The growth factor BMP11 is required for the development and evolution of a male exaggerated weapon and its associated fighting behavior in a water strider.

Exaggerated sexually selected traits, often carried by males, are characterized by the evolution of hyperallometry, resulting in their disproportionate growth relative to the rest of the body among individuals of the same population. While the evolution of allometry has attracted much attention for centuries, our understanding of the developmental genetic mechanisms underlying its emergence remains fragmented. Here we conduct comparative transcriptomics of the legs followed by an RNA interference (RNAi) screen to identify genes that play a role in the hyperallometric growth of the third legs in the males of the water strider Microvelia longipes. We demonstrate that a broadly expressed growth factor, Bone Morphogenetic Protein 11 (BMP11, also known as Growth Differentiation Factor 11), regulates leg allometries through increasing the allometric slope and mean body size in males. In contrast, BMP11 RNAi reduced mean body size but did not affect slope either in the females of M. longipes or in the males and females of other closely related Microvelia species. Furthermore, our data show that a tissue-specific factor, Ultrabithorax (Ubx), increases intercept without affecting mean body size. This indicates a genetic correlation between mean body size and variation in allometric slope, but not intercept. Strikingly, males treated with BMP11 RNAi exhibited a severe reduction in fighting frequency compared to both controls and Ubx RNAi-treated males. Therefore, male body size, the exaggerated weapon, and the intense fighting behavior associated with it are genetically correlated in M. longipes. Our results support a possible role of pleiotropy in the evolution of allometric slope.

Introduction

Extravagant ornaments and weapons, often carried by males in a wide range of lineages, represent some of the most striking outcomes of directional sexual selection driven by female choice or male competition [1-4]. Exaggerated sexually selected traits are often characterized by their extreme growth and hypervariability among individuals of the same population [5,6]. These features derive from changes in scaling relationships, especially the elevation of the allometric coefficient (or allometric slope), which results in certain structures growing disproportionately larger relative to the rest of the body among individuals of the same population [5-8]. Variation in the size of morphological characters (Y) is often correlated with variation in body size (X) through a scaling relationship that follows a power law distribution such as Y = aX [9,10]. When log transformed, this allometric equation becomes linear with log(a) representing the intercept, the relative size of trait Y, and b the slope, representing the proportional growth of trait Y relative to trait X. When b = 1, the 2 traits grow proportionally to one another, and when b is different from 1, there is a disproportionate growth between the 2 traits (hypoallometry with b < 1 or hyperallometry with b > 1). Despite the longstanding interest in the study of allometry, the developmental genetic mechanisms underlying covariation in growth patterns between traits and body among individuals of the same population are unclear [6,11-14].

Allometric slope is known to be relatively stable during evolution, and several studies suggest that this stasis could be the consequence of genetic correlation with other traits [9-13,15]. In the context of trait exaggeration (e.g., hyperallometry), Fisher predicted a genetic correlation between male exaggerated ornaments and female preference for the exaggeration—a process known as runaway selection [1]. Other studies, for example, in the ruffs or horned beetles, also suggest a genetic link between the elaboration of male exaggerated traits and other secondary sexual traits such as male body size and reproductive behavior [16-18]. However, empirical tests of these predictions remain difficult to achieve. Determining the genes that functionally influence slope and testing their correlation with other traits would therefore greatly advance our understanding of the evolution of scaling relationships [6,11-14].

We address these questions in the water strider Microvelia longipes, where some males exhibit extremely long third legs due to a hyperallometric relationship with body size [19]. Among the over 200 known species in this genus, M. longipes is the only species found where males simultaneously exhibit large body size and high variation in both body and leg length [19,20]. Males of other species that exhibit either large body, such as Microvelia sp. (see below), or high variation in body size, such as Microvelia pulchella, a sister species to M. longipes, do not display exaggerated leg length, suggesting that the evolution of hyperallometry may be linked to the evolution of increased mean and variation in body size [19]. In addition, males of M. longipes use the exaggerated legs to fight and dominate egg-laying sites where they intercept and mate with gravid females [19]. Compared to M. pulchella, M. longipes males fight over an order of magnitude more frequently, indicating that the evolution of the exaggerated weapon is also associated with the increase in competition intensity between males [19]. Here we use comparative transcriptomics combined with RNA interference (RNAi) as a tool to study gene function during development and determine the genes regulating the scaling relationship between the exaggerated legs and the body in M. longipes and 2 additional species. We found that the developmental gene Bone Morphogenetic Protein 11 (BMP11), also known as Growth Differentiation Factor 11 (GDF11), evolved a multifaceted role modulating the hyperallometry of the exaggerated legs, increased body size, and fighting intensity in males, thus revealing a genetic correlation between these traits in the male of M. longipes.

Results

Growth and gene expression differences in the legs during M. longipes nymphal development

In M. longipes, third leg length is variable among males and this variability is higher in males compared to females (Fig 1A and 1C) [19]. A developmental growth curve (Fig 1B; S1 Fig) revealed that leg length is similar between individuals at the first to third instars but starts to diverge between the sexes and also among males at the fourth instar (Fig 1B). The length of the third legs continues to increase between fifth instar and adult males in a higher rate when compared to females’ legs or to the 2 other male legs during this last developmental stage (Fig 1B; S1 Table). These differences in leg length between individuals are independent of any variation in developmental time from first instar to adult (ANCOVA with body length as covariate, F-value = 0.156, p-value = 0.69; S2 Fig). Therefore, the growth burst and its variation at the end of postembryonic development in male exaggerated legs accounts for both the sexual dimorphism and the hypervariability among males (Fig 1A–1C). Ontogenetic allometry (i.e., scaling relationship between developmental stages of an individual) also confirmed the exaggerated growth rate of male third legs during development (S2 Table). The first and second legs in males and females also grew faster at the final instars, although their growth rate was not as extreme as the third legs (Fig 1B). Based on these observations, we hypothesize that the striking variation in leg growth dynamics within and between the sexes is attributable to variation in gene expression at the end of nymphal development.

Fig 1

Growth dynamics of male exaggerated third legs in M. longipes.

(A) Adult large and small males and a female showing final growth phenotypes. (B) Leg length variation during postembryonic development (nymphal development) in males and females. Each line represents leg growth dynamics of a single individual during nymphal development until adulthood. (C) Static allometries between body length and the 3 pairs of legs in adult males and females. Power law regressions were fitted to the raw data to represent the static allometric equation y = bxa. Differences in static allometry parameters were tested on log-transformed data in [19]. The data underlying this figure may be found in S1 and S2 Tables.

Comparative transcriptomics of the legs at the fifth nymphal instar, the developmental stage where we detected a burst of growth, first revealed a global difference in gene expression profiles between the 3 pairs of legs in both sexes at this developmental window (S3A Fig). When comparing the first versus third legs, we found that the third legs diverged more from the first legs in males than they did in females, both in terms of number of differentially expressed genes and in the degree of differential expression (S3 and S4 Figs). Overall, these data indicate that leg exaggeration in M. longipes males is associated with a higher number and higher degree of differential expression of leg-biased genes in males compared to females. We used these datasets to select and test the role of candidate genes in leg exaggeration.

Key developmental genes control the scaling relationship of male exaggerated legs

Extravagant signals and weapons, such as M. longipes third legs, are predicted to occur only in good condition individuals because of the high fitness cost they impose [4,21,22]. Proposed mechanisms for the development of exaggerated phenotypes include increased sensitivity or local production of growth factors by the exaggerated organ [4,23]. To test this hypothesis, we examined the developmental genetic pathways that are enriched in the exaggerated legs through analyzing gene ontology of leg-biased genes (S3 Fig; S3 Table). Genes that are up-regulated in the third legs of males (in comparison with the first legs of males) were enriched in metabolic processes, transmembrane transport, and growth signaling pathways. Genes that are up-regulated in the third legs of females (in comparison with first legs of females) were enriched in signaling and metabolic pathways as well as response to stimuli and cell communication processes. Among the genes and pathways that were hypothesized to play a role in the regulation of growth-related exaggerated traits [3], many were identified in our datasets as differentially expressed between legs within the same sex or between the sexes (S4 Table; see [24] for sex-biased genes).

To test the role of these genes in trait exaggeration, we conducted an RNAi screen targeting about 30 candidates representing transcripts that were either leg-biased, sex-biased (up-regulated in male’s or female’s third legs), broadly expressed, or tissue-specific (S5 Table). This functional screen identified 2 genes, BMP11 (also known as GDF11) and Ultrabithorax (Ubx), the depletion of which unambiguously and reproducibly resulted in altered scaling relationships in the males. Some of the other genes tested were lethal or did not produce any detectable effect, thus preventing conclusive analyses of scaling relationships (S5 Table). Ubx is a tissue-specific Hox protein known to be confined to the second and third thoracic segments in water striders [25-28]. In our comparative transcriptomics dataset, Ubx is absent from the first legs, lowly expressed in the second legs, and highly expressed in the third legs of both sexes (S5A Fig), thus confirming its tissue-specific expression in M. longipes.

BMP11 is a known secreted growth regulator with a systemic effect in vertebrates [29,30]; however, its expression in waters striders is unknown. A phylogenetic analysis of the BMP gene family clustered this sequence from many water striders with BMP11 of humans and mice, thus confirming their homology (S6 Fig; S1 and S2 Data). Our comparative transcriptomics analysis showed that BMP11 is expressed in all legs and its levels of expression matched the differences in leg length along the body axis of both males and females (S5B Fig). A quantitative reverse transcription polymerase chain reaction analysis confirmed that BMP11 is up-regulated in the third legs compared to the first legs of both males and females (two-way ANOVA; p < 0.0001; S5C Fig). Unlike comparative transcriptomics, however, qPCR revealed that BMP11 is significantly up-regulated in the third legs of males compared to females (two-way ANOVA; p < 0.0001; S5C Fig). In addition, qPCR analysis revealed that BMP11 is expressed in the body of both sexes but with significantly higher levels in the females (two-way ANOVA; p < 0.0005; S5C Fig). This analysis confirms that BMP11 is broadly expressed in M. longipes.

