A novel highly divergent protein family identified from a viviparous insect by RNA-seq analysis: a potential target for tsetse fly-specific abortifacients

Abstract

In tsetse flies, nutrients for intrauterine larval development are synthesized by the modified accessory gland (milk gland) and provided in mother's milk during lactation. Interference with at least two milk proteins has been shown to extend larval development and reduce fecundity. The goal of this study was to perform a comprehensive characterization of tsetse milk proteins using lactation-specific transcriptome/milk proteome analyses and to define functional role(s) for the milk proteins during lactation. Differential analysis of RNA-seq data from lactating and dry (non-lactating) females revealed enrichment of transcripts coding for protein synthesis machinery, lipid metabolism and secretory proteins during lactation. Among the genes induced during lactation were those encoding the previously identified milk proteins (milk gland proteins 1-3, transferrin and acid sphingomyelinase 1) and seven new genes (mgp4-10). The genes encoding mgp2-10 are organized on a 40 kb syntenic block in the tsetse genome, have similar exon-intron arrangements, and share regions of amino acid sequence similarity. Expression of mgp2-10 is female-specific and high during milk secretion. While knockdown of a single mgp failed to reduce fecundity, simultaneous knockdown of multiple variants reduced milk protein levels and lowered fecundity. The genomic localization, gene structure similarities, and functional redundancy of MGP2-10 suggest that they constitute a novel highly divergent protein family. Our data indicates that MGP2-10 function both as the primary amino acid resource for the developing larva and in the maintenance of milk homeostasis, similar to the function of the mammalian casein family of milk proteins. This study underscores the dynamic nature of the lactation cycle and identifies a novel family of lactation-specific proteins, unique to Glossina sp., that are essential to larval development. The specificity of MGP2-10 to tsetse and their critical role during lactation suggests that these proteins may be an excellent target for tsetse-specific population control approaches.

Figure 1. Fold changes in transcript expression for contigs based on RNA-seq analysis.

Green indicates expression higher in lactating flies and red indicates higher expression in dry flies. (A) Relative expression of each contig with at least 50 mapped reads. (B) Contigs with significantly different expression values from A (Kal's test with Bonferroni correction, P<0.05).

Figure 2. Gene ontology enrichment analysis.

Reads in dry and lactating flies that mapped to genes with specific metabolic function.

Figure 3. Summary of specific genes that are differentially expressed in lactating compared to dry flies.

(A) Sphingomyelinase genes, asmase1–4 and nsmase. (B) Milk gland proteins genes, mgp1–10. (C) Iron-associated genes, non-hemecontaining ferritin, ferritin light, ferritin heavy and transferrin. (D) Ribosomal RNAs, 18S rRNA and 28S rRNA. *, significantly different expression values from B, Kal's test with Bonferroni correction, P<0.05.

Figure 4. Mapping of reads to lactation-specific genes and fold changes in transcript expression for contigs after removal of lactation-specific genes based on RNA-seq analysis.

(A) Left, number of reads mapping to individual genes coding for the 12 milk-specific proteins. Numbers above columns are the percent of total sample that mapped to the specific gene. Right, Sum of reads from all 12 milk-specific proteins. Number above columns are percent of total samples. (B) Relative expression of each contig with at least 50 mapped reads after removal of Illuminia reads for milk-specific contigs. (C) Contigs with significantly different expression values from B, Kal's test with Bonferroni correction, P<0.05.

Figure 5. GO enrichment and genes identified following RNA-seq analysis after milk-specific gene removal.

(A) Reads in dry and lactating flies that mapped to genes with specific metabolic function. *, indicating a significantly higher level in lactating flies based on chi-square test. (B) Select genes identified as increased during lactation following RNA-seq analysis after milk-specific gene removal. *, indicates significantly different between lactating and dry flies based on Kal's test followed by Bonferroni correction.

Figure 6. Validation of specific highly abundant proteins within the milk proteome.

FB/MG, fat body/milk gland analyzed from lactating flies, 24 hours after birth and 48 h after birth along with 3rd instar larval gut. Transcript levels were determined by qPCR using a CFX PCR detection system (Bio-Rad, Hercules) and data were analyzed with CFX manager software version 3.1 (Bio-Rad). Data represents the mean ± SE of three replicates and was normalized to tubulin. a (lower) and b (higher), denotes significant difference by ANOVA with Tukey's test at P<0.01 in comparison to the other samples.

Figure 7. Selective pressure acting upon mgp gene family.

(A) Site-specific dN/dS analysis along a multiple alignment of MGP coding sequences. Residues identified as subject to negative selection under FEL analysis (posterior probability cutoff = 95) are indicated in red. Regions are described as conserved or variable according to visual examination of the multiple alignment and corroborated by dN/dS anslysis. (B) Specific codons and their corresponding amino acid sequence under negative selection. (C) Percent amino acid and nucleotide homology, average dN/dS ratio and selection tests for specific regions of MGP2–10. Positive, neutral and purifying test were conducted with codon-based Z-test in MEGA5 .

Figure 8. Genome localization of Glossina morsitans and mgp2–10 phylogeny of genes for novel Glossina morsitans morsitans milk gland proteins.

(A) Initial sequence alignment was completed using ClustalX , and formatted with BioEdit . Evolutionary analyses were conducted in MEGA4/5 , , and displayed as a neighbor-joining tree. (B) Percent amino acid similarity (Bottom) and amino acid differences (Top) between MGP2–10.

Figure 9. Temporal and spatial expression of milk gland protein genes.

(A) Tissue specific RT-PCR. Data represents three replicates. ( B) Time course of mgp1–10 and asmase1 expression during the first two tsetse gonotrophic cycles. Transcript levels were determined by qPCR using a CFX PCR detection system (Bio-Rad, Hercules) and data were analyzed with CFX manager software version 3.1 (Bio-Rad). Data represents the mean ± SE of three replicates and was normalized to tubulin.

Figure 10. Phenotypes in lactating females following injection of MGP-specific siRNA.

(A) Transcript levels determined by qPCR after siRNA injection, mean ± SE of three groups of 3 combined flies normalized to tubulin. (B) Duration of the 1st gonotrophic cycle after siRNA injection, mean ± SE of three groups of 30 flies. (C) Duration of the 2nd gonotrophic cycle after siRNA injection, mean ± SE of three groups of 30 flies. (D) Number of pupae deposited by 20 females over 40 d (Only those from the first two gonotrophic cycles were counted), mean ± SE of four groups of 20 flies. (E) Example of lipid separation from an unstable tsetse milk emulsion. (F) Rate of emulsion separation after MGP7–9 knockdown, mean ± SE of ten assays. *, denotes significant difference from siGFP-injected control following ANOVA with Tukey's test at P<0.01.

Figure 11. Summary of the results from our tsetse fly lactation study.

The cross section of the milk gland tubules was adapted from Yang et al. and modified according to Ma et al. to represent tubules in a lactating fly, characterized by secretory vacuoles full of milk and condensed nuclei, and in the milk gland of a dry fly, characterized by exhausted secretory vacuoles and expanded nuclei.