Evolution of a chordate-specific mechanism for myoblast fusion

Zhang

et al

Sci. Adv.

8

, eadd2696 (2022) 2 September 2022

SCIENCE ADVANCES

|

RESEARCH ARTICLE

1 of 15

EVOLUTIONARY BIOLOGY

Evolution of a chordate-specific mechanism

for myoblast fusion

Haifeng Zhang

, Renjie

Shang

1,2

, Kwantae

Kim

, Wei

Zheng

4,5

, Christopher J.

Johnson

, Lei

Sun

Xiang Niu

, Liang

Liu

8,9

, Jingqi

Zhou

, Lingshu

Liu

, Zheng

Zhang

, Theodore A.

Uyeno

Jimin Pei

, Skye D.

Fissette

, Stephen A.

Green

, Sukhada P.

Samudra

, Junfei

Wen

Jianli Zhang

, Jonathan T.

Eggenschwiler

, Douglas B.

Menke

, Marianne E.

Bronner

Nick V.

Grishin

11,15

, Weiming

, Kaixiong

2,9

, Yang

Zhang

4,5

, Alberto

Stolfi

*, Pengpeng

1,2

Vertebrate myoblast fusion allows for multinucleated muscle fibers to compound the size and strength of mono-

nucleated cells, but the evolution of this important process is unknown. We investigated the evolutionary origins

and function of membrane-coalescing agents Myomaker and Myomixer in various groups of chordates. Here, we

report that

Myomaker

likely arose through gene duplication in the last common ancestor of tunicates and verte-

brates, while

Myomixer

appears to have evolved de novo in early vertebrates. Functional tests revealed a complex

evolutionary history of myoblast fusion. A prevertebrate phase of muscle multinucleation driven by Myomaker

was followed by the later emergence of Myomixer that enables the highly efficient fusion system of vertebrates.

Evolutionary comparisons between vertebrate and nonvertebrate Myomaker revealed key structural and mecha-

nistic insights into myoblast fusion. Thus, our findings suggest an evolutionary model of chordate fusogens and

illustrate how new genes shape the emergence of novel morphogenetic traits and mechanisms.

INTRODUCTION

A fundamental step in vertebrate muscle development is the fusion

of mononucleated myoblasts to form multinucleated myofibers (

Generation of syncytial myofibers allows concerted power outputs

to fulfill complex locomotor functions and therefore was likely

instrumental for the adaptive radiation of vertebrates. Myomaker

(MymK) and Myomixer (MymX) are two recently identified muscle-

specific fusogens that drive plasma membrane coalescence during

vertebrate myoblast fusion (

–

). Deletion of either gene causes

perinatal lethality of mice due to fusion defects resulting in muscle

malfunction (

). Moreover, forced expression of this duo confers

fusogenic activity even onto fibroblasts, which are not normally

capable of undergoing cell fusion (

Here, we report the identification and characterization of MymX

and MymK orthologs outside of jawed vertebrates. We demon

strate that the fusogenic activity of MymK likely evolved in the

last common ancestor of tunicates and vertebrates (Olfactores) and

therefore predates the origin of MymX, which appears to have evolved

de novo specifically in the vertebrate lineage to facilitate the massive

multinucleation of skeletal muscles. Coculturing mammalian cells

expressing either vertebrate or tunicate MymK revealed that MymK/

MymX synergy primarily depends on the presence of either compo

nent on a different cell (i.e., in trans). Together, our study provides

a crucial insight into the still poorly understood evolutionary and

molecular mechanisms underlying vertebrate myogenesis.

RESULTS

Evolutionary origins of

MymK

The phylum Chordata is composed of vertebrates together with two

nonvertebrate subphyla: Tunicata and Cephalochordata (Fig. 1A).

Cephalochordates have mononucleated muscles indicating no myo

blast fusion (fig. S1A) (

), whereas tunicates exhibit limited multi-

nucleation of certain muscles (

) and vertebrates have extensive,

obligatory multinucleation (fig. S1B). Therefore, we reasoned that

comparative gene function studies of these closely related animal

groups might shed insights into the evolutionary history and cellular

mechanisms of myoblast fusion.

Originally known as

Tmem8c

MymK

belongs to a gene family

that in vertebrates also contains two other paralogs:

Tmem8a

and

Tmem8b

. Homology-guided searches revealed that multiple tuni-

cate species have both

MymK

and a

Tmem8a/b-

like gene (herein

named

Tmem8

-related) (Fig. 1, A and B). In the cephalochordate

Branchiostoma floridae

, only a single

Tmem8

family gene could be

identified.

Tmem8

sequences are found in diverse eukaryotes, in-

cluding the unicellular filasterean

Capsaspora owczarzaki

(Fig. 1B

and fig. S2A). Comparisons of these proteins revealed an epidermal

growth factor–like domain that exists in all Tmem8 family proteins

except MymK, which appears to have lost this domain (fig. S2B).

Therefore, the duplication of an ancestral

Tmem8

gene likely gave

rise to

Tmem8a/b

and

MymK

before tunicates and vertebrates di-

verged (fig. S2C). Although the exact timing of this duplication

event cannot be determined (fig. S2C), the lack of multinucleated

Center for Molecular Medicine, University of Georgia, Athens, GA, USA.

Department

of Genetics, University of Georgia, Athens, GA, USA.

