Larval desperation and histamine: how simple responses can lead to complex changes in larval behaviour.

SUMMARY Some marine invertebrate larvae expand the range of settlement cues to which they will respond as they age. How do relatively simple larvae achieve such complex changes in behaviour? Larvae of the Australian sea urchin Holopneustes purpurascens settle and metamorphose specifically in response to a settlement cue, dissolved histamine, produced by the host alga Delisea pulchra. Older H. purpurascens larvae appear to accept a wider range of host algae, which contain far less histamine than D. pulchra, than newly competent larvae. We tested the hypothesis that older H. purpurascens larvae accept a greater range of host algae by metamorphosing in response to lower concentrations of histamine. We compared the response of newly competent and older larvae to a range of histamine concentrations in settlement assays. Larval age strongly affected the minimum concentration of histamine that induced metamorphosis in H. purpurascens, with older larvae responding to lower concentrations of histamine than newly competent larvae. Older larvae were more sensitive to lower concentrations of histamine yet still maintained a stringent requirement for exposure to histamine in order to metamorphose. In addition, older larvae metamorphosed after shorter exposure periods to histamine than did younger larvae. By using histamine concentration as a proxy for specific habitat cues, H. purpurascens larvae appear to expand their range of settlement preferences with age by simply changing their sensitivity to a single settlement cue. Overall, our results show that marine invertebrate larvae can exhibit surprisingly complex changes in behaviour via simple changes in their response to a single cue.


Introduction Larval desperation and histamine
alga Vaucheria longicaulis (Krug and Manzi, 1999;Krug and Zimmer, 2000), older larvae do not metamorphose indiscriminately. Rather, older, starved larvae of Alderia sp. became more sensitive to settlement cues from the host alga, but older, fed larvae did not (Botello and Krug, 2006). They suggest that the DLH should not apply to host-dependant species because metamorphosis in the absence of such hosts/prey would likely be fatal (Botello and Krug, 2006), and a theoretical consideration of the problem supports this suggestion (Elkin and Marshall, 2007).
Despite the increased interest in the DLH and its apparent applicability to a wide range of species (reviewed in Elkin and Marshall, 2007), we have little idea as to the mechanism of decreased selectivity with respect to settlement cues (i.e. 'desperation'). Desperate larvae are rarely indiscriminate; for example, larvae of the opisthobranch Haminaea callidegenita react to a greater variety of cues as they age; young larvae metamorphose in response to a single inducer but older larvae metamorphose in response to Zostera marina and a green alga but not biofilmed sediment (Gibson, 1995). How do relatively simple larvae increase the range of cues that they react to whilst continuing to reject other cues? To examine this question, we used the lecithotrophic larvae of the Australian sea urchin Holopneustes purpurascens. H. purpurascens is a relatively specialised herbivore that lives on and consumes kelp, Ecklonia radiata, during the adult benthic phase (Steinberg, 1995;Williamson et al., 2004). As a newly settled juvenile, it is found predominantly on the red alga Delisea pulchra and coralline turfing algae Amphiroa anceps and Corallina officinalis (Swanson et al., 2006). The observed distribution of new recruits in the field matched settlement choices in the laboratory, where larvae metamorphosed in response to D. pulchra and coralline algae but not E. radiata (Swanson et al., 2006). H. purpurascens larvae metamorphose specifically in response to dissolved histamine, a naturally occurring settlement cue produced in high quantities by the red foliose alga D. pulchra (Swanson et al., 2006;Swanson et al., 2004). The settlement cue(s) from coralline algae are unknown but seem to be produced by the surface-associated microbial community (biofilm) and may be bacterial histamine (Swanson et al., 2006).
H. purpurascens appears to be a typical species to which the DLH applies: red algae (D. pulchra, A. anceps, C. officinalis) are common hosts of new recruits of H. purpurascens (Swanson et al., 2006), and these algae induced 20-100% metamorphosis of newly competent (6-day-old) larval H. purpurascens . Brown algae (Sargassum vestitum, E. radiata) induce minimal metamorphosis of newly competent larvae; however, as they age, their response to brown algae increases . S. vestitum and E. radiata contain histamine but at much lower levels than the preferred host of new recruits, D. pulchra (Swanson et al., 2006). Thus, older H. purpurascens larvae may expand the range of host algae by metamorphosing in response to lower concentrations of histamine. Alternatively, they may be responding to different cues present in brown algae as they age and use entirely different response pathways. We sought to distinguish among these hypotheses.