Ubx RNAi knockdown during nymphal development resulted in a 16% average reduction in males’ third leg length (t-test: t = −3.5135, df = 62.904, p-value = 0.0008; S6 Table). However, this effect of Ubx RNAi was uniform across the range of male body size variation such that the intercept was shifted but the slope remained unaffected (Fig 2C; S6 Table for statistics). We also detected a small but significant effect in the second legs, and no effect in the first legs, consistent with the expression of Ubx in these tissues (Fig 2A and 2B; S6 Table). We did not detect any effect on average body length in males resulting from Ubx knockdown (S6 Table). Ubx knockdown in females also altered the intercept for the third legs but to a lower extent than in males despite the similar expression of Ubx in females (Fig 2C; S6 Table). This result first demonstrates that, in M. longipes, changes in leg length can be genetically decoupled from changes in body length, consistent with previous findings for other exaggerated sexually selected traits [4]. Second, this tissue-specific gene is associated with a change in intercept (i.e., increases the relative size of the third and second legs in a proportionate manner across individuals) but not in slope.

Fig 2

Effect of Ubx RNAi on scaling relationships between leg length and body length in M. longipes.

(A) Scaling relationships are not altered in males or females by Ubx RNAi. (B) There is a subtle decrease in the intercept of second legs upon Ubx RNAi. (C) Ubx RNAi causes a significant decrease in intercept of the third legs in males. Note that body size is not affected by Ubx RNAi. Linear regressions were fitted to the log-transformed datasets. Statistical analyses were also performed on log-transformed data and can be found in S6 Table. Sample sizes are as follows: control males n = 33; control females n = 44; Ubx RNAi males n = 32; Ubx RNAi females n = 18. The data underlying this figure may be found in S6 and S7 Tables. Ctl, xxxx; RNAi, RNA interference; Ubx, Ultrabithorax.

RNAi knockdown of BMP11 during nymphal development affected the growth of multiple traits, consistent with the broad expression of this gene (Fig 3). M. longipes females have higher expression of BMP11 in the body (S5C Fig) and larger mean body length (S6 Table). Depletion of BMP11 RNA decreased the growth of the pronotum and significantly reduced mean body length by 8% in males and 12% in females (Fig 3A, 3B, 3G and 3H; S6 Table). Importantly, RNAi against BMP11 resulted in a significantly reduced slope of male third legs from 3.1 to 2.0 (ANOVA, F-value = 9.6327, p-value = 0.002468) (Fig 3M; S6 Table). The slope for the second legs of males, which are also exaggerated, was significantly lower in individuals treated with BMP11 RNAi compared to the controls (S7 Fig; S6 Table). Because small males generated through artificial selection or through poor diet treatment maintain the same allometric coefficient as large males [19], we conclude that BMP11 is necessary to increase the allometric coefficient of the third and second legs as well as mean body size in the males of M. longipes. Just like in males, BMP11 knockdown reduced mean body size in females (Fig 3M; S6 Table). Unlike males, however, BMP11 knockdown in females did not affect either slope or intercept, despite the up-regulation of BMP11 in females’ third legs compared to the 2 other legs (Fig 3M; S5 Fig; S6 Table). Altogether, these results indicate that BMP11 is a key growth regulator in M. longipes, playing a role in leg allometry (i.e., increases the relative size of the third and second legs in a disproportionate manner relative to the body) in males and mean body size in both males and females during development.

Fig 3

Effect of BMP11 RNAi on the growth and scaling relationships of various traits in 3 species of Microvelia.

The pronotum in (A) M. longipes and (C) M. sp., but not in (E) M. pulchella, covers the entire mesonotum (brackets in A and C). BMP11 RNAi characteristically reduces the size of the pronotum in M. longipes and M. sp. (brackets in B and D). (E, F) This effect is absent in M. pulchella, consistent with the absence of this trait. (G–L) Only M. sp. males have a set of spines in the third legs (arrowheads in I), which are removed or reduced due to BMP11 RNAi (arrowheads in J). The lengths of the third legs are reduced in adult males and females. Statistical analyses can be found in S6 Table. M. pul stands for M. pulchella. (M–O) Effect of BMP11 RNAi on adult male and female scaling relationships between leg length and body length of M. longipes (M), M. sp. (N), and M. pulchella (O). Statistical analyses can be found in S6 Table. Comparisons of adult static allometries for each species between controls and BMP11 RNAi for males and females are shown in the keys below graphs. Linear regressions were fitted to the log-transformed datasets. Samples sizes for each species are indicated in the panels. Statistical analyses were also performed on log-transformed data and can be found in S6 Table. The data underlying this figure may be found in S6 and S7 Tables. BMP11, Bone Morphogenetic Protein 11; RNAi, RNA interference.

Evolution of BMP11’s role in regulating scaling relationships across Microvelia

Trait exaggeration, whereby second and third legs grow in disproportion to the growth of the body in males, is a derived state that separates M. longipes from other Microvelia spp. [19,31]. We therefore sought to determine whether the role of BMP11 in regulating scaling relationships is also a derived state in the lineage leading to M. longipes. To test this hypothesis, we examined the effect of BMP11 RNAi on 2 additional sister species, namely, M. pulchella, a sister species of M. longipes and M. sp. (a species in the process of being described and for which we have not chosen a name yet) which is sister to M. longipes/M. pulchella [19]. Just like in M. longipes, BMP11 RNAi also reduced the growth of the pronotum in both sexes of M. sp. (Fig 3 compare C to D; S6 Table). In M. pulchella, however, the pronotum does not cover the mesonotum, and BMP11 RNAi did not cause any detectable reduction in the growth of this trait (Fig 3E and 3F). In addition, M. sp. is the only species in this sample where males’ third legs possess a prominent set of spines that are not found in the females, indicative of sexually antagonistic selection [32,33] (Fig 3I). Surprisingly, BMP11 RNAi reduced or removed these spines (Fig 3J), indicating that this gene could also play a role in the development of discrete sexually dimorphic traits. Analyses of scaling relationships revealed that BMP11 RNAi did not affect slope either in M. pulchella or M. sp. (Fig 3N and 3O; S7 Table; S7C, S7D, S7E, and S7F Fig). However, BMP11 RNAi significantly reduced mean body size in both sexes of M. sp. and M. pulchella (Fig 3N and 3O; S6 Table). The size of the body and that of all appendages was reduced by BMP11 RNAi in comparable proportions between males and females (S6 Table). We conclude that the role of BMP11 in increasing mean body size is ancestral and that its role in increasing allometric coefficient is derived in M. longipes. In addition to its role in regulating quantitative sexual characters such as hyperallometric scaling relationships, BMP11 also plays a role in the development of discrete male-specific traits such as the spines found on the third legs of M. sp.

Link between BMP11 and fighting behavior in M. longipes males

We have shown that the evolution of exaggerated third legs, which are used as weapons by M. longipes males, was accompanied with a drastic increase in male fighting frequency to over an order of magnitude more compared to its sister species M. pulchella [19]. We have already shown that leg allometry and body size were genetically correlated through the role of BMP11 during development. To test whether the evolutionary concurrence between the exaggerated weapon and the increased male fighting behavior in M. longipes could also be the result of genetic correlation, we compared the frequency of fights between controls and individuals treated with BMP11 RNAi in male–male competition assays (see Material and methods; S1 and S2 Videos). Similar to the controls, BMP11 RNAi-treated males stood on egg-laying sites and signaled through ripples to attract females (S1 and S2 Videos; [19]). However, unlike the controls, these BMP11 RNAi-treated males avoided fighting when other males approached egg-laying sites. Instead, all these males gathered around the female and tried to copulate instead of engaging in fights to chase rival males away (S2 Video). Sometimes, the males abandoned the site altogether without a fight. We quantified this “docile” behavior by calculating the frequency of fights on egg-laying sites and found that BMP11 RNAi-treated males fought on average 12 times less than control males (Fig 4A; GLM: F-value = 29.96, df = 6, p-value = 0.0028). BMP11 RNAi-treated females were attracted by male’s signals on floaters, but we failed to observe any mating events during our trials (S2 Video). By contrast to BMP11 RNAi- treated males, Ubx RNAi-treated males retained aggressive fighting behavior (Fig 4A), although the fight between 2 rival males lasted significantly longer compared to controls (Fig 4B; S3 Video). These results suggest that the disruption of scaling relationships of the exaggerated weapon, through RNAi, also results in the disruption of the fighting behavior associated with it. Therefore, BMP11, but not Ubx, is necessary to increase fighting frequency in M. longipes males. Collectively, our results show that in M. longipes males, BMP11 regulates the development of the exaggerated weapon and the associated fighting behavior.

Fig 4

Effects of BMP11 and Ubx RNAi on male fighting behavior.

(A) Effect of BMP11 and Ubx knockdowns on male fighting frequency (number of fights per unit of time) compared to control individuals. (B) Effect of BMP11 and Ubx knockdowns on male fighting duration compared to control individuals. The data underlying this figure may be found in S8 and S9 Tables. BMP11, Bone Morphogenetic Protein 11; RNAi, RNA interference; Ubx, Ultrabithorax.

Discussion

We have shown that leg exaggeration in M. longipes males is associated with a specific signature of gene expression. Among the genes up-regulated in the exaggerated trait, we demonstrated that a broadly expressed growth factor, BMP11, increases allometric slope, whereas a tissue-specific factor, Ubx, increases allometric intercept. Finally, we have shown that BMP11 regulates the growth of the body, the disproportionate growth of the legs, and also increases aggressiveness in males. This indicates that these 3 male traits are genetically correlated through the pleiotropic effect of BMP11. How BMP11 changes fighting behavior in M. longipes males is unknown. A possible explanation is that the docile behavior of BMP11 RNAi-treated males can be due to an adaptive response, whereby males recognize their small body size and change reproductive strategy. Another possible explanation could involve a role of BMP11 in the development of nervous system components associated with the fighting behavior.

