The Drosophila Midkine/Pleiotrophin Homologues Miple1 and Miple2 Affect Adult Lifespan but Are Dispensable for Alk Signaling during Embryonic Gut Formation

Midkine (MDK) and Pleiotrophin (PTN) are small heparin-binding cytokines with closely related structures. The Drosophila genome harbours two genes encoding members of the MDK/PTN family of proteins, known as miple1 and miple2. We have investigated the role of Miple proteins in vivo, in particular with regard to their proposed role as ligands for the Alk receptor tyrosine kinase (RTK). Here we show that Miple proteins are neither required to drive Alk signaling during Drosophila embryogenesis, nor are they essential for development in the fruit fly. Additionally we show that neither MDK nor PTN can activate hALK in vivo when ectopically co-expressed in the fly. In conclusion, our data suggest that Alk is not activated by MDK/PTN related growth factors Miple1 and Miple 2 in vivo.


Introduction
Since their identification the small secreted molecules Midkine (MDK) and Pleiotrophin (PTN) have been implicated in a multitude of developmental events including enhancement of cell growth and survival, cell migration, angiogenesis, neurite outgrowth and development (reviewed in [1], and [2]). MDK, also known as Retinoic acid-Inducible Heparin Protein (RIHP) [3], was first identified as a gene induced by retinoid acid in mouse embryo carcinoma cells [4,5] and in an independent study a chicken homologue was purified from basement membranes [6]. The related family member PTN also known as Heparin-Binding Growth-Associated Molecule (HB-GAM) [6][7][8], Heparin-Binding Neurotrophic Factor or Neurite-promoting Factor (HBNF) [9], Osteoblast Specific Protein-1 (OSF-1) [10] and Heparin Affin Regulatory Peptide (HARP) [11], was first purified from bovine uterus as a weak mitogen towards fibroblasts. Given that both MDK and PTN encode small secreted cytokines with heparinbinding properties, they have long been considered to be functional ligands for cell surface receptors.
A number of reports have specifically addressed the role of MDK and PTN signaling through the Anaplastic Lymphoma Kinase (ALK) receptor tyrosine kinase (RTK) [12,13]. In these studies it has been reported that activation of ALK by either MDK or PTN leads to activation of both ERK and PI3K pathways [14][15][16][17]. However, a number of independent studies present contradictory results showing that MDK and PTN do not activate ALK [18][19][20][21].
To date, homologues have been reported in most species within the chordate and arthropod phyla, including human, mouse, fish, chicken, frogs and insects [3,6,7,[22][23][24][25]. In Drosophila melanogaster the miple1 (also referred to as miple) and miple2 genes display significant homology to the vertebrate MDK/PTN family [25]. In Drosophila, the expression of miple1 and miple2 is dynamic throughout embryogenesis, with both genes expressed in spatially and temporally restricted patterns. The transcript of miple1 is mainly expressed in the developing CNS, while miple2 transcripts are present in a number of cells within the developing endoderm [25]. Until now, no genetic characterisation of these two genes has been accomplished and their functional importance has therefore remained elusive.
In Drosophila melanogaster the Alk RTK together with its ligand Jelly Belly (Jeb) plays a well described critical role in the development of the embryonic visceral musculature [26][27][28][29]. Alk is expressed in progenitors of the developing visceral mesoderm (VM) where it is activated by the LDL domain containing ligand Jeb, leading to activation of the MAPK signaling pathway in a restricted subset of VM cells that subsequently differentiate into muscle founder cells (FCs). Alk activation by a Jeb-like ligand is functionally conserved in the nematode C. elegans where an Alk homologue is encoded by the scd-2 gene [30,31] and the Jeb homologue hen-1 [32]. Together they regulate dauer formation in the worm by modulating TGF-b signaling and its response to dauer pheromone [33].
In addition to their role in the specification of FCs during development of embryonic visceral mesoderm, signaling mediated by the Jeb/Alk ligand-receptor pair is employed in photoreceptor axonal targeting during the late maturation of the optic lobe neuropile [34,35]. Additionally, Jeb/Alk signaling regulates development and neuronal transmission of the larvae neuromuscular junction (NMJ) [36,37] and is required to protect larval neuroblasts during starvation in a process similar to brain sparing [38]. Alk is also involved in body size regulation, memory and learning [39,40] and ethanol sensitivity [41].
While the mechanisms by which MDK/PTN mediate signaling are fundamental for understanding their role in disease as well as development, their involvement in signaling and their receptor specificity in vivo are poorly understood. Drosophila has two well conserved homologues of the MDK/PTN family, Miple1 and Miple2 in addition to the characterised Alk ligand Jeb. Therefore the fruit fly presents a unique opportunity to study and characterise their function with focus on their potential interactions with Alk. In this work we have deleted the loci encoding the two Drosophila MDK/PTN family members: miple1 and miple2, each individually and together. We conclude that miple1 and miple2 are not essential for normal development, since both single and double mutant flies are viable and fertile. Furthermore, we have examined a potential relationship of the Miple proteins with the Alk signaling pathway. Neither Miple1 nor Miple2 are able to activate Alk or rescue phenotypes caused by the loss of the Alk ligand Jeb, nor are they able to ectopically activate Alk signaling in the eye. Taken together, our data suggests that neither Miple1 nor Miple2 activate the Alk RTK in vivo during Drosophila development.

