Mutations in Caenorhabditis elegans actin, which are equivalent to human cardiomyopathy mutations, cause abnormal actin aggregation in nematode striated muscle

Actin is a central component of muscle contractile apparatuses, and a number of actin mutations cause diseases in skeletal, cardiac, and smooth muscles. However, many pathogenic actin mutations have not been characterized at cell biological and physiological levels. In this study, we tested whether the nematode Caenorhabditis elegans could be used to characterize properties of actin mutants in muscle cells in vivo. Two representative actin mutations, E99K and P164A, which cause hypertrophic cardiomyopathy in humans, are introduced in a muscle-specific C. elegans actin ACT-4 as E100K and P165A, respectively. When green fluorescent protein-tagged wild-type ACT-4 (GFP-ACT-4), is transgenically expressed in muscle at low levels as compared with endogenous actin, it is incorporated into sarcomeres without disturbing normal structures. GFP-ACT-4 variants with E100K and P165A are incorporated into sarcomeres, but also accumulated in abnormal aggregates, which have not been reported for equivalent actin mutations in previous studies. Muscle contractility, as determined by worm motility, is not apparently affected by expression of ACT-4 mutants. Our results suggest that C. elegans muscle is a useful model system to characterize abnormalities caused by actin mutations.


Introduction
Actin is an essential component of the cytoskeleton in both muscle and non-muscle cells. A number of mutations in the six human actin genes cause a wide range of diseases in various tissues (Despond & Dawson, 2018;North & Laing, 2008;Rubenstein & Wen, 2014). In muscles, actin, together with myosin, generates contractile forces, and therefore, alterations in contractile and/or structural properties of actin can cause muscle malfunction. Mutations in skeletal muscle α-actin (ACTA1) cause congenital myopathies, including nemaline myopathy and intranuclear rod myopathy, in which skeletal muscle exhibits abnormal accumulations of sarcomeric components (Clarkson et al., 2004;Laing et al., 2009;North & Laing, 2008;Ono, 2010). Many of these cytoskeletal abnormalities can be reproduced by expression of mutant actins in cultured non-muscle or muscle cells (Bathe et al., 2007;Costa et al., 2004;Domazetovska et al., 2007;Vandamme et al., 2009a;Vandamme et al., 2009b) or in transgenic mice (Lindqvist et al., 2013;Ravenscroft et al., 2011). By contrast, mutations in cardiac α-actin (ACTC1) cause hypertrophic and dilated cardiomyopathies (Mogensen et al., 1999;Olson et al., 1998). Biochemical studies indicate that these cardiomyopathy mutations of actin alter its properties to generate contractile forces (Despond & Dawson, 2018). However, abnormalities in sarcomeric or cytoskeletal structures have not been reported when the mutant actins are expressed in cultured cells (Muller et al., 2012;Vang et al., 2005) or transgenic mice (Song et al., 2010;Song et al., 2011).
In this study, we used the nematode Caenorhabditis elegans as a model to examine effects of cardiomyopathy mutations in actin. The body wall muscle of C. elegans is obliquely striated muscle with a number of functional and structural similarities to vertebrate striated muscles (Ono, 2014). Four actin genes are expressed in C. elegans muscle (Files et al., 1983;Stone & Shaw, 1993), and they are 95% identical to human cardiac and skeletal muscle α-actins (Ono & Pruyne, 2012). Since all known residues that are mutated in human cardiomyopathies are conserved in C. elegans actins, we selected two representative hypertrophic cardiomyopathy mutations and tested whether these pathogenic mutations perturb the properties of actin in C. elegans muscle in vivo. We found that the mutant actins were incorporated into sarcomeres and also accumulated in abnormal aggregates, suggesting that C. elegans muscle is a unique model system to characterize pathogenic actin mutations.

Worm culture
Worms were cultured following standard methods (Stiernagle, 2006). Wild-type C. elegans strain N2 was obtained from the Caenorhabditis Genetics Center (Minneapolis, MN) and used in this study.

