Innate immunity in insects: surface-associated dopa decarboxylase-dependent pathways regulate phagocytosis, nodulation and melanization in medfly haemocytes


Prof. V. J. Marmaras, Department of Biology, University of Patras, 26500 Patras, Greece.
Senior author: Prof. V. J. Marmaras


Phagocytosis, melanization and nodulation in insects depend on phenoloxidase (PO) activity. In this report, we demonstrated that these three processes appear to be also dependent on dopa decarboxylase (Ddc) activity. Using flow cytometry, RNA interference, immunoprecipitation and immunofluorescence, we demonstrated the constitutive expression of Ddc and its strong association with the haemocyte surface, in the medfly Ceratitis capitata. In addition, we showed that Escherichia coli phagocytosis is markedly blocked by small interfering RNA (siRNA) for Ddc, antibodies against Ddc, as well as by inhibitors of Ddc activity, namely carbidopa and benzerazide, convincingly revealing the involvement of Ddc activity in phagocytosis. By contrast, latex beads and lipopolysaccharide (LPS) did not require Ddc activity for their uptake. It was also shown that nodulation and melanization processes depend on Ddc activation, because antibodies against Ddc and inhibitors of Ddc activity prevent haemocyte aggregation and melanization in the presence of excess E. coli. Therefore, phagocytosis, melanization and nodulation depend on haemocyte-surface-associated PO and Ddc. These three unrelated mechanisms are based on tyrosine metabolism and share a number of substrates and enzymes; however, they appear to be distinct. Phagocytosis and nodulation depend on dopamine-derived metabolite(s), not including the eumelanin pathway, whereas melanization depends exclusively on the eumelanin pathway. It must also be underlined that melanization is not a prerequisite for phagocytosis or nodulation. To our knowledge, the involvement of Ddc, as well as dopa and its metabolites, are novel aspects in the phagocytosis of medfly haemocytes.


bovine serum albumin


dopa decarboxylase




focal adhesion kinase


fluorescein isothiocyanate


horseradish peroxidase


immunoglobulin G




phenylmethylsulphonyl fluoride








sodium dodecyl sulfate–polyacrylamide gel electrophoresis


small interfering RNA




Tris-buffered saline.


In insects, phagocytosis, nodulation, melanotic encapsulation and wound healing are important innate immune responses against pathogens and parasites.1–4 Phagocytosis refers to the engulfment of entities by an individual haemocyte; nodulation refers to the multicellular haemocytic aggregates that entrap a large number of bacteria; encapsulation refers to the binding of haemocytes to larger targets, such as parasites and nematodes; and melanization refers to the pathway leading to melanin formation. Melanization is central to the defense against a wide range of pathogens and participates in nodule and capsule formation in some lepidopteran and dipteran insects, such as Pseudoplusia and Drosophila.5,6

Recently, it has been reported that the prophenoloxidase (proPO) activation system, an important part of insect innate immunity, is also involved in the phagocytosis of bacteria.4 Briefly, phagocytosis in medfly haemocytes depends on RGD-binding receptors, focal adhesion kinase (FAK)/sarcoma (Src) and mitogen-activated protein kinase signalling pathways that induce the secretion of phenoloxidase activating peptides among other components that might activate haemocyte-surface-associated proPO. Latex beads and lipopolysaccharide (LPS) phagocytosis do not depend on proPO activation.4,7 The proPO activation system is composed of proteins recognizing several microbial components (pattern-recognition proteins), several serine proteases and their zymogens, proPO, as well as proteinase inhibitors that function as regulatory factors.2 ProPO is synthesized in haemocytes and appears to be distributed ubiquitously in the cytoplasm as well as on the surface of haemocytes.4,8 The proPO activation system is triggered by several microbial components, such as LPS and peptidoglycans, thus ensuring that the system will become active in the presence of potential pathogens.

Activated phenoloxidase (PO), the last component of the proPO activation system, catalyses the hydroxylation of tyrosine to dihydroxy-phenyl-alanine (dopa), an important branch-point substrate. Dopa can be oxidized by PO to dopaquinone, which, via PO and the dopachrome conversion enzyme, forms indole-5,6-quinone, ultimately resulting in melanin. Dopa may also be decarboxylated by dopa decarboxylase (Ddc), a multifunctional enzyme in insects, to form dopamine. A PO-based oxidation of dopamine leads to dopaminequinone and finally the cross-linking and melanization of proteins. Ddc catalyses the production of the neurotransmitters dopamine and serotonin and is involved in wound healing, parasite defense, cuticle hardening, melanization and in the behavior of insects.9 Ddc is found in epidermal, neural and ovarian cells and in haemocytes. A Ddc mRNA transcript was constitutively expressed at low levels in the haemocytes among other tissues, and the expression of Ddc mRNA in the haemocytes of Pseudaletia separata was enhanced by injection of an insect cytokine, growth-blocking peptide.10 Microarray analysis demonstrated that Ddc levels increased 11-fold after 3 hr of septic infection of Drosophila with a mixture of Escherichia coli and Micrococcus luteus.11