We tested the hypothesis that older H. purpurascens respond to a greater range of host algae by reducing the threshold concentration of histamine required for metamorphosis using settlement bioassays. Three batches of larvae were tested against a range of concentrations of dissolved histamine, at competence and each week thereafter for 2-3·weeks, to see if older larvae (1) became more sensitive (responsive) to low concentrations of histamine and (2) showed decreased specificity for histamine as a settlement cue. Recent findings suggest that older larvae may require less exposure to settlement cues than younger larvae to induce metamorphosis (Jackson et al., 2005). Therefore, we also tested the hypothesis that older H. purpurascens required less exposure to histamine than newly competent larvae to induce metamorphosis by exposing larvae at 7, 14 and 21·days old (same batch) to 10·mol·l -1 histamine for different time periods.

Materials and methods
Larval culture Adult Holopneustes purpurascens Agassiz 1872 (30-40) and ambient seawater (SW) were collected from Bare Island (Sydney, Australia) in July, August and October 2004. Sea urchins were held in a constant-temperature room (CTR) at 19°C in two buckets containing 15·litres of ambient seawater with continuous air in a 12·h:12·h light:dark cycle. Urchins were transferred to clean buckets with fresh ambient SW at least every other day and were maintained until they spawned and died. Buckets were checked each morning for spawn. Air was stopped to allow (fertilised) eggs to float to the surface, where they were gently collected in a 200·ml glass dish and poured into three autoclaved 2-litre beakers. Eggs were rinsed three times in sterilised SW containing antibiotics (22·g·l -1 penicillin G, 37·g·l -1 streptomycin sulfate, SSW) and were cultured in the CTR outlined above. Cultures were aerated continuously and SSW was changed every other day. Batches yielding sufficient numbers of hatched larvae (~2000) were kept for settlement assays over ensuing weeks. Each batch of larvae was of mixed parentage, with parents also differing between batches.

Settlement assays
All assays were done in the CTR in 36-mm sterile Petri dishes with 4·ml SSW under static conditions. Replicates were randomly assigned among treatments. Treatment and control dishes were prepared by first adding aliquots of concentrated stock solutions of histamine in SSW followed by the addition of SSW to bring histamine concentration to the desired test concentration. Larvae were added once all Petri dishes were prepared. Larval H. purpurascens reach competence (developmentally ready for metamorphosis, recognised by the presence of five well-developed tube-feet) at around 6 days old (6 days post fertilisation).

Dose-response of larvae of different ages
The dose-response of H. purpurascens larvae was investigated to see if older larvae become more sensitive to lower concentrations of histamine as an inducer of metamorphosis. Assays were done with three batches of larvae (A, B and C) at 7·days old (i.e. newly competent larvae), 14·days old (~1·week post initial-competence), 21·days old (~2·weeks post initial-competence) and 28·days old (~3·weeks post initialcompetence; batch A only). Histamine was tested at 1·mol·l -1 , 100·nmol·l -1 and 10·nmol·l -1 in each assay as an inducer of metamorphosis. Histamine was tested at 10·mol·l -1 as a positive control because this concentration induced maximal metamorphosis of larvae in previous studies (Swanson et al., 2004;Swanson et al., 2006). SSW was tested in all assays as a measure of spontaneous metamorphosis. Twenty-five larvae from batches A and B were used per treatment (five dishes with five larvae), and 100 larvae from batch C were used per treatment (10 dishes with 10 larvae). Percent metamorphosis was scored at 1, 24, 48 and 72·h.
A repeated-measures ANOVA was used to examine the effects of larval age on the sensitivity of larvae to lower concentrations of histamine as an inducer of metamorphosis (SYSTAT ® 10.0 for Windows). Only histamine concentrations of 1·mol·l -1 , 100·nmol·l -1 and 10·nmol·l -1 were included in the analysis because 10·mol·l -1 histamine induced maximal metamorphosis and SSW induced minimal metamorphosis, regardless of larval age. Histamine was treated as a categorical factor rather than a co-variate because of the small range of values but the outcome of the analysis was the same in either case (D.J.M., unpublished analysis). Finally, values were pooled across all three experimental runs (i.e. batches A, B and C) with 7-day-old, 14-day-old and 21-day-old larvae for the analysis. We also applied the alternative approach of analysing each run separately, but the results were not qualitatively different and we therefore opted for the most parsimonious approach of pooling batches. Note that we first tested for interactions between run and the factors of interest but, as these were non-significant and of no biological interest, they were omitted from the final model following Quinn and Keough (Quinn and Keough, 2002). Data from batch A were analysed separately to include the response of 28-day-old larvae. For both analyses, Greenhouse-Geisser adjusted P-values were used as ⑀<0.75 (Quinn and Keough, 2002). The use of Greenhouse-Geisser P-values represents a conservative approach that takes into account the fact that the assumption of 'sphericity' for the repeated-measures ANOVA was not met in our analyses [Quinn and Keough (Quinn and Keough, 2002), p. 337].