Transgenic strains
An expression vector for GFP-ACT-4(wild-type: WT) was constructed by inserting ACT-4 cDNA at the EcoRI-NheI sites of pPD118.20 (provided by Andrew Fire, Stanford University) in-frame with the 3'-end of the GFP coding sequence. Briefly, first-strand cDNAs were reverse-transcribed from total RNAs from the N2 strain using oligo-dT by a Maxima H -First Strand cDNA Synthesis Kit (Thermo Fisher Scientific). The ACT-4 cDNA with added EcoRI and NheI sites in the primer sequences was amplified from the pool of cDNAs by polymerase chain reaction using Pfu DNA polymerase (Agilent Technologies), digested with EcoRI and NheI, and ligated with pPD118.20 that had been cut with EcoRI and NheI. Expression vectors for GFP-ACT-4(E100K) and GFP-ACT-4(P165A) were generated by site-directed mutagenesis using a QuickChange Site-directed Mutagenesis Kit (Agilenet Technologies). Sequences of the inserts were verified by DNA sequencing. Transgenic nematodes were generated by microinjection of DNA vectors into the distal gonads as described previously (Mohri et al., 2006). Transgenic worms were selected by expression of GFP as observed by fluorescence microscopy, and the transgenes were maintained as extrachromosomal arrays. Strains used in this study are ON16, ktEx6[Pmyo-3::GFP::ACT-4(WT)]; ON209, ktEx154[Pmyo-3:: GFP::ACT-4(E100K)]; and ON212, ktEx157[Pmyo-3::GFP:: ].
Worm motility assay Worm motility was determined by counting swinging motions of worms for 30 seconds in M9 buffer as described (Epstein & Thomson, 1974;Ono et al., 1999).

Fluorescence microscopy
Fixation and staining of worms with rhodamine-phalloidin were performed as described previously (Ono, 2001). GFP was observed by its own fluorescence. Specimens were observed by epifluorescence using a Nikon Eclipse TE2000 inverted microscope with a CFI Plan Fluor ELWD 40x (Dry; NA 0.60) objective. Images were captured by a SPOT RT monochrome CCD camera (Diagnostic Instruments) and processed by IPLab 4.0 imaging software (BD Biosciences) and Adobe Photoshop CS3. Figure 1A were generated using PyMol 2.1.0 (Schrödinger), and texts added using Adobe Photoshop CS3.

Statistical analysis
The data used in Figure 1C were analyzed by Student's t-test using SigmaPlot 14.0 (Systat Software, Inc.). The data used in Figure 1G were analyzed by one-way ANOVA with Turkey test using SigmaPlot 14.0. The data used in Figure 1H were analyzed by one-way ANOVA with pairwise multiple comparison using the Student-Newman-Keuls method using SigmaPlot 14.0.

Results
We constructed an expression vector for GFP-tagged ACT-4, an actin isoform that is expressed in the body wall muscle (Stone & Shaw, 1993), under the control of the myo-3 promotor (Pmyo-3) (Okkema et al., 1993). The ACT-4 sequence was fused to the C-terminus of GFP with a 9-residue linker sequence (SPQALEFSS) to minimize the interference of actin function by GFP (Aizawa et al., 1997). We selected two missense mutations, E99K and P164A in human cardiac α-actin (Despond & Dawson, 2018;Olson et al., 2000), that dominantly cause hypertrophic cardiomyopathy. The E99K mutation weakens actinmyosin interaction (Bookwalter & Trybus, 2006) and increases the critical concentration of actin (Mundia et al., 2012). In a transgenic mouse model, E99K increases calcium sensitivity of the thin filaments and causes abnormal heart functions (Song et al., 2011). In contrast, the effect of P164A mutation remains unclear. Although P164A causes alteration in protein folding in vitro (Vang et al., 2005), an equivalent mutation in yeast actin does not change its basic biochemical properties (Wong et al., 2001). C. elegans ACT-4 is 95% identical in amino acid sequence to human cardiac α-actin, and E99 and P164 are conserved as E100 and P165, respectively ( Figure 1A). Therefore, we introduced E100K and P165A mutations in GFP-ACT-4 and examined their effects on the sarcomeric structures in C. elegans body wall muscle.

