ABSTRACT
Salivary gland homogenates of the sand fly Phlebotomus papatasi contain large amounts of adenosine and 5′-AMP, of the order of 1 nmol per pair of glands, as demonstrated by liquid chromatography, ultraviolet spectrometry, mass spectrometry and bioassays. These purines, 75–80 % of which are secreted from the glands following a blood meal, have vasodilatory and anti-platelet activities and probably help the fly to obtain a blood meal. Salivary 5′-AMP is also responsible for the previously reported protein phosphatase inhibitor in the salivary glands of P. papatasi, which is shown to be artifactual in nature as a result of allosteric modification by AMP of the phosphatase substrate used (phosphorylase a).
Introduction
Hematophagy evolved independently in several orders of insects and ticks. Perhaps for this reason a variety of salivary anti-hemostatic compounds are found in this diverse group of arthropods. Many different salivary anti-clotting and anti-platelet agents and vasodilators have been described. Salivary vasodilators range from inorganic nitric oxide to prostaglandins, various different peptides and enzymes that destroy norepinephrine. Several anti-clotting peptides, acting against different clotting factors, have been characterised, as have many anti-platelet substances acting on different pathways of platelet aggregation (Champagne, 1994; Ribeiro, 1995).
Many salivary compounds that interfere with vertebrate hemostasis have additional effects on the host immune system, and these may in turn affect the fate of parasites transmitted during the act of blood feeding. For example, the New World sand fly Lutzomyia longipalpis has a very potent vasodilator, maxadilan (Lerner et al., 1991), in its saliva, and this also has immunosuppressive activities against macrophages and lymphocytes (Qureshi et al., 1996; Soares et al., 1998). This immunosuppression may be the reason that sand fly saliva potentiates the transmission of Leishmania parasites (Theodos et al., 1991; Titus and Ribeiro, 1988; Warburg et al., 1994).
The Old World sand fly Phlebotomus papatasi does not have the vasodilator maxadilan, but its saliva nevertheless potentiates Leishmania transmission (Theodos et al., 1991). A protein phosphatase inhibitory activity has recently been found in this fly (Waitumbi and Warburg, 1998). P. papatasi salivary protein phosphatase inhibitor has been implicated in the down-regulation of nitric oxide production by murine macrophages (Waitumbi and Warburg, 1998). In an attempt to find the salivary protein phosphatase inhibitor, we discovered large amounts (approximately 1 nmol per pair of glands) of 5′-AMP and adenosine in P. papatasi salivary glands, two substances with known vasodilatory (Collis, 1989) and anti-platelet (Edlund et al., 1987) functions. In the present study, salivary adenosine monophosphate is shown to be the previously observed salivary protein phosphatase inhibitor of P. papatasi.
Materials and methods
Reagents
A protein phosphatase assay kit was obtained from Gibco-BRL, Life Technologies. [γ-32P]ATP (6000 Ci mmol−1) was purchased from Amersham Pharmacia Biotech. Other reagents were purchased from Sigma Chemical Co.
Sand fly rearing and collection of salivary gland lysate
Phlebotomus papatasi (Scopoli) were reared as described previously (Modi and Tesh, 1983). Salivary glands from 3-to 6-day-old female flies, maintained with a cotton swab containing 10 % sucrose solution, were dissected in phosphate-buffered saline, transferred to conical 1.5 ml plastic tubes (in a final volume no larger than 100 μl) and stored at −70 °C. Before use, the glands were ultrasonically disrupted by immersing the tube, held by forceps, in water under the tip of a Branson Sonifier 450 (Danbury, CT, USA) sonicator with the energy level set to just under cavitation. Approximately 20 intermittent cycles of 50 % duration were sufficient for complete disruption of the glands. The tubes were centrifuged at 14 000 g for 2 min after sonication, and the supernatant was used for the experiments.
