Proteomics insights: proteins related to larval attachment and metamorphosis of marine invertebrates

The transition in an animal from a pelagic larval stage to a sessile benthic juvenile typically requires major morphological and behavioral changes. Larval competency, attachment and initiation of metamorphosis are thought to be regulated by intrinsic chemical signals and specific sets of proteins. However, the molecular mechanisms that regulate larval attachment and metamorphosis in marine invertebrates have yet to be fully elucidated. Despite the many challenges associated with analysis of the larvae proteome, recent proteomic technologies have been used to address specific questions in larval developmental biology. These and other molecular studies have generated substantial amount of information of the proteins and molecular pathways involved in larval attachment and metamorphosis. Furthermore, the results of these studies have shown that systematic changes in protein expression patterns and post-translational modifications (PTM) are crucial for the transition from larva to juvenile. The degeneration of larval tissues is mediated by protein degradation, while the development of juvenile organs may require PTM. In terms of application, the identified proteins may serve as targets for antifouling compounds, and biomarkers for environmental stressors. In this review we highlight the strengths and limitations of proteomic tools in the context of the study of marine invertebrate larval biology.

The transition in an animal from a pelagic larval stage to a sessile benthic juvenile typically requires major morphological and behavioral changes. Larval competency, attachment and initiation of metamorphosis are thought to be regulated by intrinsic chemical signals and specific sets of proteins. However, the molecular mechanisms that regulate larval attachment and metamorphosis in marine invertebrates have yet to be fully elucidated. Despite the many challenges associated with analysis of the larvae proteome, recent proteomic technologies have been used to address specific questions in larval developmental biology. These and other molecular studies have generated substantial amount of information of the proteins and molecular pathways involved in larval attachment and metamorphosis. Furthermore, the results of these studies have shown that systematic changes in protein expression patterns and post-translational modifications (PTMs) are crucial for the transition from larva to juvenile. The degeneration of larval tissues is mediated by protein degradation, while the development of juvenile organs may require PTM. In terms of application, the identified proteins may serve as targets for antifouling compounds, and biomarkers for environmental stressors. In this review we highlight the strengths and limitations of proteomic tools in the context of the study of marine invertebrate larval biology.

IMPORTANCE OF MARINE INVERTEBRATES: BARNACLES, BRYOZOANS AND POLYCHAETES
Marine invertebrates are important players in various marine ecosystems, including in biofouling, intertidal and sediment ecosystems. The barnacle Balanus amphitrite, the bryozoan Bugula neritina and the serpulid polychaete Hydroides elegans are the most dominant fouling species, are distributed worldwide, and are commonly used in larval settlement (including both attachment and metamorphosis) studies and antifouling research (Ten Hove, 1974;Khandeparker and Anil, 2007;Aguirre et al., 2008;Cohen, 2011). Marine polychaete worms, including Capitella teleta, Pseudopolydora vexillosa, and Neanthes arenaceodentata, occur in habitats ranging from subtropical to inshore waters, and are used as indicators of marine pollution and as toxicological test animals (Reish and Gerlinger, 1997;Greg and Kristian, 1998;Rouse and Fauchald, 1998;Glasby and Timm, 2008;Blake et al., 2009). Because of high sensitivity to marine contamination mussels have been used as indicators of marine pollution (Romeo et al., 2003). For instance, mussels and barnacles were used as biomonitors in coastal waters (Phillips and Rainbow, 1988) Abalone Haliotis diversicolor supertexta was used for investigating contamination of marine ecosystems by toxic chemicals (Zhou et al., 2010). Calcifying marine invertebrates including polychaetes and barnacles, are capable of physiological and behavioral adaptation to various environmental stressors and thus serve as model organisms in climate change research (Dupont and Thorndyke, 2009;Taylor et al., 2010). Abalone, gastropod mollusk, is used for food and their shells are often used for ornaments. Several invertebrates provide habitat to other species. For example, the mussel beds provide living space to the species living on them. Peracaridan species utilize B. neritina as living space (Conradi et al., 2000).