Ethical statement
All rabbit immunisations were approved by the Umeå Ethical Board of Animal Research, Umeå, Sweden, according to the declaration of Helsinki (permit number A56-04).

Generation of UAS-miple transgenes and expression vectors
To generate UAS-miple1 transgenic flies, DNA encoding the open reading frame (ORF) of miple1 was generated by PCR and subcloned into pUAST. UAS-miple2-YFP.venus and UAS-miple2.6Xmyc transgenic flies were generated by cloning of miple2 from the RE73766 EST (DGRC) into pUAST-C-terminal Venus (pTWV, 1092, DGRC) and pUAST-C-terminal 6xMyc (pTWM, 1108, DGRC) respectively using Gateway Technology. UAS-MDK transgenes were generated by EcoRI and XhoI excision from the BC011704 (ATCC) cDNA clone, prior to ligation into pUAST vector. UAS-PTN was generated by PCR from the BC005916 (ATCC) cDNA clone using the primer pair 59-TATGAATTCTTGCAACAAAGGCAG-39 and 59-ATACTC-GAGTATAAGCCCCTACTGG-39, followed by cloning into pUAST using EcoRI and XhoI. Constructs were confirmed by DNA sequencing and used for generation of transgenic fly strains (BestGene Inc.). The pcDNA3-MDK, and pcDNA3-PTN constructs was generated as described above.

Generation of Miple2 antisera
For preparation of Miple2 antisera, DNA encoding amino acids 21-279 of Miple2 was subcloned into the pGEX-T2 GST fusion protein expression vector. The reading frame of the GST-Miple2 sequence was subsequently analysed by DNA sequencing. GST-Miple2 fusion protein was induced and purified from E. coli (BL21(DE3)) bacterial lysates by standard protocols using glutathione sepharose beads (GE Healthcare). GST-Miple2 recombinant protein was proteolytically cleaved with thrombin (4 units/ ml). After cleavage, thrombin was removed with p-aminobezamidine agarose and the sample dialysed overnight, with a cut-off of 15 kDa, followed by concentration. The resulting Miple2 recombinant protein was used for immunisation of rabbits (Agrisera AB, Vä nnä s, Sweden). Antisera was IgG purified (PIERCE Biotechnology) and subsequently affinity purified by standard protocols using glutathione sepharose beads. The previously generated Miple1 antisera [25] was IgG purified (PIERCE Biotechnology) and then affinity purified by standard protocols using Ni-NTA beads (Qiagen).