School of Biological Sciences,

Georgia Institute of Technology, Atlanta, GA, USA.

Department of Computational

Medicine and Bioinformatics, University of Michigan, Ann Arbor, MI, USA.

Depart-

ment of Biological Chemistry, University of Michigan, Ann Arbor, MI, USA.

The

Fifth People’s Hospital of Shanghai, and Shanghai Key Laboratory of Medical

Epigenetics, Institutes of Biomedical Sciences, Fudan University, Shanghai, China.

Tri-Institutional Program in Computational Biology and Medicine, Weill Cornell

Medical College, New

York, USA.

Department of Statistics, University of Georgia,

Athens, GA, USA.

Institute of Bioinformatics, University of Georgia, Athens, GA, USA.

Department of Biology, Valdosta State University, Valdosta, GA, USA.

Howard

Hughes Medical Institute, University of Texas Southwestern Medical Center, Dallas,

TX, USA.

Department of Fisheries and Wildlife, Michigan State University, East

Lansing, MI, USA.

Division of Biology and Biological Engineering, California Institute

of Technology, Pasadena, CA, USA.

College of Engineering, University of Georgia,

Athens, GA, USA.

Department of Biophysics, University of Texas Southwestern

Medical Center, Dallas, TX, USA.

*Corresponding author. Email: pbi@uga.edu (P.B.); alberto.stolfi@biosci.gatech.edu

(A.S.)

The Authors, some

rights reserved;

exclusive licensee

American Association

for the Advancement

of Science. No claim to

original U.S. Government

Works. Distributed

under a Creative

Commons Attribution

NonCommercial

License 4.0 (CC BY-NC).

Downloaded from https://www.science.org at California Institute of Technology on May 08, 2023

Zhang

et al

Sci. Adv.

8

, eadd2696 (2022) 2 September 2022

SCIENCE ADVANCES

|

RESEARCH ARTICLE

2 of 15

muscles in cephalochordates and other deuterostomes (fig. S2A)

suggested a functional link between muscle multinucleation and the

presence of

MymK

in olfactorians (tunicates + vertebrates). Although

multinucleation is also a prominent feature of arthropod muscula-

ture (

–

), the

MymK

gene is absent from this phylum, suggesting

convergent evolution of myoblast fusion through different molecu-

lar mechanisms.

Functional comparisons between vertebrate

and

nonvertebrate MymK proteins

We identified

MymK

orthologs in various tunicates, including

benthic ascidians and pelagic thaliaceans, ranging from ~26 to 30%

amino acid identity in alignments with human MymK (fig. S3, A

and B). When expressed in human myoblasts (fig. S3C), tunicate

MymK protein can be specifically detected in the membrane fraction

(fig. S3D). To examine their functional conservation as fusogens,

we devised a heterologous rescue approach (Fig. 1C). First, we used

CRISPR to inactivate

MymK

in myoblasts isolated from different

vertebrate species (human, Fig. 1, D to F; mouse, fig. S4, A to D;

lizard, fig. S4, E to H), which completely abolished syncytializations

(Fig. 1E and fig. S4, C and H). We then expressed tunicate MymK

proteins and assayed their ability to rescue the fusion of these

MymK-deficient cells. All tunicate MymK orthologs tested can con-

sistently rescue the fusion of

MymK

−/−

myoblasts, albeit with lower

levels of efficiency than vertebrate proteins (Fig. 1F and figs. S4, D

and H, and S5). Consistent with the neofunctionalization of

MymK

tunicate Tmem8-related and cephalochordate Tmem8 proteins did

not elicit fusogenic activity (fig. S6).

Although MymK is necessary and sufficient for vertebrate myo-

blast fusion, a second membrane protein called MymX synergistically

Fig. 1. Tunicate MymK orthologs have weak fusogenic function in vertebrate myoblasts.

(

) Phylogenetic relationships of various chordate clades used to deduce

the evolutionary origins of the

MymK

gene (also known as

Tmem8c

). Asterisks represent two potential duplication events of

Tmem8

genes that give rise to

8-related

and

members. Sequence identities of MymK orthologs were compared to human MymK protein. Scale at the top shows approximate date in millions of years (Ma) ago.

(

) Phylogeny of the Tmem8 gene family inferred by a distance-based method (neighbor joining). The bootstrap percentages were obtained from 1000 replicates.

O. anatinus

Ornithorhynchus anatinus

;

P. mammillata

Phallusia mammillata

. Tmem8 gene members from jawless vertebrates were highlighted in blue and tunicates in

green. Extended phylogenetic analysis is seen in fig. S2A. (

) Schematic of experimental design to test the fusogenic function of tunicate MymK proteins in human

MymK

−/−

myoblasts. (

) Human

MymK

gene structure, sgRNA positions, and genotyping results that showed biallelic frameshift mutations induced by CRISPR/Cas9. bp, base pair;

UTR, untranslated region. (

) Myosin immunostaining of human

MymK

−/−

myoblasts transfected with MymK orthologs. Muscle syncytia (outlined) were observed in

nonvertebrate (

Styela

and

Ciona

) MymK expression groups, although smaller than the syncytia induced by vertebrate MymK proteins. Scale bar, 100



m. (

) Measure-

ments of myoblast fusion after 4 days of myogenic differentiation. E. shark, elephant shark. Data are means ± SEM. **

< 0.01 and ***

< 0.001, compared to control group,

one-way analysis of variance (ANOVA).

