Evidence of Compositional and Ultrastructural Shifts During the Development of Calcareous Tubes in the Biofouling Tubeworm, Hydroides Elegans

Journal of Structural Biology 189 (2015) 230–237

Contents lists available at ScienceDirect

Journal of Structural Biology

journal homepage: www.elsevier.com/locate/yjsbi

Evidence of compositional and ultrastructural shifts during the development of calcareous tubes in the biofouling tubeworm, Hydroides elegans

Vera Bin San Chan a, Olev Vinn b, Chaoyi Li a, Xingwen Lu c, Anatoliy B. Kudryavtsev e,f, ⇑ J. William Schopf d,e,f, Kaimin Shih c, Tong Zhang c, Vengatesen Thiyagarajan a, a The Swire Institute of Marine Science and School of Biological Sciences, The University of Hong Kong, Pokfulam, Hong Kong Special Administrative Region b Department of Geology, University of Tartu, Ravila 14A, 50411 Tartu, Estonia c Department of Civil Engineering, The University of Hong Kong, Pokfulam, Hong Kong Special Administrative Region d Department of Earth, Planetary and Space Sciences, Molecular Biology Institute, University of California, Los Angeles, CA 90095, USA e PennState Astrobiology Research Center, University Park, PA 16802, USA f University of Wisconsin Astrobiology Research Consortium, Madison, WI 53706, USA article info abstract

Article history: The serpulid tubeworm, Hydroides elegans, is an ecologically and economically important species whose Received 14 October 2014 biology has been fairly well studied, especially in the context of larval development and settlement on Received in revised form 8 January 2015 man-made objects (biofouling). Nevertheless, ontogenetic changes associated with calcareous tube com- Accepted 9 January 2015 position and structures have not yet been studied. Available online 16 January 2015 Here, the ultrastructure and composition of the calcareous tubes built by H. elegans was examined in the three early calcifying juvenile stages and in the adult using XRD, FTIR, ICP-OES, SEM and Raman spec- Keywords: troscopy. Ontogenetic shifts in carbonate mineralogy were observed, for example, juvenile tubes con- Hydroides elegans tained more amorphous calcium carbonate and were predominantly aragonitic whereas adult tubes Biomineralization Biofouling were bimineralic with considerably more calcite. The mineral composition gradually shifted during the Calcification tube development as shown by a decrease in Sr/Ca and an increase of Mg/Ca ratios with the tubeworm’s Serpulid age. The inner tube layer contained calcite, whereas the outer layer contained aragonite. Similarly, the Tube ultrastructure tube complexity in terms of ultrastructure was associated with development. The sequential appearance of unoriented ultrastructures followed by oriented ultrastructures may reflect the evolutionary history of serpulid tube biominerals. As aragonitic structures are more susceptible to dissolution under ocean acidification (OA) conditions but are more difficult to be removed by anti-fouling treatments, the early developmental stages of the tubeworms may be vulnerable to OA but act as the important target for biofouling control. Ó 2015 Elsevier Inc. All rights reserved.

1. Introduction and organic matrix, both of which are produced primarily by calcium-secreting glands located in the collar region (Fig. 1), on each Calcifying marine tubeworms of the family Serpulidae (phylum side of the peristomium and the ventral epithelial mucous cells Annelida: class Polychaeta: order Sabellida) represent a unique (Hedley, 1956a,b, 1958). The serpulid tubes show enormous struc- taxa for calcification and biomineralization study (Chan et al., tural diversity and organizational complexity (Smith et al., 2013; 2012). The serpulids are the only taxa of marine polychaetes that Vinn and Kupriyanova, 2011; Vinn et al., 2008c). Aggregation of produce exclusively calcium carbonate skeletons and control bio- this species provide a three-dimensional ecosystem that can medi- mineralization like many other marine calcifiers (Simkiss and ate undesirable physical environments, providing shelter, food and Wilbur, 1989). Serpulid tubes is composed of calcium carbonate substrates for many benthic species. Therefore, tubeworms are rec- ognized as the ‘‘ecosystem engineers’’ which enhance biodiversity, especially in temperate regions (Poloczanska et al., 2004). Despite ⇑ Corresponding author at: The Swire Institute of Marine Science and School of Biological Sciences, Rm: 3N-20, Kadoorie Building, The University of Hong Kong, their ecological importance and complexity, our understanding of Pokfulam Road, Hong Kong Special Administrative Region. the serpulid mineralogy is confined to a few extant species E-mail address: [email protected] (V. Thiyagarajan). (Bornhold and Milliman, 1973; Smith et al., 2013; Vinn et al., http://dx.doi.org/10.1016/j.jsb.2015.01.004 1047-8477/Ó 2015 Elsevier Inc. All rights reserved. V.B.S. Chan et al. / Journal of Structural Biology 189 (2015) 230–237 231