Survival analysis
Homozygous virgin females of indicated genotype were crossed to Df(3L)BSC125/TM3 actinGFP males in vials and transferred to new vials every 24 hours. After scoring genotype and sex of hatched adult progeny, the average percentage between homozygous (mutant or control over deficiency) and heterozygous (mutant over balancer), was calculated from three biological replicates.

Electron & Stereo microscopy
Adult flies were fixed in 2.5% formaldehyde, 2.5% glutaraldehyde in PBS before standard processing for Scanning Electron Microscopy (SEM). For light imaging a Nikon SMZ150 stereo microscope was employed. To generate a merged z-stacked image the EDF module in the NIS-BR software (Nikon) was used.

Pupal size measurements
Parental flies were kept in bottles for at least two generations and subsequently used for crosses of 10 males and 10 females in vials. Dark pupae were scored by absence of sex combs and sex was further confirmed after hatching. Sample pupae were aligned on sticky tape and measured using the Zeiss Axio Imager.Z2 microscope and Axio vision Release 4.8 software.

Life span and fecundity measurements
Parental flies were kept in bottles for at least two generations and then used for crosses (.90 for over-expression experiments and .113 for mutant experiments). Flies of age 1-5 days were used for lifespan experiments performed in an incubator with controlled conditions of light/dark cycles, humidity and temperature (12 h/12 h, 60% and 25 degrees). 25 flies/vial of the same gender were transferred to new food vial every 2-3 days and dead flies were counted. Fecundity experiments were modified from [54] and [55]. 1-2 days old females were mated to w 1118 males until age 4-5 days, after which males were removed and the number of embryos produced by females during each 24 hour period were analysed for 5 days.

In situ hybridisation
The entire coding regions of miple1 and miple2 were cloned into the dual PCR II promoter TOPO vector (Invitrogen) and used as template to generate DIG-labeled anti-sense and sense probes using SP6/T7 RNA polymerase (Roche) and NTP/DIG-UTP mix (Roche). In situ hybridisation of larval tissues was performed according to [56], while for adult brains a protocol including Proteinase K treatment was employed [57]. Samples were mounted on polylysine coated slides and visualised using DIC with Zeiss Axio Imager.Z2 microscope and Axio vision Release 4.8 software.
Results miple1 and miple2 encode small secreted heparin binding proteins that are dispensable for Drosophila development To investigate the functional roles of the Drosophila MDK/ PTN homologues miple1 and miple2 in vivo, we generated deletion mutants in the miple1 and miple2 loci, which are tandemly situated at 61B3 on the left arm of the third chromosome ( Figure 1A). To do this we performed an imprecise excision screen of a P-element (P{EPgy2} EY04126 ) situated between miple1 and miple2 ( Figure 1A). This approach led to the identification of single miple1 (miple1 D194 ) and miple2 (miple2 D396 ) null mutants (Figure1 B, C) that do not produce detectable RNA (data not shown). Examination of both miple1 D194 and miple2 D396 mutants revealed no obvious gross abnormalities, and flies were healthy, fertile and viable as homozygous stocks.
Given that Miple1 and Miple2 proteins display rather distinct expression patterns during embryogenesis, it is reasonable to expect that these small secreted molecules may be functionally redundant. This has been observed previously in Drosophila for example in the case of the FGF-like proteins Thisbe and Pyramus whose overlapping role is only seen when both genes are inactivated [58]. In order to delete miple1 and miple2 simultaneously we performed the excision screen in mutants lacking the DNA helicase DmBlm (Blm D2 /Blm D3 background). This approach led to the isolation of miple1,miple2 D104 mutant animals with a complete deletion of both miple1 and miple2 genomic regions ( Figure 1B). The miple1,miple2 D104 deletion also partially deletes the downstream neighboring gene tiny tim 2 (ttm2), a gene reported to be exclusively expressed in testis [59]. All generated deletion mutants of the miple genes are viable, although the miple1,miple2 D104 males are sterile due to loss of Ttm2 (see below).
We further analysed the survival of progeny from females lacking the maternal contribution of miple mRNA by crossing homozygous mutant females to deficiency males, obtaining the expected Mendelian rations in both male and female progeny ( Figure S1).