The data proposing the involvement of Ddc in cellular immunity, the expression of Ddc in haemocytes, its participation in tyrosine metabolism and the involvement of proPO activation system in phagocytosis, nodulation and melanization processes4,11,12 prompted us to explore the involvement of Ddc in phagocytosis, nodulation and melanization, major defense mechanisms in insects. In the present study, we demonstrated the haemocyte surface expression of Ddc and its involvement in the regulation of these three unrelated defense procedures, in the medfly Ceratitis capaitata.

Materials and methods

Antibodies, inhibitors and activators

Polyclonal anti-human Ddc (Cat# AB1569; Chemicon International, Temecula, CA) were used to investigate the role of C. capitata Ddc because of the high identity and similarity of its amino acid sequences. Bioinformatics show that the Ddc gene is highly conserved in phylogenetically diverged species, such as insects and humans. Comparison analysis of the predicted amino acid sequences between C. capitata (NCBI sequence viewer, accession no.: AAM92163) and human Ddc (NCBI sequence viewer, accession no.: CAG33005) showed 62% identity and 76% positivity in amino acid sequences of medfly and human Ddc. Furthermore, conservation of the Ddc sequence is supported by the fact that the binding region for the coenzyme pyridoxal 5-phosphate is almost completely conserved because the predicted sequence in C. capitata is -NLNPHKW-, whereas in human Ddc it is -NFNPHKW-. Polyclonal antibodies against Manduca sexta proPO were kindly offered by Prof. Michael R. Kanost (Department of Biochemistry, Kansas State University, Manhattan, KS, USA).

Goat anti-rabbit immunoglobulin G (IgG) conjugated to horseradish peroxidase (HRP) was purchased from Cell Signalling Technology (Beverly, MA, USA), and goat anti-rabbit IgG conjugated to fluorescein isothiocyanate (FITC), and anti-tubulin were from Santa Cruz Biotechnology (Santa Cruz, CA, USA). The Ddc water-soluble inhibitors carbidopa and benzeraside, LPS–FITC, latex beads–FITC and anti-phospho-FAK were obtained from Sigma (St Louis, MO, USA). E. coli–FITC was prepared according to Mavrouli et al.4 Amerlex-M second antibody reagent was purchased from Amersham Life Science (Bucks., UK). Other reagents were obtained as indicated.

Collection of haemocytes and cell-free haemolymph

Haemolymph from third instar larvae was collected and centrifuged (2000 g, 10 min, 4°). Sedimented haemocytes were washed three times with Ringer's solution (128 mm NaCl, 18 mm CaCl2, 1·3 mm KCl, 2·3 mm NaHCO3, pH 7·0). The viability of haemocytes was assessed, under microscopy, by exclusion of Trypan blue dye (Sigma).


Haemocytes were lysed in lysis buffer [50 mm Tris, pH 7·4, 150 mm NaCl, 5 mm EDTA, 1% Triton X-100, 1 mm sodium orthovanadate, 5 mm NaF, 1 mm phenylmethylsulphonyl fluoride (PMSF), 10 µg/ml of leupeptin and 10 U/ml of aprotinin] and protein crude extract was collected. For immunoprecipitation, 2 µg of anti-Ddc was incubated with 40 µl of Amerlex-M secondary antibody reagent (goat anti-rabbit IgG conjugated on paramagnetic beads) for 30 min at 37° and the sedimented beads with the bound anti-Ddc were isolated. In another tube, 300 µg of protein crude extract, in 300 µl of lysis buffer, was incubated with 40 µl of Amerlex-M secondary antibody reagent for 30 min at 37° and centrifuged for 10 min at 10 000 g to remove non-specific bound proteins onto the beads. The supernatant was collected and supplemented with the paramagnetic beads with the bound anti-Ddc and incubated for 2 hr at 37°. Immune complexes on the paramagnetic beads were washed four times with Tris-buffered saline (TBS) (10 mm Tris-HCl, pH 7·5, 100 mm NaCl). Proteins were eluted from the beads by boiling samples for 3 min in 50 µl of electrophoresis sample buffer. Samples were electrophoresed and immunoblotted with polyclonal anti-Ddc.