Downloaded from https://www.science.org at California Institute of Technology on May 08, 2023

Zhang

et al

Sci. Adv.

8

, eadd2696 (2022) 2 September 2022

SCIENCE ADVANCES

|

RESEARCH ARTICLE

3 of 15

enhances the fusogenic activity of MymK during myogenesis (

Despite extensive searching (Materials and Methods), we were not

able to identify MymX homologs in tunicates or any other non-

vertebrate species. Thus, we postulated that fusogenic activity of

tunicate MymK is independent of MymX.

To test this idea, we gen-

erated human

MymX

MymK

double knockout (KO) myoblasts by

CRISPR.

In the absence MymX, tunicate and human MymK induce

comparable levels of human myoblast fusion, supporting its con-

served function (fig. S7). However, a functional difference between

human and tunicate MymK was unmasked by resupplying MymX.

Specifically, coexpression of human MymX + human MymK induced

massive fusion (fig. S7). Such synergy was not observed when

human MymX was paired with tunicate MymK (fig. S7). These re-

sults suggest that, although the fusogenic role of MymK predates

the emergence of vertebrates and the

MymX

gene, vertebrate-

specific changes to MymK were essential for the evolution of func-

tional synergy with MymX.

Temporally and

spatially restricted expression of

MymK

drives multinucleation program of

Ciona

muscle

Having established the fusogenic activity of tunicate MymK in ver-

tebrate cells, we next asked whether it plays a role in the develop-

ment of multinucleated myofibers in the laboratory model tunicate,

Ciona robusta.

The presence of

MymK

in tunicates was intriguing

because these nonvertebrate chordates also have multinucleated

muscles (

). While the tail muscles from tunicate larvae are

mononucleated, the siphon and body wall muscles of postmeta-

morphic juveniles and adults are composed of multinucleated fibers

(

). The presence of multinucleated siphon muscles in other tuni-

cate species also correlates with the presence of the

MymK

gene

(Fig. 2B). Notably,

MymK

is absent from appendicularians, which

have secondarily lost multinucleated siphon and body wall muscles

(

). In contrast,

MymK

orthologs were found in all other tunicate

species with multinucleated siphon or body wall muscles (Fig. 2B).

Ciona

, expression of

MymK

was observed exclusively in multi

nucleated juvenile muscles by in situ hybridization (Fig. 2C) and by

MymK

promoter green fluorescent protein (GFP) reporter (Fig. 2D

and fig. S8, C and D). In contrast,

MymK

expression was not ob-

served in any other cell type including mononucleated larval tail

muscle cells (fig. S8A). This was further confirmed by reanalyzing

published single-cell transcriptome data (

) collected at differ-

ent developmental stages, in which we detected

MymK

expression

specifically in multinucleated muscle precursor cells (Fig. 2, E and F,

and fig. S8B). Together, these results suggest that expression of the

MymK

gene is highly specific to multinucleated muscles in tunicates.

We then performed

MymK

loss-of-function experiments in

Ciona

using tissue-specific CRISPR mutagenesis (

) in the cardiopharyngeal

mesoderm lineage that gives rise to the multinucleated muscles of

the atrial siphon (Fig. 2G and fig. S9A) (

). In control juveniles,

circular atrial siphon myofibers invariably formed as orderly rings

with occasional longitudinal myofibers emanating from the siphon

region (Fig. 2G, fig. S9B, and movie S1). In contrast,

MymK

CRISPR

resulted in highly disorganized atrial siphon muscles (Fig. 2G;

fig. S9, C and D; and movie S2). Moreover, there was a reduction in

the frequency of binucleated atrial siphon/longitudinal myofibers

MymK

CRISPR juveniles (Fig. 2H), suggesting that

MymK

required for myoblast fusion in

Ciona.

However, overexpression of

MymK in mononucleated larval tail muscle cells did not promote

obvious multinucleation [fig. S10; 16 hours post-fertilization (hpf)

control, movie S3; 16 hpf

MRF

MymK

, movie S4]. This suggests

that, as in vertebrates (

), the fusogenic activity of MymK in

Ciona

likely requires other factor(s) present in juvenile but not larval tail

muscle cells.

A distantly related

MymX

sequence from

lamprey genome

can replace its mammalian orthologs in

enhancing

myoblast fusion

Cyclostomes such as lampreys and hagfish diverged from jawed

vertebrates (gnathostomes)

∼

500 million years ago (

). Histological

analysis revealed extensive multinucleation of sea lamprey (

Petromyzon

marinus

) muscle (Fig. 3A), which can host up to several hundred

myonuclei per fiber, a stark contrast to maximally a few dozen in

tunicates (

). We hypothesized that a protein with MymX function

exists in lamprey to robustly induce myoblast fusion in cooperation

with MymK.