Loss of ttm2 but not miple1 and miple2 results in male sterility
To investigate a possible function of the ttm2 gene and its contribution to the sterility phenotype of miple1,miple2 D104 animals we analysed progeny from crosses with either one of the two deficiency strains Df(3L)BSC125 (that uncovers miple1, miple2, ttm2 and 5 additional genes) and Df(3L)BSC126 (which uncovers miple2 and ttm2 and 6 additional genes). Transheterozygous Df(3L)BSC125/Df(3L)BSC126 males deficient in both miple2 and ttm2 are male sterile (Table 1, Table S1). Male sterility is also observed in miple1,miple2 D104 homozygous flies  cannot be rescued by ectopic over-expression of Miple2 (miple2-Gal4.UAS-miple2) in the miple1,miple2 D104 mutant background (Table 1, Table S1), suggesting that the miple1,-miple2 D104 mutants lack functional Ttm2 protein. A putative ttm2 allele, a P-element insertion (PBac{3PHy+}C205), positioned downstream of miple2 and upstream of ttm2 ( Figure 1A), was also examined. Homozygous PBac{3PHy+}C205 males mated to virgin female PBac{3PHy+}C205 or to control w 1118 or control rev657 flies, were also sterile (Table S1). We further analysed this stock in allelic combinations with various miple alleles and deficiencies, concluding that this stock is male sterile and supporting a role for Ttm2 in male fertility in Drosophila in keeping with its previously described expression pattern. Analysis of the early fecundity of miple1,miple2 D104 females in comparison to control (control rev657 ) females suggests that loss of Miple proteins does not affect female fertility ( Figure S2).

Characterisation of miple1 and miple2 expression in larval tissues
The embryonic mRNA expression pattern of miple1 and miple2 mRNA has been described previously [27]. Briefly, miple1 RNA is mainly expressed in the CNS, while miple2 RNA is expressed in a broader pattern in both the developing endoderm and in the CNS in addition to the presence of a strong maternal contribution. We extended the analysis of miple mRNA expression to L3 larvae, where we detected strong expression of miple1 RNA in the neuropile (data not shown) of early L3 larvae, and in the optic lobes of late L3 larvae ( Figure 1D).
Weak miple1 mRNA expression can also be detected in imaginal discs ( Figure 1E and F) and in a distinct population of cells in the anterior gut ( Figure 1K) although no detectable expression could be seen in any other region of the larval gut ( Figure 1J). The miple1 mRNA pattern is further supported by the activity of a miple1-Gal4 transgenic driver that displays similar activity in the larval gut ( Figure 1N and O, see arrow). Robust expression of miple2 mRNA was detected in all L3 larval imaginal discs ( Figure 1H and I), while the CNS ( Figure 1G) does not show any detectable miple2 expression. The early L3 larval gut shows discrete miple2 RNA expression in a subset of cells ( Figure 1L and M, see arrows), particularly in potential adult midgut progenitor cells (AMPs), a pattern supported by miple2-Gal4 driven Red-Stinger (nRFP) which is expressed in only smaller but not in large nuclei ( Figure 1P and Q, see arrows). The identity of these cells as AMPs was further confirmed with the AMP marker esg-lacZ (data not shown). These results suggest that both genes are strongly expressed after embryonic stages and present a combined pattern of broad expression in L3 larval tissues.