2008c). Additionally, the focus of existing literatures have mainly 2. Materials and methods focused the adult stages, and the early stage of serpulid mineralization has hardly been addressed. 2.1. Tube sample collection It still remains a challenging goal to describe the mechanisms of biomineralization, i.e. how mineral ions concentrate and Mature adult H. elegans (tubeworms) were collected from a fish nucleate to form the final mineral products (Dorvee and Veis, farm in Yung Shue O (22°270N, 114°230W) in Hong Kong. Sperms 2013). However, by studying the mineral succession pattern in and oocytes were obtained from more than 100 individuals to pro- the early calcification events, the intermediate stages of calcifica- duce a sufficient genetic diversity in the study population. Embryos tion may be captured and may provide insights into the crystal- were divided into four groups and were maintained independently lization pathway. At metamorphosis, the free-swimming larvae using standard larval culture procedures in 5 L culture containers of serpulid worms undergo a pelagic-benthic transformation (until 25 day post-metamorphic stage). To minimize the influences when they attach to suitable substrates and actively produce to Sr/Ca and Mg/Ca caused by physical conditions (Schöne et al., CaCO3. The calcium secreting glands control the calcification pro- 2013; Shen et al., 1996; Klein et al., 1996), the independent culture 2+ cess (Bubel, 1983; Hedley, 1956a,b), using HCO3 and Ca as the containers were maintained in uniform physical conditions with calcification reactants (Roleda et al., 2012). The early calcification ambient seawater (23 °C, 33ppt, pH 8.1, initial density of 10 products commonly involve complex succession, for example, embryos per mL) (Qian, 1999). Under these optimal culture condi- formation of the larval spicules in sea urchins and the early lar- tions, larvae were ready to attach and metamorphose (compe- val shells in mussels, reorganized from highly the disordered tency) 6 days after fertilization. Natural biofilm was developed amorphous calcium carbonate (ACC) phase into mature crystal- for 7 days on acrylic plates in unfiltered seawater, and were line minerals (Beniash et al., 1997; Medakovic´, 2000; Addadi deployed as substrates to stimulate attachment, metamorphosis et al., 2003; Politi et al., 2008). The structure of the ACC lacks and initiation of post-metamorphic calcification (Pechenik et al., any long-range order, but commonly contains short-range order 2007). Tubeworms were maintained for 25 days post-metamor- of the designated mineral product, i.e. calcite or aragonite. While phosis in optimal culture conditions. Seawater and algal food, the use of an ACC precursor seems to be a common strategy in Isochrysis galbana,(105 cells mL1, provided ad libitum) were marine invertebrate biomineralization (Weiner and Addadi, renewed every other day during the larval phase; and renewed 2011), this phenomenon is yet to be demonstrated in the serpu- every 2–3 days during the post-metamorphic stage (Qiu and lid calcification model. Qian, 1997). Tube samples were collected by removing one plate Elevated anthropogenic CO2 in the seawater is causing ocean from each independent culture container (n = 3–4 replicates) at acidification, a trend that will continue in the coming centuries four sampling times. Tube materials were also collected at 4, 11, (Feely et al., 2008). In current and future coastal waters, the serpu- day 18 and 25 days post-metamorphosis. Tube materials collected lids may be able to produce their tubes under elevated CO2 envi- on day 4 (post-metamorphosis) represented the earliest mineral ronments (Lane et al., 2013). However, recent studies suggest bearing post-metamorphic stage because tube calcification begins that aragonite-producing organisms are more susceptible to OA at the collar region immediately after completion of attachment due to the higher solubility of aragonite and ACC compared to cal- and metamorphosis (Fig. 1; Hedley, 1958), whereas samples col- cite (Andersson et al., 2008; Beniash et al., 1997; Ragazzola et al., lected on day 25 represented mature adults. The collected tubes 2012; Ries, 2011a). Determining the mineral composition and were rinsed with MilliQ water to remove adhered salts and then structural changes of calcified tubes during development in serpu- dehydrated at room conditions prior to biomineral analyses. The lid tubeworms, therefore, is a crucial step to understand their sur- spectroscopic analyses of the post-metamorphic changes in the vival in the future oceans. The tubeworm, Hydroides elegans,isan mineral composition and ultrastructures of the tubes were com- ecologically and economically important species in warm water pleted within 3 months. areas because it is a dominant biofouler and ubiquitous reef-forming species in the benthic habitat (Nedved and Hadfield, 2009). By culturing a population of H. elegans, the initial stages of calcifica- 2.2. Aragonite/calcite analysis by X-ray diffraction (XRD) tion and the succession of mineral features during development were examined using Fourier transform infrared spectroscopy, X- To compare of the calcite/aragonite ratio at the different post- ray diffraction and coupled plasma-optical emission spectropho- metamorphic stages, 10–20 individual tubeworms were detached tometry. The ultrastructures of tube minerals were documented from the substrate of each replicate (n = 3–4) and submerged in by scanning electron microscopy, and the locations of biominerals 5% bleach (NaOCl, Clorox™) for 30 min to remove soft tissues. In were observed using a Raman spectroscopy. total there were 12 samples (4 stages 3 replicates). The tubes were rinsed twice with MilliQ water and air-dried before being ground into a fine powder for X-ray diffraction (XRD). The XRD pattern of each sample was collected on a Bruker D8 Advance X-ray powder diffractometer equipped with a Cu Ka radiation and a LynxEye detector (Bruker, Germany). The system was calibrated with Standard Reference Material 660a (lanthanum hexaboride,

LaB6, National Institute of Standards and Technology, USA) for the line position. The diffractometer operating parameters were 40 kV, 40 mA, a 2h scan range from 10° to 110°, with a step size of 0.02° and a scan speed of 0.3 s/step. Calcite and aragonite were identified by matching the powder XRD patterns with those retrieved from the standard powder diffraction database of the International Centre for Diffraction Data (ICDD PDF-2 Release 2008) using Eva XRD Pattern Processing software (Bruker Co., Ltd.). The ratio of each mineral was quantified and processed by Fig.1. Light microscopy image of the post-metamorphic tubeworm, Hydroides elegans. The early tube materials were used for biomineral analyses. The collar the TOPAS (version 4.0) crystallographic program using the Riet- region, i.e. the tube forming region, is labeled with a dark arrow. veld refinement method (Larson and Von Dreele, 1994). 232 V.B.S. Chan et al. / Journal of Structural Biology 189 (2015) 230–237

