Characterization of trade-offs between immunity and reproduction in the coral species Astrangia poculata

Introduction

The ability to successfully defend against a variety of pathogenic threats is an essential component of organismal fitness (Lochmiller & Deerenberg, 2000). However, most immune defenses which prevent or mitigate damage from pathogens are energetically costly (Lochmiller & Deerenberg, 2000). This is particularly important as organisms operate using a fixed budget of energetic resources that must be allocated to competing demands (i.e., growth, reproduction, defense, etc; Stearns, 1989). Consequently, allocation of resources to one demand often comes at the cost of others, creating trade-offs (Melbinger & Vergassola, 2015). Optimal organismal fitness therefore requires careful allocation of resources across these demands to minimize negative effects of trade-offs and maximize reproductive output. Notably, the high costs of maintaining immune systems and mounting immune responses often result in trade-offs with other traits (Lochmiller & Deerenberg, 2000; Martin, Weil & Nelson, 2008; Rauw, 2012; Rigby & Jokela, 2000; van der Most et al., 2010). Among the most well documented immune-related trade-offs are those between reproduction and immunity, with organisms most often able to exhibit high immune capacity or high reproductive ability, but not both (Adamo, Jensen & Younger, 2001; Brokordt et al., 2019; Gwynn et al., 2005; Hosken, 2001; Schwenke, Lazzaro & Wolfner, 2016). However, despite frequent empirical observation of immune-reproductive trade-offs and robust proposed theory, many questions still remain regarding the generalizability of resource allocation theory and immune-related trade-offs. In some systems, consistent immune-reproductive trade-offs have not been observed (Kelley et al., 2021; McNamara et al., 2013; Miyashita et al., 2019; Syed et al., 2020; Xu, Yang & Wang, 2012). Furthermore, it remains unclear how dynamic changes in resources/energetic budget might impact these trade-offs. It is theorized that since total energy and relative allocation are co-dependent, variation in energy budget may have an impact on trade-offs such as those between immunity and reproduction, however limited empirical evidence exists to support or refute this theory (Cotter et al., 2010; Descamps et al., 2016; Simmons, 2012; Stahlschmidt et al., 2013). Combined, these gaps in existing theory make it difficult to predict how resources are allocated among traits, requiring further study to improve understanding of resource allocation and immune trade-offs in complex contexts. This is particularly important in light of the rising prevalence of epizootic outbreaks affecting vulnerable species across the globe (Croquer & Weil, 2009; Glenn & Pugh, 2006; Kilburn et al., 2010).

Scleractinian or reef building corals are arguably one of the taxa most impacted by increases in epizootic outbreaks (Bruno et al., 2007; Precht et al., 2016; Ruiz-Moreno et al., 2012). Disease outbreaks have been one of the largest drivers of coral declines in recent decades, affecting almost all major reef building species (Alvarez-Filip et al., 2022; Bourne et al., 2009; Cróquer, Weil & Rogers, 2021; Moriarty et al., 2020; Sharma & Ravindran, 2020; Sutherland, Porter & Torres, 2004). Still, despite the severity of these outbreaks, many gaps remain in our understanding of associated disease processes. For example, significant inter- and intra-specific variation has been observed in coral disease resilience (Alvarez-Filip et al., 2022; Miller et al., 2019; Palmer et al., 2011a). While some studies have pointed to the importance of divergence in certain cellular processes in driving this variation (Beavers et al., 2023; Fuess et al., 2017; MacKnight et al., 2022; Pinzon et al., 2014a; Traylor-Knowles et al., 2021), little is known regarding how broader ecological processes, including resource allocation, might contribute to variation in disease resistance. To date, most studies of cnidarian resource allocation have focused on stress associated trade-offs. The negative effects of thermal anomalies on coral reproductive output have been well documented (Michalek-Wagner & Willis, 2001; Nielsen et al., 2020; Paxton et al., 2016). In contrast, only a handful of studies have documented reproductive trade-offs involving immunological processes, all of which had focused on the negative impacts of disease outbreaks on coral reproduction (Fuess et al., 2018; Weil, Croquer & Urreiztieta, 2009). No studies have directly linked reproductive investment and constitutive immunity in cnidarians, in part due to the historic lack of tractable model cnidarian systems.

