Hummingbirds in flight exhibit the highest metabolic rate of all vertebrates. The bioenergetic requirements associated with sustained hovering flight raise the possibility of unique amino acid substitutions that would enhance aerobic metabolism. Here, we have identified a non-conservative substitution within the mitochondria-encoded cytochrome c oxidase subunit I (COI) that is fixed within hummingbirds, yet exceedingly rare among other vertebrates. This unusual change is also rare among metazoans, but can be identified in several clades with diverse life histories. We performed atomistic molecular dynamics simulations using bovine and hummingbird COI models, thereby bypassing experimental limitations imposed by the inability to modify mtDNA in a site-specific manner. Intriguingly, our findings suggest that COI amino acid position 153 (bovine numbering system) provides control over the hydration and activity of a key proton channel in COX. We discuss potential phenotypic outcomes linked to this intriguing alteration encoded by the hummingbird mitochondrial genome.

Figure 1A, Figure 1B, and Table S1:

Mitochondria-encoded protein sequences were downloaded from RefSeq (release 92) and bird sequences were extracted. Sequences from Bos taurus were also included. Polypeptides sequences were arrayed and concatenated [Figure 1A and Table S1 and Figure 1B (Accession_by_Species_Aves)_RefSeq_JAN7_2019.xlsx] [Figure 1A and Table S1 and Figure 1B (Sequences_by_Species_Aves)_RefSeq_JAN7_2019.xlsx]. Duplicates or egregiously misaligned concatenate sequences were removed. Alignments were performed using T-coffee in regressive mode and taking advantage of the MAFFT G-INS-i strategy [Figure 1A and Table S1 and Figure 1B (Aves_mtProt_NoDups_Clean_TCOFF_G)_RefSeq_JAN7_2019.fasta]. FastTreeMP was subsequently used to generate the NWK tree file [Figure 1A and Table S1 and Figure 1B [(Aves_mtProt_NoDups_Clean_TCOFF_G)_RefSeq_JAN7_2019.nwk] for downstream analysis by PAGAN and our custom-made scripts (available on GitHub at https://github.com/corydunnlab/hummingbird) [Figure 1A and Table S1 and Figure 1B (Aves_mtProt_NoDups_Clean_TCOFF_G_PAGAN).anctree; Figure 1A and Table S1 and Figure 1B (Aves_mtProt_NoDups_Clean_TCOFF_G_PAGAN).fas].

Figure 1C and Table S2:

Mitochondria-encoded protein sequences were downloaded from RefSeq (release 92) and bird COI sequences were extracted. The COI sequence from Bos taurus was subsequently included. Sequences were aligned by MAFFT using the G-INS-i strategy [Figure 1C and Table S2 (Aves_COX1_extracted_from_RefSeq_JAN7_2019_w_bovine_MAFFT_G-INS-i).fasta].

Table S3: 

In BOLD Systems (February 12, 2020), the query "Apodiformes" was used. FASTA entries were downloaded and aligned by MAFFT using the L-INS-i strategy. Samples unaligned or uninformative (no character at selected positions) were removed [Table S3 (Apodiformes_bold_MAFFT_L-INS-i_correct_translation).fasta]. Samples without a specific species assignment were not processed further.

Table S4: 

In BOLD Systems (February 12, 2020), the query "Aves" was used. FASTA entries were aligned using MAFFT under the 'auto' setting [Table S4 (Aves_BOLD_MAFFT_auto).fasta].

Table S5:

Mitochondria-encoded protein sequences were downloaded from RefSeq (release 92) and Chordata sequences were extracted. FASTA entries were aligned using MAFFT under the 'auto' setting [Table S5 (Chordata_COX1_MAFFT_auto).fasta], with subsequent analysis removing unaligned or uninformative sequences.

Next, BOLD samples were sought for the relevant genera (February 28, 2019) using the terms "Ambystoma", "Serrivomer", "Pangio", and "Polypterus". Downloaded FASTA entries were aligned using MAFFT under the 'auto' setting: [Table S5 [Ambystoma_BOLDSYSTEMS_MAFFT_auto].fasta; Table S5 [Serrivomer_BOLDSYSTEMS_MAFFT_auto].fasta,; Table S5 [Pangio_BOLDSYSTEMS_MAFFT_auto].fasta; Table S5 [Polypterus_BOLDSYSTEMS_MAFFT_auto].fasta]. Subsequent analysis removed unaligned or uninformative samples.

Table S6: 

Mitochondria-encoded protein sequences were downloaded from RefSeq (release 92), and metazoan COI protein sequences were extracted and aligned by MAFFT under the 'auto' setting [Table S6 (Metazoan_COI_MAFFT_auto).fasta].

For BOLD analysis, FASTA sequences were sought for each of the relevant genera (February 29, 2020). Files were merged and limited to those entries obtained by the COI-5P barcode primers. Sequences were then aligned by MAFFT using the FFT-NS-2 option [Table S6 (BOLD_combined_COI-5P_MAFFT_FFT_NS_2).fasta]. Unlike other analyses performed with BOLD, assignment to specific species was not required to be counted for Table S6. Translation of these divergent entries was accomplished with the help of the gappyout function of trimal.

Table S7: 

BOLD samples were obtained for bee families using the queries "Andrenidae", "Apidae", "Colletidae", "Halictidae", "Megachilidae", and "Melittidae" (April 4, 2019). FASTA entries were aligned using MAFFT under the 'auto' setting: [Table S7 (Andrenidae_BOLD_COI-5P_MAFFT_auto).fasta; Table S7 (Apidae_BOLD_COI-5P_MAFFT_auto).fasta; Table S7 (Colletidae_BOLD_COI-5P_MAFFT_auto).fasta; Table S7 (Halictidae_BOLD_COI-5P_MAFFT_auto).fasta; Table S7 (Megachilidae_BOLD_COI-5P_MAFFT_auto).fasta; Table S7 (Melittidae_BOLD_COI-5P_MAFFT_auto).fasta].

Table S8: 

BOLD samples were obtained for hoverfly genera using the queries "Anasimyia", "Chasmomma", "Eristalinus", "Eristalis", "Helophilus", "Lejops", "Mallota", "Mesembrius", "Palpada", "Parhelophilus", "Phytomia", and "Senaspis" (April 4, 2019). FASTA entries were aligned using MAFFT under the 'auto' setting: [Table S8 (Anasimyia_BOLD_MAFFT_auto).fasta; Table S8 (Chasmomma_BOLD_MAFFT_auto).fasta; Table S8 (Eristalinus_BOLD_MAFFT_auto).fasta; Table S8 (Eristalis_BOLD_MAFFT_auto).fasta; Table S8 (Helophilus_BOLD_MAFFT_auto).fasta; Table S8 (Lejops_BOLD_MAFFT_auto).fasta; Table S8 (Mallota_BOLD_MAFFT_auto).fasta; Table S8 (Mesembrius_BOLD_MAFFT_auto).fasta; Table S8 (Palpada_BOLD_MAFFT_auto).fasta; Table S8 (Parhelophilus_BOLD_MAFFT_auto).fasta; Table S8 (Phytomia_BOLD_MAFFT_auto).fasta; Table S8 (Senaspis_BOLD_MAFFT_auto).fasta].