2.3. Amorphous calcium carbonate analysis by Fourier transform objectives of 50 or 100 magnification. For each region, Raman infrared (FT-IR) spectroscopy spectra were obtained at each spot (2 lm 2 lm) to generate a mineralogy map for each of the (1) posterior region (76 lm 6 lm, The relative amorphous calcium carbonate (ACC) contents of 20 s scan) and (2) anterior region (102 lm 6 lm, 10 s scan). Each the tube samples were determined using FT-IR spectral analysis. spectrum was obtained in the range 150–2940 cm1, with the win- For the FTIR spectra, dried and ground tube samples from the 3 dow centered at 1650 cm1, covering the major peaks necessary to or 4 replicates per developmental stage were prepared as a KBr identify calcite and aragonite phases (Kontoyannis and Vagenas, pellet (KBr dehydrated at 98 °C overnight, 13 mm diameter pellet, 2000). The typical laser power of 8–21 mW at each spot was well 9 tons for 2 min). Infrared absorption spectra were obtained using below the threshold for causing radiation damage to the specimens a Fourier transform infrared spectrometer (FT-IR, L120-000B, Per- (Taylor, 2008). The region analyzed in the Raman mapping was kin Elmer, USA) to record spectra in the range of 500–2000 cm1 imaged at 20 and 4 magnification under a bright field micro- with 1 cm1 resolution. Infrared spectra were baseline corrected scope to correlate with the scanning electron microscope (SEM) 1 before measuring the peak heights at 855 cm (m2) and analysis of the tube ultrastructures. 1 713 cm (m4), which correspond to the internal vibration modes of the carbonate ions. The intensity ratio of m2 and m4 indicated 2.6. Tube ultrastructure detected by scanning electron microscopy the relative amorphous contents (Beniash et al., 1997; Long et al., 2012). Therefore, during the transformation from ACC into crystal- Tube sections near the tube opening were observed using scan- line minerals, the relative quantities of amorphous calcium carbon- ning electron microscopy (SEM). Intact tubes were first secured in ate (ACC) was quantified using the infra-red absorption ratios 4% agar and then dislodged from the acrylic plates before undergo- between the m2 and the m4 carbonate peaks (Beniash et al., 1997). ing dehydration and resin infiltration steps. The tube samples were treated with 70% ethanol for 2 h; 95% ethanol for 2 h; 3 100% 2.4. Elemental ratio analysis using inductively coupled plasma-optical ethanol for 2 h; 3 propylene oxide for 2 h; propylene oxide: resin emission spectrophotometry (ICP-OES) (3:1) for 4 h; propylene oxide: resin (1:1) for 4 h; propylene oxide: resin (1:3) for 4 h; 3 resin for 8 h), and finally embedded in resin Quantification of Sr, Mg and Ca contents in the tubes were per- polymerized at 60 °C in an oven. Tubes were manually sectioned formed using ICP-OES, and the contents of Sr and Mg in the tube across the tube opening using histological blades. The resin around were normalized to the CaCO3 mass, i.e. the quantity of Ca. Ele- the cross-sectional surface was trimmed away before subsequent mental ratios of magnesium/calcium (Mg/Ca) and strontium/cal- ultramicrotome sectioning. The resin-embedded tubes were man- cium (Sr/Ca) in the tubes were quantified for each sample at ually trimmed with a glass knife to remove excess resin. Finally, each developmental stage. Tubeworms (8–10 samples) were dis- the sectional surfaces were smoothed by sectioning with a dia- lodged from each acrylic plate and were immersed in a 5% bleach mond knife (DiATOME Ultra 45°), using an ultramicrotome (Ultra- (NaOCl, Clorox™) for about 30 min to remove organic soft tissues. cut S, Leica, Germany). The sectioned specimens were etched for The tubes were twice rinsed with double distilled water to remove 1 min in 0.5 M EDTA solution to reveal the tube ultrastructures. excess bleach. The mineral content was digested in 6 mL of 2% This embedding, sectioning and etching procedure allowed good nitric acid. The analytes were characterized for their relative inten- quality cross sections that allow most tube ultrastructures to be sities for calcium (Ca, 396.847 nm), magnesium (Mg, 285.213 nm) clearly distinguished (Vinn et al., 2008c), the mineral integrity and strontium (Sr, 407.771 nm) using inductively coupled plasma- was similar to the fractured samples (Chan et al., 2012). The sec- optical emission spectrophotometry (ICP-OES, PE Optima 8300). tioned specimens were mounted sectioned side up, onto aluminum For each metal, the analytes were prepared at different dilutions stubs with carbon tape. The surrounding embedded resin surfaces (1-fold, 10-fold, 100-fold dilutions), to quantify them within the were covered with silver paint to minimize electron charging, and concentration range of the standard calibration curve (Andreasen the sectional surfaces were then sputter coated with a 50 nm thick et al., 2006). All containers used in the sample analysis were gold–palladium alloy. To allow comparison of the ultrastructures soaked in a 10% v/v HCl acid bath overnight, twice rinsed with dou- with the spatial distribution of mineralogy detected by Raman ble-distilled water and dried at 80 °C. spectroscopy, the same specimen that was used in the Raman analysis was used for SEM analyses. Tube ultrastructures were visual- 2.5. Spatial distribution of aragonite and calcite by Raman ized at 5 kV using field emission SEM (Leo 1530, Germany). spectroscopy