Macrophages
Macrophages for use in protein phosphatase assays were obtained from 8-to 12-week-old C3H/HeN female mice as described previously (Waitumbi and Warburg, 1998). Mice were stimulated with 2.0 ml of 3 % thioglycollate injected intraperitoneally. Four days later, peritoneal exudate cells were collected by peritoneal lavage using 10 ml of RPMI 1640 (Gibco-BRL), and were then washed in Ca2+- and Mg2+-free Hank’s balanced salt solution (HBSS) before being resuspended in RPMI 1640 containing 10 % fetal bovine serum. 107 cells in 5 ml were seeded in a 50 ml culture flask and incubated at 37 °C, 5 % CO2 and 95 % humidity for 90 min. Non-adherent cells were removed by washing the cells three times with ice-cold HBSS containing 1 mmol l−1 sodium vanadate. Adherent cells are hereafter referred to as macrophages.
Preparation of lysates for protein phosphatase assays
Macrophage lysates for use in protein phosphatase assays were prepared as described previously (Dong et al., 1995). Briefly, macrophages were scraped off the plastic culture flask, centrifuged and resuspended in ice-cold lysis buffer containing 20 mmol l−1 Tris–HCl, pH 7.0, 250 mmol l−1 sucrose, 20 mmol l−1 KCl, 2.5 mmol l−1 MgCl2, 30 mmol l−1β-mercaptoethanol, 2 mmol l−1 EDTA, 1 mmol l−1 phenylmethylsulfomyl fluoride, 10 μg ml−1 leupeptin, 10 μg ml−1 aprotinin, 25 % glycerol and 1 % Triton X-100. The cells were left on ice for 30 min with occasional pipetting. Nuclear and cell debris were removed by centrifugation at 1000 g for 15 min at 4 °C, and the total protein concentration in the cytosolic fraction was determined using the Bradford method (Bradford, 1976). For preparation of Leishmania major lysates, 2×107 logarithmic-stage promastigotes were washed three times in HBSS and lysed in 0.5 ml of lysis buffer as described above for macrophages.
Protein phosphatase assays
Macrophages and Leishmania major lysates prepared as described above were used for measurements of protein phosphatase activity as described previously (Waitumbi and Warburg, 1998). The assay was carried out using a kit from by Gibco-BRL, which uses radioactive ATP and phosphorylase kinase to transform phosphorylase b to its phosphorylated form, phosphorylase a, which is then used as a substrate for the protein phosphatases 1 (PP1) and 2A (PP2A) contained in the macrophage and Leishmania major lysates.
High-performance liquid chromatography
A CM 4100 pump and SM4100 dual-wavelength detector from ThermoSeparation Products were used (Rivera Beach, FL, USA). A diode array detector (model SPD-M10AV, Shimadzu Corp., Columbia, MD, USA) was used to monitor the light absorption of the column effluent in the range 200–350 nm, as indicated. The following columns were used. (1) Molecular sieving Superdex Peptide HR 10/30 from Amersham Pharmacia Biotech (10 mm×300 mm), run with 30 % acetonitrile and 0.1 % trifluoroacetic acid (TFA). This volatile solvent mixture was chosen to allow complete evaporation and increased concentration of the samples for bioassays. This solvent lacks salts that disrupt ionic interactions with the stationary matrix, and this may result in some charged compounds failing to elute in a truly molecular sieving mode. (2) Alltech (Deerfield, IL, USA) C18 macrosphere (4.3 mm×250 mm) run isocratically with 0.1 % TFA in water (prior to each run, the column was washed with 10 vols of 95 % acetonitrile and equilibrated with 10 vols of eluting solution). (3) An anion-exchange TSK-DEAE column (4.5 mm×150 mm) obtained from BioRad (USA), perfused isocratically with 50 mmol l−1 sodium phosphate adjusted to pH 3.5 with phosphoric acid. The eluate was monitored either at 220 and 280 nm or at 260 and 280 nm, as indicated. Under these conditions, the adenosine and AMP content of individual salivary gland pairs could be measured. Ratios of absorption at 260 nm and 280 nm were between 6.5 and 8.2, consistent with the absorption ratios of adenosine and AMP standards.