Establishment of transgenic strains and expression quantification
We established at least three independent transgenic strains for each of the transgenes, GFP-ACT-4(wild-type: WT), GFP-ACT-4(E100K), and GFP-ACT-4(P165A), and examined expression levels of GFP-ACT-4 variants by western blot. We selected one strain each, which expressed the GFP-ACT-4 variants at similar levels ( Figure 1B, C) for further analysis. Western blot analysis using anti-actin antibody showed that all the GFP-ACT-4 variants were expressed at much lower levels than endogenous actin in total worm lysates ( Figure 1B). The level of GFP-ACT-4(WT) was roughly estimated by densitometry to be lower than 10% of that of total endogenous actin, although strong saturated signals for endogenous actin made precise quantification difficult. Considering that body wall muscle is the major tissue expressing actin as a sarcomeric component, the expression level of GFP-ACT-4 should be still much less than that of endogenous actin within the body wall muscle cells. Raw uncropped western blots, alongside all other raw data, are available on Figshare (Hayashi et al., 2019).

Subcellular localization of GFP-ACT-4 variants and motility of worms
GFP-ACT-4(WT) was incorporated into sarcomeres in body wall muscle cells ( Figure 1D). Staining of F-actin in fixed animals with rhodamine-phalloidin showed a nearly identical localization pattern to GFP-ACT-4(WT) ( Figure 1D). Motility of the worms expressing GFP-ACT-4(WT) (81.5 ± 7.5 beats/30 sec, n = 20), as determined by beating frequency in liquid, was slightly slower than that of wild-type worms with no transgene (94.8 ± 11 beats/30 sec, n = 20), suggesting that GFP-ACT-4 (WT) has a weak negative effect on contractility of the body wall muscle.
Both GFP-ACT-4(E100K) and GFP-ACT-4(P165A) were incorporated into sarcomeres but also formed spherical aggregates in the cytoplasm of the body wall muscle cells ( Figure 1E, F). Staining with rhodamine-phalloidin showed that sarcomeric organization of actin filaments were somewhat disorganized by expression of GFP-ACT-4(E100K) ( Figure 1E) but not GFP-ACT-4(P165A) ( Figure 1F). However, motility of the worms expressing GFP-ACT-4(E100K) or GFP-ACT-4(P165A) was not significantly different from that of wild-type worms ( Figure 1G), suggesting that these actin mutants did not disturb muscle contractility. These aggregates resemble F-actin aggresomes induced by inhibitors of actin dynamics (Lázaro-Diéguez et al., 2008). However, the aggregates of GFP-ACT-4(E100K) or GFP-ACT-4(P165A) were not recognized by rhodaminephalloidin, a specific probe for F-actin ( Figure 1E, F). In addition, we could not detect these aggregates by immunofluorescence using anti-actin monoclonal or polyclonal antibodies, even after attempts to expose antigens using guanidine hydrochloride (Peränen et al., 1993) or microwave (Shi et al., 1991), suggesting that the mutant forms of actin were present in an inclusion-body-like state and not readily accessible to the actin probes. Such aggregates were not detected in worms expressing GFP-ACT-4(WT) ( Figure 1D, H), while variable numbers (0 -36 per cell) of aggregates were found in worms expressing GFP-ACT-4(E100K) or GFP-ACT-4(P165A) ( Figure 1E, F). In randomly selected worms (n = 30), GFP-ACT-4(E100K) (median = 9.5 aggregates per cell) induced significantly more aggregates than GFP-ACT-4(P165A) (median = 6.0 aggregates per cell) ( Figure 1H). These aggregates were randomly located in the cytoplasm but not within the nucleus. Thus, we conclude that the missense mutations in ACT-4 induced the formation of abnormal cytoplasmic aggregates in muscle cells.

Discussion
Formation of actin aggregates by E99K (E100K in worm) or P164A (P165A in worm) mutation in actin has not been reported in human patients or other experimental systems. When cardiac α-actin mutants (E99K and P164A) are expressed in COS-7 cells, these actin mutants are not incorporated in the non-muscle actin cytoskeleton with no detectable aggregate formation (Vang et al., 2005). When E99K cardiac α-actin is . The results were analyzed by one-way ANOVA (n=30) and significant difference was found between the data for GFP-ACT-4(E100K) and GFP-ACT-4(P165A) (**p = 0.006). expressed in the mouse heart, the mutant actin is incorporated in the cardiac thin filaments and causes disarray of cardiomyocytes but with no detectable aggregate formation (Song et al., 2011). Thus, effects of these actin mutations appear to be dependent on cellular contexts. Formation of actin aggregates by these actin mutations might be specific to the nematode muscle. We also cannot exclude the possibility that the aggregate formation is artificially enhanced by the GFP tag. Nonetheless, we were able to detect actin aggregates because of the GFP tag and might not have been able to detect the aggregates if a fluorescent tag was absent. Abnormal protein aggregates have been reported in idiopathic dilated cardiomyopathy (Gianni et al., 2010;Subramanian et al., 2015)  This collection contains the following underlying data: • Uncropped western blots • Unprocessed microscopy images • Dataset 1-3 (containing western blot quantification, and raw data for worm motility and number of aggregates per cell) Data are available under the terms of the Creative Commons Zero "No rights reserved" data waiver (CC0 1.0 Public domain dedication).