LARVAL DEVELOPMENT: COMPETENCY, ATTACHMENT, AND METAMORPHOSIS
The population dynamics of marine invertebrates are controlled by recruitment and post-settlement survival. The choice of attachment site is crucial because attachment to an unfavorable substrate can be fatal (Bishop et al., 2006). Factors that influence attachment and metamorphosis, including settlement cues, competency, and habitat, are selective for the larvae of a particular species. For instance, larval metamorphosis in barnacles is a dynamic and rapid process that involves substantial tissue remodeling and differentiation (Thiyagarajan et al., 2007;Maruzzo et al., 2012). Barnacle species have six nauplius stages and a cyprid larval stage, and larva-juvenile transitions may or may not require a metamorphosis inducer (Thiyagarajan et al., 2005). In the absence of cues, the larvae of B. neritina commonly www.frontiersin.org October 2014 | Volume 1 | Article 52 | 1 settle within several minutes (Woollacott and Zimmer, 2005).
In H. elegans, the fertilized eggs hatch into young trochophore larvae, which develop into competent larvae and metamorphose into young juveniles. The larvae will not settle without suitable settlement cues, and the larval-juvenile transition requires substantial tissue reorganization and major morphological changes (Hadfield et al., 2001). In contrast, the young larvae of P. vexillosa can take 7-10 days to develop competency, and larval metamorphosis is a gradual process that does not require substantial development of juvenile organs (Mok et al., 2008). This type of metamorphosis in the absence of specific inducers is referred to as "spontaneous metamorphosis." Although past studies have contributed to our understanding of the basic underpinnings of the morphological, ecological and behavioral patterns of larval settlement, the molecular mechanisms that regulate larval attachment and metamorphosis are poorly understood. In recent years, it has been hypothesized that attachment and metamorphic events are regulated by changes in protein expression patterns and intrinsic signaling pathways that influence larval morphology and behavior. To test this hypothesis, proteomics tools are being used increasingly to identify proteins and pathways involved in larval development, despite technological limitations associated with the advancement of larval biology research in this area. Hence, the goal of this review is to describe the progress in marine larval proteomic research in relation to changes in protein expression patterns in different larval stages of marine invertebrates. Specifically, this review highlights: (i) technical challenges in larval proteomics research; (ii) proteome and phosphoproteome insights during larval attachment and metamorphosis; (iii) the involvement of signaling pathways during larval metamorphosis; and (iv) the usefulness of commonly expressed proteins in antifouling applications.

PROTEIN EXPRESSION PATTERNS DURING LARVAL DEVELOPMENT: TWO-DIMENSIONAL GEL ELECTROPHORESIS (2-DE) PROTEOMICS
Proteomics tools and pipelines used in the study of marine larval biology are outlined in Figure 1. A list of studies on larval proteomics is shown in Table 1. In the majority of these studies, the larval proteome was characterized by gel-based (2-DE) or gelfree proteomics, followed by mass spectrometry (MS) analysis. These approaches enabled documentation of changes in protein expression patterns and protein modification status at various stages in the larval lifespan. López et al. (2005) was the first group to use 2-DE proteomics approach to identify larval-specific proteins in the mussel Mytilus galloprovincialis. DeBoer et al. (2007) then conducted comparative proteome and transcriptome analyses to reveal symbiosis-related differences in coral larvae and dinoflagellates and found no significant changes in the protein profiles of the larvae. Thiyagarajan and Qian (2008) used 2-DE proteomics to document the global proteome expression pattern in the larvae of B. amphitrite. They optimized protein extraction and iso-electric focusing (IEF) conditions for a better resolution of protein spots on 2-DE gels and showed that the proteome pattern of cyprid larvae was very distinct from that of attached larvae and newly-metamorphosed juveniles. Another study (Mok et al., 2009) found that there are proteomic changes in the polychaete worm P. vexillosa during larval metamorphosis. They found that the proteome of competent larvae of P. vexillosa was similar to that of juveniles, whereas the proteomes of precompetent and  competent larvae were more distinct. Availability of genome sequencing allowed a better characterization of the proteins involved in different biological processes during the development in ascidians (as reviewed in Inaba et al., 2007). Nomura et al. (2009) found the involvement of vitellogenin, elongation factors, actin-binding proteins and chaperones during embryonic development of Ciana intestinalis. In the abalone Haliotis asinina, the first proteomic study characterized protein domains from calcified shell, which were rich in glutamine, glycine, alanine, and aspartate (Marie et al., 2010). All the studies mentioned above have relied on low-resolution immobilized pH gradient (IPG) strips (13 cm in length; pH 3-10) for the first-dimension separation and Coomassie-brilliant blue (CBB) staining of the gels. Aggregation or overlapping of proteins spots limited the resolution of many spots on the 2-DE gels and could not really quantify actual changes in protein abundance. To overcome these shortcomings, Qian et al. (2010) and Zhang et al. (2010a) used combination of different methods to probe the larval proteomes of B. neritina and B. amphitrite, including pre-fractionation of the larval proteomes prior to 2-DE to reduce the sample complexity, sequential fluorescent dye gel staining to increase the detection limit (∼0.5-5 ng) and dynamic range (Panfoli et al., 2012) and implementation of narrow-range IPG strips (pH 4-7) to enhance resolution of the protein spots on 2-DE gels. In these studies, they used used a Zoom IEF fractionator (Invitrogen, Carlsbad, CA, USA) to fractionate the total proteome into two subproteomes of pH 3.0-5.4 and pH 5.4-7.0 and high-resolution narrow-range IPG (pH 4-7) strips followed by gel staining with SYPRO® Ruby (Molecular Probes, Eugene, OR, USA) to achieve better protein separation. The number of protein spots detected was more than that detected for unfractionated samples.