Mass spectrometry measurements
Liquid chromatography mass spectrometry analyses were performed on a Hewlett Packard 1100 electrospray mass spectrometer system. The sample was injected onto a Zorbax SB300 C3 column (2.1 mm×150 mm). The elution was started with 5 % acetic acid with a gradient to 100 % acetonitrile. The mass spectrometer was scanned from m/z 60 to 1000 (where m is mass and z is valence) using the fragmentor voltages indicated. Accurate mass measurements were made using fast atom bombardment with a polyethylene glycol matrix.
Platelet aggregation assay
Platelet aggregation assays were performed on a Thermomax microplate reader (Molecular Devices, MenloPark, CA, USA) with a kinetic module (Bednar et al., 1995) with further modifications. Human citrated platelets were obtained by plasmaphoresis and were further diluted with Tyrode’s solution without Ca2+ containing 1 mg ml−1 bovine serum albumin (T-CaBSA) to give a final absorption (turbidity) of 0.1 optical density units at 650 nm. The indicated volumes of dry chromatographic fractions resuspended in T-CaBSA were incubated in a 96-well flat-bottomed plate (Falcon 3912; Becton Dickinson, Oxnard, CA, USA) in a total volume of 95 μl. The plate was incubated at 37 °C for 2 min, and ADP was added in volumes of 5 μl to give a final concentration of 15 μmol l−1. The plate was shaken for 5 s on a microplate shaker (Cole Parmer series 4732, Vernon Hills, IL, USA), and aggregation was monitored at 11 s intervals with shaking of the plate between readings.
Smooth muscle bioassay
Rabbit aorta was purchased from Spring Valley Labs, Woodbine, MD, USA. Organs were shipped in ice-cold Hepes-buffered Tyrode’s solution (see below for details). Rings could be used up to 3 days after organ removal. Rabbit aorta smooth muscle ring bioassays were performed isometrically using Hepes (10 mmol l−1) buffered Tyrode’s solution (Webster and Prado, 1970) at 37 °C bubbled continuously with 95 % O2 and 5 % CO2. Dexamethasone (0.1 μmol l−1) was added to the bath to prevent induction of inducible NO synthase (Moncada et al., 1991), while EDTA (30 μmol l−1) was added to chelate heavy metal contaminants. Aortic rings (2–3 mm wide) were attached to small (0.5 ml final volume) temperature-controlled (37 °C) cuvettes by means of stainless-steel hooks. Another hook, linked to a polyester thread, connected the ring to an isometric (force) transducer. The isometric force transducer could be moved up and down using a gear mechanism from Harvard Apparatus (MA, USA). This gear mechanism was used to stretch the ring to an initial tension of 3 g. Rings were allowed to stabilise for 90 min, with changes in the saline solution at 15–20 min intervals, when the tension was readjusted if necessary. Experiments were started by the addition of the agonist norepinephrine (1 μmol l−1), which induced a contraction of the ring, measured in units of grams of force. When a stable tension was achieved, samples (25 μl) from the evaporated and reconstituted column fractions were added to the bath to test for vasodilative (tension relaxation) activity. Isometric transducers were obtained from Harvard Apparatus (MA, USA), and their output was recorded in a computer through an analog-to-digital card converter.
Statistical tests
Statistical tests were carried out using Sigmastat Statistical Package version 2.0 (Jandel, San Rafael, CA, USA). When variances were homogeneous between groups, t-tests were performed; otherwise, Mann–Whitney rank sum tests were performed. Values for IC50, the dose giving half-maximal inhibition, were calculated from the interpolation of the linear regression lines of the percentage inhibition of phosphorylase a dephosphorylation against the logarithm of AMP concentration.