Protein determination

Protein concentration was determined with a modified solution containing 10% (w/v) Coomassie G250 (Merck, Darmstadt, Germany) in 5% (v/v) ethanol, 10% (v/v) H3PO4 and 1 mg/ml of bovine serum albumin (BSA), in TBS, as a standard.13 The absorbance was recorded at 595 nm.

Sodium dodecyl sulfate–polyacrylamide gel electrophoresis and immunoblot analysis

Sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) was performed on 10% acrylamide and 0·10% bisacrylamide slab gels. Samples were electrophoretically analysed and electroblotted onto Immobilon P polyvinylidene fluoride membranes (Millipore Corp., Bedford, MA, USA). Membranes were incubated in SuperBlock™ Blocking Buffer (Pierce, Rockford, IL, USA) for 1 hr at room temperature. Subsequently, membranes were incubated overnight at 4° with polyclonal anti-Ddc and then diluted 1 : 2000 in TBS (10 mm Tris-HCl, pH 7·5, 100 mm NaCl) containing 10% (v/v) SuperBlock™ Blocking Buffer and 0·05% (v/v) Tween 20. Membranes were washed with TBS containing 0·05 (v/v) Tween 20, followed by incubation with horseradish peroxidase-linked secondary antibody (Cell Signalling Technology) for 1 hr at room temperature. Immunoreactive proteins were visualized on X-ray film by enhanced chemiluminescence (ECL) (Amersham). In the case of phospho-FAK and tubulin immunoblot analyses, antibodies were diluted 1 : 1000 and 1 : 5000, respectively. Prestained Protein Markers, broad range, were used to indicate the size of the protein bands (Cell Signalling Technology).


Isolated haemocytes were suspended in 100 µl of Grace's medium (5 × 105 cells) and allowed to attach on glass slides for 10 min at 25°. Slides were washed with Ringer's solution to remove non-adherent haemocytes. The resulting monolayers were fixed with 4% formaldehyde solution for 10 min. To investigate intracellular Ddc, slides were embedded in 0·01% Triton X-100 for 5 seconds and immediately washed in TBS. Haemocytes were treated with protein blocking agent (Pierce) for 10 min to reduce non-specific binding. Slides were incubated with polyclonal anti-Ddc (1 : 100 dilution) for 1 hr at 25° in a humid atmosphere. Following antibody treatment, slides were washed with TBS and further incubated with goat anti-rabbit IgG–FITC conjugate for 10 min. Cells were washed with TBS and observed under ultraviolet (UV) microscope. Non-immune rabbit serum and goat antirabbit IgG–FITC conjugate staining were used as negative controls (Fig. 1a,b).

Figure 1.

 Demonstration of surface dopa decarboxylase (Ddc) on medfly haemocytes. Ddc was identified on the haemocyte surface (c) and inside haemocytes (d) with indirect immunocytofluorescent staining. Control staining was performed with either non-immune serum (a) or secondary antibody (b) to detect a non-specific signal. Ddc was also identified in haemocyte protein crude extract by immunoprecipitation and immunoblot analysis (e) and on the surface of haemocytes by flow cytometry (f), using specific antibodies. Ab”, non-specific immunoprecipitation caused by the secondary antibody; E, endogenous fluorescence; IP, immunoprecipitation; FITC, fluorescein isothiocyanate; L, haemocyte lysate; MW, molecular weight; S, supernatant after precipitation.

Flow cytometry analysis

To determine surface Ddc, larval haemocytes (5 × 105 cells) were incubated in 100 µl of Grace's insect medium with anti-Ddc (1 : 100 dilution) for 10 min at 8°. Subsequently, FITC-labelled antirabbit IgG (Santa Cruz Biotechnology) was added and incubated for 10 min at 8°. Incubation of haemocytes with only FITC-labelled anti-rabbit IgG was used to reveal non-specific binding. To study E. coli phagocytosis, larval haemocytes (5 × 105 cells) were incubated in 100 µl of Grace's insect medium containing E. coli–FITC (10 bacteria per haemocyte) for 10 min at 25°, in the presence or absence of anti-Ddc (1 : 100 dilution), anti-proPO (1 : 100 dilution), phenylthiurea (PTU) (25 µm) or Ddc inhibitors (carbidopa 25 µm and benzeraside 25 µm). Internalized E. coli–FITC was measured by quenching attached E. coli–FITC with 200 µl of Ringer's solution containing Trypan blue (4%). Approximately 20 000 cells from each sample were analysed by flow cytometry using a Coulter EPICS-XL-MCL cytometer (Coulter, Miami, FL), and data were processed using the xl-2 software (Coulter).