Recent advances in the development of cnidarian model and study systems have greatly expanded the scope of cnidarian research. One such emergent model species is Astrangia poculata (common name: northern star coral), a non-reef building stony coral that can be found along the Atlantic Coast of the United States ranging from Cape Cod, MA to the Texas Gulf Coast (Dimond et al., 2012). Importantly, unlike tropical corals which are dependent on symbiotic relationships for survival, A. poculata associates facultatively with a single species of symbiont, Breviolum psygmophilum (Lajeunesse, Parkinson & Reimer, 2012) with colonies ranging from high symbiont density (“brown”) to low symbiont density (“white”; Sharp et al., 2017). Classification of symbiotic states is based on color, approximate chlorophyll concentration, and symbiont density (“brown” >10⁶ cells cm⁻²; “white” 10⁴–10⁶ cells cm⁻²; Sharp et al., 2017). A key benefit of the symbiotic relationship between corals and their symbionts is the exchange of organic nutrients to the host (Kirk & Weis, 2016). Consequently, facultative symbiosis may serve as a natural system for exploring the effects of variable resource budget on immune-associated trade-offs in corals, though the exact effects of variation in symbiont density on host energetic budget are poorly studied (Szmant-Froelich & Pilson, 1980). In contrast, recent studies have highlighted significant impacts of variation in A. poculata symbiont density on host immunity (Changsut et al., 2022; Harman et al., 2022), suggesting that variation in host-symbiont dynamics have significant consequences for broader organismal physiology.

Here we used the tractable A. poculata study system to investigate cnidarian resource allocation and potential trade-offs between reproduction and immunity (i.e., negative associations between reproduction and immunity resulting from resource limitation). We first assessed general effects of variability in symbiont density on host physiology (reproductive output, energetic budget, and immune activity). Next, we assessed the effects of energetic budget on both reproductive output and immune activity generally, independent of symbiont density. Finally, we tested for reproductive-immune trade-offs. Our study provides a preliminary framework for future studies considering the inter-connected roles of symbiosis, resource allocation, and immunity in cnidarians.

Materials and Methods

Portions of this manuscript were previously published as a preprint (Villafranca et al., 2023).

Results

Hypothesis testing

Symbiont density was significantly higher in colonies classified as “brown” than those classified as “white” (Wilcoxon test, W=91, p < 0.001; Fig. S1). However, there was no significant association between symbiont density and carbohydrate (linear model, coefficient estimate = −3.18e⁻⁹, std err = 7.25e⁻⁹, t val. = −0.439, p = 0.666) or lipid concentration (linear model, coefficient estimate = −2.65e⁻¹⁰, std err = 6.50e⁻⁹, t val. = −0.041, p = 0.968), nor was there a difference in either carbohydrates (linear model, coefficient estimate = 0.00133, std err = 0.00178, t val = 0.749, p = 0.463) or lipids (linear model, coefficient estimate = 0.000383, std err = 0.00131, t val = 0.293, p = 0.773) between white and brown colonies. Finally, the only significant predictor of sperm density was carbohydrate concentration, which was positively associated with normalized sperm density (Fig. 1, Tables 1, S1).

Table 1:

Linear model results for sperm density.

Predictors	Estimates	SE	t	p value
(Intercept)	12.3	0.687	17.9	<0.001 ***
Carbohydrates	224	84.5	2.65	0.0167*

DOI: 10.7717/peerj.16586/table-1

Note:

Best-fit linear models for sperm density when including symbiont density, carbohydrate concentration, lipid concentration as predictors. All possible models were compared and model averaging was used where appropriate (AIC delta < 2). Bold font indicates significant p-values; asterisks represent significance: *p < 0.05, **p < 0.01, ***p < 0.001.