2.7. Statistical analysis The specific location of aragonite and calcite in the thicker bimineralic tubes of the adult worm was examined using Raman The tube composition, in terms of relative levels of amorphous imaging spectroscopy. Adult tubes at day 25 were dried and calcium carbonate, polymorph composition, Mg/Ca and Sr/Ca mounted on glass slides with Crystalbond™ embedding media. ratios, were measured in different days after the tubes deposition The embedded tubes were sectioned along the long axis using a and growth began at metamorphosis. Their relationship were plot- razor blade under a dissection microscope. The tube section sur- ted in a scatter plot, with each dot representing a measurement faces were cleaned in 5% NaOCl (Clorox™) to remove organic mate- from each independent replicate with multiple individuals. The rials. Two regions on this longitudinal sectional surface were hypothetical linear relationships between the two continuous vari- analyzed according to the reported growth mode of the serpulid ables (tube composition and time) were tested using Pearson cor- tube (Hedley, 1958), where (1) the posterior region represents the relation, when p-value is lower than 0.05 the linear relationship earlier mineralization period, and (2) the anterior region near the was regarded as significant. tube opening represents the more recently accreted materials. The analysis was performed on a T64000 triple-stage laser-Raman system (JY Horiba, Edison, NJ, USA) with macro-Raman and confocal 3. Results micro-Raman capabilities. Reflected white light was focused on the area of interest, and specimens were centered on the path of the The succession in biomineral composition and the ultrastruc- laser beam (excitation at 458 nm) projected through an Olympus tures of the post-metamorphic tubeworm, H. elegans, were charac- BX41 microscope (Olympus, Center Valley, PA, USA) with the terized. There were considerable compositional changes over time V.B.S. Chan et al. / Journal of Structural Biology 189 (2015) 230–237 233 in the construction of tube materials, in terms of the amorphous spherulitic prismatic (SPHP) layer composed the outermost layer calcium carbonate (ACC) content, the calcite and aragonite content, (Fig. 3f and g). In a field collected adult specimen (age unknown, and the elemental ratios of Mg/Ca and Sr/Ca. The alteration in the Supplementary Fig. 3a) with a thicker tube, an additional type of composition can be demonstrated in the correlation relationships oriented ultrastructure known as the lamello-fribrillar (LF) struc- summarized in a scatter plot (Fig. 2). As required by the X-ray pow- ture was observed (Supplementary Fig. 3b and c). The IOP ultra- der diffraction technique, the day 4 tubeworms represented the structures were always found in the innermost layer, whereas earliest stage that produces a characterizable amount of tube the more complex and advanced SPHP structures only appeared materials to distinguish the quantities of calcite and aragonite in the outermost tube layer in the later development stages (e.g., (Supplementary Fig. 1). It can be shown that the juvenile tubes first day 25). The cross (Supplementary Fig. 4a) or longitudinal (Supple- begin with almost entirely aragonite, then the tubes are incorpo- mentary Fig. 4b) sections were similarly effective ways to observe rated with more low Mg-calcite, position of XRD main peak at and distinguish these ultrastructures (Vinn et al., 2008c). The SEM h = 29.6° (Ries, 2011a) as the tubeworms grow (Fig. 2a; r = 0.676, observation did not provide information about the amorphous n = 14, p < 0.01). Among the four independent samples in the ultrastructure, probably because EDTA treatment would lead to a 25 day, the tubeworms produce greater amounts of calcite and ara- loss of ACC minerals, especially in the day 4 tube structures. gonite, with their ratio as high as 1:1. Longitudinally sectioned specimens from the day 25 were ana- The infrared spectra of the juvenile tube structures had two lyzed by Raman spectroscopy to find out the distribution of calcite peaks located at 856 cm1 and 713 cm1, they correspond to the and aragonite. Raman spectra showed two peaks that identify ara- 1 1 1 m2 and m4 absorption bands of the carbonate ions, respectively gonite at 704 cm and 1084 cm and calcite at 716 cm and 1 (Supplementary Fig. 2). The absence of a prominent m4 absorption 1087 cm (Fig. 4e). The mapping results summarize the spatial band from day 4 indicate the production of ACC in the early min- distribution of aragonite and calcite, in red and blue respectively eralization process. The negative correlation between Imaxm2/Imaxm4 (Fig. 4b and d). As can be seen from Fig. 4e the band at ratio and age of tubeworms (Fig. 2b; r = 0.709, n = 14, p < 0.01) 1085 cm1 is noticeably broader in the Raman spectrum of calcite demonstrates that ACC contributes less to the biomineralization versus the one of aragonite, what possibly can be an evidence of process as the tubeworms age. The ratios of Sr/Ca and Mg/Ca are Mg incorporation into the calcite crystal structure. Tube materials proxies for aragonite and calcite contents. The Mg/Ca ratio nearer the posterior tip was entirely aragonite (Fig. 4a and b), increased significantly with age (Fig. 2c, r = 0.605, n = 14, while the anterior side near the wider tube opening has aragonite p < 0.05), corroborating the higher calcite content in the later bio- as the outer layer, and calcite as the inner layer (aragonite in red, mineralization process. On the other hand, the Sr/Ca ratio decline calcite in blue; Fig. 4c and d). This result indicates that the more with age (Fig. 2d, r = 0.622, n = 14, p < 0.05, p < 0.05), similar to recently deposited materials at the anterior side are bimineralic, the observed reduction in aragonite content during the worm with a calcitic inner layer. However, from the same longitudinally development. sectioned sample at the same anterior (Fig. 5d) and posterior The common serpulid ultrastructures were visible after the region (Fig. 5a), the ultrastructures of the two different biomineral etching procedure (Vinn et al., 2008c). The tubes of older individu- materials showed similar IOP ultrastructures (Fig. 5b, c, e and f). als had complex ultrastructures (Cross-sectional views: day 4, The succession of tube ultrastructures is schematically illustrated Fig. 3a; day 11, Fig. 3b; day 18, Fig. 3d; day 25, Fig. 3f). On day 4 using line drawings according to (Vinn et al., 2008c). of biomineralization, only the simplest ultrastructure, round homogeneous crystals (RHC), was produced (Fig. 3a). The irregu- 4. Discussion larly oriented prismatic (IOP) structure was found near the lumen as the innermost layer on day 11 after biomineralization (day 11, The tube mineral in mature Hydroides elegans is bimineralic, Fig. 3c; day 18, Fig. 3e; Fig. 3h). Oriented ultrastructure appeared containing comparable proportion of aragonite and calcite. on day 25, when an oriented ultrastructure known as the Although the spatial distribution of Mg was not examined in

Fig.2. The changes in tube composition of Hydroides elegans over developmental stages (4, 11, 18 and 25 days) is shown in scattered plots, each dot represents an independent measurement from a replicate container for animal maintenance. (a) Imaxm2/Imaxm4 ratio, a proxy of amorphous CaCO3 content; (b) calcite/aragonite ratio; (c) Sr/ Ca ratio; and (d) Mg/Ca ratio. All correlation relationships are statistically signiﬁcant, r- and p-values are shown. 234 V.B.S. Chan et al. / Journal of Structural Biology 189 (2015) 230–237