POST-TRANSLATIONAL MODIFICATIONS (PTMs) OF PROTEINS DURING LARVAL METAMORPHOSIS
The identification and characterization of protein modifications among marine invertebrates have received substantial attention in recent years. It has been hypothesized that the early phase of metamorphic events depends on PTMs of proteins. For example, in B. neritina, the rapid initial phase of metamorphosis was facilitated by PTMs of existing proteins . In fact, protein modification often plays a crucial role in many biological processes (Zhou et al., 2007). In particular, www.frontiersin.org October 2014 | Volume 1 | Article 52 | 5 protein phosphorylation is a rapid process that regulates many cellular processes and signal transduction pathways (Hunter, 1995). Thiyagarajan et al. (2009) demonstrated the usefulness of multiplex staining and 2-DE to probe the larval phosphoproteome of B. amphitrite and B. neritina. In their study, specific fluorescent dyes including Pro-Q Diamond and Sypro Ruby, were used to detect phosphoproteins and total proteins on a single 2-DE gel. This method enabled parallel analyses of total protein and phosphoprotein expression. The similar tools were used to study phosphoproteome patterns in other marine species. In H. elegans, the larval phosphoproteome pattern changed significantly during competency, whereas the proteome pattern was not affected (Zhang et al., 2010b). In Capitella teleta (this refers to Capitella sp. I in early publication) phosphorylated proteins were essential for the elongation of body segments, production of hooded hooks, and preparation of juvenile tissues during larval metamorphosis (Chandramouli et al., 2011a). Protein enrichment methods have enabled the separation of proteins according to their PTM (e.g., phosphorylation) (Gilchrist et al., 2006;Puente et al., 2006;Cantin et al., 2008). This approach was used successfully by Chandramouli et al. (2011b) to improve the detection and identification of phosphoproteins during larval metamorphosis of P. vexillosa, revealing that competent larvae exhibit a higher degree of phosphorylation than precompetent larvae and juveniles. However, MS analysis of phosphopeptides remains challenging (Leitner et al., 2011). Using immobilized metal affinity chromatography (IMAC) and liquid chromatography (LC)-MS/MS to enrich phosphorylated peptides, Chandramouli et al. (2012a) identified nine phosphopeptides and phosphorylation sites (Thr/Ser) that are crucial for the early development (ova and larval stages) of N. arenaceodentata. This enrichment strategy illustrated an effective method for the identification of phosphoproteins and phosphopeptides in a larval proteome. On the other hand, glycoproteomics is an emerging proteomic technology and can be used to detect specific glycosylation patterns and to identify glycoproteins in complex proteomes (Pan et al., 2011). This method has the advantage of sequential staining, whereby ProQ Emerald fluorescent dye is used to selectively detect glycoproteins on 2-D gels. Using thie method, Chandramouli et al. (2012b) identified 19 abundant glycoproteins involved in the stress response and metabolic processes during larval metamorphosis of three marine invertebrate species. They found that the degree of protein glycosylation in B. amphitrite and P. vexillosa was greater prior to metamorphosis, whereas glycosylation in B. neritina increased right after metamorphosis. Overall, these studies clearly demonstrated emerging proteomics technologies can be used to document both glycosylation and phosphorylation dynamics during larval metamorphosis.