Miple1 and Miple2 are secreted heparin binding proteins
The residues identified in human MDK and PTN required for these proteins to bind heparin are not fully conserved in the Drosophila Miple1 and Miple2 [25]. However both Miple proteins are enriched in basic amino acids such as arginine and lysine, which suggests that the Drosophila Miple proteins may also bind heparin and subsequently have the capacity to bind heparin sulphate proteoglycan (HSPG) proteins. We employed HEK293 cells to express secreted Miple1 and Miple2 and examined binding to heparin-agarose in vitro, using human MDK and PTN as positive controls ( Figure S3). These results show that both Drosophila Miple1 and Miple2 proteins exhibit heparin binding properties in vitro and may therefore share a conserved HSPG binding ability previously described for human MDK and PTN.
Neither Miple1 nor Miple2 are critical for the activation of Alk signaling in the developing visceral mesoderm The genetically tractable Drosophila model offers an excellent opportunity to explore the suggested role for the miple homologues PTN and MDK as ligands for the ALK receptor tyrosine kinase in vertebrates. We performed a detailed analysis of the embryonic visceral mesoderm in single and double miple mutant embryos, including activation of MAPK signaling (dpERK) mediated by Alk signaling in FCs [27][28][29], the founder cell specific marker Org-1 [52] and the fusion competent myoblast specific factor Vrp1 [53].
To remove the maternal and zygotic contributions of both miple1 and miple2, double mutant miple1,miple2 D104 , homozygous females were crossed with Df(3L)BSC125 males which harbor a deficiency removing the entire miple1 and miple2 loci. Examination of visceral mesoderm morphology in miple1,-miple2 D104 /Df(3L)BSC125 was performed in both, single miple1 and miple2 mutants as well as in double miple1,miple2 mutants. In early stage embryos we clearly detect the layer of columnar FCs and the more round FCMs in the developing VM indicating proper specification of the two cells types in both mutant and control embryos (Figure 2, compare B, C and D with control A, arrows indicate FCs). Similarly, in late stage embryos we detect no obvious abnormalities in gut formation, as the four midgut chambers are properly formed and organised in embryos of all  Figure 2, compare B9, C9 and D9 with control in A9). The specification of visceral FCs is dependent on Alk signaling and one protein that is expressed in a restricted pattern in FCs is the transcription factor Org-1 [28,52]. We analysed the Org-1 expression as well as the expression of Verprolin 1 (Vrp-1/WIP/ Sltr) [53,60,61], a co-regulator of the Arp2/3 complex that is expressed exclusively in FCMs, to analyse if cell type determination is affected in miple mutants. Org-1 expression is restricted to FCs in both double mutant and control embryos (Figure 2, compare F9 with control in E9, see arrow), while Vrp-1 is exclusively observed in FCMs (Figure 2, compare E99 with control in F99, see arrow). Activation of Alk leads to MAPK activation which is visualised by detection of phosphorylation of MAP-K(ERK) (pERK) in FCs. We observed robust MAPK activation in FCs of miple1,miple2 D104 mutant embryos (Figure 2, compare H9 with control in G9, see arrow). Thus, we conclude that Miple proteins are not required for Alk signaling during embryogenesis in vivo as measured by founder cell specification in the developing Drosophila visceral mesoderm.