Exaggerated traits represent striking cases of differential growth between various organs, and we are just beginning to understand the underlying developmental genetic processes [23,34]. Studies in various insects suggest that traits would achieve exaggeration via their increased sensitivity to systemic growth factors such as Juvenile Hormone or Insulin [4,7,8,35,36]. Our data also show that exaggeration, as seen in the second and third legs of M. longipes males, can be induced by growth regulating molecules other than hormones, such is the case for BMP11. These molecules are produced in excess by the exaggerated tissue, indicating an alternative path to trait exaggeration where overproduction of the growth factor occurs locally rather than through heightened sensitivity to a systemic factor. Whether and how this increased tissue-specific expression of growth factors is connected to systemic growth pathways remains to be tested [8].

Our experiments describe a common effect of BMP11 on mean body length in both sexes. By contrast to males, the effect of BMP11 on body length is decoupled from its effect on leg length in females. It is possible that increased body size has been favored in both sexes, through increased competitiveness in males and increased fertility in females [37,38]. Another explanation could be that smaller females were disfavored due to reproductive incompatibility as body size increased in males. On the other hand, the lack of correlation in leg length between the sexes may be a consequence of sexual conflict resolution through dimorphism [32].

We have previously shown that closely related members of the Microvelia genus differ in their slope, intercept, mean body length, and the extent of variation of leg length and body length [19]. Among the over 200 known Microvelia, M. longipes is the only species that combines large mean body length and high variation in both body and leg length. This suggests that the size of the legs and the size of the body can evolve the ability to grow independently or in tight scaling relationship. Our data show that tissue-specific developmental regulators, such as Ubx, regulate trait growth independently of body growth, resulting in a change in intercept. BMP11 on the other hand, which is expressed in the legs and in the body, is necessary for systemic growth. Interestingly, our RNAi experiments show that the role of BMP11 in regulating the growth of the body is a conserved feature in this sample of species and this role does not seem to be sex-specific. However, the role of BMP11 in increasing slope is a derived state in M. longipes. This indicates that the input provided by developmental genes, likely in the context of their interactions within genetic networks, can evolve and result in various phenotypic outcomes, from discrete sex-specific phenotypes (as is the case of male spines in Microvelia. sp.) to quantitative variable traits such as body size and leg length.

Studies of scaling relationships have reported a strong stasis of the allometric coefficient at the microevolutionary time scale [11-13]. Even at the macroscale, slope is known to evolve slower compared to other scaling parameters such as intercept [11,12]. Several hypotheses were formulated to explain such evolutionary stasis, including a lack of genetic variation, strong stabilizing selection acting on slope, or a constraining effect of pleiotropy [11-14]. Our results, implicating BMP11 in regulating growth of body size and all 3 legs, lend support to pleiotropy as a major driver of evolutionary stasis in allometric coefficient. In contrast, Ubx, which regulates the size of the third legs but not body size, has a role in changing the intercept. This suggests that the genetic architecture of slope may have more pleiotropic effects than that of intercept, possibly explaining the observed differences in evolutionary stasis between these 2 scaling parameters. Whether this conclusion can be generalized to other traits requires additional comparative studies between species that have evolved different scaling relationships.

An important question is how trait exaggeration evolves despite the evolutionary stasis of allometric slope. We have shown that the evolution of hyperallometry in M. longipes is associated with increased mean body size and male aggressiveness compared to its sister species M. pulchella [19]. These changes are also associated with strong selection on males due to intense competition for access to females [19]. In this context, selection is expected to favor traits that increase male competitiveness. Strikingly, in M. longipes, BMP11 regulates mean body size, allometric slope, and male aggressiveness, all of which increase male competitiveness and likely fitness. It is possible that directional selection favored genotypes that regulate all 3 traits at the same time leading to the correlated evolution of these male traits in M. longipes. Pleiotropy would therefore become both a promoting factor for the evolution of exaggerated weapons and a constraining factor involved in the stasis of slope.

Material and methods

Establishment of growth curves

M. longipes is a hemimetabolous insect that develops through successive molts which represent exuvia of the previous nymphal instar. Using a population collected in Crique Patate near Cayenne in French Guyana, we raised each individual separately from first nymphal instar to adulthood. Throughout this process, each individual produced 5 exuvia corresponding to the 5 nymphal instars. We then collected the exuvia and the adult from each individual and measured their leg lengths (S1 Fig).

Comparative transcriptomics: Experimental design and Principal Component Analyses (PCA)

Comparative transcriptomics was performed on RNA extracted from individual legs belonging to 2 inbred populations, raised through over 20 generations of brother–sister crosses in laboratory conditions. During inbreeding, each of these 2 populations was selected either for long-legged (long-legged selected line) or short-legged males (short-legged selected line), respectively [19]. The sampling was as follows: 2 lines, 2 sexes, the 3 pairs of legs each replicated 3 times. Each replicate represents a pool of 20 individuals, randomly selected at the fifth nymphal instar. The raw Illumina reads were mapped to the genome of M. longipes [24].

We performed an analysis of gene expression variance of all genes expressed in the 3 pairs of legs. For this we selected genes with FPKM > 2 in at least 1 of the 3 legs. The filtering process and PCA analysis were performed in males and females separately (S3 and S4 Figs). For the PCA, we also corrected both for line and replicate effects that are intrinsic to our experimental design [24]. The latter effect matched the days where RNA was extracted from each sample. We corrected for both effects, using Within Class Analysis, with the R package ‘ade4’ version 1.7–15 [39].

Identification of leg-biased genes

The number of reads per “gene” was used to identify differences in expression among the different legs using DESeq2, with the R package DESeq version 1.39.0 [40]. To identify the genes underlying the growth differences in the pairs of legs, we looked for genes that were differentially expressed between leg pairs (first versus third leg, first versus second leg, and third versus second leg) in each sex separately. All differential expression analyses were performed on the 2 lines combined as we aimed to identify genes involved in allometric slope, which is a common feature of both lines. We first filtered transcripts for which expression was lower than 2 FPKM in more than half of the samples after combining the 2 inbred populations and the 3 replicates per condition (2 lines × 2 legs × 3 replicates = 12 samples total). Transcripts with average expression that was lower than 2 FPKM in the 2 legs being compared were also discarded. Finally, we ran the differential expression analysis by taking into account the line and replicate effects. We called differentially expressed genes any gene with a fold-change > 1.5 and a Padj < 0.05.

Gene Ontology analysis

Gene names and functions were annotated by sequence similarity against the NCBI “nonredundant” protein database using Blast2GO. The Blast2GO annotation was then provided to detect Gene Ontology terms enrichment (p-value < 0.05) using the default method of TopGO R package version 2.34.0.

Nymphal RNAi

Double-stranded RNA (dsRNA) was produced for BMP11 and Ubx. T7 PCR fragments of the 2 genes were amplified from complementary DNA (cDNA) template using forward and reverse primers both containing the T7 RNA polymerase promoter. The amplified fragments were purified using the QIAquick PCR purification kit (Qiagen, France) and used as a template to synthesize the dsRNA as described in [28]. The synthetized dsRNA was then purified using an RNeasy purification kit (Qiagen) and eluted in Spradling injection buffer [41] at a concentration of 6 μg/μl. For primer information, see S5 Table. Nymphal injections were performed in the line selected for long-legged males [19] at the fourth instar as described in [27]. In the first experiment, we injected nymphs in parallel with Ubx or BMP11 dsRNA or with buffer as negative controls. For Ubx RNAi, we compared adults obtained from Ubx dsRNA injection to those obtained from buffer injection. For BMP11 RNAi, 2 additional experiments were added to ensure RNAi reproducibility and increase sample size. Those experiments were performed in parallel with nymphs isolated but not injected as negative controls. Nymphs were placed in water tanks (22 cm long, 13 cm high, 10 cm wide), for which the water was changed every day, and fed ad libitum with crickets until they developed into adults. To control for RNAi specificity and efficiency, we used various methods as follows: (1) We performed quantitative reverse PCR on individuals injected with BMP11 dsRNA and confirmed that BMP11 mRNA levels were significantly lower compared to normal individuals (S8 Fig). (2) We used negative control individuals consisting in nymphs that were injected with buffer or noninjected (but reared in parallel to treatments). (3) We injected M. longipes with 3 different dsRNA preparations based on 3 different fragments of the BMP11 gene (two of which are nonoverlapping) and obtained the same effect as assessed through measurements of pronotum size (S7 Table).

Absolute quantitative reverse transcription polymerase chain reaction

This method was used to compare the expression of BMP11 in the legs and body of males and females and to confirm the efficiency of RNAi knockdown. For BMP11 expression, the 3 pairs of legs and the body from 30 fifth instar male and female nymphs were used, separately, to extract total RNA. To confirm the efficacy of RNAi treatment, 4 male and 4 female whole nymphs from both untreated and BMP11 dsRNA-injected samples were used for total RNA extraction. After DNAse treatment, 400 nanograms (expression) or 1 microgram (RNAi) per sample were used for reverse transcription to produce cDNA. The qPCR was conducted using 2 BMP11 primer pairs using cDNA from the legs and the body in 3 replicates. To determine the concentration of BMP11 in each sample, we first determined the efficiency of 2 BMP11 primer pairs to be 94.23% and 96.63%, respectively (see primer sequences in S5 Table). Second, we build a concentration curve using a BMP11 PCR-amplified DNA template with incremental (factor of ×2) concentrations from 0.0488 femtograms to 100 femtograms. This curve yielded a slope of 3.66 and 3.60 and intercept of 14.9 and 14.5 for the first and second primer pairs, respectively. The concentrations of BMP11 in the various samples was calculated using the data from the curve. qPCR and RNA-sequencing data used to construct S5 and S8 Figs can be found in S10–S12 Tables.