Effect of duration of exposure to histamine on inducing metamorphosis H. purpurascens larvae were exposed to 10·mol·l -1 histamine for varying time periods at 7·days old, 14·days old and 21·days old to determine the duration of exposure that was required to induce settlement, a reversible process, and metamorphosis, an irreversible process, in larvae of different ages. Histamine was tested at 10·mol·l -1 in these experiments as this concentration induced maximal settlement/ metamorphosis of larvae in previous studies (Swanson et al., 2004;Swanson et al., 2006). Herein, settlement is defined as the attachment of larvae to a surface via tube-feet, and metamorphosis is defined as the morphological transformation to the juvenile form. It was noted that attachment of H. purpurascens larvae immediately triggered the retraction of the larval body and the partial extrusion of juvenile spines and podia from the vestibule. However, these initial morphological changes in settled larvae were potentially reversible if they were removed from histamine exposure and placed in SSW. That is, a proportion of settled larvae reverted back to swimming larvae when placed in SSW depending on the duration of histamine exposure.
Seven-day-old larvae were exposed to 10·mol·l -1 histamine for 15·min, 20·min, 30·min, 45·min, 1·h, 2·h, 3·h, 4·h and 5·h, and percent settlement was scored before transferring individuals to SSW. Percent metamorphosis of individuals was scored (in SSW) 24·h after commencing the experiment. Larvae were continuously exposed to 10·mol·l -1 histamine or SSW for 24·h in control treatments. A further set of procedural control dishes was included, transferring larvae from 10·mol·l -1 histamine dishes to another set of 10·mol·l -1 histamine dishes at each exposure time, to determine if the transfer process alone affected the metamorphosis of larvae. The assay was repeated with 14-day-old and 21-day-old larvae; however, not all exposure times could be tested due to limited numbers of larvae. Larvae that were 14·days old were exposed to 10·mol·l -1 histamine for 1·h, 2·h, 3·h, 4·h and 5·h, while 21-day-old larvae were exposed to histamine for 30·min, 45·min, 1·h and 2·h (shorter times were selected as the majority of older larvae were metamorphosing after 1·h exposure). Fifty larvae were used for each treatment (five dishes with 10 larvae). From these data, a 'reversion score' was calculated, which is the difference between the proportion of larvae that were metamorphosed at 24·h (in SSW) and the proportion of larvae deemed to have settled during exposure to 10·mol·l -1 histamine, before transferring to SSW. Thus, a positive score indicates that all larvae that had settled during exposure to histamine metamorphosed after transfer to SSW and that additional larvae also metamorphosed in the 24·h period. Conversely, a negative score indicates that a proportion of larvae that had settled during histamine exposure did not metamorphose after transfer to SSW but reverted back to swimming larvae. A two-factor (age = fixed factor, exposure time = fixed factor) analysis of variance (ANOVA) was used to examine the effects of larval age and exposure time on the induction of metamorphosis in H. purpurascens. The 'reversion scores' of newly competent larvae (7·days old) and older larvae (21·days old) in the 30·min, 45·min, 1·h and 2·h exposure treatments were compared; only these treatments were analysed as not all exposure times were tested with larvae of different ages.
Larval age strongly affected the minimum concentration of histamine that induced metamorphosis in H. purpurascens, with older larvae responding to lower concentrations of histamine than newly competent larvae (Table·1, Fig.·1). Older larvae metamorphosed in greater numbers than newly competent larvae in response to low concentrations of histamine (Table·1, Fig.·1). Because batch A had more time periods (7-day-old, 14day-old, 21-day-old and 28-day-old) than the other batches, we also analysed this batch separately to include the response of 28-day-old larvae (Table·2). Older larvae from batch A also metamorphosed at lower concentrations than newly competent larvae (Fig.·1); however, in the analysis, this interaction was obscured by a significant time ϫ age ϫ concentration interaction (Table·2).