Over-expression of Miple proteins is not sufficient to drive Alk activation in vivo in the fruitfly
Our results show that endogenous Alk signaling in the developing visceral mesoderm is not impaired by loss of Miple proteins in vivo. Next, we examined whether over-expression of Miple proteins with the Gal4/UAS system [62] is sufficient to drive ectopic Alk activation. This can be tested since all cells of the developing visceral mesoderm express Alk, while only those cells exposed to the Jeb ligand activate robust Alk signaling. Thus ectopic over-expression of the Alk ligand Jeb using the twist2xPE-Gal4 driver leads to ectopic activation of MAPK in all cells of the visceral mesoderm ( Figure 3B) [28] in contrast to the normal pERK pattern which is restricted to the FC population ( Figure 3A). We next asked whether activation of MAPK could be expanded to all VM cells by ectopic expression of either Miple1 or Miple2. We did not observe any ectoptic activation of MAPK upon over-expression of Miple proteins in the visceral mesoderm ( Figure 3C and D). Consistent with the results above, overexpression of Jeb using the twist2xPE-Gal4 leads to robust activation of the FC-specific rP298-lacZ enhancer trap [46] in all Alk positive cells of the VM ( Figure 3F). In contrast, overexpression of Miple1 and Miple2 does not affect rP298-lacZ expression ( Figure 3G and H), which is comparable with the restricted founder cell expression observed in controls ( Figure 3E).
To analyse potential Alk activation by Miple proteins in a different tissue context we investigated the effects of Alk signaling on cell growth. Alk and Jeb have been reported to regulate body size determination [39,40], where over-expression of the ligand Jeb (as well as over-expression of wild type Alk and constitutively active Alk) in pan neuronal tissues using C155-Gal4 dramatically decreases body size. We therefore analysed the pupal size of miple mutants in comparison to controls. The results clearly show that loss of Miple proteins does not affect body size determination, at the level of pupal length ( Figure S4). We next asked whether ectopic expression of Miple proteins effect body size determination by ectopically expressing Miple proteins in pan-neuronal tissues. In contrast to the effect seen on Jeb expression, we did not observe any effect on size, suggesting that Miple proteins are unable to activate Alk in this cellular context ( Figure 3I, quantified in J).
In a final attempt to examine the hypothesis that Miple proteins may drive Alk activation we tried to rescue the jeb mutant embryonic gut phenotype with Miple over-expression. To do this either Miple1 ( Figure 4D) or Miple2 (not shown) were expressed in a jeb mutant background using the en-Gal4 (engrailed-Gal4) driver. Clearly, even elevated levels of Miple1 ( Figure 4D) or Miple2 (not shown) are completely unable to activate Alk and rescue the formation of the visceral mesoderm in absence of Jeb protein ( Figure 4D). Control animals overexpressing Jeb protein allowed rescue of the visceral muscle structure at stage 14 ( Figure 4C), and resulting in a portion of animals (19,5%, n = 340) that developed into adult flies ( Figure S5).
Taken together, we find no in vivo evidence for the activation of Alk signaling in developing VM by Miple proteins. First, flies lacking both Miple1 and Miple2 are viable and fertile with no apparent defects in Alk signaling in the developing visceral mesoderm. Second, over-expression of Miple proteins is unable to drive Alk signaling in two independent processes. Third, excessive levels of Miple proteins are unable to activate Alk and rescue the jeb mutant gut phenotype in vivo. Thus, our data suggest that neither of these small heparin-binding proteins function as Alk activating ligands in the Drosophila tissues we have analysed in vivo.

Miple1 and Miple2 affect adult lifespan and Miple1 is expressed in the adult brain
While miple mutant flies are viable they have a decreased lifespan, as observed in both single miple2 D396 and miple1, miple2 D104 mutants when compared with control rev657 (Figure 5A). An intermediate effect on mean lifespan was seen in  (Table S2), suggesting that elevated levels of Miple proteins are well tolerated in the fly. However, we did observe a clear prolongation effect on lifespan in animals expressing Miple proteins with actin5C-Gal4 ( Figure 5B), and the pan-neuronal C155-Gal4 driver (data not shown), indicating a novel function for these two proteins in the regulation of lifespan in vivo. While both proteins may exert similar functional effects on lifespan, the differences in their expression patterns may underlie the difference between loss of Miple1 and Miple2 animals in terms of lifespan. The observed effect on lifespan suggests that Miple1 and Miple2 have a function in adults and thus may be expressed during adult stages. Distinct expression of miple1 mRNA was detected in adult brains ( Figure 5C). This staining is absent in miple1, miple2 D104 double mutants (Figure 5E). In contrast to the robust miple1 expression we were unable to miple2 mRNA expression in adult brain ( Figure 5D).