Behavioral assays

Male competition assays were performed using artificial puddles (containers 14 cm long, 10 cm large, and 4 cm high) containing 5 egg-laying floaters. Each replicate corresponded to a population of 5 wild-type females [19] and 10 males from a treatment (either 10 control or 10 BMP11-RNAi males). Analyses were performed on 4 replicates for each condition. Male and female interactions were recorded on a Nikon digital camera D7200 with an AF-S micro Nikkor 105 mm lens. Video acquisitions were taken a couple of hours after the bugs were transferred to the puddle. We defined a fight as an interaction between a focal male already present on a floater and a challenging male approaching the floater and inducing the 2 males to turn back to back and to engage into kicking with their third legs (S1 Video). If 2 contestants stopped fighting for more than 5 seconds, we counted the new interaction as a separate fight. For each replicate, video recordings last about 80 minutes (details about video record, number, and duration of fights are provided in S8 and S9 Tables).

Statistical analyses

All statistical analyses were performed in RStudio 0.99.486. Comparisons of mean trait size and trait variance were performed on log-transformed data and used for scaling relationship comparisons. Winged individuals are rare and have different scaling relationships. We therefore excluded them from the analysis. We used standardized major axis (SMA) as well as linear models (ANCOVA with body size as covariate) to assess differences in scaling relationships (using R package ‘smatr’ version 3.4–8 and ANOVA in R, respectively, [42,43]). Coefficients of variation were calculated using the R package ‘goeveg’ version 0.4.2.

To test for differences in fight frequency and fight duration, we used a generalized linear model with Gamma distribution, considering replicates as batch effects. Fights that lasted less than 1 second were standardized to 0.5 second.

Data deposited in the Dryad repository: https://datadryad.org/stash/share/2Q9tb7wD4R_iWkDSRAiOY2NsFuXCZE99o60C6Oeydk4 [43].

Representative picture of the nymphal molts left by an individual during its postembryonic development in M. longipes (here a male).

These molts were used to build a growth curve for each individual during nymphal development.

(TIF)

Click here for additional data file.

Covariation between male third leg length and duration of postembryonic development (nymphal development) until adulthood.

R-squared and p-value of the linear regression are indicated. The data underlying this figure may be found at https://datadryad.org/stash/share/2Q9tb7wD4R_iWkDSRAiOY2NsFuXCZE99o60C6Oeydk4.

(TIF)

Click here for additional data file.

Signature of trait exaggeration among leg-biased genes.

(A) PCA analysis of expressed genes in male and female legs separately. (B) MA plots of transcripts differentially expressed in the third and first legs in both sexes. Gray circles represent unbiased genes, while colored circles (blue in males, purple in females) represent genes significantly differentially expressed between the 2 pairs of legs at an adjusted p-value of <0.05. Venn diagrams illustrate the number of leg-biased genes (fold-change > 1.5) shared in females (purple) and males (blue), both for up-regulated genes in the third (top) and first legs (down). The data underlying this figure may be found at https://datadryad.org/stash/share/2Q9tb7wD4R_iWkDSRAiOY2NsFuXCZE99o60C6Oeydk4. MA, Log ratio and Mean average; NS, Non significant; PCA, Principal component analysis.

(TIF)

Click here for additional data file.

Boxplot showing differences in fold change (log2FoldChange) between males and females among the leg-biased genes.

The data underlying this figure may be found at https://datadryad.org/stash/share/2Q9tb7wD4R_iWkDSRAiOY2NsFuXCZE99o60C6Oeydk4.

(TIF)

Click here for additional data file.

Levels of expression of Ubx and BMP11 transcripts in M. longipes as revealed by comparative transcriptomics (A and B) and by quantitative RT-PCR (C). Note that for qPCR we also determined that BMP11 is significantly biased in female body (two-way ANOVA; p < 0.0001). The data underlying this figure may be found in S10 and S11 Tables. BMP11, Bone Morphogenetic Protein 11; RT-PCR, reverse transcription polymerase chain reaction; Ubx, Ultrabithorax.

(TIF)

Click here for additional data file.

Phylogeny of BMP family (protein and nucleotide sequences) comparing BMP sequences in multiples species of water striders with insects and vertebrates.

BF: Brachymetra furra, HH: Husseyella halophyla, ML: Microvelia longipes, MP: Microvelia pulchella, OB: Oiovelia brasiliensis, PB: Platyvelia brachialis, GB: Gerris buenoi, Pyr: Pyrrhocoris apterus, Tribolium: Tribolium casteneum, Droso: Drosophila melanogaster, H.sp1: Hebrus sp1, chicken: Gallus gallus, human: Homo sapiens, mouse: Mus musculus. Nucleotide and protein alignment files can be found in S1 and S2 Data. The data underlying this figure may be found at https://datadryad.org/stash/share/2Q9tb7wD4R_iWkDSRAiOY2NsFuXCZE99o60C6Oeydk4.

(TIF)

Click here for additional data file.

Scaling relationships of the first and second legs in the males and females of M. longipes, M. sp. (From Cayenne, French Guiana) and M. pulchella.

There is no effect of BMP11 (GDF11) RNAi on slope in any of these tissues except the exaggerated second legs of M. longipes (B). The data underlying this figure may be found in S6 and S7 Tables. BMP11, Bone Morphogenetic Protein 11; Ctl, Control; GDF11, Growth Differentiation Factor 11; RNAi, RNA interference.

(TIF)

Click here for additional data file.

Quantitative RT-PCR confirms that BMP11 RNAi significantly reduces the levels of expression of BMP11 both in males and females (two-way ANOVA; p < 0.0001).

The data underlying this figure may be found in S12 Table. BMP11, Bone Morphogenetic Protein 11; RNAi, RNA interference; RT-PCR, reverse transcription polymerase chain reaction; WT, wild type.

(TIF)

Click here for additional data file.

Leg measurements of males and females over the 5 nymphal instars and adult stage.

(CSV)

Click here for additional data file.

GLM statistics on ontogenetic leg allometries.

(XLSX)

Click here for additional data file.

Tables summarizing Gene Ontology terms for leg-biased genes across legs and sexes.

(XLSX)

Click here for additional data file.

Summary of genes/pathways known or suspected to be involved in regulating exaggerated trait growth.

(XLSX)

Click here for additional data file.

Identifiers, names, expression patterns, fold changes, and phenotypes of the genes screened by RNAi and primer sequences.

(XLSX)

Click here for additional data file.

Summary statistics of BMP11 and Ubx knockdown effects in leg lengths, body length, and static allometries across the 3 Microvelia species.

(XLSX)

Click here for additional data file.

Raw measurements of control and RNAi individuals used in this study.

(XLSX)

Click here for additional data file.

Summary table of the competition assays between M. longipes controls and knockdowns.

(CSV)

Click here for additional data file.

Raw data number and duration of fights between M. longipes controls and knockdowns.

(CSV)

Click here for additional data file.

Results of qPCR in M. longipes male and female legs and body.

(CSV)

Click here for additional data file.

Counts of BMP11 and Ubx expression using RNA-seq data in M. longipes male and female legs.

(XLSX)

Click here for additional data file.

qPCR test results of the efficiency of BMP11 knockdown using RNAi.

(XLSX)

Click here for additional data file.

Representative video of the fighting behavior in M. longipes control individuals.

(MOV)

Click here for additional data file.

Representative video of the fighting behavior in M. longipes BMP11 individuals.

(MOV)

Click here for additional data file.

Representative video of the fighting behavior in M. longipes Ubx individuals.

(MOV)

Click here for additional data file.

BMP11 nucleotide alignment.

(FASTA)

Click here for additional data file.

BMP11 protein alignment.

(FASTA)

Click here for additional data file.

3 Apr 2020

Dear Abdou,

Thank you for submitting your manuscripts "Pleiotropy promotes male exaggerated weapon and its associated fighting behaviour in a water strider" and "Impact of male trait exaggeration on sex-biased gene expression and genome architecture in a water strider" for consideration as Research Articles by PLOS Biology.

Your manuscripts have now been evaluated by the PLOS Biology editorial staff as well as by an academic editor with relevant expertise and I am writing to let you know that we would like to send your submission - "Pleiotropy promotes male exaggerated weapon and its associated fighting behaviour in a water strider" - out for external peer review as a Short Reports paper. Our Academic Editor thinks, and we agree, that the other submission - "Impact of male trait exaggeration on sex-biased gene expression and genome architecture in a water strider" - does not possess the conceptual advances and sufficient mechanistic depth which are required for consideration at PLOS Biology.

Before we can send your manuscript to reviewers, we need you to complete your submission by providing the metadata that is required for full assessment. To this end, please login to Editorial Manager where you will find the paper in the 'Submissions Needing Revisions' folder on your homepage. Please click 'Revise Submission' from the Action Links and complete all additional questions in the submission questionnaire.

Please re-submit your manuscript within two working days, i.e. by Apr 05 2020 11:59PM.

Login to Editorial Manager here: https://www.editorialmanager.com/pbiology

During resubmission, you will be invited to opt-in to posting your pre-review manuscript as a bioRxiv preprint. Visit http://journals.plos.org/plosbiology/s/preprints for full details. If you consent to posting your current manuscript as a preprint, please upload a single Preprint PDF when you re-submit.

Once your full submission is complete, your paper will undergo a series of checks in preparation for peer review. Once your manuscript has passed all checks it will be sent out for review.

Given the disruptions resulting from the ongoing COVID-19 pandemic, please expect delays in the editorial process. We apologise in advance for any inconvenience caused and will do our best to minimize impact as far as possible.