Fig.3. The identification of tube ultrastructures of tube sections of Hydroides elegans over developmental stages (4, 11, 18 and 25 days). Abbreviations: HM, higher magnification; LM, lower magnification; RHC, round homogeneous crystals; IOP, irregularly oriented prismatic structures; SPHP, structure and spherulitic prismatic structures. (a) Day 4 ultrastructure shows RHC structures; day 11 tube at (b) the lower magnification and (c) the higher magnification views show the outer RHC and the inner IOP layers; day 18 ultrastructure shows (d) the lower magnification and (e) the higher magnification views show the outer RHC and the inner IOP layers; day 25 tube (f) at lower magnification and at higher magnification which shows (g) outer SPHP layer and (h) middle RHC layer and the inner IOP layer. details, an increasing content of Mg may stabilize the ACC phase Therefore, further studies of the different biomineral groups in (Politi et al., 2009), and may increase the hardness of calcite the serpulids may verify the observation that the mineral succes-

(Wang et al., 1997). As MgCO3 and SrCO3 are isostructural to calcite sion observed in H. elegans is a common characteristic of serpulids. and aragonite respectively, incorporation of Mg and Sr into the The developmental pattern of mineralogy can be extended to

CaCO3 lattice structure is proportional to the amThe American compare with the pattern of the serpulid biomineral evolution. Geophysical Unionount of calcite and aragonite present (Mucci, The earliest ancestral serpulids produced aragonitic as the plesio- 1983). Therefore, the pattern of Sr and Mg content normalized to morphic mineral in the Aragonite II sea of the Triassic Era Ca, i.e. Sr/Ca and Mg/Ca ratios, may eventually account for the con- (Stanley, 2006; Taylor, 2008; Taylor and Vinn, 2006; Vinn et al., trol of the calcium carbonate polymorphs, i.e. aragonite and calcite 2008b; Weedon, 1994; Weedon et al., 1994). Calcite producing ser- content over the succession of tube mineralization. pulids possibly appeared later, during the calcite II sea of the Juras- A pattern of biomineral succession over development has been sic and Cretaceous (Vinn pers. obs.). The experimental seawater of reported in a few groups of marine invertebrates. For example, the the present study has an Mg/Ca ratio that is aragonite-inducing. urchin and mollusk larvae are known to have an amorphous pre- The ancestral trait of the aragonite mineralizing phase may exist cursor prior to the formation of calcite and aragonite mineral in all the earliest mineralization events of the serpulids, while cal- (Beniash et al., 1997; Medakovic´, 2000). Larval oysters, for exam- cite production may be an expanded function in during develop- ple, produce aragonitic shells and switch to producing calcite after ment. The question how the serpulid family acquires a diverse metamorphosis in the sessile juvenile stage (Medakovic´ et al., carbonate mineral composition potentially over multiple parallel 1997). The carbonate composition of serpulid tubes shows major events of neofunctionalization would be a future topic of research variability and speciﬁcity among species (Smith et al., 2013). (Cañestro et al., 2007; Smith et al., 2013). Evolutionarily, the Hydroides dianthus, H. norvegicus and H. spongicola share similar appearance of calcite in the innermost layer of H. elegans is inter- compositional features with H. elegans, all producing bimineralic esting, as the worm would have better protected its tube against tubes containing both aragonite and calcite (Ries, 2011a; Tanur dissolution with an external layer of calcite instead. Although the et al., 2009; Vinn et al., 2008a). Unlike H. elegans which produces outer tube surface is not covered by organic materials, the inner low Mg-calcite, H. dianthus produces high Mg-calcite (Tanur tube wall is covered by a layer of organic lining which may count et al., 2009), and the tubes of H. spongicola can contain as much as a strategy that protects the tube lumen against dissolution as 95% aragonite. Mg incorporation into calcite would lead to peak (Tanur et al., 2009). Similar to brachiopods and bryozoans, the broadening in Raman spectroscopy (Bischoff et al., 1985). more advanced serpulids tube ultrastructures with complex V.B.S. Chan et al. / Journal of Structural Biology 189 (2015) 230–237 235

the underlying weaker unoriented structures from external attacks (Vinn and Kupriyanova, 2011), as well as preventing crack propagation that the early stage structures (without an oriented SPHP cover) can be prone to (Vinn and Kupriyanova, 2011). SEM and Raman spectroscopy analyses shows that the mineralogy cannot be easily distinguished using mineral ultrastructures, both calcite and aragonite regions were formed in IOP ultrastructures, similar to observations of H. dianthus (Tanur et al., 2009). In addition, the calcitic lamello-fibrillar (LF) structures can be seen in an H. elegans tube collected from the field (Tanur et al., 2009)(Supple- mentary Fig. 3), which suggests that the calcite content of the tube will increase in the later stages of tube mineralization. Calcifiers demonstrate a mixed response to ocean acidification (Ries et al., 2009), mineralogy is one of the many factors influenc- ing the response of calcifiers (Ries, 2011a,b). Mineralogy of the juvenile tubeworm suggests an inherent susceptibility to ocean acidification during its early life stage, because early juvenile tubes contain a notable amount of amorphous calcium carbonate. The solubility of ACC is higher than crystalline forms of calcium carbonate (Brecˇevic´ and Nielsen, 1989). In addition, the juveniles produce the crystalline carbonate in the form of aragonite which is about 30% more soluble than calcite (Morse, 1983). As observed from the ultrastructure, the early juvenile tubes are built without the thick SPHP, and may be prone to crack propagation (Vinn and Kupriyanova, 2011). Notably, this may be one of the reasons why the majority of H. elegans larvae fail to metamorphose and cal- cify in an environmentally realistic low pH value of 7.7 (Lane et al., 2013; Mukherjee et al., 2013). The succession pattern of mineral products may represent a general phenomenon in serpulid biomineralization. However, the ontogenetic calcification patterns of serpulids may differ between groups that build their tubes with different minerals, using either calcite or aragonite, or with both calcite and aragonite. The mineralization tissues have been described and may play a role in determining the carbonate polymorph in serpulid mineralization (Hedley, 1956b). As calcification in serpulid tubeworms occurs in the collar regions at the anterior end of the tube (Hedley, 1956a), Fig.4. Optical photomicrographs, Raman images, and Raman spectra of longitudi- the changes in mineralization over age may indicate a change in nally sectioned Hydroides elegans tube samples collected on day 25. (a and c) Photomicrographs of younger and older regions, respectively, of a juvenile tube. The collar cells, but how such a shift in the biocalcification mechanism black rectangles denote the areas analyzed by laser Raman imagery. (b and d) Two- towards production of more calcite awaits future research. The dis- dimensional Raman images of the areas from (a) and (c) showing the distribution of tinctive juvenile stages and mechanical properties of their calcare- aragonite (red region) and the distribution of calcite (cyan region). (e) Raman ous structures should be further studied in greater details because spectra of aragonite and calcite measured in this specimen. The Raman images were acquired in a spectral window centered on the secondary Raman lines of aragonite of the potential implications on biofouling (Nedved and Hadfield, (at 704 cm1) and calcite (at 716 cm1), because the major Raman bands of 2009), reef formation (Poloczanska et al., 2004; Sun et al., 2012), aragonite (1084 cm1) and calcite (1087 cm1) are too close to each other. The and geochemical cycling (Haines and Maurer, 1980; Medernach Raman images show the younger region of the tube (a) is composed of aragonite (b) et al., 2000). whereas the more mature region (c) contains both aragonite and calcite. (d) The calcite occurs in the innermost parts of the tube. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) 4.1. Conclusion

