially the class I nucleases, from germinating plant seedlings (15,17). The similarity to these plant enzymes is based on a comparison of their enzymatic properties to those described by Wilson (15) in his classification of plant nucleases. These enzymes, like the crithidial 3'nucleotidase/nuclease, are capable of hydrolyzing the 3'-phosphate group from nucleoside 3'-monophosphates, principally ribonucleotides, as well as hydrolyzing nucleic acids. The hydrolysis of RNA by this group of enzymes is greater than that of denatured DNA, which is greater than that of native DNA. The enzymes are inhibited by EDTA and have a requirement for divalent cations, frequently Zn2+, for activity and/or stability. The plant nucleases I are generally recognized as having pH optima in the range of 5.0-6.5 (15,17). However, the plant nucleases usually have been assayed in the presence of varying, and frequently undisclosed, amounts of sulfhydryl reagents. Although the crithidial enzyme displays its greatest activity at pH 8.0, it does exhibit a lower pH optimum when assayed in the presence of thiols. The thiol-dependent shift in pH optima is due to the fact that thiols inhibit enzyme activity to a greater extent at alkaline pH than at acid pH in both the plant nucleases (30,31) and in crithidia. Similar effects are observed with certain divalent cations. For example, Zn2+, which is necessary for maintenance of enzyme activity during storage at acid pH, is inhibitory when present during enzyme assays at pH 8.0 and to a lesser extent at pH 6.0. In addition to enzymatic properties, the enzymes from trypanosomatid protozoa and from plants are similar in size, with apparent molecular weights of approximately 40,000. Despite the numerous similarities there are a number of properties of the protozoa1 enzyme that distinguish it from the analogous enzymes and make it an important subject for further investigation.
The trypanosomatid 3'-nucleotidase/ nuclease is the first such enzyme described from organisms incapable of synthesizing purines de novo. It is also the first enzyme of this group which has been definitively localized to the surface membrane of a cell (2,32,33).
Despite the suggestion that the plant enzymes may be particle-or membraneassociated (15), nucleases I are generally thought to be soluble and indeed some have been obtained from supernatant fluids of plant cell cultures (29).
Based upon its surface membrane localization and its ability, in some cases in combination with a 5'-nucleotidase activity, to generate nucleosides from extracellular nucleotides and nucleic acids, the protozoa1 enzyme functions in the acquisition of required purines from exogenous sources (9, 10) which are not capable of being transported across the surface membrane. This enzymatic activity may therefore be essential to parasite replication and hence survival. In leishmanial parasites, which reside in the digestive tracts of sandfly vectors as well as the phagolysosomal system of mammalian macrophages, the enzyme may enable the parasite to compete with its hosts for purines. As species of the genus Leishmania are important agents of human disease, this enzyme may be a suitable target for chemotherapeutic interventions for disease control. Indeed, purine metabolism has been recognized as an appropriate approach to controlling these organisms and purine analogs are being tested for their effectiveness (34).
A remarkable feature of the crithidial 3'-nucleotidase/nuclease, and an important area of further investigation, is the regulation of its activity. The availability of purified crithidial enzyme should allow for the development of suitable reagents for the further analysis of this system. Antibodies currently are being developed which should enable us, in conjunction with metabolic labeling, to determine if the increased enzymatic activity is due to the synthesis of new enzyme protein or if the increase results from modification of a pre-existing, catalytically inactive protein. Antibodies should also prove useful in screening libraries in efforts to clone the gene encoding the enzyme. The purified protein should also lead to the generation of amino acid sequences which will provide the basis for the production of synthetic oligonucleotide probes which will also allow us to analyze the regulatory events at the molecular level. As this is the only known example in this group of organisms of such a regulatory phenomenon, it may serve as a model system to elucidate mechanisms of differentiation events associated with life cycle changes. Such developmental processes are critical for parasite survival as they allow the organisms to adapt to different host environments.