Effect of duration of exposure to histamine on inducing metamorphosis
Larval age strongly affected the induction of metamorphosis of H. purpurascens larvae, with older larvae requiring less exposure to 10·mol·l -1 histamine than younger larvae to induce metamorphosis (Table·3). Over 60% of newly competent larvae (7·days old) had settled after 20·min exposure to 10·mol·l -1 histamine (Fig.·2A). However, most settled larvae reverted back to swimming larvae once removed to SSW when exposed to histamine for 1·h or less, as indicated by the negative reversion scores (approximately -0.6) (Fig.·3A). Following 2·h exposure to 10·mol·l -1 histamine, approximately half of the settled larvae metamorphosed into juveniles after transfer to SSW. Newly competent larvae required 3·h continuous exposure to 10·mol·l -1 histamine to induce metamorphosis of >90% of settled larvae after transfer to SSW (Fig.·2A,·Fig.·3A). Conversely, most 14-day-old larvae that had settled during 1·h (or longer) exposure to 10·mol·l -1 histamine metamorphosed after transfer to SSW, generating reversion scores close to zero (Fig.·2B,·Fig.·3B). Although fewer 21-day-old larvae overall had settled after 30-45·min exposure to 10·mol·l -1 histamine, most settled larvae metamorphosed after transfer to SSW (Fig.·2C,·Fig.·3C). Larvae of all ages that were continuously exposed to 10·mol·l -1 histamine were metamorphosed after 24·h. None of the 7-day-old or 14-day-old larvae had metamorphosed in control SSW after 24·h; however, approximately 10% of 21-day-old larvae had metamorphosed in control SSW after 24·h. All larvae in the procedural control treatment (transferred from 10·mol·l -1 histamine to 10·mol·l -1 histamine) were metamorphosed.

Discussion
Larval age strongly affected the minimum concentration of histamine that induced metamorphosis in H. purpurascens, with older larvae metamorphosing in response to much lower concentrations than younger larvae. Our study is the first of its kind to directly examine the effect of a pure form of a naturally occurring settlement cue on the dose-response curve of metamorphosis in young versus older larvae. This study supports the observations of Botello and Krug, who found that older larvae were more sensitive to lower concentrations of settlement cue (present in extracts of the alga V. longicaulis) than younger larvae (Botello and Krug, 2006). H. purpurascens joins a growing list of species that show shifts in the response of larvae as they age (reviewed in Elkin and Marshall, 2007). Given that H. purpurascens recruits are found on a number of different algae (Swanson et al., 2006), it seems reasonable to assume that this species, whilst having settlement preferences, is not a strict specialist at settlement. Thus, whilst larvae that settle and metamorphose on preferred host algae may have higher fitness, larvae that settle and metamorphose on other algae (i.e. brown algae) in the habitat may still survive. The shift to a broader range of host algae to facilitate metamorphosis of aging non-feeding larvae is predicted by a theoretical model (Elkin and Marshall, 2007). As larvae accumulate direct (planktonic mortality) and indirect (reduced energetic stores) costs in the plankton, the benefits of metamorphosing in a suboptimal habitat increase (Elkin and Marshall, 2007).
Despite older H. purpurascens larvae reacting to lower and . Data from 7-day-old, 14-day-old and 21-day-old larvae were pooled from three batches (N=150 larvae per treatment); however, data shown for 28-day-old larvae are from batch A only (N=25 larvae per treatment).