Grant information
The C. elegans strain N2 was provided by the Caenorhabditis Genetics Center, which is funded by the National Institutes of Health Office of Research Infrastructure Programs (P40 OD010440). This work was supported by a grant from the National Institutes of Health (AR048615) to S.O.

The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Expression of WT GFP-ACT4 results in a negative effect on worm motility. This is mentioned by the authors in the results (Subcellular localization of GFP-ACT-4 variants and motility of worms). Such an observation is not unprecedented in mammalian cells and is basically due to the GFP-moiety possibly interfering with cellular functions. This was further investigated in Agbulut (2007 ) and it was shown et al. that GFP on its own interferes with myosin-based contractility which is potentially also the case here. In this respect should the authors not interpret the worm motility relative to the GFP-ACT-4 control rather than relative to WT worms? I realise it is more difficult to reconcile disturbed muscle structure (D-F) and the expected dominant-negative phenotype with increased worm motility (G) but the authors indicate the motility of the worms is significantly different between mutant and WT-GFP actin variants and this should be commented on in the results or the discussion section.
In the discussion the authors interpret their results solely in terms of cardiomyopathy based on the origin of the mutations. However the experimental system used is body wall muscle and interestingly the cellular phenotype is more reminiscent of nemaline myopathy (dotted patterns and irregular fibers). Functional differences between alpha-cardiac actin and alpha-skeletal muscle actin in human are unclear because the proteins are highly similar (only 4 amino acids different and these changes are even conservative) and differential phenotypes of disease mutants likely result from differential expression in the respective tissues (in agreement with the authors' statement: "Thus, effects of these actin mutations appear to be dependent on cellular contexts"). These ACTC1 mutants do not cause aggregates in COS cells (Vang et , 2005, mentioned in the discussion), therefore I suggest to also make the parallel with nemaline al. myopathy where dot-like patterns in cellular context of (fused) myoblasts are frequently documented.
The dots appear to be phalloidin-negative as judged by visual inspection (Figure 1D and 1E); perhaps this 1 The dots appear to be phalloidin-negative as judged by visual inspection (Figure 1D and 1E); perhaps this should be mentioned in the main text.

Other minor points:
Introduction: In the two sentences: "Many of these cytoskeletal abnormalities can be reproduced by expression of mutant actins in cultured non-muscle or muscle cells (Bathe ., 2007;Costa ., 2004;et al et al Domazetovska ., 2007;Vandamme ., 2009a;Vandamme ., 2009b) 1998)." a contrast is mentioned although the two items do not contrast each other. Methods: Correct: Agilent. Figure 1A: upon printing (and in the 100% view option of PDF) these figures are difficult to interpret. I suggest to make them larger by rearranging this figure (quite some white space left and right).

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? Partly
No competing interests were disclosed.