The search for MymX orthologs is intrinsically challenging due to

the small size (<100 amino acids) and high frequency of substitutions

(Fig. 3B). Nonetheless, iterative BLAST (basic local alignment search

tool) searches identified one hit from a genome shotgun sequence

(GenBank: AEFG01021847.1) of the sea lamprey. Alignment of RNA

sequencing (RNA-seq) reads revealed a single-exon open reading

frame (ORF) that encodes 583 amino acids including the hydrophobic

AxLyCxL motif (

) that is essential for mammalian MymX function

(Fig. 3, B and C, and fig. S11, A to C). A homologous sequence was also

found from arctic lamprey (

Lethenteron camtschaticum

, APJL01015224),

revealing an ORF of 595 amino acids that shares 93% identity with the

sea lamprey sequence (fig. S12). The complete ORF of sea lamprey

MymX was codon-optimized, cloned by gene synthesis, and expressed

in human myoblasts. Western blot readily detected a 70-kDa band

specifically from the membrane fraction (Fig. 3D). Immunofluores-

cence revealed the presence of lamprey MymX on the cell surface

(Fig. 3E and fig. S11D), suggesting a function in this compartment.

Given the low sequence identity between lamprey and gnatho-

stome MymX, we examined its function in a heterologous rescue

experiment. While

MymX

−/−

myoblasts are weakly fusogenic due to

the residual activity of MymK in these cells (

), the expression of lam-

prey MymX robustly enhanced cell fusion (human, Fig. 3, F and G;

mouse, fig. S13). Similar to mammalian MymX, this fusion-promoting

activity of lamprey MymX strictly requires MymK for function be-

cause it failed to induce fusion when

MymK

was deleted from

human myoblasts (fig. S14).

Lamprey-specific C terminus from

MymX is indispensable

for

optimal fusogenic activity

For lamprey MymX, only a short region of 52 amino acids at the

N terminus (N52) can be aligned to conventional orthologs

(Fig. 3B). However, expression of the N52 polypeptide failed to

induce myoblast fusion (Fig. 4, A to C). A similarly deleterious

effect was observed when the conserved AxLyCxL motif was re-

moved (Fig. 4, A to C). We continued to dissect the function of its

unusually long extracellular C-terminal sequence by generating a

series of sea lamprey

MymX

mutants (Fig. 4A). As the region of

deletions enlarged, MymX function, quantified as nuclei number

per syncytium, gradually diminished (Fig. 4C and fig. S15). There-

fore, the optimal activity of sea lamprey MymX requires its large,

nonconserved C-terminal structure.

We next investigated the expression pattern of

MymX

and

MymK

during lamprey muscle development. Sea lampreys have a complex

Downloaded from https://www.science.org at California Institute of Technology on May 08, 2023

Zhang

et al

Sci. Adv.

8

, eadd2696 (2022) 2 September 2022

SCIENCE ADVANCES

|

RESEARCH ARTICLE

4 of 15

life cycle that involves a freshwater-based larval period of 2 to 10 years,

followed by metamorphosis into a marine-based adult stage. It is

unclear when myoblast fusion occurs in lampreys, although it was

reported that muscle cells from young larvae remained mononucleated

(

). Because of their complex life history, we could not examine

larvae of defined age or stage to identify the temporal window of myo-

blast fusion in sea lamprey. Instead, we compared groups of larvae

of uncertain age (estimated 2 to 3.5 years of age) but of different

sizes (Fig. 4D), assuming that multinucleation might be occurring

primarily during muscle growth (

). Moderate multinucleation

was consistently observed in muscles of both groups (Fig. 4D),

although a higher number of nuclei per myofiber was associated with

larger muscle and body size (Fig. 4,

D′

and

D′′

). Last, expression of

both

MymX

and

MymK

was detected in larval muscles of both sea

and arctic lampreys but not in other tissues or adult muscles (Fig. 4E

and fig. S16).

Together, our expression and functional data suggest that these

distantly related lamprey sequences are true orthologs of MymX

Fig. 2. MymK is required for multinucleation of postmetamorphic muscles in the tunicate

Ciona

(

) Diagram of biphasic life cycle of ascidians (sessile tunicates) like

Ciona.

The motile larvae have strictly mononucleated tail muscles during the dispersal phase. After settlement and metamorphosis, tail muscle cells undergo programmed

cell death and are reabsorbed, while dedicated muscle progenitors set aside in the larva differentiate to form the multinucleated siphon and body wall muscles of the

juvenile. Muscles surrounding and emanating from the oral and atrial siphons are derived from distinct cell lineages in the larva. Only those from the atrial siphon are

derived from the Mesp

B7.5 lineage [in (E)]. (

) Cladogram of extant chordates showing correlation between the presence of

MymK

gene and muscle multinucleation in

different clades. (

) Whole-mount mRNA in situ hybridization showing

MymK

expression in developing atrial siphon muscle (ASM) and oral siphon muscle (OSM) cells in

metamorphosing juveniles. Smaller arrows indicate autofluorescent tunic cells. (

)

C. robusta

juvenile developed from a zygote transfected with a

MymK

promoter reporter

plasmid, labeling ASMs and longitudinal body wall muscles (LoM). (

) Diagram of the B7.5 lineage in

C. robusta

, based on conclusions from (

). FC, founder cell; TVC, trunk

ventral cell; ATM, anterior tail muscle cell; STVC, secondary TVC; FHP, first heart precursor; SHP, second heart precursor; ASMF, atrial siphon muscle founder cell; ASMP,

atrial siphon muscle precursor; oASMP, outer ASMP; iASMP, inner ASMP.

Asterisk indicates that both FCs give rise to identical lineages.

MRF

myogenic regulatory factor

(

MyoD

ortholog). (

)

-distributed stochastic neighbor embedding (tSNE) plots based on information from (

) showing

MymK

expression mapped onto TVC progeny

clusters at 20 hpf.

