Purification and cloning of the salivary nitrophorin from the hemipteran Cimex lectularius

The saliva of blood-sucking arthropods contains a large array of antiplatelet, anticlotting and vasodilatory compounds that assist feeding by these animals (Ribeiro, 1987; Law et al. 1992; Champagne and Valenzuela, 1996). Because blood-feeding behavior developed independently several times, even within the same insect order (Sweet, 1979), salivary antihemostatic substances are diverse. For example, vasodilatory substances found in blood-sucking arthropods include enzymes that destroy vasoactive amines, several different peptides, prostaglandins and nitric oxide (NO) (Ribeiro, 1995).

Two blood-sucking Hemiptera, Rhodnius prolixus and Cimex lectularius, use salivary NO as their main vasodilator. In both insects, the unstable and volatile gas NO is stored and transported from the salivary glands to the host skin by a heme protein that we named nitrophorin (Ribeiro and Walker, 1994; Champagne et al. 1995; Valenzuela et al. 1995). In R. prolixus, this Fe(III) heme protein releases NO on dilution at neutral or alkaline pH (Ribeiro et al. 1993) and on binding to histamine (Ribeiro and Walker, 1994). Although R. prolixus and C. lectularius belong to the order Hemiptera, they are from different families (Reduvidae and Cimicidae, respectively) and are thought to have evolved independently to hematophagy (Sweet, 1979; Cobben, 1979). We therefore wanted to determine the relationship between these two insect nitrophorin proteins.

Cimex lectularius colonies were maintained as previously described (Valenzuela et al. 1995). Salivary glands of insects 8–10 days after blood feeding were dissected and stored in sodium acetate saline buffer (10 mmol l⁻¹ sodium acetate at pH 5.0 with 150 mmol l⁻¹ NaCl) at −75 °C.

Absorption spectra of 60 μl of salivary gland homogenate (in 10 mmol l⁻¹ sodium acetate, pH 5.0) were measured using a lambda 19 spectrophotometer (Perkin Elmer Cetus, Norwalk, CT, USA) before and after the addition of 8 μl of 0.1 mol l⁻¹ sodium phosphate solution (to give a final pH of 7.2) and exposure of the salivary homogenate to argon for 45 min.

Argon is a noble and inert gas that cannot react with any biological molecule, but can physically displace other gases, such as NO. Nitric oxide was obtained from Matheson Gas Products, Inc. (Baltimore, MD, USA). Lines from which NO gas was obtained were purged for 3 h with argon that had passed through an oxygen scrubber column. The NO lines were also bubbled twice through water to remove NO anhydrides. A saturated solution of NO was obtained by bubbling previously argon-purged water with NO gas for 15 min using a glass test tube with a rubber cap.

High-performance liquid chromatography (HPLC) was carried out using a CM4-100 pump and SM4-100 dual-wavelength detector obtained from Thermo Separations (Riviera Beach, FL, USA) with computer analysis as described previously. Salivary glands (200 pairs in 1 ml of buffer in conical 1.5 ml tubes) were homogenized and centrifuged at 10 000 g for 5 min; the supernatant was then injected into a TSK-gel DEAE-5PW HPLC column (7.5 cm×7.5 mm) (Supelco Inc., Bellefonte, PA, USA). Buffer and gradient conditions were as follows: solution A, 20 mmol l⁻¹ Tris/HCl, pH 8.3; solution B, solution A plus 2 mol l⁻¹ NaCl. A linear gradient from 100 % solution A to 75 % solution A and 25 % solution B over a 30 min period at a flow rate of 0.5 ml min⁻¹ was used. Eluting material was detected at 280 nm for total proteins and at 400 nm for heme proteins. Although the wavelength of 220 nm is more sensitive for detecting peptide bonds, the salt buffers are not transparent at this wavelength and, for this reason, the wavelength of 280 nm was used. The wavelength of 400 nm was chosen as being intermediate between the NO-bound and NO-unbound absorption peaks of nitrophorin. Fractions containing heme protein were injected into a Hamilton reverse-phase octadecyl PRP-∞ HPLC column (10 cm×4.4 mm). Buffer and gradient conditions were as follows: solution C, 10 % acetonitrile in water plus 0.1 % trifluoroacetic acid; solution D, 60 % acetonitrile in water plus 0.1 % trifluoroacetic acid. A linear gradient from 100 % solution C to 100 % solution D over a period of 60 min at a flow rate of 0.5 ml min⁻¹ was used. Eluting material was detected at 220 nm and 280 nm, and eluting fractions were collected at 1 min intervals.