Feel free to email us at plosbiology@plos.org if you have any queries relating to your submission.

Kind regards,

Di

PLOS Biology

12 May 2020

Dear Dr Khila,

Thank you very much for submitting your manuscript "Pleiotropy promotes male exaggerated weapon and its associated fighting behaviour in a water strider" for consideration as a Short Reports at PLOS Biology. Your manuscript has been evaluated by the PLOS Biology editors, an Academic Editor with relevant expertise, and by three independent reviewers.

In light of the reviews (below), we will welcome re-submission of a much-revised version that will address all of the reviewers' concerns. You will need to comprehensively modify the manuscript in a way that it will stand alone, and you can reference your genome study which is deposited in BioRxiv. We cannot make any decision about publication until we have seen the revised manuscript and your response to the reviewers' comments. Your revised manuscript is also likely to be sent for further evaluation by the reviewers.

We expect to receive your revised manuscript within 2 months. Please note given lab shutdown due to the COVID-19 pandemic, we are flexible regarding turnaround time for revision.

Please email us (plosbiology@plos.org) if you have any questions or concerns, or would like to request an extension. At this stage, your manuscript remains formally under active consideration at our journal; please notify us by email if you do not intend to submit a revision so that we may end consideration of the manuscript at PLOS Biology.

**IMPORTANT - SUBMITTING YOUR REVISION**

Your revisions should address the specific points made by each reviewer. Please submit the following files along with your revised manuscript:

1. A 'Response to Reviewers' file - this should detail your responses to the editorial requests, present a point-by-point response to all of the reviewers' comments, and indicate the changes made to the manuscript.

*NOTE: In your point by point response to the reviewers, please provide the full context of each review. Do not selectively quote paragraphs or sentences to reply to. The entire set of reviewer comments should be present in full and each specific point should be responded to individually, point by point.

You should also cite any additional relevant literature that has been published since the original submission and mention any additional citations in your response.

2. In addition to a clean copy of the manuscript, please also upload a 'track-changes' version of your manuscript that specifies the edits made. This should be uploaded as a "Related" file type.

*Re-submission Checklist*

When you are ready to resubmit your revised manuscript, please refer to this re-submission checklist: https://plos.io/Biology_Checklist

To submit a revised version of your manuscript, please go to https://www.editorialmanager.com/pbiology/ and log in as an Author. Click the link labelled 'Submissions Needing Revision' where you will find your submission record.

Please make sure to read the following important policies and guidelines while preparing your revision:

*Published Peer Review*

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out. Please see here for more details:

https://blogs.plos.org/plos/2019/05/plos-journals-now-open-for-published-peer-review/

*PLOS Data Policy*

Please note that as a condition of publication PLOS' data policy (http://journals.plos.org/plosbiology/s/data-availability) requires that you make available all data used to draw the conclusions arrived at in your manuscript. If you have not already done so, you must include any data used in your manuscript either in appropriate repositories, within the body of the manuscript, or as supporting information (N.B. this includes any numerical values that were used to generate graphs, histograms etc.). For an example see here: http://www.plosbiology.org/article/info%3Adoi%2F10.1371%2Fjournal.pbio.1001908#s5

*Blot and Gel Data Policy*

We require the original, uncropped and minimally adjusted images supporting all blot and gel results reported in an article's figures or Supporting Information files. We will require these files before a manuscript can be accepted so please prepare them now, if you have not already uploaded them. Please carefully read our guidelines for how to prepare and upload this data: https://journals.plos.org/plosbiology/s/figures#loc-blot-and-gel-reporting-requirements

*Protocols deposition*

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosbiology/s/submission-guidelines#loc-materials-and-methods

Thank you again for your submission to our journal. We hope that our editorial process has been constructive thus far, and we welcome your feedback at any time. Please don't hesitate to contact us if you have any questions or comments.

Sincerely,

Di Jiang

PLOS Biology

*****************************************************

REVIEWS:

Reviewer #1 (Arnaud Le Rouzic, signed): Review of "Pleiotropy promotes male exaggerated weapon and its associated fighting behaviour in a water strider" by Dr Toubiana and colleagues, PBIOLOGY-D-20-00810R1.

This manuscript reports an interesting set of results on the genetic architecture of the leg length in a species of water strider. In this species, the third pair of legs is much longer in males than in other related species, which is strongly suggestive of sexual selection. The authors show that (i) the leg length difference is due to a change in the limb development at the end of the nymphal stage, (ii) a set of genes which expression is associated with the third legs length could be identified, (iii) two genes, BMP11 and Ubx, could be validated by RNAi knockdown; the knockdown had a strong effect in allometric slope for BMP11 but not for Ubx; (iv) RNAi genotypes were characterised by a different fighting behaviour. The authors conclude that this interaction pattern between body size, leg length, and pleiotropy could explain the emergence of this specific, sexually selected trait.

I am rather balanced about the manuscript. In the one hand, I found the result on the genetic architecture of allometric slopes very interesting. Whether or not allometry participates to the selection response (and even whether or not it is evolvable at all) remains an open question, and the results presented in the paper suggest that allometry is indeed evolvable, and that in addition allometric slopes and intercepts might be affected by different genes. As far as I can tell, this genuinely improves our understanding of the genetic bases of allometric relationships. The evidence based on RNAi "only" (not underestimating the complexity of genetic studies in non-model organisms) might need to be confirmed independently, but this seems to match perfectly the scope of Plos Biology short reports ("Short Reports are research articles that may be preliminary, based on a small number of experiments that might not completely flesh out the biological phenomenon under study"). On the other hand, the manuscript suffers from numerous weaknesses and inaccuracies, some being benign, others being more embarrassing, up to make peer-review uncomfortable due to the lack of information (or to the lack of consistency between different sources of information).

*** Major concerns ***

1. I have the feeling that the manuscript may have been deeply re-organised several times. This is not a problem per se, but the reader often lacks basic information to understand the authors' reasoning. Moreover, the manuscript abuses references to a companion paper, which was provided as supplementary material. Yet, this reference should not preclude self-consistency of the current manuscript, and I was rather confused about the lack of basic information about the experimental protocol (how many individuals measured? What is the biological material (population, inbred lines, wild-caught vs lab-raised individuals? There are a couple of references to selected lines from ref 19, line 270 mentions "the two inbred populations" (?), this was genuinely confusing). How many populations / samples / replicates for the transcriptome analysis? Line 295 refers to "controls" (what are they?). Even more embarrassingly, I could not find any of the supplementary tables (in the supplementary PDF, there are 11 sup figures, and the captions of the supplementary tables and movies; I had a link to the movies but not to the supplementary tables), and the figure resolution was so low in the main document that I could barely read the captions. When digging into the companion paper or counting the data points in the figures, sample sizes and replication conditions seem adequate in most cases, so I don't think that the authors are hiding details on purpose; I assume that this rather comes from copy-paste etc. across several papers without checking that the necessary information was provided. Paradoxically, the manuscript gave a fair amount of unnecessary details (such as the focal of the camera for the behaviour movies) without making the experiment reproducible (how large were the artificial puddles, what distance was considered to define that the fight occurred "near a floater", how long were the recordings?). Note that the statistics are not reproducible (I did not find any link to the code, and key analyses (such as the power-law fit) was not documented beyond the fact that the authors used RStudio).

2. Allometry is the central topic of the paper (and I think the title should be changed, see below). Yet, the introduction does not introduce basic facts about allometry; the reader is thus expected to be familiar with the manipulation of allometric slopes and intercepts, which may not match the requirements of a wide-audience journal. In addition, some sentences appear to be very confusing, e.g. line 51, "a genetic link between [...] traits and other sexual traits such as male body size": isn't it exactly what allometry is about? , or line 62, "the evolution of hyperallometry may be linked to the evolution in increased mean [...] body size": does this make sense?. To be honest, I thought at that point that the authors were simply confused about what allometry is about; Details in the methods and in some figures made me changed my mind; at least, allometry was probably measured properly, and discussed in a reasonable way in some parts of the manuscript. Yet, the authors made the strange choice of working on the trait scale and fit non-linear (power law) model, which makes everything more complicated and more difficult to interpret. Why not running the whole analysis on log traits? This would (partly) fix the heteroscedasticity that is obvious from figure 3A and 3C, make allometric relatioships linear, make it possible to use a traditional linear models ( ~ sex + leg + genotype + interactions) to compare allometric slopes and/or intercepts with t-tests... If necessary, the figure axes could be log axes so that the original meaning of the trait measures is not lost. At least, working on a log scale would have avoided what is probably an interpretation error of fig 1B, as the authors observe "a dramatic increase in leg length [...] between the 5th instar and adult males" (line 81) (from the figure: from about 3000 to 6000 µm). However, the length at the 4th instar seems to be about 1500 µm, which would mean that the leg length is simply doubling betweeen larval stages, including the last transition, and that there is nothing specific about this particular developmental stage.

3. The manuscript reports three experimental results, which have different strengths.

(i) The RNAi + length measurement experiment seems the most convincing, even if it probably lacks replication. (ii) The behavioral experiment is also convincing, although here again the possibility that the causal effect of RNAi is not really demonstrated. (iii) The transcriptomics analysis is very descriptive (and probably very noisy), and I am not sure it brings a lot of understanding to the main question. It might be that the only interesting result of the transcriptomics approach to the question raised in the introduction is to shortlist a set of genes that are overexpressed specifically in male third legs, but is remains unclear whether the list of genes from sup table 5 was derived from the transcriptomics approach. To be fair, reporting transcriptomics results that does not seem to be informative nor reliable is the norm in the field, but I found the ratio explanatory power/effort to understand fig 2 particularly disappointing.