MymK

is expressed exclusively in ASMPs and especially enriched in outer ASMPs. Abbreviations same in (E). (

) Representative Z-projection confocal

fluorescence images of 84 hpf negative control (transfected with

Mesp > Cas9

only, no sgRNAs) juveniles alongside same-age juveniles in which

MymK

was targeted for

mutagenesis specifically in the B7.5 lineage.

MymK

CRISPR: zygotes transfected with

Mesp > Cas9

and

MymK-

sgRNA vectors. Muscle plasma membranes and nuclei

labeled by

MRF > CD4::GFP

and

MRF > H2B::mCherry

, respectively. Arrows in negative control panels showing development of typical binucleated myofibers that is inhibited

upon

MymK

CRISPR. (

) Data from scoring of juveniles represented in (G) showing reduced frequency of binucleated atrial siphon/longitudinal myofibers in

MymK

CRISPR

juveniles.

, numbers of juveniles assayed for each condition. Scale bars, 50



Downloaded from https://www.science.org at California Institute of Technology on May 08, 2023

Zhang

et al

Sci. Adv.

8

, eadd2696 (2022) 2 September 2022

SCIENCE ADVANCES

|

RESEARCH ARTICLE

5 of 15

and that the functional cooperativity between MymX and MymK

in myoblast fusion is likely to be conserved in cyclostomes. Because

MymX does not appear to share homology with any other protein

and because our extensive in silico TBLASTN (translated BLAST)

search did not identify an AxLyCxL motif containing sequence of

interest from genome/transcriptome of multiple tunicate species

or any invertebrates, we propose that

MymX

is a vertebrate-specific

orphan gene encoding a core molecular component of myoblast

fusion that arose de novo before the split between jawed and jawless

vertebrates (fig. S17).

Insights into mechanisms of

myoblast fusion obtained

from evolutionary comparisons

We next sought to leverage these newly identified muscle fusogens

from different chordate groups for insights into the mechanisms

of myoblast fusion, which remain poorly understood. Because

Fig. 3. Discovery of the unusual

MymX

genes from lampreys.

(

) Histological staining and immunofluorescence of muscle tissues dissected from adult sea lamprey

(

P. marinus

). Multinucleated myofibers (arrows) are observed from the longitudinal sections. (

) Cross-species homology of lamprey MymX aligned with its orthologs from

jawed vertebrates. Only a few residues from the AxLyCxL motif and the N terminus can be aligned. x denotes leucine, valine, or isoleucine, and y denotes serine, threonine,

or glycine. The numbers below the consensus sequence refer to the positions in sea lamprey MymX (only the N-terminal 52 amino acids are shown). (

) RNA-seq tracks

that confirmed transcription of

MymX

gene in sea lamprey (sequence read archive accession: PRJNA497902). No splicing junction was detected in the hypothetical

ORF.

The purple rectangle highlights the coding region of the N-terminal 52 amino acids of lamprey MymX shown in (B). (

) Western blot analyses of cytosolic (c) and

membrane (m) fractions of human myoblasts transfected with C-tagged lamprey MymX.

C-tag is a small four–amino acid peptide tag E-P-E-A.



-Tubulin blot was used as

a positive control of cytosolic proteins. Insulin receptor



(INSR-



) blot was used as a positive control of membrane proteins. (

) Human myoblasts transfected with

lamprey MymX-GFP fusion protein. Nuclei were counterstained with Hoechst dye. The fluorescence intensity cross the plasma membrane (white bar in the image) was

measured. Membrane and cytosol targeting GFPs were added as measurement controls (see images in fig. S11D). (

) Myosin immunostaining of human

MymX

−/−

myoblasts

transfected with MymX orthologs. Note that sea lamprey MymX can rescue fusogenic defects of human

MymX

−/−

cells that formed larger muscle syncytia (outlined) than

control (empty vector). E. shark, elephant shark. Scale bar, 100



m. (

) Measurement of myoblast fusion in (F) after 4 days of differentiation. Data are means ± SEM.

***

< 0.001, compared to control group, one-way ANOVA.

Downloaded from https://www.science.org at California Institute of Technology on May 08, 2023

Zhang

et al

Sci. Adv.

8

, eadd2696 (2022) 2 September 2022

SCIENCE ADVANCES

|

RESEARCH ARTICLE

6 of 15

the emergence of MymX and its functional cooperativity with

MymK represent a key step in the evolution of vertebrate myo-

genesis, we focused on dissecting the mechanism of MymX-MymK

synergy.

Our previous fusion reconstitution assays revealed that MymK

protein is needed in the plasma membrane of both cells undergoing

fusion, whereas adding MymX to only one side is sufficient to boost

the efficiency of fusion (Fig. 5A) (

). This raised the question of

whether MymX promotes the function of MymK in trans (between

the two cells’ membranes) or cis (in the same membrane). The dis-

covery of tunicate MymK proteins that are unable to synergize with

mammalian MymX allowed us to design a novel experiment testing

the cis/trans basis of the MymX-MymK synergy between myoblasts

expressing either tunicate or human MymK (Fig. 5A). Unexpectedly,

we found that MymX was only capable of significantly promoting

cell fusion when expressed in trans to a cell expressing mamma-

lian MymK protein. In contrast, fusion was not significantly en-

hanced when mammalian MymK and MymX were expressed in cis

(Fig. 5, B and C). This potential in trans synergy is consistent with

the topology of MymX protein where the conserved AxLyCxL motif

found in all MymX proteins is located on the extracellular side,

where it might be able to interact with MymK and/or other factors

on the opposing membrane in trans.