Silencing of dopa decarboxylase

To silence Ddc, isolated haemocytes (5 × 105 cells) were incubated in 100 µl of Grace's insect medium for 7 hr. Every 75 min, medium was supplemented with 50 nì small interfering RNA (siRNA) for Ddc (Ambion, Austin, TX, USA). During this period, Ddc protein molecules were degraded and the synthesis of new ones was blocked by the added siRNA. To investigate surface Ddc silencing, or haemocyte phagocytosis, cells were processed as described earlier and analysed by flow cytometry. The sense sequence of Ddc siRNA was GUUACAUGCCUACUUCCCCUU, which was designed and provided by Ambion according to the C. capitata Ddc mRNA nucleotide sequence (NCBI sequence viewer, accession no.: Y08388). The antisense sequence was GGGGAAGUAGGCAUGUAACUU. An irrelevant siRNA provided by Ambion (Cat# 4635) was used as a negative control.


Haemocyte surface expression of Ddc

We have previously demonstrated the presence of medfly haemocyte-surface-associated proPO and its involvement in the processes of phagocytosis, nodulation and melanization.4 In the present study, we demonstrated the presence of medfly haemocyte-surface-associated Ddc and we explored its involvement in the above-mentioned unrelated procedures.

In this study, antibodies against a region of human Ddc, which is homologous to medfly Ddc, as well as siRNA corresponding to C. capitata Ddc, were used to investigate Ddc in medfly haemocytes (see the Materials and methods). Immunocytochemical experiments were performed to observe the localization of Ddc in freshly isolated primary haemocytes. Immunostaining revealed that Ddc was distributed ubiquitously in the plasmatocytes of the third instar larvae, the major type of circulating medfly haemocytes. Ddc was found inside (Fig. 1d), as well as on the surface of (Fig. 1c), haemocytes. Immunoprecipitation of Ddc in haemocyte lysate, followed by immunoblot analysis, also demonstrated the presence of Ddc in medfly haemocytes (Fig. 1e). Flow cytometry analysis clearly showed, in a more convincing manner, that Ddc was present on the surface of nearly all haemocytes (Fig. 1f).

Interestingly, changes in the level of haemocyte-surface Ddc were observed during larval development, with a peak in the white pupae (Fig. 2a). A similar developmental profile has been observed for haemocyte-surface-associated proPO, the key enzyme of the proPO activation system, which is involved in phagocytosis.4 These results, however, do not clarify whether haemocyte-surface Ddc is a strongly bound protein in the haemocyte membrane or whether free Ddc molecules bind weakly onto the haemocyte surface. To gain an understanding concerning the association of Ddc with the haemocyte surface we treated haemocytes with 0·6 m NaCl, and its effect on haemocyte-surface-associated Ddc expression was monitored by flow cytometry. This analysis showed that NaCl did not affect the level of haemocyte-surface-associated Ddc (Fig. 2b), supporting that Ddc is rather strongly bound to the haemocyte surface. Consequently, Ddc is constitutively present on the haemocyte surface during larval development, supporting a role for haemocyte physiology and, probably, a role in mediating immune and inflammatory responses.

Figure 2.

 Developmental profile of dopa decarboxylase (Ddc) at the third instar larval stage and its stable association with the cell membrane. Flow cytometry analysis revealed the presence of Ddc on the haemocyte surface at the wandering stage, followed by a peak at the white pupa and a decrease 2 hr later (a). Indirectly fluorescein isothiocyanate (FITC)-labelled Ddc remains the same on the haemocyte surface after incubation with 0·6 m NaCl (b). AFU, arbitrary fluorescent units = (median sample fluorescence value) − (median endogenous fluorescence value).

Because it was repeatedly observed in flow cytometry experiments that during short-term incubation the phagocytic haemocytes take up antibodies and other abiotic components readily, we initially quenched surface-exposed FITC-labelled antibodies with Trypan blue to establish whether phagocytic haemocytes had taken up Ddc antibodies. Indeed, ≈ 10–13% of FITC-labelled Ddc antibodies were internalized in haemocytes (data not shown). As in all experiments the percentage of FITC-labelled Ddc antibodies internalized in haemocytes remained rather trivial and constant, the rest of the experiments were performed without Trypan blue quenching.

Medfly Ddc siRNA reduces haemocyte phagocytosis

Haemocytes respond to infections by phagocytosis, nodule or capsule formation and melanization. Given the fact that the activated haemocyte-surface-associated proPO is involved in the uptake of bacteria by medfly haemocytes,4 we focused our interest on whether haemocyte-surface-associated Ddc, another key enzyme of tyrosine metabolism, is also a good candidate to mediate the phagocytosis process.