Yet, none of them provide formal support to the conclusion from the manuscript title "pleiotropy promotes male exaggerated weapon"; this is rather a reasonable hypothesis that can be derived from the results. Note that the story of BMP11 being an example of a "magic" gene that may influence several traits that all need to evolve simultaneously under sexual selection might very well be true, but the results presented in this paper do not prove that it is the case (and even less that BMP11 actually evolved under sexual selection in this species).

I am a bit concerned that by trying to juxtapose results of different strengths (especially by including transcriptomics results that are difficult to interpret in the frame of the evolution of allometry) to address a more ambitious question, the authors may have weakened the paper.

4. Some methodological details of the paper make me uncomfortable

4.1. The exclusion of putative "unaffected RNAi" individuals from phenotypic measurements (line 299) is dubious. Sup fig 10 clearly shows that pronotom/mesonotum ratios are not so good at discriminating, and no indication about the correlation between this ratio and body size is provided. Validating the exclusion by checking that excluded phenotypes behaves as control individuals looks rather circular, and I think the authors should consider measuring more accurately the size of the bias induced by such a sorting procedure. Not knowing the number of removed individuals is a bit stressful here -- due to the lack of details, here again I will have to assume that this was reasonable.

4.2. There is an issue with the behavioural results: most of the observed differences are not significant. However, when looking at figure 4, the distributions of observations are totally non-overlapping, which makes me think that the statistics are flawed, and that the authors' approach is deeply under-powered. My interpretation is that the authors have recorded four replicates of the behavioural experiment involving five females and ten males for each treatment. During each experiment, many fights have been recorded (if the cryptic "frequency" Y axis in fig 4A means "number of fights", then almost a thousand fights were recorded). But in the t-test for fight duration, it seems that the authors have just included the average duration of fights, as if there was only one fight in each experiment, and n=4 observations per replicate. I think that a much better statistical framework would be to run a generalised linear model with a Poisson family for the number of fights (fig 4A), which should increase the power of the test (no need to estimate the variance for count data, as the variance is expected to be the same as the mean); the duration of fights should be analysed based on individual fights, possibly considering the replicate as a random effect to account for possible batch effects. This should definitely display significant differences across treatments, and avoid highlighting a dubious difference between "non significant" and p = 0.08 (cf figure 4B).

4.3. I found the interpretation of transcriptomics results unclear and questionable. The authors found a convincing "legs" effect (fig 2A) in both sexes from the full expression pattern, but the fact that they propose two independent analyses for males and females suggests that the pattern might be confusing when both sexes are considered together (in this context, I don't understand how the authors could conclude that the first axis is similar across sexes line 101, but I might have misunderstood something). Then they isolate genes that are over or under-expressed in the third legs in both males and females, which (as almost always in transcriptomics studies) generate a lot of sets of genes with little overlap. Gene ontology returns different patterns in males and females, which seems rather odd, and the clustering illustrated in fig 2 does not seem to be very insightful (for instance, genes overexpressed in female third legs display the most similar pattern when comparing male third legs and female first legs...). Some conclusions from the text (e.g. line 122, "the third legs express a considerable set of common genes between males and females") are vague (what is "considerable"?) and not supported by the figures (is it really surprising that the same genes are expressed in the legs of both sexes anyway?).

*** Details ***

* Line 61, "large body size and high variation": in the only other species that is analysed (e.g. black in fig 3), the body size variation seems even larger than in the focal species. The variation is legs size could simply be due to a change in allometry.

* Line 67, and other places in the ms: the link between exaggerated weapon size and allometry is far from clear. Disproportionate organ size can either be due to a change in allometry and/or a change in size without a change in the allometric coefficient (if > 1).

* Line 105: "the third legs diverged more form the first legs in males than in females": this looks in contradiction to a previous statement line 98, "we failed to detect any correlation between such variation in morphology and variation in gene expression". I am not sure to understand why selecting a set of genes differentially expressed between legs 1 and 3 generates a pattern that was not present in the full set of genes (is it just a problem of statistical power?).

* Line 114, 63% of leg-biased genes were common between the sexes (fig 2C): fig 2C indicates 92+157+232=481 genes overexpressed in the first legs vs third legs, 157 of them (32%) being common between sexes. I am not sure to understand what were the reported 63%.

* Line 126, "extravagant signals [...] are predicted to occur in good condition individuals": how does this observation combine with the allometric argument? If high fitness males are also larger, then it is perfectly compatible with the fact that there is positive allometry involved (in which case the fitness argument is not really central, since body and organ sizes are constrained by the allometric relationship, and the association large weapon <-> high fitness is unavoidable). If this argument ignores allometry, then it means that organ size is (at least partly) independent from body size, and this suggests that the main question about allometry is not relevant (allometry becomes a statistical association which does not derive from biological constraints).

* Line 144: I am not sure to understand why BMP11 is defined relative to its vertebrate homolog (including in the full phylogeny of Sup fig 7), there is no ortholog in insects (including D. melanogaster?).

* Line 168 and 192: good practices in statistics require to report full p-values (and not only p < 0.05, which makes meta-analyses more complicated).

* Line 209: this is an interesting (but confusing) use of "plastic response", as no environmental response is involved here.

* Line 255: there is no fig 5C.

* Line 240: I am confused by the claim that trait exaggeration evolves despite the evolutionary stasis of slope. This is clearly not the case here.

* Line 262, "analysis of variance", do you mean "principal component analysis"?

* Lines 350, 361: in the references, some author names are not capitalised.

* Figure 2, I found the Venn diagrams rather confusing (the overlapping part is not proportional, and the diagrams scale is different in fig B and C. All diagrams seem to be scaled relative to the female count, which does not seem very intuitive). The color scale (low <-> high) is particularly uninformative.

* Fig 3: state that Mpul stands for M. pulchella in the caption of figs 3A and 3B. It would be great to have the colour code indicated in the figure (perhaps replacing the R^2, which is not informative in a figure since it is easy to eyeball the fit?).

* Fig 4A: "frequency of fights" is unclear (is it "number of fights"?).

Signed: Arnaud Le Rouzic

Reviewer #2: This manuscript was coreviewed by a PI and a postdoc.

Toubiana et al. provide here a very interesting study that combines descriptive, observational data about gene expression with experimental, functional data about the developmental roles of important genes. Their results show a very satisfying connection between the development of an exaggerated sexually dimorphic trait and on male behavior, indicating a pleiotropic role for one of the genes studied, BMP11. Although not yet understood mechanistically, this link between behavior and morphology is expected to have interesting evolutionary consequences. The authors have done a fine job tying their results to outstanding questions about the major hypotheses relating to the evolution of exaggerated weaponry. However, we had to struggle in multiple places to figure out what analyses were done and thus we believe that the manuscript would benefit from clarification of some key components prior to publication. So modified, this would make it an appropriate and very accessible article for PLoS Biology's broad audience.

Our substantive concerns with the paper are primarily about places where the paper does not make clear what comparisons are being discussed and what exactly the underlying data are. These concerns likely stem in part from the short format of the manuscript, in which methods have been very condensed.

Data availability:

Please ensure that the bioproject PRJNA610161 is the correct accession number and that the project is made public by the time of publication. As of 4/27/20, this accession number does not resolve to a public project. (You can release the project without releasing the data.)

Additionally, we did not have access to the supplemental tables and thus we were unable to check whether they included the data we would expect—i.e., tables that include the data necessary for reanalysis of both allometry and gene expression data. (A request was submitted to the PLoS Biology team on 4/24/20, and we received a response that the request has been forwarded, but we have not received the tables; we did get to enjoy the supplemental videos and figures.)

Aims/analyses/interpretations:

Phrasing of main aim: "determining the genetic architecture) (l. 54) and "determining the genetic basis (l. 69) both lead readers to expect an approach that identifies causal genetic variation, whereas the research approach used here identifies genes that are differentially expressed and functionally important, but may or may not be the loci responsible for the evolutionary changes (even in the context of these selected lines). Thus, we suggest rephrasing the aims.

Experimental design: The methods should briefly state the sampling scheme for the transcriptomic data: i.e., that the starting material was lines selected for larger differences in leg size, and that gene expression in 2 lines * 2 sexes * 3 replicates *3 legs were characterized, rather than refer to the companion manuscript for it..

PCA analysis (results on l.96-102): We question the interpretation of this analysis and recommend dropping it from the manuscript. If it is maintained, then a clearer explanation (including methods: e.g., what genes (all? leg-biased? different subsets in the 2 sexes?), how was PCA implemented) is needed both of what was done and of how the result supports the interpretation. We do not think it is meaningful to compare the % of variance explained by axis 1 between males and females if these are separate, independently run analyses (which is our impression). Even if total variation (not just %) were greater in one than the other, additional work would be needed to demonstrate that leg3 was the cause and further to demonstrate that a specific set of genes is involved (e.g., loading information).

Differential expression analyses (results on l. 103-123; methods on l. 267-275):

It was difficult to figure out which comparisons were used to identify genes as 'leg-biased', even after having read the relevant sections of the companion paper. We concluded that the analyses were done on pairs of legs (1 vs. 3 and 1 vs. 2) from 2 lines * 3 replicates in each sex separately. If that's correct, then including sampling design in methods and changing l. 103-105 as follows would clarify:

"To identify the genes underlying the growth differences in serial homologs, we identified genes that were differentially expressed between leg pairs (first vs. third leg and first vs. second leg) in each sex separately."

(While the term 'leg-biased' is parallel to 'sex-biased', it may be more prone to being misunderstood as implying a comparison between legs and some other tissue. Thus, we wonder if this term could be avoided—but we didn't come up with a good alternative.)