lower concentrations of histamine, only a very small proportion (<5%) of very old larvae (28-day-old) metamorphosed spontaneously in the absence of any algal cues. Thus, older larvae do not metamorphose indiscriminately but become more sensitive to the settlement cue. Given that histamine is found at lower concentrations in other algae in the habitat, this increased sensitivity is likely to result in older larvae accepting a broader range of host algae. Our results appear therefore to be a novel example of expanding settlement preferences by becoming more responsive to a single settlement cue, which can occur here because of quantitative differences in the concentration of histamine among algae in the field (Swanson et al., 2006). By using histamine concentration as a proxy for a general habitat cue, complex changes in the settlement behaviour of older larvae can result from a simple change in their response to a single cue. A number of studies have found that older polychaete, gastropod and abalone larvae metamorphose faster than younger larvae (Barlow, 1990;Botello and Krug, 2006;Knight-Jones, 1953;Pechenik and Cerulli, 1991). Similarly, older larvae of Haliotis asinina metamorphosed after shorter exposure periods to inductive algae than did younger larvae (Jackson et al., 2005). H. purpurascens larvae settled rapidly when exposed to 10·mol·l -1 histamine regardless of age (R.L.S., personal observation). Some morphological changes were evident in settled larvae such as the retraction of the larval body and the partial extrusion of juvenile spines and podia from the vestibule (R.L.S., personal observation). These initial changes in morphology were reversible in a proportion of younger larvae when exposed to histamine for less than 3·h, which suggests that components of the morphogenetic pathway were not activated sufficiently to trigger metamorphosis. Over   90% of 7-day-old larvae had settled after 45·min exposure to 10·mol·l -1 histamine; however, most of these reverted back to swimming larvae when placed in SSW. Newly competent larvae required 3·h of continuous exposure to 10·mol·l -1 histamine to induce metamorphosis of all settled larvae. On the other hand, 21-day-old larvae required less than 1·h of exposure to 10·mol·l -1 histamine to trigger metamorphosis of all settled larvae. Thus, older larvae are induced to metamorphose soon after exposure to the settlement cue whereas younger larvae appear to be more flexible at the time of settlement. Newly competent H. purpurascens larvae that settle on an alga have the potential to reject the site for a limited period after attachment if the settlement cue is no longer perceived [for an alternative process in ascidian larvae, see Jacobs et al. (Jacobs et al., 2006)].
The increased sensitivity of H. purpurascens larvae to histamine occurred gradually with age, suggesting that there is a progressive decrease in the stimulus-threshold required to induce an aging larva to metamorphose, as noted for other invertebrate larvae (Coon et al., 1990;Crisp, 1974;Gibson, 1995;Knight-Jones, 1953;Pechenik, 1980;Rittschof et al., 1984). The mechanisms in older larvae that lead to increased sensitivity to histamine and a reduction in exposure time required for metamorphosis are unclear but are likely to have the same basis. Jackson et al. refer to 'competence factors' reaching a critical level in order for larvae to attain competency (Jackson et al., 2005). These factors may include chemoreceptors (Trapido-Rosenthal and Morse, 1986), components of internal signalling pathways (Clare et al., 1995;Knight et al., 2000) or transcription factors (Jackson et al., . The reversion scores of (A) 7-day-old, (B) 14-day-old and (C) 21-day-old Holopneustes purpurascens larvae after exposure to 10·mol·l -1 histamine for varying time periods. The reversion score was calculated as the difference between the proportion of larvae that had metamorphosed in SSW, 24·h after commencing the histamine exposure, and the proportion of larvae deemed to have settled during exposure to 10·mol·l -1 histamine before transfer to SSW. Thus, a negative score indicates that a proportion of settled larvae reverted back to swimming larvae after transfer to SSW. A positive score indicates that all settled larvae metamorphosed in SSW along with additional larvae over 24·h. *, not tested; **, all settled larvae metamorphosed.
2005). These competence factors may continue to accumulate in older larvae, meaning they are 'primed' to respond to settlement cues, both to lower concentrations and more quickly than younger larvae (Jackson et al., 2005). For example, changes in endogenous levels of neurotransmitters may affect sensitivity to cues. Catecholamines, such as dopamine and its precursor L-DOPA, appear to modulate competency and control metamorphosis in the gastropods Crepidula fornicata (Pechenik et al., 2002;Pires et al., 2000b) and Phestilla sibogae (Pires et al., 2000a). Increasing the endogenous dopamine concentrations of competent P. sibogae larvae made them more sensitive to the settlement cue present in coral extract [Porites compressa (Pires et al., 2000a)]. In the gastropod Ilyanassa obsoleta, metamorphosis only occurs following the removal of endogenous levels of nitric oxide (NO), a specific inhibitor of metamorphosis (Leise et al., 2001). Bath application of serotonin induced metamorphosis of I. obsoleta but not in the presence of two NO donors, and metamorphosis of I. obsoleta was induced, in the absence of serotonin or any other inducer, when endogenous NO production was inhibited (Leise et al., 2001). Other studies suggest that internal energy reserves also affect the sensitivity to settlement cues. Botello and Krug found that unfed larvae were more sensitive to settlement cues than fed larvae (Botello and Krug, 2006), and Marshall and Keough found that larger larvae (with more nutritional reserves) can delay settlement in the absence of cues for longer than smaller larvae . Ultimately, each of these factors may also contribute to the increased sensitivity of older H. purpurascens larvae to histamine. However, it seems that the expansion of settlement preferences in older H. purpurascens larvae  is based on simple changes in their response to a single settlement cue.