The present study demonstrates the transformation of tube orientation are calcitic (Vinn, 2012), which contrasts with mollusks mineral composition and ultrastructures in the calcifying tube- which have a greater number of advanced aragonitic structures worm, H. elegans, through an integral study combining spectro- (Carter, 1990). scopic analysis and high resolution observation. The succession The succession of ultrastructures over age may result from an of tube materials in the early biomineralization of the serpulid increasing ability to produce the particular macromolecules neces- tubeworm is described for the ﬁrst time. The nascent tube materi- sary for more complex ultrastructures. These serpulid tubeworms als produced by juvenile worm contain mostly aragonite and a begin by mineralizing unoriented RHC and IOP layers, followed considerable amount of amorphous calcium carbonate (ACC). The by the oriented SPHP layer in the later stage (Fig. 6). This sequence tube mineralogy undergoes a succession over time, with a decreas- of mineral ultrastructural complexity resembles the order of serpu- ing amount of ACC and an increasing amount of calcite. The inner- lid evolution (Vinn, 2013; Vinn et al., 2008a). The unoriented RHC, most layer deposited in the anterior tube opening of the adult was IOP structures are plesiomorphic characteristics, whereas the calcitic. The elemental measurement of Mg and Sr standardized to ordered SPHP structures belongs to the more taxonomically Ca, suggest that Mg/Ca and Sr/Ca ratios may eventually account for advanced clades. The ability to construct spherulitic structure also the changes in calcite and aragonite amounts, respectively. The signiﬁes the ability to mineralize rapidly (Addadi and Weiner, tube shows a structural advancement with growth, with the ultra- 1992). The SPHP structures have a high packing density to protect structures becoming more oriented and complex (See Fig. 6). 236 V.B.S. Chan et al. / Journal of Structural Biology 189 (2015) 230–237

Fig.5. The correlative ultrastructures study of the longitudinally sectioned Hydroides elegans tube from the day 25 (same individual as studied by Raman spectroscopy), reveal the ultrastructure of (a) the posterior region and (d) the anterior region. The ultrastructures of the outer layer (b) and (e) and the inner layer (c) and (f) showed similar ultrastructure as the IOP structures. IOP, irregularly oriented prismatic structures. The observation indicates the mineralogies are indistinguishable purely from the morphological features.

Acknowledgments

The authors would like to thank Y.Y. Chui at the Prince Philip Den- tal Hospital for providing hard tissue sectioning facilities and ult- ramicrotomy training, HKU-EMU for SEM imaging facilities, VYL Fung for assisting with the FT-IR analysis, and Paul Taylor for his help with the Raman spectroscopy analysis. We also wish to thank Ackley Lane, Gray Williams, David Dudgeon, and Andrew Mount for their Fig.6. The succession of tube morphology is schematically illustrated according to valuable discussions during the course of this project. This study ultrastructure representations. The RHC structure precedes the appearance of the IOP structure, SPHP structure appeared in the later stages of tube development. was funded by three GRF grants from the HKSAR-RGC to V.T (Grant Nos. 780510M, 705511P and 705512P). O.V. is indebted to the Sep- koski Grant (Paleontological Society), Estonian Science Foundation Author’s contributions Grant ETF9064, Estonian Research Council Grant IUT20-34 and Pal- aeontological Association Research Grant for ﬁnancial support. V.B.S.C., C.L. and V.T.: conducted and designed the experiment. V.B.S.C. and O.V., wrote the ﬁrst draft of the manuscript. X.W.L. Appendix A. Supplementary data and K.S. contributed to XRD data analysis. V.B.S.C. and T.Z.: contributed to FTIR data analysis. A.B.K. and J.W.S contributed to Raman Supplementary data associated with this article can be found, in data analysis and interpretation. the online version, at http://dx.doi.org/10.1016/j.jsb.2015.01.004. V.B.S. Chan et al. / Journal of Structural Biology 189 (2015) 230–237 237