We remain unclear on what the set of "upregulated genes restricted to male third legs" is; we could not match the possibilities we considered for the number of genes given (489) and were unclear what 'restricted' means here. Figure (2B) shows 273+56=329 genes in the Venn diagram up regulated in the male T3 leg. Thus, we would have expected the 489 to be either 329 + 56 or 329 or some smaller number if genes upregulated in other contexts (e.g., male T2) didn't count. Is this a typo, or are you referring to a different defined set of genes in the text versus the figure? Clarifying this should also help clarify what is meant by the statement on l. 109-111 that the male regulated genes are "similarly expressed" in females. (This hierarchical clustering result may be more or less forced by the way in which genes were chosen for analysis—i.e., the set of genes displays maximal differences between male T1 and T3 (and perhaps isn't DE between female legs? Despite the clustering pattern, Figure 2B (right) gives an overall impression that expression levels are not similar between female T1 and T3).

l. 114: where is the 63% figure coming from. Based on figure 2C, I would have expected the figure to be 157/(92+157+232) = 33%

Supp. Fig. 2 (and associated ANCOVA test on l. 83): We would expect that development time was potentially predictive of leg length, and thus that it should be on the x-axis (rather than the other way around). Also, if the ANCOVA was applied to this analysis, please make clear what the covariate was.

We encourage a supplemental methods section that covers in more detail the methods you're referencing from the Genomics companion paper in supplemental methods for this paper. Yes, it's duplicative, but if someone has access to one and not the other, they will thank you for the duplication. Additionally, sharing the analysis code would be helpful.

l. 265 (and elsewhere): Please include the R package name and the version number in the main methods text for all R packages used, as was done for pvclust (including also the version of R, which is separate from the version of RStudio).

Figure 2 comments: We recommend that in genes be sorted by expression values for the focal leg type. For example, if the male heatmap in 2B were sorted by expression of T3 leg (high > low) it would be much more obvious that the female T3 leg expression is very different from male T3 leg.

Minor points/typos, etc.

1) Line 278: How many genes were included in this hierarchical clustering analysis?

2) Line 300-301: Since you are discarding some data points (legitimately!) I recommend including the percentage of injected adults that had the short pronotum phenotype versus those that didn't for both controls and experimentals.

3) Lines 127-129 ("Proposed mechanisms for the development of exaggerated phenotypes include increased sensitivity or the production of growth factors by the exaggerated organ [4, 25].") That sounds to me like you're talking about two distinct hypotheses. But it seems like the next sentence is referring to a single hypothesis ('this hypothesis'). Please clarify—it may be as simple as changing 'this hypothesis' to 'these hypotheses'

4) Lines 150-166: To make this more accessible to a general bio audience -- In these sentences that sum up the results, can you rephrase what change in intercept and coefficient means in terms of relative and absolute sizes of legs and bodies?

5) Line 156-158: Perhaps rephrase sentence as follows 'The allometric coefficients for the first and second legs of males were also lower than in the corresponding controls, but the changes were smaller than for the third legs.'

6) Line 164: Should be "….closer to those of its sister species…"

7) Line 230: "Studies of scaling relationships have reported" rather than just "reported"

8) Figure 1 - Line 439 "Power low regression" should be power law regression.

9) Figure 4: - legend, the word "defects" should be "effects".

Reviewer #3: Review of Toubiana et al

This is a really nice and very interesting paper describing the role the BMP and Ubx play in regulating the morphological scaling of the third leg in male water strider of the species, Microvelia longipes. Males of this species use their exaggerated third legs to compete for access to females. The authors used transcriptomics to look at differences in gene expression between the isometric first and hyperallometric third leg, in males and females. Using these data they constructed a list of 30 candidate genes that they functionally tested using RNAi. They found that two, BMP11 and Ubx had an effect on scaling. Specifically BMP11 reduced body size in males and females, and reduced the slope of the 3rd leg-body scaling relationship in males, while Ubx had no effect on body size, but reduced the intercept of the 3rd leg-body in males and females. At the same time knock-down of BMP11 reduced the aggressiveness of males.

The paper is significant in that it is one of the few that identifies a developmental mechanism that controls the slope and intercept of a morphological scaling relationship. I certainly think that it is of sufficient general interest to be published in PLoS Biology. Nevertheless, I have a number of suggestions about how the paper can be improved. The tables were not included in my review copy of the supplementary material, so some of my concerns may be addressed there.

1) A key finding of the paper is that knock-down of BMP11 changes the slope of the 3rd leg-body scaling relationship. Looking at figure 3C, it is possible that the 3rd-leg-body scaling relationship of the BMP11-RNAi males may be an extension of the wild-type scaling relationship but in smaller males. That is, among small wild-type males, the slope of the 3rd-leg-body scaling relationship is shallower. If this were the case, then it is possible that BMP11.RNAi simply reduces body size, with no additional effect on scaling. To test this, the authors should look at scaling in males that are small, for example by reducing access to food during development. If the author's hypothesis were correct, then these small males should have the same body size as BMP11-RNAi males, but a steeper leg:body scaling relationship. These data would also confirm that the pronotum/mesonotum ratio does not increase with an increase in body size. If it does, then exclusion of BMP11-RNAi individuals with high pronotum/mesonotum ratios from the analysis of the effects on BMP11 on 3rd-leg-body scaling is not reasonable: the authors are simply excluding larger individuals with steeper leg:body scaling.

2) In several places, the authors confound the necessity with sufficiency. For example: L158: "Therefore, BMP11 increases both the allometric coefficient of all legs and mean body size in the males", and L197: "These results suggest that BMP11, but not Ubx, increases fighting frequency in M. longipes males." Both these statements argue implicitly that BMP11 is sufficient to increase the allometric coefficient or fighting frequency. Since the authors have not increased BMP11 expression, but only decreased it, they can only be certain that BMP11 is necessary for the hyperallometry of the legs, and for increased fighting behavior, but not that it is sufficient.

3) The authors do not provide evidence that BMP-RNAi or UBX-RNAi knocks down expression of both these genes. They should do so.

4) I did not have access to the table showing the statistics used to test the hypotheses described in the paper. They need to be included for a complete review. In particular, I am not sure how the authors tested for the effect of BMP-RNAi on variance (L150)

Minor points:

5) The section on transcriptomics of the different legs in males and females is difficult to follow. In particular, the authors do not appear to be using the data to test a clear hypothesis about how gene expression is regulated in the legs in males and females. What do these patterns of gene expression mean? would these patterns look like under different hypothetical mechanisms of leg exaggeration? What would you expect the pattern of PC1 and PC2 to look if

morphological divergence in serial homologs between the sexes was the "result of global change in gene expression but rather from specific sets of genes" (L101)? The title is also misleading: "Leg exaggeration is associated with a specific signature of leg-biased gene expression". Given that only gene expression in the legs was measured, of course the gene expression is leg-biased. This section should be re-written or excluded.

6) Including color and shape keys in the figures would help the reader interpret the charts.

7) For Figure 3, the authors should indicate what M.pul means. It took me a while to realize it meant M. pulchella.

18 Jan 2021

Submitted filename: Response_to_Reviewers_Jan15_2021.docx

Click here for additional data file.

12 Feb 2021

Dear Abdou,

Thank you for submitting your revised Short Report entitled "The growth factor BMP11 is required for the development and evolution of a male exaggerated weapon and its associated fighting behavior" for publication in PLOS Biology. I've now obtained advice from two of the original reviewers and have discussed their comments with the Academic Editor.

Based on the reviews, we will probably accept this manuscript for publication, provided you satisfactorily address the remaining points raised by the reviewers. Please also make sure to address the following data and other policy-related requests.

IMPORTANT:

a) Please change the title to include a mention of the organism studied; we suggest: "The growth factor BMP11 is required for the development and evolution of a male exaggerated weapon and associated fighting behavior in a water strider"

b) Please attend to the remaining requests from reviewers #1 and #2.

c) Please address my Data Policy requests (see further down).

As you address these items, please take this last chance to review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the cover letter that accompanies your revised manuscript.

We expect to receive your revised manuscript within two weeks.

To submit your revision, please go to https://www.editorialmanager.com/pbiology/ and log in as an Author. Click the link labelled 'Submissions Needing Revision' to find your submission record. Your revised submission must include the following:

-  a cover letter that should detail your responses to any editorial requests, if applicable, and whether changes have been made to the reference list

-  a Response to Reviewers file that provides a detailed response to the reviewers' comments (if applicable)

-  a track-changes file indicating any changes that you have made to the manuscript.

NOTE: If Supporting Information files are included with your article, note that these are not copyedited and will be published as they are submitted. Please ensure that these files are legible and of high quality (at least 300 dpi) in an easily accessible file format. For this reason, please be aware that any references listed in an SI file will not be indexed. For more information, see our Supporting Information guidelines:

https://journals.plos.org/plosbiology/s/supporting-information

*Published Peer Review History*

Please note that you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out. Please see here for more details:

https://blogs.plos.org/plos/2019/05/plos-journals-now-open-for-published-peer-review/

*Early Version*

Please note that an uncorrected proof of your manuscript will be published online ahead of the final version, unless you opted out when submitting your manuscript. If, for any reason, you do not want an earlier version of your manuscript published online, uncheck the box. Should you, your institution's press office or the journal office choose to press release your paper, you will automatically be opted out of early publication. We ask that you notify us as soon as possible if you or your institution is planning to press release the article.

*Protocols deposition*

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosbiology/s/submission-guidelines#loc-materials-and-methods

Please do not hesitate to contact me should you have any questions.

Best wishes,

Roli

Roland G Roberts, PhD,

Senior Editor,

rroberts@plos.org,

PLOS Biology

------------------------------------------------------------------------

DATA POLICY:

You may be aware of the PLOS Data Policy, which requires that all data be made available without restriction: http://journals.plos.org/plosbiology/s/data-availability. For more information, please also see this editorial: http://dx.doi.org/10.1371/journal.pbio.1001797

Note that we do not require all raw data. Rather, we ask that all individual quantitative observations that underlie the data summarized in the figures and results of your paper be made available in one of the following forms:

1) Supplementary files (e.g., excel). Please ensure that all data files are uploaded as 'Supporting Information' and are invariably referred to (in the manuscript, figure legends, and the Description field when uploading your files) using the following format verbatim: S1 Data, S2 Data, etc. Multiple panels of a single or even several figures can be included as multiple sheets in one excel file that is saved using exactly the following convention: S1_Data.xlsx (using an underscore).