To explore whether Ddc might play a role in haemocyte phagocytosis, Ddc-siRNA was initially transfected into haemocytes of wandering-stage larvae, to test whether it silences Ddc expression. Suspended haemocytes were incubated in Grace's insect medium for 7 hr with an siRNA corresponding to C. capitata Ddc, or with an irrelevant siRNA (negative control) with no significant sequence similarity to the medfly genome, both obtained from Ambion. A control experiment was performed without the presence of any siRNA. The viability of hemocytes was 65–75% after 7–8 hr of culture. The silencing of Ddc expression on the medfly haemocyte surface was monitored by flow cytometry analysis, which clearly showed that the medfly haemocyte-surface-associated Ddc was decreased by about 90 ± 4% in haemocytes transfected with Ddc siRNA, compared to the control experiment (Fig. 3a). By contrast, transfection with the irrelevant siRNA did not reduce the haemocyte-surface-associated Ddc (Fig. 3b). Therefore, these data demonstrate convincingly, once again, the presence of Ddc at the haemocyte surface and that siRNA for Ddc blocks almost completely the expression of Ddc.

Figure 3.

 Silencing of dopa decarboxylase (Ddc) with small interfering RNA (siRNA). Haemocytes were incubated for 7 hr with or without siRNA for Ddc. Flow cytometry showed a 90 ± 4% reduction of surface Ddc using specific antibodies (a). A control experiment was performed with an irrelevant siRNA and showed no reduction of surface Ddc (b). Non-specific binding of the secondary antibody appeared to be not significant. FITC, fluorescein isothiocyanate.

The above results posed a challenge for us to explore whether Ddc gene silencing affects the haemocyte phagocytosis process. For this purpose, suspended medfly haemocytes in Grace's medium were incubated for 7 hr with siRNA corresponding to medfly Ddc, in the presence of E. coli–FITC, latex beads–FITC or LPS–FITC, and phagocytosis/endocytosis was monitored by flow cytometry (Fig. 4a). Irrelevant siRNA was used as a negative control. The results clearly showed a decrease of E. coli phagocytosis in the presence of Ddc siRNA of about 30 ± 6%, compared with the controls (Fig. 4a). On the contrary, the presence of Ddc siRNA had no effect on the phagocytosis of latex beads or LPS (Fig. 4a). Irrelevant siRNA-transfected haemocytes or untransfected haemocytes with Ddc siRNA, showed a similar degree of phagocytosis, thus demonstrating the specificity of Ddc siRNA function (Fig. 4).

Figure 4.

 Dopa decarboxylase (Ddc) is necessary for the phagocytosis of Escherichia coli. Haemocytes were incubated with fluorescein isothiocyanate (FITC) conjugated to E. coli, lipopolysaccharide (LPS) or latex beads, with or without pre-incubation with Ddc small interfering RNA (siRNA) or anti-Ddc. Ddc siRNA and anti-Ddc reduced the phagocytosis of E. coli to about 68% and 40%, respectively, but they did not influence the uptake of LPS or latex beads. A control experiment with an irrelevant siRNA showed no influence on haemocyte phagocytosis potential. The percentages represent the mean value of three experiments.

In a second series of experiments (Fig. 4b), when haemocytes were pretreated with anti-Ddc instead of Ddc siRNA before challenge with E. coli, a greater decrease of E. coli phagocytosis (about 60 ± 6%) was observed, compared with the siRNA results (about 30 ± 6%). Evidently because the depletion of siRNA was only partial owing to the short-term incubation because the viability of haemocytes decreases in long-term incubations (Fig. 4b). Similarly, antibodies against Ddc did not affect the phagocytosis of latex beads or LPS, as is the case with siRNA (Fig. 4b). Consequently, the haemocyte-surface-associated Ddc is an important determinant of bacteria phagocytosis by medfly haemocytes.

Ddc activity is required for E. coli phagocytosis

A key question that immediately arises is whether the haemocyte-surface-associated Ddc affects phagocytosis of bacteria because of its activity or because of its possible functional and/or physical association with some known intracellular signalling pathways involved in phagocytosis.4,7,14,15 To have an understanding on this issue, we repeated the experiments of Fig. 4, but instead of Ddc siRNA or anti-Ddc, haemocytes of the wandering-stage larvae were suspended in Grace's insect medium supplemented with benserazide or carbidopa, specific inhibitors for Ddc activity, before challenge with E. coli. The results clearly show that both inhibitors block E. coli phagocytosis by about 60–65%(Fig. 5). As expected, neither inhibitor had any affect on the uptake of LPS or latex beads (Fig. 5).