References Pechenik, J.A., Pearse, J.S., Qian, P.Y., 2007. Effects of salinity on spawning and early development of the tube-building polychaete Hydroides elegans in Hong Kong: not just the sperm’s fault? Biol. Bull. 212, 151–160. Addadi, L., Weiner, S., 1992. Control and design principles in biological Politi, Y., Metzler, R.A., Abrecht, M., Gilbert, B., Wilt, F.H., Sagi, I., Addadi, L., Weiner, mineralization. Angew. Chem. Int. Ed. Engl. 31, 153–169. S., Gilbert, P., 2008. Transformation mechanism of amorphous calcium Addadi, L., Raz, S., Weiner, S., 2003. Taking advantage of disorder: amorphous carbonate into calcite in the sea urchin larval spicule. Proc. Natl. Acad. Sci. calcium carbonate and its roles in biomineralization. Adv. Mater. 15, 959–970. USA 105, 17362–17366. Andersson, A.J., Mackenzie, F.T., Bates, N.R., 2008. Life on the margin: implications of Politi, Y., Batchelor, D.R., Zaslansky, P., Chmelka, B.F., Weaver, J.C., Sagi, I., Weiner, S., ocean acidification on Mg-calcite, high latitude and cold-water marine Addadi, L., 2009. Role of magnesium ion in the stabilization of biogenic calcifiers. Mar. Ecol. Prog. Ser. 373, 265–273. amorphous calcium carbonate: a structure–function investigation. Chem. Andreasen, D.H., Sosdian, S., Perron-Cashman, S., Lear, C.H., deGaridel-Thoron, T., Mater. 22, 161–166. Field, P., Rosenthal, Y., 2006. Fidelity of radially viewed ICP-OES and magnetic- Poloczanska, E., Hughes, D., Burrows, M., 2004. Underwater television observations sector ICP-MS measurement of Mg/Ca and Sr/Ca ratios in marine biogenic of Serpula vermicularis (L.) reefs and associated mobile fauna in Loch Creran, carbonates: are they trustworthy together? Geochem. Geophys. Geosyst. 7, Scotland. Estuarine Coastal Shelf Sci. 61, 425–435. Q10P18. Qian, P.Y., 1999. Larval settlement of polychaetes. Hydrobiologia 402, 239–253. Beniash, E., Aizenberg, J., Addadi, L., Weiner, S., 1997. Amorphous calcium carbonate Qiu, J.W., Qian, P.Y., 1997. Combined effects of salinity, temperature and food on transforms into calcite during sea urchin larval spicule growth. Proc. R. Soc. early development of the polychaete Hydroides elegans. Mar. Ecol. Prog. Ser. 152, Lond. B Biol. Sci. 264, 461. 79–88. Bischoff, W.D., Sharma, S.K., Mackenzie, F.T., 1985. Carbonate ion disorder in Ragazzola, F., Foster, L.C., Form, A., Anderson, P.S.L., Hansteen, T.H., Fietzke, J., 2012. synthetic and biogenic magnesian calcites: a Raman spectral study. Am. Ocean acidification weakens the structural integrity of coralline algae. Global Mineral. 70, 581–589. Change Biol. 18, 2804–2812. Bornhold, B.D., Milliman, J.D., 1973. Generic and environmental control of carbonate Ries, J.B., 2011a. Skeletal mineralogy in a high-CO world. J. Exp. Mar. Biol. Ecol. 403, mineralogy in serpulid (polychaete) tubes. J. Geol. 81, 363–373. 2 54–64. Brecˇevic´, L., Nielsen, A.E., 1989. Solubility of amorphous calcium carbonate. J. Cryst. Ries, J.B., 2011b. A physicochemical framework for interpreting the biological Growth 98, 504–510. calcification response to CO -induced ocean acidification. Geochim. Bubel, A., 1983. A fine structural study of the calcareous opercular plate and 2 Cosmochim. Acta 75, 4053–4064. associated cells a polychaete annelid. Tissue Cell 15, 457–476. Ries, J.B., Cohen, A.L., McCorkle, D.C., 2009. Marine calcifiers exhibit mixed Cañestro, C., Yokoi, H., Postlethwait, J.H., 2007. Evolutionary developmental biology responses to CO -induced ocean acidification. Geology 37, 1131–1134. and genomics. Nat. Rev. Genet. 8, 932–942. 2 Roleda, M.Y., Boyd, P.W., Hurd, C.L., 2012. Before ocean acidification: calcifier Carter, J.G., 1990. lossary of skeletal biomineralization. Skeletal Biomineralization: chemistry lessions. J. Phycol. 48, 840–843. Patterns, Processes and Evolutionary Trends. 1. The American Geophysical Schöne, B.R., Radermacher, P., Zhang, Z., Jacob, D.E., 2013. Crystal fabrics and Union, pp. 609–671. element impurities (Sr/Ca, Mg/Ca, and Ba/Ca) in shells of Arctica islandica – Chan, V.B.S., Li, C., Lane, A.C., Wang, Y., Lu, X., Shih, K., Zhang, T., Thiyagarajan, V., implications for paleoclimate reconstructions. Palaeogeogr. Palaeoclimatol. 2012. CO -driven ocean acidification alters and weakens integrity of the 2 Palaeoecol. 373, 50–59. calcareous tubes produced by the serpulid tubeworm, Hydroides elegans. PLoS Shen, C.C., Lee, T., Chen, C.Y., Wang, C.H., Dai, C.F., Li, L.A., 1996. The calibration of One 7, e42718. D[Sr/Ca] versus sea surface temperature relationship for Porites corals. Dorvee, J.R., Veis, A., 2013. Water in the formation of biogenic minerals: peeling Geochim. Cosmochim. Acta 60, 3849–3858. away the hydration layers. J. Struct. Biol. 183, 278–303. Simkiss, K., Wilbur, K.M., 1989. Biomineralization: Cell Biology and Mineral Feely, R.A., Sabine, C.L., Hernandez-Ayon, J.M., Ianson, D., Hales, B., 2008. Evidence Deposition. Academic Press, San Diego. for upwelling of corrosive "acidified" water onto the continental shelf. Science Smith, A.M., Riedi, M.A., Winter, D.J., 2013. Temperate reefs in a changing ocean: 320 (5882), 1490–1492. skeletal carbonate mineralogy of serpulids. Mar. Biol. 160, 2281–2294. Haines, J., Maurer, D., 1980. Quantitative faunal associates of the serpulid Stanley, S.M., 2006. Influence of seawater chemistry on biomineralization polychaete Hydroides dianthus. Mar. Biol. 56, 43–47. throughout phanerozoic time: paleontological and experimental evidence. Hedley, R., 1956a. Studies of serpulid tube formation I. The secretion of the Palaeogeogr. Palaeoclimatol. Palaeoecol. 