The molecular mass and purity of purified protein were determined by SDS–PAGE (8 % to 25 % gel) using a Phast system (Pharmacia, Piscataway, NJ, USA) and by matrix-assisted laser desorption mass spectroscopy (MALDMS). SDS–PAGE standards were cytochrome c, myoglobin, carbonic anhydrase and ovalbumin. MALDMS, amino acid analysis, tryptic digestion and partial protein sequencing of the purified protein were carried out at the Harvard Microchemistry Facility (Cambridge, MA, USA).

Heme content was determined by reverse-phase HPLC using the protocol described above for the final purification step of the heme protein. Authentic hemin (Sigma, St Louis, MO, USA) was used to calibrate a standard curve.

Inositol trisphosphate phosphatase activity was measured in 100 mmol l⁻¹ Tris/HCl, pH 7.5, 1 mmol l⁻¹ MgCl₂, and orthophosphate was measured with a microtiter plate modification of the method of Ernster et al. (1950). We followed the exact protocol, except that all volumes were reduced in order to obtain a final volume of 120 μl per well, and the microplate was read in an enzyme-linked immunosorbent assay (ELISA) plate reader.

C. lectularius salivary gland mRNA was isolated from 500 gland pairs (dissected 5–7 days after a blood meal). A cDNA library was prepared following the specifications from the ZAP-cDNA synthesis kit (Stratagene, La Jolla, CA). The unamplified library has a complexity of 3.5×10⁵ recombinants.

Another cDNA library was subsequently prepared from salivary glands of C. lectularius mRNA using the SMART cDNA synthesis kit (Clontech, Palo Alto, CA, USA) from 150 ng of mRNA. This library, which is based on primers that recognize the cap and poly(A) region of the cDNA, could be made with a greatly reduced number of salivary glands (60 pairs instead of 500 pairs) allowing us to obtain full-length clones that contained the amino-terminal region of the C. lectularius nitrophorin clone.

For polymerase chain reaction (PCR) experiments, we isolated mRNA from C. lectularius salivary glands (100 pairs). The mRNA was then reverse-transcribed to cDNA using SuperscriptII RNase H⁻ reverse transcriptase (Gibco-BRL, Gaithersburg, MD, USA) following the manufacturer’s protocol. Degenerate oligonucleotide primers were designed from the peptides obtained by tryptic digestion of the purified nitrophorin (see Results).

A 200 base pair (bp) PCR clone labeled with dUTP-digoxigenin (Genius System; Boehringer Mannheim, Indianapolis, IN, USA) was used to screen a C. lectularius salivary gland cDNA library.

The insert of the isolated plasmid was sequenced with an ABI model 373 cyclosequencer, first using the T3 and T7 primers and then custom-designed primers constructed from the internal sequence of the nitrophorin clone.

To obtain the sequence of the 5′ region of the nitrophorin cDNA clone, cDNA obtained by PCR amplification containing the SMART sequence was used as a template for the PCR reaction. The primers used in this reaction were the 5′ primer that recognizes the SMART sequence (Clontech) and a primer designed from the upstream region of the non-full-length C. lectularius nitrophorin cDNA clone (CN253R; 5′-CAT AGT CTC TGT GAT GGT GT-3′).

Analysis of the predicted protein sequence was performed using the BLAST programs (http://www.ncbi.nlm.nih.gov/BLAST/) and the Sequence Analysis Services programs (http://molbio.info.nih.gov) at the computational molecular biology Internet site from the National Institutes of Health.