We next sought to better understand the structural basis of both

conserved and divergent MymK functions across Chordata. By

applying a deep neural network-based structure assembly method

(

–

), we obtained seven structural models for MymK proteins

from representative species of vertebrates and tunicates (fig. S18,

A and B). MymK proteins from all taxonomic groups share >80%

similarity of the overall structure (fig. S18C) in which seven trans-

membrane (TM) helices are arranged in an anticlockwise manner

when viewed from the extracellular space (Fig. 5D). As part of TM1,

the N-terminal residues protrude toward the outside of the cell and

form the extracellular face together with three extracellular loops

(fig. S18, D and E). The structure of C-terminal residues is disordered

and forms the intracellular face together with three intracellular loops

(fig. S18, D and E). The TM helices of MymK enclose an internal cavity

that goes through the entire structure with a small intracellular

Fig. 4. Mutation and expression analysis of lamprey MymX protein.

(

) Hydrophobicity map of sea lamprey MymX and a schematic of mutants. Red lines highlight

deleted regions. HH, hydrophobic helix; MH, membrane-anchor helix; aa, amino acid; GAPDH, glyceraldehyde-3-phosphate dehydrogenase. (

A′

) Western blot results that

confirmed expression of lamprey MymX mutants in human myoblasts. MymX was detected by blotting a diminutive C-tag fused at the C terminus of target. The predicted

and detected molecular weights are labeled on the schematics and Western blots, respectively. The four–amino acid epitope tag (E-P-E-A) is 0.4 kDa. Quantifications of

blots are seen in fig. S15B. EV, empty vector. (

) Myosin immunostaining of human

MymX

−/−

myoblasts transfected with full length [wild type (WT)] or truncated lamprey

MymX proteins. (

) Measurement of myoblast fusion in (B) after 4 days of myogenic differentiation. (

) Staining of longitudinal sections of muscles dissected from sea

lamprey larvae of two different size groups to identify muscle fusion stage. Measurements of larva body length and weight (

D′

) and nuclei number per myofiber (

D′′

(

) Reverse transcription PCR results that validated the muscle-specific expression pattern of

MymX

and

MymK

genes in sea lamprey larval muscle tissues. m, muscle cDNA;

n, nonmuscle (intestine and liver) cDNA.

Downloaded from https://www.science.org at California Institute of Technology on May 08, 2023

Zhang

et al

Sci. Adv.

8

, eadd2696 (2022) 2 September 2022

SCIENCE ADVANCES

|

RESEARCH ARTICLE

7 of 15

opening and a larger extracellular opening (Fig. 5E). Unsupervised

comparison clustered the predicted structures into consistent taxo-

nomic groups (fig. S18C). The major structural differences between

tunicate and vertebrate MymK are on the protein surfaces (fig. S18D)

and the orientation of TM5 helix, which is tilted by 11° relative to

TM5 in tunicates (Fig. 5F), hinting at a potentially important role of

this structural adaption for the synergy with MymX.

Last, our comparative three-dimensional protein modeling pre-

dicted a close resemblance between MymK and adiponectin receptor

(AdipoR) structures (Fig. 5D and fig. S18F) (

). Stabilization of

Fig. 5. Evolutionarily distinct Mymk proteins reveal mechanistic insights into structure function and synergy.

(

) Schematic of experiment design. Note that the

basal level of myoblast fusion requires MymK to be present in both cells, while the expression of vertebrate MymX can only boost vertebrate MymK (e.g., human) but not

nonvertebrate (e.g.,

Ciona

) MymK activity. The natural uncoupling between MymX synergy and fusogenicity observed in tunicate MymK permits the test of vertebrate

MymX/MymK synergy using cell mixing cultures. dKO, double knockout. (

) Representative fluorescence images of human myoblasts after mixing culture as illustrated in

(A). Arrows point to the EdU

nuclei inside GFP

cells formed from fusion. Scale bar, 100



m. (

) Measurement of heterologous fusion by counting EdU

nuclei inside GFP

syncytia. Data were normalized to the “no MymX” group. Data from the same replicate were highlighted in the same color.

= 10. **

< 0.01; ns, not significant. (

) Ribbon

representation of the predicted human MymK structure. TM, transmembrane helix. The conserved histidine and cysteine residues on human MymK model are highlighted.

Zinc-binding motif of adiponectin receptor 1 (AdipoR1; PDB ID: 6KRZ) was shown on the right. (

) Side views of the predicted cavities inside MymK proteins. (

) Super-

impositions of the overall structural models for MymK proteins from vertebrates (human, mouse, zebrafish, and elephant shark) and tunicates (

Phallusia

Ciona

, and

Styela

The orientations of TM5 show obvious shifts between the two taxonomic groups. (

) Myosin immunostaining of human

MymK

−/−

myoblasts that revealed the fusogenic

activity of human MymK mutants. Cells were differentiated for 4 days. Scale bar, 100



m. (

) Measurement of myoblast fusion in (G) after myogenic differentiation and

compared to WT expression group. Data are means ± SEM. ***

< 0.001, one-way ANOVA.