Multivariate statistics

Immunity did not vary significantly as a result of symbiont state (MANOVA, Pillai’s Trace = 0.482, F (5/11) = 2.047, p = 0.1497). Principle component analysis was in congruence with this finding (Fig. 2); no clear spatial separation between white and brown colonies was observed. Furthermore, forward model selection for RDA analysis failed to identify any significant predictors. Finally, we identified several significant correlations among our continuous variables (Fig. 3). Carbohydrate concentration was significantly positively correlated to both antibacterial activity (Pearson correlation, r = 0.59, p = 0.0158) and lipid concentration (Pearson correlation, r = 0.51, p = 0.0459). Symbiont density was significantly correlated to melanin concentration (Pearson correlation, r = 0.54, p = 0.0317). Sperm density was significantly positively correlated to total phenoloxidase activity (Pearson correlation, r = 0.5, p = 0.0494).

Univariate statistics

Based on the limited observed significant multivariate associations, we also considered the impacts of each of our predictors (symbiont density/state, energetic budget, sperm density) on activity of our immune metrics using general linear models. Only melanin was significantly impacted by any of predictors (Tables 2, S2). Melanin concentration was significantly positively associated with both carbohydrate concentration (p = 0.00927) and symbiont density (p = 0.0176; Fig. 4).

Table 2:

Linear model results for individual immune parameters.

Catalase
Predictors	Estimates	SE		SEadj		z		p value
(Intercept)	1,386	134.5		141.5		9.80		<0.001 ***
Sperm	−1.10E−5	2.02E−5		2.10E−5		0.529		0.597
Carbohydrates	9,829	18,200		18,820		0.522		0.602
Peroxidase
Predictors	Estimates	SE		SEadj		z		p value
(Intercept)	1.875	0.315		0.334		5.62		<0.001 ***
Symbiont	−2.74E−6	2.98E−6		3.08E−6		0.889		0.374
Antibacterial activity
Predictors	Estimates	SE		SEadj		z		p value
(Intercept)	3.35	0.187		0.196		17.2		<0.001 ***
Carbohydrates	13.5	20.3		20.9		0.647		0.518
Lipids	5.32	15.0		15.5		0.343		0.732
Total phenoloxidase
Predictors	Estimates	SE		SEadj		z		p value
(Intercept)	−1.39	0.144		0.153		9.12		<0.001 ***
Sperm	4.06E−8	3.50E−8		3.77E−8		1.08		0.282
Melanin
Predictors	Estimates		SE		t		p value
(Intercept)	2.61E−5		1.61E−5		1.621		0.126
Carbohydrates	5.44E−3		1.82E−3		2.984		0.00927**
Symbiont	1.97E−10		7.37E−11		2.67		0.0176*

DOI: 10.7717/peerj.16586/table-2

Note:

Best-fit linear models for each immune parameter when including symbiont density, carbohydrate concentration, lipid concentration, and sperm produced per polyp as predictors. All possible models were compared and model averaging was used where appropriate (Δ AIC < 2). Bold font indicates significant p-values; asterisks represent significance: *p < 0.05, **p < 0.01, ***p < 0.001.

Discussion

Coral reefs globally are undergoing rapid, unprecedented declines due to anthropogenic climate change and associated disease outbreaks (Hoegh-Guldberg et al., 2007; Johnston, Clark & Bruno, 2022; Precht et al., 2016). Corals exhibit both inter- and intra-specific variation in susceptibility to disease and infection, but the ecological factors that drive susceptibility are poorly understood (Fuess et al., 2017; Mydlarz et al., 2016). Few studies have investigated trade-offs between key life-history traits and immunity in cnidarians (Alvarez-Filip et al., 2022; Pinzon et al., 2014b; Schlecker et al., 2022; Weil, Croquer & Urreiztieta, 2009). Here, we used the facultatively symbiotic coral Astrangia poculata to investigate potential reproductive-immune trade-offs in scleractinian corals, and the impacts of variation in symbiont density and energy budgets on these trade-offs.

