Genomic insights into metabolic flux in hummingbirds

Abstract

Hummingbirds are very well adapted to sustain efficient and rapid metabolic shifts. They oxidize ingested nectar to directly fuel flight when foraging but have to switch to oxidizing stored lipids derived from ingested sugars during the night or long-distance migratory flights. Understanding how this organism moderates energy turnover is hampered by a lack of information regarding how relevant enzymes differ in sequence, expression, and regulation. To explore these questions, we generated a chromosome-scale genome assembly of the ruby-throated hummingbird (A. colubris) using a combination of long- and short-read sequencing, scaffolding it using existing assemblies. We then used hybrid long- and short-read RNA sequencing of liver and muscle tissue in fasted and fed metabolic states for a comprehensive transcriptome assembly and annotation. Our genomic and transcriptomic data found positive selection of key metabolic genes in nectivorous avian species and deletion of critical genes (SLC2A4, GCK) involved in glucostasis in other vertebrates. We found expression of a fructose-specific version of SLC2A5 putatively in place of insulin-sensitive SLC2A5, with predicted protein models suggesting affinity for both fructose and glucose. Alternative isoforms may even act to sequester fructose to preclude limitations from transport in metabolism. Finally, we identified differentially expressed genes from fasted and fed hummingbirds, suggesting key pathways for the rapid metabolic switch hummingbirds undergo.

Figure 1.

Experimental design and genome/transcriptome assembly overview. (A) Overview of experimental design and methods for producing a genome assembly and transcriptome of A. colubris. (B) Step plot of scaffold length for the 33 A. colubris chromosomes. (C) Number of isoforms per gene in the C. anna NCBI Liftoff annotation versus the StringTie2 hybrid transcriptome assembly. (D) The SLC2A5 gene locus in the C. anna NCBI Liftoff annotation and the StringTie2 hybrid transcriptome assembly.

Figure 2.

Positive selection in nectivory. (A) Upset plot illustrating the number of shared orthogroups between the 19 species. Only bars with more than 50 orthogroups are shown. (B) ShinyGO Gene Ontology plot showing a hierarchical clustering tree to summarize the correlation among significant pathways. Pathways with many shared genes are clustered together. Dot size is representative of P-value significance.

Figure 3.

Hummingbird sugar transporters. (A) Normalized gene count plots for ruby-throated hummingbird muscle and liver tissue. (B) Isoform expression of SLC2A5 in the liver (left) and muscle (right). FPKM of each isoform color-coded according to the top scale bar. (C) Ribbon representations of the two protein models for SLC2A5 isoforms predicted by AlphaFold2 based on the mammalian SLC2A5 ortholog. Left is the X1 isoform; right is the X2 isoform that is missing exon 3. In both atomic models, N- and C-terminal TM bundles are colored blue and red, respectively. Regions of the X1 isoform that are missing in the X2 isoform (TM3/TM4) are depicted in light blue. Arginine–glutamate salt bridges at the intracellular tips of TMs are green. (ICHs) Intracellular helices. (D) Tracer oxidation rate over 20 min when birds were fed ¹³C on either the glucose or the fructose portion of the sucrose disaccharide. (E) Peak oxidation time of glucose and fructose in minutes (P = 0.02, paired t-test).

Figure 4.

Differential expression. (A) Volcano plot displaying differentially expressed genes (DEGs) from A. colubris liver tissue comparing fasted to fed. (B) Summary heatmap of top 12 enriched GO terms in the liver. (C) Volcano plot displaying DEGs from A. colubris muscle tissue comparing fasted to fed. (D) Summary heatmap of top 13 enriched GO terms in the muscle.