Downloaded from https://www.science.org at California Institute of Technology on May 08, 2023

Zhang

et al

Sci. Adv.

8

, eadd2696 (2022) 2 September 2022

SCIENCE ADVANCES

|

RESEARCH ARTICLE

8 of 15

AdipoR structure requires a zinc ion coordinated by three histidine

(His) residues (Fig. 5D). We identified a similar motif in MymK, near

the outer lipid layer of the membrane (His

, His

180

, and His

184

). In

addition, two cysteine (Cys) residues from the TM2 (Cys

) and

extracellular loop 1 (Cys

) are predicted to form a disulfide bond.

These histidine and cysteine residues are perfectly conserved in all

tunicate and vertebrate MymK orthologs, suggesting a crucial con-

tribution to the structure and function of MymK.

Mutating these

residues in human MymK drastically affected its fusogenic activity

(Fig. 5, G and H). In summary, by looking at both conserved and

divergent features, we have gained new insights into the mechanisms

of MymK and MymX function in myoblast fusion.

DISCUSSION

Our comparative study of MymK and MymX in multiple chordate

(vertebrate and nonvertebrate) species sheds light on the evolution

of myoblast fusion (Fig. 6). Whereas the

MymK

gene was certainly

generated through duplication of an ancestral

Tmem8

gene,

MymX

as an orphan gene, might have arisen de novo. Our data suggest that

tunicate MymK can promote myoblast fusion in both tunicate and

vertebrate cells but is unable to synergize with vertebrate MymX

proteins to augment fusion levels. In contrast, lamprey MymX

function is conserved enough to synergize with human MymK,

despite its highly divergent length and sequence. Together, our data

are consistent with a de novo origin of MymX after the tunicate-

vertebrate split.

One scenario for the origin of the

MymX

gene could be through a

transitory protogene that produced a short polypeptide, given the

short length (<100 amino acids) of MymX proteins in most verte-

brates. After cyclostomes and gnathsostomes had diverged, MymX

may have been secondarily elongated in lampreys (583 amino acids

in sea lamprey and 595 amino acids in arctic lamprey). Alternatively,

the ancestral MymX protein was closer in size to that of extant

lampreys but was secondarily reduced in length in jawed vertebrates.

Notably, possibly attributed to the fact that the hagfish (

Eptatretus

burgeri

) genome is not complete (

), a

MymX

ortholog has yet to

be identified in this species.

This potential stepwise evolution of myoblast fusion in chordates

lends support to an updated “new head/new heart” hypothesis

(

), which postulates that the active predatory lifestyle of early

vertebrates was made possible due to increased sensory capabilities,

a chambered heart, and a muscularized pharynx, all derived from

mostly cephalic neural crest or cardiopharyngeal progenitor cells.

In this context, the evolution of MymK may have been a key inno-

vation of the last common ancestor of vertebrates and tunicates.

Fig. 6. Evolution and mechanism of chordate-specific control system of myoblast fusion.

The emergence of

MymK

gene after duplication of the eukaryotic

Tmem8

gene allowed the multinucleation of muscle cells for common ancestors of tunicates and vertebrates. The emergence of MymX and structural adaptions of MymK proteins

drove the extensive and obligatory fusion in vertebrates. Note that Tmem8a/b is called Tmem8-related in tunicates simply due to poor phylogenetic resolution, according

to tunicate gene nomenclature rules. In vertebrates, this gene became duplicated again to give rise to Tmem8a and Tmem8b. Evolutionary comparison of vertebrate and

tunicate MymK revealed that MymK/MymX synergy primarily depends on the presence of either component on a different cell (i.e., in trans).

Downloaded from https://www.science.org at California Institute of Technology on May 08, 2023

Zhang

et al

Sci. Adv.

8

, eadd2696 (2022) 2 September 2022

SCIENCE ADVANCES

|

RESEARCH ARTICLE

9 of 15

Both tunicates and vertebrates have a prominently muscularized

pharynx, which may have been facilitated by the emergence of MymK

as a driver of myoblast fusion. However, additional enhancement of

the myoblast fusion pathway in vertebrates would have been made

possible by the de novo evolution of MymX after the split from

tunicates. The higher number of nuclei per myofiber seen in verte-

brates compared to tunicates suggests that a more “active lifestyle”

may have required stronger, longer myofibers formed by more con-

stituent myoblasts. However, considering gene loss is a pervasive

source of genetic variation in evolution (

), it is also possible

that

MymX

(and by extension, MymK-MymX synergy) arose earlier

in chordate evolution but was later lost in tunicates as this group

secondarily evolved a biphasic life cycle with a sessile adult phase.

In addition to clarifying the evolutionary history of MymK and

MymX, our comparative approach also yielded key insights into

the molecular mechanisms of these chordate-specific fusogens.

Taking advantage of the natural uncoupling between fusogenicity

and cooperativity with MymX observed with tunicate MymK, our

cis/trans tests suggest that the functional synergy between MymK

and MymX is likely most important in trans. We propose that, while

MymK is needed in both cells undergoing fusion, MymX synergizes

with MymK in trans and not in cis. Last, our structural modeling and

sequence alignment of diverse MymK proteins revealed key residues

that are hyperconserved across all chordates and crucial for MymK

function. Although it remains unknown whether the functional

synergy of MymX/MymK involves a direct physical interaction or

indirectly through engaging additional fusion-promoting factors in

the same pathway (

), our structure-function analysis of both factors

should provide a key basis for the complete understanding of the

biophysical mechanism of myoblast fusion.