232, 214–236. calcareous and organic components of the tube by Pomatoceros triqueter.Q.J. Sun, Y., Ten Hove, H.A., Qiu, J.-W., 2012. Serpulidae (Annelida: Polychaeta) from Microsc. Sci. 3, 411–419. Hong Kong. Zootaxa 3424, 1–42. Hedley, R., 1956b. Studies of serpulid tube formation II. The calcium-secreting Tanur, A., Gunari, N., Sullan, R., Kavanagh, C.J., Walker, G., 2009. Insights into the glands in the peristomium of spirorbis, Hydroides, and Serpula. Q. J. Microsc. Sci. composition, morphology, and formation of the calcareous shell of the serpulid 3, 421–427. Hydroides dianthus. J. Struct. Biol. 169, 145–160. Hedley, R., 1958. Tube formation by Pomatoceros triqueter (Polychaeta). J. Mar. Biol. Taylor, P.D. 2008. Seawater chemistry, biomineralization and the fossil record of Assoc. UK 37, 315–322. calcareous organisms, pp. 21–29 Origin and Evolution of Natural Diversity: Klein, R.T., Lohmann, K.C., Thayer, C.W., 1996. Sr/Ca and 13C/12C ratios in skeletal Proceedings of the International Symposium, The Origin and Evolution of calcite of Mytilus trossulus: covariation with metabolic rate, salinity, and carbon Natural Diversity, held from 1–5 October 2007 in Sapporo, Japan. 21st Century isotopic composition of seawater. Geochim. Cosmochim. Acta 60, 4207–4221. COE for Neo-Science of Natural History, Hokkaido University. Kontoyannis, C.G., Vagenas, N.V., 2000. Calcium carbonate phase analysis using XRD Taylor, P.D., Vinn, O., 2006. Convergent morphology in small spiral worm tubes and FT-Raman spectroscopy. Analyst 125 (2), 251–255. (‘Spirorbis’) and its palaeoenvironmental implications. J. Geol. Soc. 163, 225– Lane, A.C., Mukherjee, J., Chan, V.B.S., Thiyagarajan, V., 2013. Decreased pH does not 228. alter metamorphosis but compromises juvenile calcification of the tube worm Vinn, O., 2012. Calcite in the Skeletons of Annelids. Nova Science Publishers. Hydroides elegans. Mar. Biol. 160, 1983–1993. Vinn, O., 2013. SEM study of semi-oriented tube microstructures of Serpulidae Larson, A.C., Von Dreele, R.B., 1994. GSAS. General Structure Analysis System. (Polychaeta, Annelida): implications for the evolution of complex oriented LANSCE, MS-H805, Los Alamos, New Mexico. microstructures. Microsc. Res. Tech. 76, 453–456. Long, X., Nasse, M.J., Ma, Y., Qi, L., 2012. From synthetic to biogenic Mg-containing Vinn, O., Kupriyanova, E.K., 2011. Evolution of a dense outer protective tube layer in calcites: a comparative study using FTIR microspectroscopy. Phys. Chem. Chem. serpulids (Polychaeta, Annelida). Carnets de Géologie-Notebooks on Geology Phys. 14 (7), 2255–2263. CG2011_L05, pp. 137–147. Medakovic´, D., 2000. Carbonic anhydrase activity and biomineralization process in Vinn, O., Tenhove, H.A., Mutvei, H., 2008a. On the tube ultrastructure and origin of embryos, larvae and adult blue mussels Mytilus edulis L. Helgol. Mar. Res. 54, 1– calcification in sabellids (Annelida, Polychaeta). Palaeontology 51, 295–301. 6. Vinn, O., Jager, M., Kirsimae, K., 2008b. Microscopic evidence of serpulid affinities of Medakovic´, D., Popovic´, S., Grzˇeta, B., Plazonic´, M., Hrs-Brenko, M., 1997. X-ray the problematic fossil tube ‘Serpula’ etalensis from the Lower Jurassic of diffraction study of calcification processes in embryos and larvae of the Germany. Lethaia 41, 417–421. brooding oyster Ostrea edulis. Mar. Biol. 129, 615–623. Vinn, O., Ten Hove, H.A., Mutvei, H., Kirsimae, K., 2008c. Ultrastructure and mineral Medernach, L., Jordana, E., Grémare, A., Nozais, C., Charles, F., Amouroux, J., 2000. composition of serpulid tubes (Polychaeta, Annelida). Zool. J. Linn. Soc. 154, Population dynamics, secondary production and calcification in a 633–650. Mediterranean population of Ditrupa arietina (Annelida: Polychaeta). Mar. Wang, R., Addadi, L., Weiner, S., 1997. Design strategies of sea urchin teeth: Ecol. Prog. Ser. 199, 171–184. structure, composition and micromechanical relations to function. Philos. Trans. Morse, J.W., 1983. The kinetics of calcium carbonate dissolution and precipitation. R. Soc. Lond. B 352, 469–480. Rev. Mineral. Geochem. 11, 227–264. Weedon, M.J., 1994. Tube microstructure of recent and Jurassic serpulid polychaetes Mucci, A., 1983. The solubility of calcite and aragonite in seawater at various and the question of the Palaeozoic ‘‘spirorbids’’. Acta Palaeontol. Pol. 39, 1–15. salinities, temperatures, and one atmosphere total pressure. Am. J. Sci. 283, 780. Weedon, M.J., Taylor, P.D., Railsback, L.B., 1994. Original mineralogy of Mukherjee, J., Wong, K.K.W., Chandramouli, K.H., Qian, P.-Y., Leung, P.T.Y., Wu, carboniferous worm tubes: evidence for changing marine chemistry and R.S.S., Thiyagarajan, V., 2013. Proteomic response of marine invertebrate larvae biomineralization: comment and reply. Geology 22, 281–282. to ocean acidification and hypoxia during metamorphosis and calcification. J. Weiner, S., Addadi, L., 2011. Crystallization pathways in biomineralization. Annu. Exp. Biol. 216, 4580–4589. Rev. Mater. Res. 41, 21–40. Nedved, B.T., Hadfield, M.G., 2009. Hydroides elegans (Annelida: Polychaeta): a model for biofouling research. In: Flemming, H.-C. et al. (Eds.), Marine and Industrial Biofouling. Springer, Berlin Heidelberg, pp. 203–217.