Sand fly infection and Leishmania differentiation
Le.infantum growth in the Lu. longipalpis sand fly midgut followed a typical and expected pattern [21]. Briefly, a median of 3,000 parasites detected early at 4d Pi increased to 6,000 parasites by day 6 and reached about 126,000 parasites at 14d. The proportion of metacyclic stage parasites increased from 0% on 6d to 92% on 14d.
Expanding the Lu. longipalpismidgut repertoire of putative proteins
We obtained a total of 53,683,499 high quality Lu. longipalpis midgut-specific reads from the de novo assembly of seven libraries prepared from RNA extracted from uninfected midguts at 1d, 2d, 4d, 6d, 8d, 12d and 14d. High quality reads were assembled in 57,016 contigs that were further filtered to 13,841 putative contigs based on the presence of an open reading frame (ORF) and on similarities to proteins deposited at Refseq invertebrate, NCBI Genbank or SwissProt databases (Additional file 1: Table S1). We also searched for putative secreted proteins where a signal peptide was predicted. The contigs or transcripts from these libraries had a mean size of 1,498bp, with the shortest comprising 150 bp and the longest 27,627 bp. Overall, 72% were categorized to a functional class after BLAST analysis (e< 10E-6) against nine distinct databases (Additional file 2: Fig. S1 and Additional file 3: Table S2). Unknown contigs accounted for 28% of contigs, but only for 6.56% of transcriptome abundance.
The annotations of all the 13,841 sand fly midgut transcripts from this work are described in Table S1 (Additional file 1: Table S1), where the best match of each sequence to NCBI, KOG, and Swiss protein databases are shown in different columns. Table S1 also describes if the protein is potentially secreted by giving the term “SIG” in the SigP Result column as a result of SignalP analysis of all transcripts. Using this information, we classified the transcripts into functional categories. The most represented functional categories were secreted proteins with 25.9% of transcripts per million (TPM), protein synthesis (15.2% of TPM), metabolism (14.3% of TPM) and protein modification (11.8% of TPM; Fig. 1 and Additional file 3: Table S2). Importantly, only 0.59% of all transcripts did not have a match in any of the databases tested, indicating that more than 99% of the transcripts from this transcriptome are insect specific transcripts (Additional file 1: Table S1).
The search for transcripts encoding protein families potentially participating in biological processes important for midgut physiology and Leishmania development resulted in 740 sequences encoding proteins associated with immune responses, digestion, and chitin metabolism (Additional file 4: Table S3).
Among the immune-related genes (194 transcripts), the major components of the Toll-like receptor and Imd pathways participating in recognition (GNBPs and PGRPs), signal transduction (Spatzle and TAK1), regulation (cactus and caspar), transcription factors (dorsal and relish), as well as effector molecules (antimicrobial peptides) were identified (Additional file 4: Table S3). Members of the Reactive Oxygen Species (ROX)-producing MAP kinase pathway, such as DUOX, were also identified Additional file 4: Table S3). Multiple transcripts encoding proteins associated with ROX metabolism (oxidative stress), cell death (JNK pathway and apoptosis) epithelium regeneration (JAK-STAT pathways) were also pinpointed (Additional file 4: Table S3).
Regarding the digestive enzymes (348 transcripts), 142 serine protease-encoding transcripts were identified, of which 33 and 22 transcripts encoded trypsins and chymotrypsins, respectively (Additional file 4: Table S3). Amongst the 55 carbohydrases, 24 amylases and 6 glucosydases were identified (Additional file 4: Table S3). Furthermore, 49 carboxypeptidases, 41 aminopeptidases, 56 lipases, and 5 nucleotidases were identified (Additional file 4: Table S3).
The 198 transcripts related to chitin metabolism are possibly involved with the synthesis, scaffolding, modification, and degradation of the PM (Additional file 4: Table S3). Regarding chitin synthesis, two transcripts encoding chitin synthase were identified (Additional file 4: Table S3). Transcripts encoding PM-scaffolding proteins encompassed 28 peritrophins and 7 of 25 chitin-binding protein of the CPAP subgroup (Cuticular proteins analogous to peritrophins), which displayed relevant expression levels in the midgut (Additional file 4: Table S3). Regarding PM modification, 4 transcripts encoding chitin deacetylases were pinpointed (Additional file 4: Table S3). For PM degradation and chitin digestion, 15 and 6 transcripts encoding chitinases and N-acetyl-glucosaminidases, respectively, were identified (Additional file 4: Table S3).
Compared to the annotation of the P. papatasi cDNA libraries of whole bodies and multiple stages and time points post-blood feeding [20], multiple additional homolog transcripts were identified in the Lu. longipalpis midgut RNA-Seq libraries, including transcripts encoding PGRPs, C-type lectins, and lysozymes amongst the immune genes. Also, the RNA-Seq libraries displayed more matches of sequences encoding all classes of digestive enzymes (except aminopeptidases), multiple peritrophins, and twice as many chitinase sequences (Table 1).
Sand fly midgut gene expression
The obtained midgut transcriptome dataset was used to determine the sand fly midgut differential expression caused by Leishmania infection. All transcripts used for this analysis can be found in Table S1 (Additional file 1: Table S1.).