Results
The salivary protein phosphatase inhibitor of Phlebotomus papatasi has been previously characterised as a relatively small molecule because of its capacity to pass through membrane filters with 3 kDa limiting pores. Furthermore, the activity resisted boiling, RNAase, DNAase and proteases and was soluble in ethanol (Waitumbi and Warburg, 1998). To obtain further information on the molecular nature of this activity, we submitted 40 pairs of homogenised salivary glands to molecular sieving chromatography using a 10 kDa exclusion column. The absorbance of the eluate was monitored at both 220 and 280 nm. Samples of the fractions were evaporated and assayed for protein phosphatase inhibitory activity. The chromatogram indicated late elution of two major peaks having a relatively large 280 nm absorption compared with absorption at 220 nm (280 nm absorption being approximately one-third of that at 220 nm). This is atypical of peptides, which have 10–20 times more absorption at 220 nm. Protein phosphatase inhibitory activity co-eluted with the second of the two large 280 nm absorbing, late-eluting, peaks (Fig. 1).
Attempts to use octadecyl reverse-phase protocols to purify the Phlebotomus papatasi protein phosphatase inhibitor (PpPPI), using gradients from 10 % to 60 % acetonitrile in 0.1 % TFA in water were unsuccessful, because the activity was eluted in the solvent front (not shown). However, the activity was retained in the same column if elution was performed isocratically using 0.1 % TFA in water. We submitted the homogenate of 40 pairs of salivary glands to ultrafiltration on a Centricon 3 (3 kDa cut-off), and applied the filtrate to this isocratic reverse-phase protocol. The eluate was monitored in a diode array detector. Two major peaks eluted, both having ultraviolet absorption maxima near 258–260 nm. Protein phosphatase inhibitory activity co-eluted with the first major peak (Fig. 2).
Fractions corresponding to peak 1 and peak 2 were evaporated to dryness and submitted to mass spectrographic analysis. These fractions are named peak 1 and peak 2 throughout this paper. Although only peak 1 showed inhibitory activity in our assay, peak 2 was analysed because it had similar ultraviolet spectral properties to peak 1 and its mass spectrum could help in the identification of peak 1. Peak 1 gave a predominant molecular ion of m/z=348.1 (Fig. 3A), whereas peak 2 produced a molecular ion of mass 268.1 (Fig. 4A), the difference corresponding to 80 mass units, the same mass as one HPO3 radical. The observed masses were hypothesised to derive from adenosine monophosphate (protonated mass 348.2) for peak 1 and adenosine (protonated mass 268.2) for peak 2. Accurate mass measurements (fast atom bombardment with polyethylene glycol matrix) of the adenosine phosphate peak confirmed the formula of the peak 1 compound. The protonated mass of the parent ion ([M-H]−) was 346.0563 (calculated formula C10 H13N5 O7 P, 346.0553). To identify the compounds within peaks 1 and 2 as AMP and adenosine, both were submitted to electrospray mass spectrometry at an increased electrospray voltage to produce fragmentation of the compounds. In both cases, the expected protonated adenine mass of 136.1 was obtained as the parental masses of 348.1 and 268.1 were destroyed (Figs 3B, 4B). Reverse-phase chromatography of authentic 5′-AMP and adenosine gave the same retention times (not shown) as peaks 1 and 2, respectively, and the same ultraviolet spectra (Fig. 5). The retention time of 3′-AMP was intermediate between those of AMP and adenosine, and it was therefore was excluded as a candidate for peak 1.
To confirm the role of 5′-AMP as the inhibitor of phosphorylase a dephosphorylation in our assay, we tested the ability of authentic 5′-AMP to inhibit both the Leishmania major and murine macrophage protein phosphatase activities, as determined using the phosphorylase a kit. Phosphorylase a dephosphorylation by both enzyme preparations was inhibited, with IC50 values of 7.14±0.47 μmol l−1 for Leishmania major and 8.51±0.27 μmol l−1 for murine macrophages (means ± S.E.M., N=3) (Fig. 6).