Our study found limited links between variation in symbiont density and our measured physiological traits (lipid/carbohydrate concentration, immune activity, reproductive output). Most notably, we observed no significant difference in total lipid and carbohydrate content as a result of variable symbiont density in adult coral colonies. While this is in agreement with a previous study of A. poculata (Szmant-Froelich & Pilson, 1980), it is in opposition to conventional thinking regarding symbiosis, which would suggest that increased symbiont density results in increased energetic budget (Changsut et al., 2022; Harman et al., 2022; Hughes et al., 2010; Pupier et al., 2019). We propose two alternative hypotheses to explain these patterns: first, it is possible corals with lower symbiont density compensate via increased heterotrophic feeding. Unlike tropical corals, Astrangia is commonly found in low light, nutritionally rich waters. Consequently, these corals may obtain a significant portion of their nutrition from heterotrophy, especially colonies with low symbiont densities (Szmant-Froelich, 1981; Wuitchik et al., 2021). Heterotrophic feeding has previously been demonstrated to have a significant role in the efficiency of physiological traits in Astrangia and other facultatively symbiotic corals (Aichelman et al., 2016; Dimond & Carrington, 2007; Ferrier-Pagès et al., 1998; Miller, 1995). Thus, it is possible that our colonies with lower symbiont densities were compensating via heterotrophic feeding and actually not nutrient limited. The absence of nutrient limitation due to heterotrophic feeding may also explain the lack of trade-offs observed (see below). Alternatively, total lipid and carbohydrate assays may not fully captured host energetic budget, as coral energetic budgets are exceptionally complex (Lesser, 2012). Other approaches which account for carbon flux, utilize stoichiometric approaches, or assess diverse metabolites in both the presence and absence of heterotrophy, may provide further insight regarding the relationship between symbiont state and energetic budget.

Similar to energetic budget, associations between symbiont density and immunity were limited. Our multivariate analyses failed to detect any differences in immunity between white and brown colonies, nor as a result of symbiont density. However, correlational analysis and univariate linear modeling detected a positive association between melanin concentration and symbiont density (but not symbiont state). Previous A. poculata studies have noted a significant positive association between melanin synthesis and symbiotic state/symbiont density (Changsut et al., 2022; Harman et al., 2022). This broad consensus regarding the positive association between melanin and symbiont density is most likely reflective of the dual roles of melanin synthesis in immunity and symbiont regulation. Melanin production can be used as a mechanism of symbiont shading in response to UV, reducing organismal stress (Palmer, Bythell & Willis, 2011). Further investigation of response of A. poculata to pathogens/immune threats, as opposed to measurement of constitutive immunity, will provide insight regarding the roles of melanin in immune responses, and clarify the mechanisms of symbiont-melanin associations.

Next, we considered associations between symbiont density, carbohydrate and lipid concentration and sperm density. Only carbohydrate concentration was positively associated with sperm density, suggesting that sperm release is at least in part resource limited. As gametogenesis in A. poculata begins earlier in the year (Szmant-Froelich, Yevich & Pilson, 1980), this association may be more indicative of the energetic costs of sperm release (spawning), rather than sperm production. Nutrient limitation has significant negative effects on multiple ejaculate traits (Macartney et al., 2019), reflective of the costs of sperm release. Further analysis of the association between energetic budget and sperm density throughout gametogenesis will provide improved understanding of temporal shifts in energetic allocation to reproduction.

In addition to sperm density, carbohydrate concentration was also positively associated with two metrics of immunity: melanin concentration and antibacterial activity. These associations are likely indicative of the high costs of both of these immune metrics. Melanin production is the result of a complex pathway; beginning with pathogen recognition and continuing through a series of protein cascades leading to the production of melanin (Cerenius et al., 2010; Palmer, Mydlarz & Willis, 2008). The complexity of the pathway is likely associated with high metabolic costs. To such end, previous studies have documented negative associations between lipid reserves and melanin concentration following immune stimulation (Sheridan et al., 2014). Similarly, antibacterial activity as measured here is the result of the action of many unique compounds. Corals and other cnidarians secrete a wide diversity of often complex antibacterial compounds (Mitchell et al., 2019; Mydlarz et al., 2016; Palmer & Traylor-Knowles, 2012), which may require significant energy to produce. In contrast, other metrics of immunity measured here, specifically antioxidants, are single molecules/compounds, and likely require significantly less metabolic input (Traylor-Knowles & Connelly, 2017). The low cost of investment in these components could explain the lack of strong correlation between our metrics of energetic budget and these immune parameters.