Figure 5.

 Dopa decarboxylase (Ddc) activity is necessary for the phagocytosis of Escherichia coli. Haemocytes were incubated with fluorescein isothiocyanate (FITC) conjugated to E. coli, lipopolysaccharide (LPS) or latex beads, with or without pre-incubation with the specific Ddc inhibitors, carbidopa or benserazide. Both inhibitors reduced the phagocytosis of E. coli to ≈ 61–65% but they did not influence the uptake of LPS or latex beads. The percentages represent the mean value of three experiments.

In another series of experiments, we explored whether Ddc affects phagocytosis, via a possible functional association with a known signalling molecule involved in haemocyte phagocytosis. The documented functional association of FAK, a key signalling molecule, with extracellular bacteria through RGD-binding receptors in medfly haemocytes,7,14 prompted us to explore any functional association of FAK with Ddc. For this purpose, suspended haemocytes in Grace's medium were supplemented with benserazide or carbidopa, inhibitors of Ddc activity, or with Ddc siRNA, before challenge with E. coli, and we observed FAK phosphorylation at Y397. Figure 6 clearly shows that FAK phosphorylation at Y397 remains unaltered in the presence of either inhibitors or siRNA for Ddc, indicating that haemocyte-surface Ddc does not functionally associate with the FAK pathway. The above data led us to postulate that haemocyte-surface-associated Ddc evidently affects phagocytosis through its activity.

Figure 6.

 There is no functional relationship between focal adhesion kinase (FAK) and dopa decarboxylase (Ddc). Haemocytes were incubated in Grace's medium with Escherichia coli. Haemocytes were pretreated with either Ddc small interfering RNA (siRNA) or Ddc-specific inhibitors (benserazide or carbidopa). Immunoblot analysis of haemocyte lysates with specific anti-phospho-FAK showed an increase of FAK phosphorylation in E. coli-treated haemocytes (lane 2) compared with untreated haemocytes (lane 1), which did not change in the presence of Ddc siRNA (lane 3), benzeraside (lane 4) or carbidopa (lane 5). Tubulin immunostaining was used as a control for equal amounts of protein loaded into each well. MW, molecular weight.

Dopa-derived metabolites and phagocytosis

The participation of PO and Ddc in phagocytosis encouraged us to search for their downstream targets responsible for phagocytosis, in tyrosine metabolism. For this purpose, suspended haemocytes were incubated in two different media: in Ringer's salt solution, which lacks all amino acids, including tyrosine (a precursor of dopa); and in Grace's complete medium, which includes all amino acids and hence tyrosine. As seen in Fig. 7(a), phagocytosis was apparent only in Grace's medium. However, when Ringer's solution was supplemented with tyrosine before challenge with E. coli, phagocytosis was observed (Fig. 7b). Similar results were obtained using certain tyrosine metabolites – dopa or dopamine – instead of tyrosine (Fig. 7c,d). By contrast, norepinephrine, another tyrosine metabolite, did not promote phagocytosis of E. coli (Fig. 7e). Therefore, certain tyrosine-derived metabolites, such as dopa and dopamine, and at least the enzymes PO and Ddc, which convert tyrosine to dopa and dopa to dopamine, are required for E. coli phagocytosis.

Figure 7.

 The active role of dihydroxy-phenyl-alanine (Dopa) and dopamine in Escherichia coli phagocytosis. Haemocytes were incubated with fluorescein isothiocyanate (FITC)-labelled E. coli in Ringer's salt solution (R). Experiments in Grace's medium (G) were performed in parallel, as positive controls of phagocytosis. Flow cytometry analysis showed no phagocytosis of E. coli in plain Ringer's salt solution (a), compared with Grace's medium, whereas E. coli was internalized in haemocytes in Ringer's salt solution when supplemented with tyrosine (b), dopa (c) or dopamine (d). No phagocytosis took place in the presence of norepinephrine (e).