Because approximately one pair of homogenised salivary glands is needed to produce 50 % phosphatase inhibition (Waitumbi and Warburg, 1998), the IC50 values measured for AMP suggest that each pair of glands should contain at least 0.6 nmol of AMP to produce a final nucleotide concentration of approximately 10 μmol l−1 in the 60 μl reaction volume used in the protein phosphatase inhibition assay. We then dissected 10 pairs of salivary gland from 5-day-old female P. papatasi that had never been fed on blood and another 10 pairs from flies that just received a blood meal. The adenosine and AMP contents of individual gland pairs were measured by anion-exchange chromatography. The results indicated the presence of 0.93±0.08 nmol of adenosine and 1.2±0.09 nmol of AMP in the fly salivary glands (means ± S.E.M, N=10). These values were significantly reduced (by 75–80 %) following a blood meal (P<0.001 in both cases, t-test) (Fig. 7). We conclude that a pair of salivary glands possesses the amount of AMP required to produce the observed inhibition of phosphorylase a dephosphorylation in our assay.
Adenosine is a strong inhibitor of platelet aggregation (Edlund et al., 1987), so we tested the ability of peak 1 and peak 2 fractions to inhibit human citrated platelet aggregation induced by 15 μmol l−1 ADP. As expected, the peak 2 fraction significantly inhibited platelet aggregation (P=0.002, Mann–Whitney rank sum test, N=6) (Fig. 8, and a second similar experiment) at all concentrations tested, which included concentrations of the inhibitor equivalent to 1.25, 2.5 and 5 pairs of glands in a 100 μl assay. The same equivalents from the peak 1 fraction gave no significant inhibition of platelet aggregation.
Both AMP and adenosine are known vasodilators (Lewis et al., 1994; Rose’Meyer and Hope, 1990). We therefore tested the effects of the peak 1 and peak 2 fractions on a rabbit aortic ring contracted with norepinephrine. A small but reproducible relaxation was observed when either of the two fractions was added to the smooth muscle preparation (Fig. 9). The results were reproduced with material derived from another column and with a second aortic preparation.
Discussion
The results presented in this paper indicate that the previously described protein phosphatase inhibitor of the salivary glands of Phlebotomus papatasi (Waitumbi and Warburg, 1998) is 5′-AMP, as identified by chromatographic, ultraviolet spectrographic and mass spectrographic techniques (Figs 2–5). The amounts of 5′-AMP and adenosine in the salivary glands (approximately 1 nmol per pair of glands) are sufficient to account for all the inhibitory activity found in salivary gland homogenates. This is the first time that adenosine and AMP have been found associated with saliva or salivary gland homogenates of blood-sucking arthropods, although fire ant Pseudomyrmex triplarimus venom has been shown to have a platelet inhibitor identified as adenosine (Hink et al., 1989). The role of these purines, the levels of which decrease by 75–80 % in the salivary glands during a blood feed, may relate to their inhibition of platelet aggregation and vascular smooth muscle contraction.
The concentrations of adenosine and AMP in P. papatasi salivary glands may be calculated assuming that the volume of one pair of salivary glands is 35 nl (Adler and Theodor, 1926) and that the amounts of adenosine and AMP are of the order of 1 nmol per gland pair. This gives a concentration of approximately 30 mmol l−1 of each of these purines within the salivary gland. Considering that 80 % of this amount is discharged during a blood meal, and assuming that a volume of 1 mm3 represents the range of diffusion of the salivary compounds into the host skin during feeding, concentrations of approximately 0.8 mmol l−1 (0.8 nmol μl−1) would be achieved for adenosine and AMP in the feeding cavity.
The role of AMP and adenosine in helping the sand fly to obtain a blood meal may derive from purinergic receptors found in both blood vessels and platelets. Adenosine is a potent platelet aggregation inhibitor, with a reported IC50 value for inhibiting ADP-induced platelet aggregation of 1–2 μmol l−1 (Dionisotti et al., 1992). Adenosine increases platelet cyclic AMP concentration and thus prevents platelet aggregation by all agonists, not only by competition with ADP, as is the case for AMP. In addition, both adenosine and AMP are vasodilatory substances, with IC50 values of >50 μmol l−1 for relaxation of aortic preparations (Gordon and Martin, 1983; Lewis et al., 1994; Rose’Meyer and Hope, 1990; Urquhart and Broadley, 1991). Assuming, from the calculation in the previous paragraph, that the concentrations of secreted purines in the vicinity of the mouthparts are approximately 0.8 mmol l−1, we conclude that pharmacologically active amounts of purines are produced by P. papatasi during a blood meal. Indeed, inhibition of platelet aggregation in a 100 μl assay was obtained with the adenosine (peak 2) equivalent of one pair of glands, which should give an adenosine concentration of approximately 10 μmol l−1 in the assay shown in Fig. 8. This concentration is 5–10 times larger than the reported IC50 for adenosine of 1–2 μmol l−1 (Dionisotti et al., 1992), explaining the saturation of the curves for peak 2 shown in Fig. 8. A modest vasodilatory activity was found with both nucleotide peaks when the equivalent of the AMP or adenosine content of 25 gland pairs was added to a 500 μl assay, producing a final concentration of approximately 50 nmol ml−1 or 50 μmol l−1.