Competing Interests:
Reviewer Expertise: actin cytoskeleton I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above. The specific aim of this Research Note is to demonstrate that can be used as a useful C. elegans in vivo model to study mutations in actin associated with human myopathies, specifically in this case, cardiomyopathies. The authors have thus expressed only in the muscle tissue of C. elegans GFP-chimeras of wild type (WT) actin and two actin mutants (E100K and P165A) equivalent to human actin mutants E99K and P164A both of which are known to cause hypertrophic cardiomyopathy in humans. The authors then study by fluorescence microscopy the localization of the actin-GFP in the body wall muscle, the sarcomeric organization using fluorescent phalloidin to visualize F-actin, and worm motility. Their main findings are that all of the GFP-actin proteins incorporate into the sarcomere with only the E100K mutant causing sarcomeric disarray and that both mutant actins, but not WT, form GFP-positive aggregates that do not stain with fluorescent phalloidin or with an array of actin antibodies even after implementing antigen retrieval protocols. From that, the authors suggest the relevance of C.
to study actin-related cardiomyopathies. However, actin aggregates have not been detected in elegans mouse heart from animals expressing the E99K mutant nor have aggregates been reported in human cardiomyopathies from subjects with either of the mutations used here. Thus, we are left to wonder, as indeed the authors have themselves, if the aggregates are specific to nematode muscle, if aggregate formation by the mutant actins is enhanced by the presence of the GFP-tag, and what other molecules are found in the aggregates that might provide a clue to their composition.
As the authors have clearly written in their introduction, there is a need for further characterization of pathogenic mutations and expanding the available organisms to study specific diseases is always welcomed, as every model comes with their benefits and caveats. In spite of this, and the fact that the manuscript is well written and self-critical, there are some limitations in the studies presented that prevent us from approving its acceptance at this stage.
Although it is understandable why the authors choose to express the fluorescently tagged version of the actin mutants, expression of the untagged versions should also have been examined to see if the sarcomere disruption by the E100K occurs in the absence of the GFP tag. It is recognized that under these conditions it would not be possible to obtain ratios of the expressed transgene with the endogenous protein, but with a fluorescent protein expressed from a different promoter, it should be possible to obtain worms with the untagged actin mutants to determine its effect on sarcomere disruption. As recognized by the authors, the GFP tag could increase the propensity of the construct to oligomerize/aggregate. Whether this could be driven by GFP aggregation on its own is not clear. It has been reported that GFP linked via its C-terminus to a 16 amino acid "degron" peptide is not effectively degraded but forms aggregates very similar to the ones observed here when expressed in the body wall muscle of C. elegans (Link , 2006 ). Since the aggregates observed in the current study have not been found to contain et al. actin, one has to wonder if some cleavage has occurred to give rise to the same type of GFP-aggregates observed by Link Why these do not form with the GFP-WT actin is unknown, but might reflect its et al. greater stability to proteolysis. It would be useful to know if the number of aggregates correlate with total GFP intensity in a muscle cell.
Also, the absolute level of the expressed proteins relative to endogenous actin is problematic. As recognized by the authors themselves, their western blot was made with whole organism lysate and the 1 1.

2.
3. recognized by the authors themselves, their western blot was made with whole organism lysate and the signal of endogenous actin was near saturation. Although the contribution of actin from regions outside the body wall muscle is thought by the author's to be low, if it were possible to isolate only the body wall muscle, the values would be more meaningful. For the results to mimic human disease pathology, the authors need to try and reproduce the ratio between mutant and endogenous actin. As the authors demonstrate decreased worm motility when overexpressing total WT GFP-actin, it is clear that the expression of exogenous actin must be kept low to avoid potential toxic effects, although in this case it is unclear if the decreased motility arises from the presence of the GFP-tag.
While trying to explain the difference observed in their worm model with results from a previous paper (Song 2011) in which no aggregates were found with the E99K mutant, the first explanation given is et al., "Thus, effects of these actin mutations appear to be dependent on cellular context." Such an explanation actually negates the main purpose of the paper. We would thus recommend to first try to resolve the issues that have been identified in this review, because, as interesting as the starting idea is, there are some flaws that render the results difficult to interpret.

Minor questions:
How do the authors explain the lack of correlation between sarcomeric disarray and worm motility?
Does the expression of any of the constructs lead to an increased mortality or to a reduced lifespan?
The nature of the aggregates that contain the GFP should be explored in more detail. Are they phase separated protein clusters such as one finds in stress granules and which can be identified by rapid diffusion within the structure if photobleached, or are these structured solid aggregates that might be able to be isolated, such as those from human from idiopathic dilated cardiomyopathy? The reason for their lack of actin staining needs to be identified since right now the title of the manuscript that states "human cardiomyopathy mutations, cause abnormal actin aggregation" cannot be used since no actin in the aggregates has been detected.

Is the study design appropriate and is the work technically sound? Yes
Are sufficient details of methods and analysis provided to allow replication by others? Yes

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? Yes
No competing interests were disclosed.

I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.
The benefits of publishing with F1000Research: Your article is published within days, with no editorial bias You can publish traditional articles, null/negative results, case reports, data notes and more The peer review process is transparent and collaborative Your article is indexed in PubMed after passing peer review Dedicated customer support at every stage For pre-submission enquiries, contact research@f1000.com