It is well known that in tyrosine metabolism in insects, dopa and dopamine are branch point substrates leading to several end-products (eumelanin and sclerotin, among others). Furthermore, PO acts in several steps of tyrosine metabolism and only Ddc acts in the conversion of dopa to dopamine.16 To specify further the tyrosine metabolic pathway(s) responsible for phagocytosis in suspended haemocytes in Grace's medium, PO activity was blocked with anti-proPO or phenylthiurea (PTU), resulting in a 49 ± 5% and 52 ± 3%, respectively, decrease of phagocytosis (Fig. 8). However, when haemocytes were supplemented with dopa or dopamine, after incubation with PTU or anti-proPO and before the challenge with E. coli, we observed an increase of phagocytosis up to that of the controls (Fig. 8). These results probably denote that dopamine-derived metabolite(s) are responsible for bacteria uptake. In addition, the results of Fig. 8 also demonstrate that the pathways leading to melanin and sclerotin, downstream of dopa or dopamine, are not required for phagocytosis. Consequently, the pathways leading to eumelanin and sclerotin and, hence, in melanization, are not involved in phagocytosis. In other words, phagocytosis and melanization are distinct pathways although they share several components of tyrosine metabolism.

Figure 8.

 Dopa decarboxylase (Ddc) and not phenoloxidase (PO) activity is necessary for Escherichia coli phagocytosis. Haemocytes were incubated in Grace's culture medium in the presence of fluorescein isothiocyanate (FITC)-labelled E. coli (a). In parallel experiments, cultured haemocytes were pretreated with either anti-prophenoloxidase (anti-proPO) (d) or phenylthiurea (PTU), a specific inhibitor of proPO (g). Flow cytometry analysis showed that blocking of the surface proPO decreased phagocytosis to about 50% (d, g). The addition of dihydroxy-phenyl-alanine (dopa) (b) or dopamine (c) in the untreated haemocytes did not alter E. coli phagocytosis. On the other hand, the addition of dopa (e, h) or dopamine (f, i) in the culture media, which blocked proPO, restored phagocytosis to its initial levels. The percentages represent the mean value of three experiments.

Ddc activity is required for nodulation and melanization

The dependence of haemocyte aggregation and melanization, in the presence of bacteria, upon PO activity,4 prompted us to explore whether these processes also depend on Ddc activity.

To investigate this hypothesis, haemocytes suspended in Grace's medium were supplemented with E. coli and observed under the microscope at 0, 10 and 60 min. Figure 9 clearly shows that after 10 min of challenge, traces of melanin, in small aggregates, can be observed in haemocytes (Fig. 9b). Furthermore, 1 hr later, well-developed haemocyte–E.coli aggregates with melanized areas were formed (Fig. 9c). On the contrary, in haemocytes pre-incubated with benserazide, carbidopa, anti-Ddc or anti-proPO, 1 hr later haemocytes appear to have reduced melanin production and they did not form any aggregates (Fig. 9d,e,f,g). Finally, haemocytes pre-incubated with anti-proPO and either dopa or dopamine, remained unmelanized, but formed small aggregates (Fig. 9h,i). Melanization appeared under the microscope as darkening areas of haemocytes. These data denote that PO (which catalyses several steps of the eumelanin pathway) and Ddc (convert dopa to dopamine) activities are key regulatory components for haemocyte aggregation and melanization in the presence of bacteria. The above results demonstrate that the metabolic pathways responsible for melanization and nodulation (multicellular aggregates) are distinct. However, these unrelated procedures share a number of substrates and enzymes, such as tyrosine, dopa, PO and Ddc.

Figure 9.

 Dopa decarboxylase (Ddc) participates in melanization and nodulation. Haemocytes were incubated in Grace's culture medium in the presence of Escherichia coli and were observed under microscope (a). A developing melanization and nodule formation was obvious after 10 min (b) and 60 min (c). The presence of the Ddc inhibitors, benserazide (d) and carbidopa (e), as well as anti-Ddc (f) or anti-proPO (anti-prophenoloxidase) (g), canceled both melanization and nodule formation during the 60-min observation. A limited aggregation appeared in haemocytes treated with anti-proPO and following addition of dihydroxy-phenyl-alanine (dopa) (h) or domapine (i).


Recently, we have demonstrated the involvement of the proPO activation system, an important part of innate immunity in insects, in medfly haemocyte phagocytosis, nodulation and melanization. Haemocyte-surface-associated PO, the last component of the proPO activation system, is synthesized as an inactive zymogen and is activated upon triggering by several microbial components, through limited proteolysis.2,4,17 Activated PO catalyses the hydroxylation of tyrosine to dopa, a key branch-point substrate. Dopa can be decarboxylated by Ddc to dopamine or oxidized by PO to dopaquinone. Dopamine is also an important branch-point substrate, because dopamine-derived metabolites, either via PO or other enzymes, are used in several metabolic pathways, participating in neurotransmission, cuticular sclerotization, cross-linking of cuticular components via quinone intermediates, wound healing and melanization in immune-reactive insects.18–20 Therefore, the functional analysis of Ddc is of considerable interest. It must be noted here that catecholamines, such as dopa and dopamine, have also been demonstrated in mouse lymphocytes21 and in peripheral human lymphocytes,22 and their presence in these cells have been correlated with neurotransmission. However, cell-surface-associated Ddc and the proPO activation system have not yet been detected in mammalian lymphocytes.