Finally, when considering the factors driving variation in immunity using both a multivariate and univariate approach, we found limited evidence for any association between sperm density and immunity. A moderate positive association between total phenoloxidase and sperm density was observed using correlational analyses. There is conflicting evidence regarding the link between reproduction and phenoloxidase in other invertebrates, with studies indicating both positive and negative associations (Castella, Christe & Chapuisat, 2009; Guo et al., 2021). Further study of immune-reproduction associations in cnidarians will help clarify this relationship.

In total, our results were in contrast with our hypothesis; we observed no significant trade-offs between constitutive immunity and reproduction. We can posit several hypotheses to explain these observations. First, the timing of our experiment may not have properly reflected peak reproductive investment. Gametogenesis in A. poculata begins in early March-April and continues through June and July, with colonies spawning in early August/September (Szmant-Froelich, Yevich & Pilson, 1980). While we observed some association between sperm density and carbohydrate concentration indicative of resource allocation, is possible that this allocation may have been higher at earlier points of gametogenesis, as opposed to at the time of sampling in August (Szmant-Froelich, Yevich & Pilson, 1980). Consequently, the most pronounced trade-offs would have occurred earlier in the season, and tapered off by the point of sampling as resources availability increased and more resources could be allocated to immunity. Second, we chose here to measure sperm density due to sampling logistics. Sperm production is typically considered to be less energetically costly than production of eggs (Hayward & Gillooly, 2011; Parker, 1970; Parker, 1982). While we did observe some association between energetic resources and sperm, it is likely that these associations are more pronounced in females and consequently reproductive trade-offs would be more evident when considering females. Finally, our study does not account for multiple spawning events, specifically spawning events that may have occurred naturally prior to sample collection. Some coral species spawn multiple times throughout a season, and exact timing of Astrangia spawning in natural environments is unknown (Szmant-Froelich, Yevich & Pilson, 1980). Thus, it is possible that the collected Astrangia colonies had previously spawned in situ, and the induced events were a secondary release with reduced sperm density not reflective of true reproductive investment. Congruent with this possibility, several of our colonies had negligible measured sperm density. Future studies combining histology with immunological assays at different points of gametogenesis in both male and female colonies may clarify presence and timing of potential trade-offs.

It is also possible that nuances in our study design did not allow us to observe trade-offs. It must be noted that trade-offs occur in a multi-dimensional trait space filled with competed demands (Lochmiller & Deerenberg, 2000; Stearns, 1989), and our experimental analyses only captured two of these demands (i.e., reproduction and immunity). It is highly likely that more complicated trade-offs, that involve other organismal demands (growth, maintenance, etc.,) may occur and were not captured by our study. Approaches that properly reflect the multi-dimensional trait space associated with resource allocation will be necessary to fully disentangle the relationship between immunity and other costly processes. Additionally, our study measured a limited number of immune metrics; it is possible that trade-offs exist between reproduction and immunity but involve other components of the immune system. The coral immune system is complex, involving many different components (receptors, signaling cascades, cellular and humoral effector responses, etc.), all of which have different relative costs (Colditz, 2008; Ivanina et al., 2018; Lochmiller & Deerenberg, 2000; Seppala & Leicht, 2013). It is therefore reasonable to assume that trade-offs are not equivalent across components, as has previously been observed in other systems (Adamo, Jensen & Younger, 2001; Albery et al., 2019; Gershman et al., 2010; Lochmiller & Deerenberg, 2000). Future studies should incorporate more comprehensive metrics of immunity using methods such as gene expression or proteomics.

Conclusions

In sum, our results fail to document notable associations between variation in symbiont density or energetic budget and metrics of host physiology (reproductive output and immunity). Furthermore, we find no evidence of trade-offs between reproductive output and immune activity. Still, our results are an important first step in broadening general understanding of resource allocation theory in cnidarians, which can be applied to other organisms facing rapid environmental changes. The information provided here provides an important preliminary framework for future studies of immunological trade-offs in marine invertebrates and the potential effects of variation in energetic budget on these patterns.

Supplemental Information