The brick-orange color of C. lectularius salivary glands suggested the presence of one or more hemoproteins (Fig. 1). The amount of this type of protein in the salivary gland is calculated spectroscopically from the ratio of heme absorbance (438 nm) to total protein absorbance (280 nm) in the crude extract (Fig. 2). According to this ratio and the equivalent ratios from other pure hemoproteins such as cytochrome c, myoglobin, horseradish peroxidase and catalase, salivary hemoproteins could account for approximately 40 % of the total soluble protein of the salivary gland. As shown previously (Valenzuela et al. 1995), the ability of C. lectularius salivary hemoprotein(s) to bind and release the vasodilatory molecule NO can be observed spectroscopically from the shift in the absorption maxima of the NO-bound form (438 nm) and the unbound form (388 nm). There is a blue shift from 438 to 388 nm in response to a change to pH 7.0, with a shoulder remaining at 438 nm (Fig. 2). On exposure of the homogenate to argon, the size of the shoulder at 438 nm diminishes, indicating displacement of NO by argon, while the size of the peak at 388 nm increases (Fig. 2), indicating the presence of unbound heme. As we demonstrated previously (Valenzuela et al. 1995), the addition of authentic NO to the sample produces a red shift from 388 nm to 438 nm (not shown). The results indicate that C. lectularius salivary hemoprotein(s) must constitute a large proportion of the total salivary protein.

C. lectularius salivary gland soluble homogenate was applied to an anion-exchange (DEAE) HPLC column. The eluate was monitored at both 280 nm (Fig. 3A) and 400 nm (Fig. 3B). A well-isolated peak absorbing at both 280 and 400 nm was observed (Fig. 3). The absorption spectrum of the fraction with a retention time of 13 min shows a maximum at 388 nm (Fig. 3B, inset, solid line), indicating the presence of the unbound form of the protein. This result suggests that NO is not bound, as expected, to the nitrophorin, because of the pH used in the eluting buffer (pH 8.3). After addition of NO to the purified material, adjusted to pH 5.0, the spectrum shifted to give a peak at 438 nm (Fig. 3B, inset, dashed line), a spectrum identical to that observed in the salivary extract (Fig. 2; see also Valenzuela et al. 1995). The direction of the shift indicates that the nitrophorin binds NO with an Fe(III) heme group (Antonini and Brunori, 1971).

After this purification step, the peaks at 388 nm or 438 nm both had a similar total area to that of the absorbance peak at 280 nm, in contrast to the absorbance of the salivary gland homogenate, in which the ratio of the area of the peak at 280 nm to that at 388 nm was 2.1:1 (Fig. 2), indicating a twofold purification of the nitrophorin(s). SDS–PAGE of fraction 13 resulted in a single band with an apparent molecular mass of 32 kDa (Fig. 4, lane 3).

Reverse-phase HPLC of fraction 13 from DEAE chromatography revealed only one major peak at 220 nm eluting at 43 min (Fig. 5A). The protein in this fraction was homogeneous and had a molecular mass of 32.983 kDa as determined by laser desorption/mass spectrometry (not shown).

The absorbance at 280 nm indicated the presence of an earlier peak eluting at 34 min (Fig. 5B). This earlier peak had an absorption spectrum and an identical retention time to those of pure hemin (not shown). The heme group was thus dissociated from the protein owing to the solvent (acetonitrile) and pH (2.3) used in this purification step, a behavior typical of hemoprotein of the cytochrome b type (Antonini and Brunori, 1975). Calculations of the amount of protein (estimated by measuring the areas under the 280 nm absorbance/time graphs standardized with bovine serum albumin and assuming a molecular mass of 33 kDa) indicated the presence of 106 pmol of protein (in the peak eluting at 43 min), and estimation of the hemin content (by similarly measuring the areas standardized with hemin) indicated the presence of 94 pmol of hemin (in the peak eluting at 34 min). A second purification run with 600 salivary gland pairs gave similar results (301 pmol of protein and 293 pmol of hemin). Taken together, these data indicate that we have isolated C. lectularius nitrophorin in pure form and that this protein binds 1 mol of heme per mol of apoprotein.

Attempts to obtain the amino-terminal sequence of C. lectularius nitrophorin were unsuccessful, indicating a blocked amino-terminal group. However, four tryptic peptides of the pure nitrophorin were sequenced (peptide 1, KDDPSDFLFWIGDLNVR; peptide 2, LFDGWTEPQVTF-KPTYK; peptide 3, IQPLSYNSLTNY; peptide 4, NT?FTIY?K).