Ectopic expression of Miple1 or Miple2 in the eye does not affect ommatidia organisation
We next analysed the eye morphology of the different miple mutants and observed that the ommatidia structure in flies lacking either one of the miple genes or both is normal ( Figure S6, compare B, C and D with control in A). Additionally, we analysed the effect of over-expression of multiple independent UAS-miple1 and UAS-miple2 transgenic lines in the developing eye with GMR-Gal4, and did not observe effects on adult eye morphology with either Miple1 or Miple2 in comparison to the control ( Figure S6, compare F, G and H with control in E). In each case Miple protein expression was confirmed by either immunoflouresence or immunoblotting ( Figure S7). Our results here are in contrast to those of [63] that detected a clear effect on eye morphology upon over-expression of Miple1 in the eye. We do not observe such an affect despite clear expression of Miple proteins from our transgenes. It is possible that the differences in results observed here are a result of differences in expression levels, or due to other currently unclear reasons.
To test whether MDK/PTN/Miple protein expression together with Alk could modify either Drosophila or human Alk in the Drosophila eye, we ectopically expressed Miple proteins together with Alk, using the sevEP-Gal4 driver. Expression of Alk alone causes a characteristic eye phenotype with loss of defined ommatidia in the anterior region of the eye ( Figure 6A). Expression of the two Miple proteins alone, as described above with GMR-Gal4, does not disrupt eye morphology ( Figure S6F, G), and no enhancement or suppression of the sevEP-Gal4.UAS-Alk phenotype was observed upon co-expression with either Miple1 or Miple2 ( Figure 6D, E compare with control 6A). Furthermore, the sevEP-Gal4.UAS-Alk induced eye phenotype is not modified by reduction of miple1 and miple2 ( Figure 6F).
To further address if this family of ligands can activate the Alk receptor in vivo, we extended these experiments to include human MDK and PTN employing both GMR-Gal4 (Figure 7) and sevEP-Gal4 ( Figure S8) together with human ALK. Overexpression of wild type hALK alone does not affect eye morphology ( Figure 7A) [64], as human ALK does not recognise the Drosophila Alk ligand, Jeb [65]. This is in contrast to the expression of a constitutively active gain-of-function hALK F1174L mutant that strongly disrupts eye morphology ( Figure 7D) [64]. We could not detect any visible effect on eye morphology upon expression of MDK or PTN together with wild type hALK ( Figure 7B and C) suggesting that the receptor is not activated in this context. Further, when expressing MDK and PTN with the gain of function hALK F1174L allele ( Figure 7E and F) we did not observe a modified eye phenotype. Protein expression by MDK and PTN transgenes was confirmed by immunofluoresence ( Figure  S7).
Together these results suggest that neither the Drosophila Alk nor the human ALK receptor are directly activated in vivo by the human growth factors MDK/PTN or the Drosophila counterparts Miple1 and Miple2.