2) Deposition in a publicly available repository. Please also provide the accession code or a reviewer link so that we may view your data before publication.

Regardless of the method selected, please ensure that you provide the individual numerical values that underlie the summary data displayed in the following figure panels as they are essential for readers to assess your analysis and to reproduce it: Figs 1BC, 2ABC, 3MNO, 4AB, S2, S3AB, S4, S5ABC, S6 (alignment files), S7ABCDEF, S8. NOTE: the numerical data provided should include all replicates AND the way in which the plotted mean and errors were derived (it should not present only the mean/average values).

I note that some of these data may already be in some of the Supplementary Tables (e.g. Tables S7, S8, S9); if so, please clarify, and re-name these as Supplementary Data files (e.g. S1_Data, S2_Data...).

IMPORTANT: Please also ensure that figure legends in your manuscript include information on where the underlying data can be found, and ensure your supplemental data file/s has a legend.

Please ensure that your Data Statement in the submission system accurately describes where your data can be found.

------------------------------------------------------------------------

REVIEWERS' COMMENTS:

Reviewer #1:

[identifies himself as Arnaud Le Rouzic]

Review of "The growth factor BMP11 is required for the development and evolution of a male exaggerated weapon and its associated fighting behavior" by Dr Toubiana and colleagues, PLOS BIOL MS D-20-00810R2.

This is a deeply revised version of MS D-20-00810R1 I had the opportunity to review last year. My opinion about the manuscript was balanced, as I found the topic and the results very interesting, but I had major concerns: 1) a lot of data and methodological information were missing, 2) it was not entirely clear that allometry was analyzed and interpreted properly, 3) transcriptomics results were difficult to interpret and not important for the paper, and 4) there were some methodological issues with some analyses.

The authors have substantially rewritten the paper and clarified many issues, both in the ms and in their response to the reviewers. The revised version is quite convincing, I liked it a lot, and I have no further reason to oppose publication.

My point #1 was completely addressed. The authors explained that two companion papers had to find different publication routes, and that the state of the previous version reflects an incomplete split.

Point #2 has been largely addressed. Yet, I am still not completely convinced by the authors' interpretation of what happens between the 5th instar and the adult stage. It seems to me that the authors' narrative relies on the idea that "something" (a physiological discontinuity) happens there (the authors chose to extract RNA for transcriptomics at that time point). I totally agree that third legs grow substantially between the 5th instar and the adult stage, more than between any previous developmental stages. However, my interpretation of the growth curves (fig 1B) and supplementary tables 1 and 2 is that the development is possibly allometric (difficult to confirm without reanalyzing the data including body size at each larval stage). This is probably a matter of interpretation, but if the development of legs follow an allometric trajectory, nothing "dramatic" occurs at the 5th instar beyond the fact that growth curves diverge more and more on the arithmetic scale (a mere consequence of plotting growth, which is essentially a geometric process, on an arithmetic scale). On the opposite, if the allometric trend breaks down for the 3rd leg (i.e. leg length does not follow a power law), then I would agree that the pattern is discontinuous and that something special happens here. This, of course, should be regarded as a minor conceptual disagreement between the authors and a reviewer, and does not require a change in the manuscript.

The focus on transcriptomics has been largely reduced, which fixes my point #3.

Finally, the authors ran new RNAi injections and ran the analysis without excluding control-looking individuals (point #4.1), redesigned the statistical analysis of the behavioral assay (point #4.2), and fixed interpretation issues from the transcriptome analysis (point #4.3).

*** Minor suggestions:

* Lines 244 and 245: it is probably necessary to indicate "allometric slope" and "allometric intercept".

* Line 321: PCA and analysis of variance are two different analyses.

* Figures 2 and 3(M, N, O): the x and y ranges are much wider than the data range, making it difficult to read the figures. I am not sure it is justified to plot all data (across sexes, legs, and species) on the same x and y range, since direct comparison is not of interest here.

Reviewer #2:

Overall, I found this revised manuscript substantially clearer and improved. It is much easier for readers to figure out what was done; the new functional data from two closely related species are an interesting addition; and the very strong effect of BMP11 on both behavior and the relationship between leg length and body size is unexpected and fascinating in the context of the evolution of sexually selected traits.

There are, however, still a few places where I think the ms. would benefit from additional clarification or where more caution in interpretation may be called for:

l. 68-69: unclear why large body size is relevant. This isn't highlighted in the background nor is it inherent to consideration of allometry. (But it's a point that's returned to as a reason this species was chosen in several places.) The relevance of large body size either needs to be explained or statements about it should be removed.

l. 259-260: Ubx was expressed at high levels in T3 legs long before the hyperallometry evolved. So I don't think this can be taken as evidence of an alternative pathway to trait exaggeration that's being argued for here. Additionally, I don't recall evidence that BMP11 is NOT upregulated in the sister species. This would need to be the case to support this central claim that it's overproduction of the growth factor that is causing increased leg growth. The data seem equally compatible with alternative interpretations about what the evolutionary change was, such as that it's a change in the expression of something downstream of BMP11 or in the responsiveness of something to BMP11 signaling. In the absence of comparative expression information, I think greater caution in interpreting the results is warranted.

Related to results interpretation, I was unclear what the relationship of the three studied species is. Among the three, if sp. nov. and pulchella are most closely related, with longipes sister to a clade that contains both of them, then the authors should be more cautious about interpretations of ancestral versus derived states (as there is not direct evidence about the polarity of a change when the difference distinguishes sister groups). One direction of change may be more plausible based on other considerations, but the distribution of traits on the tree would not be providing additional supporting evidence. (On the other hand, if one of the two additional species is more closely related to longipes, and the two additional species have the same trait value, this comment can be ignored, because the result would help polarize the change.)

Minor changes:

l. 55: clarify by writing 'Allometric slope'

l. 73: 'close relative' sounds more natural than 'closely related species'

l. 80: change 'multi-facetted' to 'multifaceted' (note: two corrections to word)

l. 94: unclear what's meant by 'fluctuation' of the growth birth. Just variation?

l. 98-99: unclear which similarity is meant: that there's a difference similar to the diff. between male and female 3rd legs in the 1st and 2nd legs? Or just that all these legs grow faster starting in the 4th instar?

l. 104-105: comparison unclear: what's the global difference between? Male legs and female legs? Different legs within a sex?

l. 114-115: need an only: predicted to occur ONLY in individuals in good condition because.....

l. 155: change male's to males'

l. 173-175: Not clear to me on what basis one would predict that higher expression would translate into a larger RNAi effect. I could present an argument for the opposite conclusion: that with low-level expression, knockdown may reduce expression to a greater degree, leading to a stronger effect.

l. 176: check number of digits in slopes; I'm guessing this should be reported as 3.1 vs. 2.0

l. 296: change 'effect' to 'effects'

transcriptomics: this methods section omits saying what the reads were mapped against. (I assume the genome from the former companion manuscript.)

l. 339: unclear why 2nd vs. 3rd leg comparison is omitted.

l. 365: should 'wo' be 'two'?

l. 383: I was confused by 'For RNAi'; should this be 'For qPCR'?

l. 410: Please specify the version of R being run by RStudio.

l. 565: delete 'neither'

l. 575: change 'in in' to 'in'

l. 580: space needed in M. pul

l. 580: change 'states' to 'stands'

Figure 2A, Ubx female b (on graph): think value should be 0.4787

Fig. 2 and Fig. 3 graphs: When there is no significant difference between slopes, I think the results are better represented by providing a single estimate of b (or an estimate that applies to a single sex)

Supplementary materials:

Quality of some of the figures (e.g., S3 & S7); I suspect Word is to blame for this, and that either embedding some other way or using pdf may solve the problem.

24 Feb 2021

Submitted filename: Response_To_Reviewers_R2.docx

Click here for additional data file.

25 Feb 2021

Dear Abdou,

On behalf of my colleagues and the Academic Editor, Andreas Hejnol, I'm pleased to say that we can in principle offer to publish your Research Article "The growth factor BMP11 is required for the development and evolution of a male exaggerated weapon and its associated fighting behavior in a water strider" in PLOS Biology, provided you address any remaining formatting and reporting issues. These will be detailed in an email that will follow this letter and that you will usually receive within 2-3 business days, during which time no action is required from you. Please note that we will not be able to formally accept your manuscript and schedule it for publication until you have made the required changes.

Please take a minute to log into Editorial Manager at http://www.editorialmanager.com/pbiology/, click the "Update My Information" link at the top of the page, and update your user information to ensure an efficient production process.

PRESS: We frequently collaborate with press offices. If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximise its impact. If the press office is planning to promote your findings, we would be grateful if they could coordinate with biologypress@plos.org. If you have not yet opted out of the early version process, we ask that you notify us immediately of any press plans so that we may do so on your behalf.

We also ask that you take this opportunity to read our Embargo Policy regarding the discussion, promotion and media coverage of work that is yet to be published by PLOS. As your manuscript is not yet published, it is bound by the conditions of our Embargo Policy. Please be aware that this policy is in place both to ensure that any press coverage of your article is fully substantiated and to provide a direct link between such coverage and the published work. For full details of our Embargo Policy, please visit http://www.plos.org/about/media-inquiries/embargo-policy/.

Thank you again for supporting Open Access publishing. We look forward to publishing your paper in PLOS Biology.

Sincerely,

Roli

Roland G Roberts, PhD

Senior Editor

PLOS Biology