Expression of a novel surfactant protein gene is associated with sites of extrapulmonary respiration in a lungless salamander

Numerous physiological and morphological adaptations were achieved during the transition to lungless respiration that accompanied evolutionary lung loss in plethodontid salamanders, including those that enable efficient gas exchange across extrapulmonary tissue. However, the molecular basis of these adaptations is unknown. Here, we show that lungless salamanders express in the larval integument and the adult buccopharynx—principal sites of respiratory gas exchange in these species—a novel paralogue of the gene surfactant-associated protein C (SFTPC), which is a critical component of pulmonary surfactant expressed exclusively in the lung in other vertebrates. The paralogous gene appears to be found only in salamanders, but, similar to SFTPC, in lunged salamanders it is expressed only in the lung. This heterotopic gene expression, combined with predictions from structural modelling and respiratory tissue ultrastructure, suggests that lungless salamanders may produce pulmonary surfactant-like secretions outside the lungs and that the novel paralogue of SFTPC might facilitate extrapulmonary respiration in the absence of lungs. Heterotopic expression of the SFTPC paralogue may have contributed to the remarkable evolutionary radiation of lungless salamanders, which account for more than two thirds of urodele species alive today.


Evidence for Duplication of SFTPC
SFTPC and SFTPC-like sequences diverge from one another within exonic regions, but not according to putative splice boundaries (Fig. S1a), which indicates that SFTPC-like is not an isoform of SFTPC. While SFTPC-like is divergent from SFTPC sequences, it is not an ortholog of a closely related BRICHOS domain-containing gene (Fig. S1b). We found SFTPC-like expressed in eleven species of lunged and lungless salamanders (Fig. S1a, b; Supplemental Data File 1); most also express SFTPC.
In Bayesian and maximum likelihood gene trees for SFTPC sequences from amphibians, amniotes, lungfish and coelacanth, tetrapod SFTPC and SFTPC-like fall into well-supported monophyletic lineages. However, support values for the split between SFTPC and SFTPC-like are low, leaving a polytomy including lungfish SFTPC, coelacanth SFTPC, salamander SFTPC-like and tetrapod SFTPC (Figs. S1, S2). Bayesian (Fig. S1b) and maximum likelihood (Fig. S2) gene trees both place coelacanth SFTPC sister to salamander SFTPC-like, but support values for a sister relationship to salamander SFTPC-like are low under the maximum likelihood approach. The presence of SFTPC in all amphibian and amniote groups and the absence of SFTPC-like in all surveyed tetrapods except salamanders argues that the most parsimonious explanation for the corresponding gene tree is that the tetrapod ortholog of SFTPC was duplicated in the salamander lineage following its divergence from frogs, followed by substantial sequence divergence between SFTPC and SFTPC-like (Fig. S1b). Low statistical support at the salamander SFTPC-like + coelacanth SFTPC node under maximum likelihood approaches should be interpreted as a polytomy, and long-branch attraction may cause an artifactual affiliation between coelacanth and lungfish SFTPC orthologs and salamander SFTPC-like. Therefore, origin of SFTPC-like from a salamander-specific gene duplication event is the most parsimonious explanation for the observed gene trees. However, alternative scenarios such as independent losses of SFTPC-like across lungfish, frogs and amniotes cannot be ruled out based purely on gene tree topology from Bayesian and maximum likelihood approaches alone.
Sequence alignments alone may not provide sufficient information to reconstruct gene evolution and patterns of gene gain and loss [16]. We applied PHYLDOG [16] to explicitly test for gene duplications of SFTPC. Given a guide tree with known phylogenetic relationships (Fig. S3), PHYLDOG predicts that SFTPC-like arose by gene duplication after the divergence of frogs and salamanders (Fig.  S4). SFTPC-like has been meiotically mapped to linkage group 6 in Ambystoma mexicanum, a lunged salamander, and is located within a region syntenic to human chromosome 15 [17]. Finally, SFTPC-like and SFTPC have been assembled to two separate genomic scaffolds from A. mexicanum [18], supporting SFTPC-like's origin via gene duplication.
Previous work identified a SFTPC gene duplication event in the lungfish Protopterus annectens [19]. Our analysis confirms that this duplication event is lineage specific and confined to P. annectens (Fig. S1b). Tables:   Table S1. Primers used to clone SFTPC and SFTPC-like from Ambystoma mexicanum, Desmognathus fuscus and Plethodon cinereus.

Protopterus annectens
Chris Amemiya [28]; [19] Supporting Information Figures: Figure S1. A novel form of Surfactant-associated protein C (SFTPC) is expressed in several species of salamanders. a, Amino acid alignment of SFTPC (yellow) and SFTPC-like (cyan) sequences reveals conservation of hydrophobic residues within the mature peptide domain. Full species names and accession numbers are listed in Supplemental Data File 1. Lungless (plethodontid) species are in red font. b, Bayesian 95% maximum clade credibility tree for SFTPC reveals SFTPC-like transcripts in 11 species of salamanders. SFTPC-like is not a related ortholog because it is nested within the SFTPC phylogeny. Node values are posterior probabilities; scale bar denotes expected changes per site. Salamander SFTPC is marked with a yellow box; salamander SFTPC-like is marked in cyan. c, Predicted secondary structure of SFTPC-like from Desmognathus fuscus. SFTPC-like structure predictions (cyan) utilizing SWISS-MODEL [29], QUARK Ab initio predictions [30] and I-TASSER [31] are aligned with the resolved SFTPC mature peptide (yellow) [32].   Blue nodes indicate divergence due to speciation events. PHYLDOG predicts that SFTPC-like (cyan box) originated due to gene duplication. Two other duplication events of SFTPClike are predicted in salamanders. While some species of salamanders appear to express only one form of SFTPC-like, several species express SFTPC-like transcripts with slight sequence differences (Supplemental Data File 1). Further work is needed to determine if these sequence differences indeed represent further duplications of SFTPC-like or instead are the result of assembly error or alternative splicing. Only one SFTPC-like sequence per species was selected for phylogenetic analysis. Salamander SFTPC is marked with a yellow box. Lungless (plethodontid) species are in red font.

Captions for Supplemental Data Files:
Supplemental Data File 1: Excel spreadsheet with all sequence data used for the study. Supplemental Data File 2: A FASTA amino acid alignment used to generate the gene tree.