We performed a Principal Component Analysis (PCA) to summarize the overall expression profiles of the infected and uninfected midgut transcripts from the seven time points that represent infected midguts enriched with a different Leishmania stage (Fig 2A and Additional file 5: Table S4) as well as amongst replicates (Additional file 6: Fig. S2A-B and Additional file 5: Table S4). The PC1 axis shows a clear separation of transcripts between the midguts in which blood digestion is ongoing (Fig. 2A left side, 1d PBM/Pi and 2d PBM/Pi) and less separation from the time points at which the blood was mostly digested (Fig. 2A right side, 4d PBM/Pi) and from the remaining time points where the midguts were clear of blood (Fig. 2A right side, 6d to 14 PBM/Pi). The PC1 accounted for 77.2% of the variance (Additional file 5: Table S4). On the other hand, the PC3, rather than PC2 (Additional file 6: Fig. S2A-B), depicted the separation between infected from the uninfected samples (Fig. 2A), accounting for 4.1% of the variance (Additional file 5: Table S4).
The expression profiles of midgut transcripts were validated by assessing the expression levels of selected midgut genes (n = 28-35; Additional file 7: Table S5) using the nCounter technology (NanoString). The mean log2 fold change (LFC) of infected over uninfected samples was compared at each time point with LFC data obtained by RNA-Seq for the same genes. Representative genes participating in chitin metabolism/peritrophic matrix scaffolding (peritrophins and chitinases), immunity (defensin, catalase, and spatzle), digestion (amylase and chymotrypsin) among others are depicted in Fig. 2B-H. The regression analyses between the expression levels obtained with nCounter and RNA-Seq were statistically significant (p < 0.0001) for all seven time points (Fig. 2B-H), and the regression coefficients were greater than 0.5 for all time points, except 6d (R2 = 0.40) and 12d (R2 = 0.47) as shown in Fig. 2B-H.
Minimal modulation of sand fly midgut gene expression by Leishmania infection
Differences in midgut gene expression between Leishmania-infected over uninfected midguts were assessed. Overall, such differences accounted for only 113 differentially expressed transcripts (1 < LFC > 1; q-value < 0.05). The number of DE transcripts gradually increased from 2 transcripts on 1d to 53 transcripts on 4d (Fig. 3A). Thereafter, the number of DE transcripts decreased to 20 transcripts on 6d, and 15 transcripts on 8d (Fig. 3A). Four days later, there was a strong increase in the number of DE transcripts (12d = 32 transcripts), which was reduced to 13 transcripts two days later at 14d (Fig. 3A).
Amongst the midgut genes differentially expressed upon Leishmania infection, some appear to play a role in specific biological processes (Table 2 and Additional File 8: Table S6). A gene encoding the transcription factor Forkhead/HNF-3 (lulogut44569) was down-regulated on 2d. Genes encoding proteins potentially involved with metabolism of steroid hormones, such as 17-beta-hydroxysteroid dehydrogenase 13-like (lulogut32574) and juvenile hormone esterase (lulogut40195) were down-regulated on 2d; a putative juvenile hormone binding protein (lulogutSigP-24104) was down-regulated on 4d; and an ecdysteroid kinase (lulogut41307) was down-regulated on 12d. Also, genes encoding a peritrophic matrix protein (lulogutSigP-40401), involved with the peritrophic matrix scaffolding, the antimicrobial peptide attacin (lulogutSigP-8812), and amino acid (lulogut16004) and trehalose (lulogutSigP-40100) transporters, were down-regulated on 4d. Amongst the up-regulated genes, multiple peptidases and proteases were up-regulated on 4d and 6d. Likewise, multiple insect allergen proteins (microvilli proteins) of unknown function were up-regulated on 4d and 6d upon Leishmania infection. From 8d onwards, multiple cytochrome p450 transcripts were upregulated.
The presence of Leishmania in the midgut led to more genes being down-regulated at d2 and up-regulated at later time points, except on 12d (Fig. 3A and Additional File 8: Table S6). On 1d, 2d, and 4d, the early time points, 1, 11, and 30 genes were down-regulated (Fig. 3A and Table 3 and Additional File 8: Table S6) whereas 1, 3, and 23 genes were up-regulated (Fig. 3A and Table 3 and Additional File 8: Table S6). On 6d and 8d, on the other hand, 20 and 15 genes were up-regulated, yet none were down-regulated (Fig. 3A). Infected midguts on day 12 displayed 13 up-regulated compared to 18 down-regulated genes (Fig. 3A). The 14d time point exhibited more up-regulated (12 genes) than down-regulated (1 gene) genes in infected over uninfected midguts (Fig. 3A).