In summary, our study closes long-existing gaps in the evolu-

tionary history of myoblast fusion, an important developmental

mechanism that independently evolved in chordates and ultimately

facilitated much of vertebrate evolution. Furthermore, we show that

by combining a broad, comparative “evo-devo” approach with genetic

interrogation of protein function and in silico protein structure

modeling, we can advance our understanding of the molecular and

evolutionary mechanisms of key developmental processes.

MATERIALS AND METHODS

Animal husbandry

Standard operating procedures for transporting, maintaining,

handling, and euthanizing of sea lamprey and hagfish were approved

by the Institutional Committee on Animal Use and Care of Michigan

State University and California Institute of Technology, Valdosta

State University, and University of Georgia and in compliance with

standards defined by the National Institutes of Health Guide for the

Care and Use of Laboratory Animals.

Sea lampreys were trapped in tributaries of Lakes Huron and

Michigan by the U.S.

Fish and Wildlife Service and Fisheries and

Oceans Canada. Captured lampreys were transported to the U.S.

Geological Survey, Hammond Bay Biological Station (HBBS),

Millersburg, Michigan and held in 200- to 1000-liter tanks that

were continually fed with ambient temperature, aerated Lake Huron

water. Adult lampreys also were transported to the California Institute

of Technology where lamprey husbandry was performed as previ-

ously described (

) in accordance with the Guide for the Care and

Use of Laboratory Animals of the National Institutes of Health, and

protocols were approved by the Institutional Animal Care and Use

Committee of the California Institute of Technology (lamprey, pro-

tocol no. 1436-17). Adult lamprey muscle was fixed in 4% para-

formaldehyde (PFA) and processed using conventional histology.

To produce sexually mature ovulated females and males for em-

bryo collection, sea lampreys were transferred to the Ocqueoc River,

Millersburg, Michigan and held in cages (0.5 m

) constructed of

polyvinyl chloride and polyurethane mesh, allowing natural sexual

maturation in a riverine environment. Sea lampreys were checked

daily for sexual maturation. Sexually mature males were identified

by applying abdominal pressure and checking milt expression (

Sexually mature females were identified by applying abdominal

pressure and checking for ovulated oocyte expression along with

visual observation of secondary sexual characteristics (

). Sexually

mature males and female lampreys were returned to HBBS where

they were held until used for collecting and culturing lamprey em-

bryos as previously outlined (

). Embryo viability was determined

using techniques established for evaluation of the sterile male release

program in the Laurentian Great Lakes (

). Embryos were checked

daily for viability, and dead embryos were removed from holding

containers. Embryos were pooled together for individual samples

according to Piavis stages.

Female Atlantic hagfishes (

Myxine glutinosa

, Linnaeus, 1758) were

used in this study (specimen/mass/length; #1/64 g/45 cm; #2/57

g/41 cm; #3/55 g/43 cm). Live specimens were collected at Shoals

Marine Laboratory (Appledore Island, ME) and transported to

Valdosta State University. Specimens were euthanized using 400 mg

of MS222 (Finquel anesthetic, Argent Chemicals, Redmond WA) and

200 mg of NaHCO

(pH buffer) mixed in 1 liter of filtered artificial

seawater. An incision was then made along the ventral midline to

collect tissue specimens for the histological analysis. Preserved

amphioxus and shark specimens were obtained from VWR (470001-

802 and 470001-486). Subsequent paraffin processing, embedding,

sectioning, and hematoxylin and eosin staining were performed by

standard procedures.

Human and

mouse cell cultures

Human myoblasts (hSkMC-AB1190) were isolated and immortal-

ized as previously published (

). These cells were cultured in 15%

fetal bovine serum (FBS) (GemCell, 100-500) and 5% growth medium

supplement mix (PromoCell, C-39365) in skeletal muscle cell basal

medium (PromoCell, C-23260) with GlutaMAX and 1% gentamicin

sulfate. Mouse 10T1/2 fibroblasts [American Type Culture Collec-

tion (ATCC), CCL-226] and C2C12 myoblasts (ATCC, CRL-1772)

were maintained in 10% FBS with 1% penicillin/streptomycin (Gibco,

15140122) in DMEM (Dulbecco’s modified Eagle’s medium–high

glucose, D5796). Myoblast differentiation medium contained 2%

horse serum in DMEM with 1% penicillin/streptomycin. Cells have

passed mycoplasma test using the Universal Mycoplasma Detection

Kit (ATCC, 30-1012 K).

Lizard cell culture and

CRISPR experiments

in lizard myoblasts

Myogenic single clones (myosin heavy chain+) were isolated from

immortalized

Anolis sagrei

embryonic cells ASEC-1 (

A. sagrei

em-

bryonic cell line 1; to be described in detail elsewhere). ASEC-1 and

clonally derived myoblasts were cultured in DMEM supplemented

with glutamine and 10% FBS (with penicillin/streptomycin and

amphotericin B) and cultured at 29°C and 5% CO

Downloaded from https://www.science.org at California Institute of Technology on May 08, 2023