Discussion
In this work we have characterised the Drosophila homologues of the MDK/PTN family of small heparin binding proteins -Miple1 and Miple2, by generating deletion mutants of both miple1 and miple2 and a larger deletion encompassing both genes. Removal of both miple genes does not affect viability, thus these genes appear to perform non-essential functions and are dispensible for normal development of the fruitfly. Similarly to  [66], N-syndecan [67,68], Low-Density-Lipoprotein (LDL) Receptor-related Protein 1 (LRP1) [69], Anaplastic Lymphoma Kinase (ALK) [14][15][16] the a4 b 1 -and a 6 b 1 -integrins [70] and membrane localised Nucleolin [71] with no clear consensus Reports concerning ALK as a candidate receptor for MDK/ PTN are contradictory [12], with some supporting MDK and PTN as functional ligands [14][15][16][17], as well as articles reporting no interaction [18,19]. A single Drosophila homologue of ALK exists [44], permitting in vivo analysis of this receptor tyrosine kinase. Here, we have addressed the question of whether Drosophila MDK/PTN homologues Miple1 and Miple2 are functional ligands for Alk. Analysis of miple mutants suggest that these factors do not play an important function in Alk signaling dependant processes such as specification of founder cells in the visceral mesoderm during embryogenesis in the fruitfly. The effect of miple on body size regulation, a process regulated by Alk signaling [39], was also examined in this work. In contrast to manipulation of the levels of the ALK ligand Jeb, neither loss of miple gene expression nor ectopic expression of Miple proteins affected the regulation of body size in flies. It is possible that a role for the Miple proteins exists in other Alk related processes such as photoreceptor axonal targeting [35], neuronal transmission [37], brain sparing [38], ethanol sensitivity [41], memory and learning [39,40] but these have not been examined in this study.
One phenotype observed in our mutants was that of male sterility, however further investigation revealed this to be due to the loss of the ttm2 locus, and was confirmed by genetic analysis. Interestingly, our findings that miple double mutant females are fertile indicates that loss of Miple proteins do not effect female fertility. This is in contrast to the infertility phenotype previously reported in female PTN/MK double knockout mice [72].
While no gross phenotypes are apparent in miple double mutant animals, we observed a reduced mean lifespan in miple2 single mutants as well in miple double mutants, suggesting a role for Miples in regulation of lifespan. This function is further strengthened by the increase in mean lifespan resulting from ectopic over-expression of Miple1 or Miple2. Robust miple1 mRNA expression was detected in brains of adult flies, while miple2 was not detected in brains. This is in agreement with the expression observed in larval tissues. The strong miple1 mRNA expression suggests a role for Miple1 in re-or degeneration of neurons, something that has been reported for both PTN [73] [74] [75] and MDK [76] [77].
The increase in lifespan may function on a number of levels, with one other hypothesis based on potentially altered immune function. This is suggested by recent studies reporting bactericidal  [78] and fungicidal [79] activities of MDK and PTN in vitro, that are conserved for Miple2 in vitro [78]. Thus increased levels of Miple proteins would not necessarily be harmful for the fly, but may exert a positive effect. Whether Miple1 and Miple2 perform such functions in vivo in the fly are currently unknown and the mechanisms underlying their effect on lifespan require further more detailed investigation. When considering other receptors proposed for MDK/PTN, and whether their fly homologues potentially function as receptors for Miple1 and Miple2, it is interesting to note that hypomorphic syndecan (sdc) alleles have been reported to have a reduced lifespan and display a number of metabolic phenotypes [55]. Further, alleles of the bPS integrin subunit, myospheroid, show the opposite effect with increased life span in heterozygous condition [80]. Like the miple mutants, the closest fly homologue of the rptpb/f receptor, encoded by ptp99A, is homozygous viable [81]. Viable mutants have also been reported for lrp-1, although no further characterisation has been described [82]. In this study, we have not experimentally addressed whether any of these candidates are functional ligands for Miple1 and Miple2, indeed future studies should address these possibilities.
Our findings in this study do not support previous reports that MDK and PTN directly activate ALK, but are in agreement with other studies that show that MDK and PTN do not activate ALK. Precisely what functional roles Miple1 and Miple2 play in the fly remains unclear, but we can conclude that they are not of critical developmental importance, despite their extensive expression during development. Future work will be required to address their in vivo targets and function. Figure S1 Deletion of miple1 and miple2 does not affect survival of maternal zygotic mutant flies. Graph showing the average percentage of progeny, with homozygous mutant (over Df(3L)BSC125, black bar) and heterozygous mutant (over TM3 actinGFP balancer, grey bar) genotype and for female (black bar) and male (grey bar) of the adult progeny, from crosses between homozygous mutant females and heterozygous Df(3L)BSC125/ TM3 actinGFP males. The ratio between homozygous/heterozygous and female/male adult progeny is comparable to miple 657rev control and is the expected mendelian ratio for both single miple mutants and double miple mutants. The average percentage of progeny of the different genotypes and sex, was calculated from 3 biological replicates for each genotype with total number flies (n) from all replicates indicated. (TIF) Figure S2 Loss of Miple does not affect early female fecundity as measured by embryo production. Graph indicating female fecundity as measured by number of eggs laid per female per day (24 hours) over 5 days. Homozygous miple1, miple2 D104 double mutant flies (two replicates, light and dark grey bars) were comparable to control rev657 (black bar). Overall fecundity was observed to decrease over 5 days in both mutant and control flies.