In addition to being vasodilatory and inhibitory to platelet aggregation, adenosine derivatives are also known to have lymphocyte-immunosuppressive properties (Apasov et al., 1995; Seegmiller et al., 1977; Webster, 1984). These molecules, either alone or in combination with other salivary compounds, could help to establish the Leishmania major parasites in the vertebrate host. The effects of adenosine receptor agonists on macrophages are not as clear cut as they are in lymphocytes: adenosine and P2 receptors have been reported either as inactivating (Hasko et al., 1996) or a shaving a stimulatory effect (Hon et al., 1997; Hu et al., 1998) on cultured macrophages. Thus, the role of P. papatasi salivary AMP and adenosine in decreasing NO production by murine macrophages (Waitumbi and Warburg, 1998) and in exacerbating Leishmania transmission remains to be investigated.
Although P. papatasi has pharmacologically active amounts of salivary adenosine and AMP, our unpublished observations indicate that this is not the case with the New World sand fly Lutzomyia longipalpis. Instead, Lutzomyia longipalpis salivary glands have large amounts of a secretory 5′-nucleotidase (R. Charlab, J. Valenzuela, E. Rowton and J. M. C. Ribeiro, in preparation) which, acting in concert with the salivary apyrase (Ribeiro et al., 1986), produces vasodilatory and anti-platelet adenosine from ADP and ATP released by the injured skin while the fly is probing for a blood meal. These differences in pharmacological strategies among flies from the same family, but from genera that diverged not earlier than the last separation of the continental plates, stresses the diversity of compounds found in the salivary glands of blood-sucking arthropods.
The observed inhibition of both Leishmania major and macrophage protein phosphatases by 5′-AMP, with IC50 values of approximately 10 μmol l−1, is somewhat surprising because 5′-AMP was not listed as a protein phosphatase inhibitor in recent reviews on the subject. Perhaps AMP is modifying the substrate used in our assay (phosphorylase a, the enzyme phosphorylating glycogen to produce glucose 1-phosphate) to render it insensitive to protein phosphatases. Indeed, phosphorylase is allosterically regulated by AMP (Gutierrez Merino et al., 1980), and AMP inhibits phosphorylase a dephosphorylation by muscle (Bot and Dosa, 1971) and hepatic PP1A/2A (Farkas et al., 1988), with IC50 values of the same magnitude as those found in our assays. Hence, the reported salivary protein phosphatase inhibitory activity of P. papatasi may be an artifact of the assay employed. AMP may act on the substrate, rendering it inaccessible to the phosphatase. However, in the process of characterising the activity, we discovered large amounts of pharmacologically active 5′-AMP and adenosine in the salivary glands of this fly.
ACKNOWLEDGEMENTS
We are grateful to Drs Rosane Charlab, Ivo Francischetti, Raymond Jacobson and Jesus Valenzuela for comments on a draft of this manuscript. This research was supported by grant 363/96-1 from The Israel Science Foundation funded by the Israeli Academy of Sciences and Humanities, The Deutsche Forschungsgemeinschaft (DFG) project on leishmaniasis in Israel and West Bank, AID/CDR Grant TA-MOU-96-C14-142. Additional support was provided by the National Institutes of Health grant AI36382 (to G. C. Lanzaro, University of Texas Medical Branch, Galveston, TX, USA).