Given the above data, the present report focuses on the involvement of Ddc in the processes of phagocytosis, nodulation and melanization in medfly haemocytes. The critical finding that prompted us to explore for Ddc participation in these processes was the demonstration of Ddc on the haemocyte surface (Fig. 1). Ddc has also been shown to be associated with membranes in mouse brain.23 Evidently, the presence of both PO and Ddc (the results of the present study) on the haemocyte surface was very good evidence for their functional association. To elucidate this hypothesis, we started to look for any functional relationship between Ddc and phagocytosis. Indeed, certain different experiments, using antibodies to Ddc (Fig. 4b), siRNA corresponding to medfly Ddc (Fig. 4a), or Ddc inhibitors (Fig. 5), convincingly demonstrated the participation of Ddc in the phagocytosis of E. coli (Fig. 4). In contrast to this situation in bacteria, Ddc is not required for the uptake of latex beads or LPS (Fig. 5). Consequently, PO and Ddc are required for bacteria phagocytosis, but not for the uptake of latex beads or LPS. In other words, the above data denote that the functional molecule(s) regulating phagocytosis act(s) downstream of dopa. In addition, the reported data show that the dopamine-derived metabolites leading to eumelanin or sclerotin formation could not be ranked as possible candidates for phagocytosis (Fig. 8). In addition, the participation of Ddc in bacteria phagocytosis, but not the uptake of abiotic components or small molecules (Fig. 5), denotes that distinct mechanisms regulate the phagocytic activity of biotic and abiotic components in insect haemocytes.

Next, we explored the involvement of Ddc in melanization and the possible relationship between phagocytosis and melanization. Melanization is the process leading to melanin formation. Melanization occurs in both haemocyte-free haemolymph as well as in haemocytes after wounding or upon invasion with pathogens. The question of interest is whether melanization and phagocytosis, two unrelated procedures, are linked and facilitate each other. The activity of Ddc has been shown to be elevated during melanotic responses in Drosophila24 and in the mosquito Armigeres subalbatus.12,25 Melanization is also a critical process in defense against bacteria, and several reports link components of the melanization process with phagocytosis.4,26–28 Moreover, studies in other invertebrates have also linked components of the melanization cascade with phagocytosis.28 It is proposed that micro-organisms might be killed by toxic reactive oxygen metabolites produced in the process of melanization.29

In the current study, however, we observed that Ddc-based melanization, an innate immune mechanism in insects, appears to be distinct from the pathway leading to phagocytosis (Figs 8 and 9). However, these two unrelated procedures share a number of substrates (tyrosine, dopa, dopamine) and enzymes (PO, Ddc). Consequently, our data contradict earlier results supporting that PO-based melanization facilitates phagocytosis.

Nodulation, as stated in the introduction, refers to multicellular haemocyte aggregates that entrap a large number of bacteria, and PO and Ddc are key enzymes in this process (Fig. 9).4 Nodules may be attached to tissue or surrounded by haemocytes. Nodule formation has not been fully characterized, although it is known that it is lectin-mediated. Therefore, the query ‘Are nodulation and melanization linked and facilitate each other?’ was raised. In an in vitro nodulation assay we demonstrated that in the presence of anti-proPO or PTU, haemocytes were capable of aggregating, but could not be melanized. By contrast, haemocytes in the presence of carbidopa or benzerazide or anti-Ddc were incapable of forming aggregates and became melanized (Fig. 9). Consequently, nodulation and melanization are two distinct pathways. However, these two unrelated mechanisms share a number of substrates and enzymes. In conclusion, phagocytosis and nodulation processes are distinct from the melanization process. However, we have not yet explored whether any branch-point substrate exists that differentiates phagocytosis from nodulation or whether they are processes in sequence. From the above data it can be proposed that phagocytosis, nodulation and melanization are three distinct innate immune mechanisms, sharing several key components of tyrosine metabolism (tyrosine, dopa, dopamine, PO, Ddc). Further in-depth investigation is necessary to clarify the key metabolite(s) of the tyrosine pathway responsible for each process and their mode of action.


We thank the European Social Fund (ESF), the Operational Program for Education and Vocational Training II (EPEAEK II), and particularly the Program PYTHAGORAS, for funding the above work.