Venn diagrams show the number of DE genes at unique time points, as compared to the number of DE genes shared by multiple time points (Fig. 3B-C and Additional File 9: Table S7). In the comparisons between early time points (1d through 6d; Fig. 3B), 1 out of the 2 DE genes on 1d was only modulated at that time point (Fig. 3B). Similarly, 13 out of the 14 genes, and 43 out of 54 genes, were uniquely DE on 2d and 4d, respectively (Fig. 3B). Only the 6d DE genes exhibited as many unique as shared with 4d DE genes (10 genes; Fig. 3B). The comparisons of DE genes between later time points (6d through 14d) showed a greater number of shared DE genes between time points (Fig. 3C). For instance, only 5 out of 15, and 5 out of 13, DE genes were unique to 8d and 14d, respectively (Fig. 3C). The 12d midguts, on the other hand, exhibited 26 uniquely expressed genes out 32, the most amongst the late time points (Fig. 3C).
Patterns of differentially expressed genes across time points
Most of the midgut genes DE by Leishmania infection were up-regulated by up to 32-fold (LFC < 5; Table 3; Additional File 10: Fig S3; Additional File 11: Table S8). These DE genes encoded multiple digestive enzymes and allergen-related peptides at early time points and detoxification-related proteins at later time points (Table 3; Additional File 10: Fig S3). On the other hand, multiple genes were down-regulated in Leishmania-infected midguts by more than 32-fold (LFC > -5; Table 4; Additional File 10: Fig S3; Additional File 11: Table S8). Regarding the midgut genes downregulated by Leishmania infection (Table 4 and Additional File 10: Fig S3), none were DE on 6d and 8d (Table 4). Such genes encode a variety of proteins of unknown function as well as proteins involved in the lipid metabolism (Table 4).
Functional profiles of the differentially expressed genes at different time points
Although the midgut genes up- and down-regulated by Leishmania infection exhibited different expression patterns across time points (Additional File 10: Fig S3), such DE genes belonged to the same functional groups for the most part (Fig. 4 and Tables 3 and 4 and Additional file 11: Table S8). Regarding the up-regulated genes, 28%, 38%, and 18% belonged to the detoxification (detox), metabolism (met), and secreted (s) protein molecular functions, respectively (Fig. 4A and Table 3). In fact, the enrichment of such molecular functions amongst the up-regulated genes was consistent through time (Fig. 4A and Table 3): between 2d through 14d for the metabolism function; and between 8d and 14d for the detoxification function. For the secreted protein category, the enrichment of up-regulated genes was more restricted to 4d and 6d (Fig. 4A and Table 3). At earlier time points (1d and 2d), the few up-regulated genes perform different functions ranging from transporter channels (tr, 1d) to proteosome machinery (prot, 2d; Fig. 4A and Table 3). Regarding midgut genes downregulated by the Leishmania infection, 34% of these genes belonged to the metabolism (22%) and secreted protein (12%) functional groups (Fig. 4B and Table 4). Both categories were consistently enriched on 4d, 12d, and 14d (Fig. 4B and Table 4). At earlier time points (1d and 2d), transporter channels (tr, 1d and 2d) and signaling transduction (st, 2d) were the most enriched molecular functions amongst the down-regulated genes (Fig. 4B and Table 4). All the molecular functions identified over all time points were matched by analogous GO terms (Additional file 12: Table S9 and Additional file 13: Table S10).
In order to investigate in-depth the functional profiles of the DE genes, we broke down the most predominant functional classes into subclasses. For the midgut DE genes belonging to the detoxification molecular function (detox), the cytochrome P450 gene family encompassed 76% of the up-regulated genes (Fig. 4C and Table 3 and Additional file 11: Table S8). Such genes were consistently upregulated between 6d and 14d (Fig. 4C and Table 3). In contrast, the down-regulated genes belonging to the detoxification molecular function were enriched in metallothioneins (4d and 12d, thio; Fig. 4D and Table 4). As far as the DE midgut genes belonging to the metabolism function, 55% of the up-regulated genes were related to the metabolism of lipids (lipd; Fig. 4E and Table 3) which was consistently the most predominant between 6d and 14d (Fig. 4E). Among the down-regulated genes performing metabolic functions (Fig. 4F and Table 4), 31% participated in the metabolism of lipids (lipd) at early time points (2d and 4d) or nucleotides (nuc) at later time point (12d and 14d, Fig. 4F and Table 4). Regarding the DE midgut genes encompassing the secreted proteins (Fig. 4G-H), 50% of those up-regulated belonged to the ‘other category’ (s, multiple protein functions) that was enriched in transcripts of insect allergen proteins (Fig. 4G; Table 3 and Additional file 11: Table S8), also known as microvilli proteins. Although the insect allergens, along with the mucins, and to a lesser extent metalloproteases (metal), were more predominant on 4d and 6d (Fig. 4G and Table 2), up-regulated transcripts encoding proteins of unknown function were enriched at 14d, a later time point (Fig. 4G and Table 3). Among the down-regulated transcripts encoding secreted proteins, 44% belonged to the unknown function (31%, uk) and “other” (17%, s) categories (Fig. 4H and Table 4). The “other” category (s) was consistently downregulated on 4d and 12d (Fig. 4H and Table 4) and was enriched in transcripts encoding juvenile hormone (JH) binding proteins as well as attacin (Table 4). Transcripts of secreted proteins related to the digestion of lipids (met-li) were down-regulated on 2d (Fig. 4H and Table 4).