Sense and antisense degenerate DNA primers prepared from the sequence DPSDFLFWI (from peptide 1) and GWTEPQVTFK (from peptide 2) generated a single 200 bp PCR product (not shown). This 200 bp PCR product was labeled with digoxigenin and used to screen a C. lectularius cDNA library. Two clones, one of 1000 bp and the other of 900 bp, were completely sequenced in both directions, and both contained the same sequence. We were unsuccessful in obtaining a full-length clone containing the initiation codon (MET) when screening this library. A novel PCR-based cDNA synthesis technique (Clontech) was used to obtain the 5′ end of the nitrophorin cDNA clone. A 400 bp PCR product containing the sequence of the 5′ end of the open reading frame and the start codon ATG of C. lectularius nitrophorin cDNA was obtained. We compared the sequence of the isolated PCR product (minus the untranslated region and initiation codon region) with the sequence of the isolated clone using the Sequence Analysis Services programs and obtained 100 % identity between these two sequences.

The full-length C. lectularius nitrophorin cDNA (Fig. 6) has a predicted open reading frame of 906 bp and codes for an unprocessed protein of 302 amino acid residues containing all four peptides sequenced from the purified protein (underlined in Fig. 6). Secreted proteins usually contain a signal peptide that is cleaved before secretion. The cleavage site of the signal peptide was predicted by the SignalP V1.1 program via the Internet at the Center for Biological Sequence Analysis at the Technical University of Denmark (signalp@genome.cbs.dtu.dk) (Nielsen et al. 1997). The program predicted cleavage between residues 20 and 21 (Fig. 6, arrow), which could generate a processed protein of 282 residues with a predicted molecular mass of 31 710 Da, which is very similar to the molecular mass of the native protein (32.9 kDa). The predicted amino acid composition from the deduced protein of the isolated clone is very similar to the amino acid composition of the native protein (not shown).

The predicted protein contains two potential glycosylation sites (Marshall, 1972) at residues 122–125 (NETI) and 199–202 (NATH). In addition, the protein contains various protein kinase C phosphorylation sites, with consensus sequences of TXK, TXR and SXK, and a potential N-myristylation site at residues 136–141 (GGIVTS). These sites could account for the additional non-peptide mass added to the parent protein and to the discrepancy of 1.2 kDa between the mass deducted from the primary sequence of the clone and that of the pure native protein.

A comparison of C. lectularius nitrophorin with R. prolixus nitrophorin yielded no significant similarities (not shown). Additionally, the BLAST program did not find R. prolixus nitrophorin when searching for homologies with C. lectularius nitrophorin.

When we searched for similarities between the C. lectularius nitrophorin sequence and other proteins in the available data bases, we found significant homology between C. lectularius salivary nitrophorin and different inositol phosphatases (P<3.2×10⁻¹⁰ indicated by the BLAST program; http://www.ncbi.nlm.nih.gov/BLAST/). Fig. 7 shows, as an example, an alignment between the 75 kDa inositol 1,4,5-trisphosphate 5-phosphatase of human platelets (Ross et al. 1991) and C. lectularius nitrophorin. Attempts to detect inositol 1,4,5-phosphatase activity in C. lectularius nitrophorin were unsuccessful (not shown).

The high level of homology between the C-terminal region of C. lectularius nitrophorin and different inositol 5-phosphatases suggests that C. lectularius nitrophorin evolved from a phosphatase able to bind heme and NO or, alternatively, that the homology reflects an ancient and conserved structural domain that has been employed in different ways in diverged lineages to give rise to distinct proteins having unique functions. In contrast, R. prolixus nitrophorin evolved from a different molecule despite having the ability to bind heme and NO in a similar way. Because C. lectularius and R. prolixus belong to two different families of Hemiptera that evolved independently to blood feeding, their nitrophorins must have arisen as a product of convergent evolution. This indicates the importance of this type of protein for insect survival by acting as the transporter of NO, a vasodilator and an inhibitor of platelet aggregation at the feeding site. The results in this paper also raise the possibility that some inositol phosphatases may be regulated by NO through a heme binding site.

We thank Dr Rosane Charlab, Dr Joseph Vinetz and two anonymous reviewers for critical reviews of the manuscript, Dr William Lane for the excellent peptide sequencing facility and Brenda Rae Marshall for preparation of the manuscript.