GEL-FREE HIGH-THROUGHPUT PROTEOMICS AND COMPREHENSIVE LARVAL PROTEOME ANALYSIS
Although gel-based proteomics provided valuable information about larval protein expression patterns, limitations associated with the identification of low-abundance larval proteins prompted the development of high throughput gel-free approaches (López, 2007;Tomanek, 2011;Slattery et al., 2012). In addition, transcriptome sequence datasets for several marine invertebrates, including B. neritina , B. amphitrite De Gregoris et al., 2011), the coral Acropora millepora (Moya et al., 2012) and P. vexillosa (Chandramouli et al., 2013b), are now available in the public domain, which allows larval biologists to conduct more gel-free proteome analysis. Label-free quantitative methods involving individual sample preparation steps followed by LC-MS/MS analysis. Quantification of the amount of proteins is based on mass spectrum counting of the identified peptides Zhang et al., 2011;Chen et al., 2014).  applied next generation RNA sequencing and high-throughput LC-MS-based proteomics to profile larval transcripts and proteins of B. neritina. They generated 131,450 high-quality transcripts which were used as a reference transcriptome for proteome analysis. About 882 proteins were identified and quantified for the swimming and attached larval stages of B. neritina. Zhang et al. (2011) identified >1000 proteins (among which 61 were differentially expressed) in swimming and in two post-attachment stages of B. neritina. These differentially expressed proteins were grouped in the several categories: energy metabolism, cytoskeleton, transcription and translation, and calcification. Chen et al. (2014) identified about 700 proteins in each stage of B. amphitrite (nauplius II, nauplius VI, cyprids and juveniles) using label-free quantitative proteomics and found that proteins related to energy metabolism, nervous system and signal transduction were up-regulated in cyprids, whereas those related to cytoskeleton, transcription and translation, cell differentiation, and biomineralization proteins were up-regulated in juveniles. Zhang et al. (2014) identified about 1500 proteins that are associated with energy metabolism, redox homeostasis, and cytoskeleton development during larval development of H. elegans. Although label-free proteomic studies provide a robust platform for direct quantitative analysis of the proteome, it is very challenging and complicated to analyze many mass spectra derived from multiple MS runs (Haqqani et al., 2008). To address this issue, peptide labeling methods have been used for relative quantification of protein abundance in marine invertebrates (Chandramouli et al., 2013a,b;Han et al., 2013;Sun et al., 2013). In fact, isobaric tags for relative and absolute quantification (iTRAQ) have been established (Pütz et al., 2005) and used to multiplex 4-8 samples (Evans et al., 2012;Martyniuk et al., 2012). Using this method, the peptides are labeled and then the samples are pooled and desalted prior to LC-MS/MS. The protein or peptide abundance is calculated based on the chromatographic peptide peak intensity, or the height of the various labeled peptide samples. In addition, an improved fractionation technique involving the separation of proteins or peptides in a multi-well device according to their isoelectric points (pI) has been developed to reduce the proteome complexity (Agilent Technologies, Böblingen, Germany; Chenau et al., 2008). Chandramouli et al. (2013b) combined peptide OFFGEL fractionation with iTRAQ labeling to identify and quantify proteins involved in key molecular processes during larval metamorphosis of P. vexillosa and identified 107 differentially expressed proteins in precompetent, competent, and juvenile stages. Among these 107 proteins, 14 and 53 proteins were differentially expressed during competency and metamorphosis, respectively. Using the same approach,

Frontiers in Marine Science | Marine Molecular Biology and Ecology
October 2014 | Volume 1 | Article 52 | 6 Chandramouli et al. (2013a) identified a set of proteins that influence the unique pattern of reproduction in the polychaete N. arenaceodentata and demonstrated that male worms have consistent protein expression patterns, whereas female worms (particularly spent females) show marked changes in the phosphoproteome before and after spawning. These studies showed that labeling methods permit the relative and absolute quantification of proteins expressed during a specific developmental stage.

MOLECULAR PATHWAYS UNDERLYING LARVAL METAMORPHOSIS IN MARINE INVERTEBRATES
Signaling pathways that are possibly regulate larval metamorphosis of marine invertebratespecies have been a hot topic in recent years. It was reported that cyclic adenosine-monophosphate (cAMP) and calcium signaling in barnacles (Clare et al., 1995;Clare, 1996), inositol-1,4,5-tris-phosphate/diacylglycerol (IP 3 /DAG) signaling in cnidarians (Fleck, 1997), nitric oxide/cyclic guanosine monophosphate (NO/cGMP), protein kinase C (PKC) in Capitella (Biggers and Laufer, 1999) and sea urchins Amador-Cano et al., 2006), and p38 mediated mitogen-activated protein kinase (MAPK) recently has been implicated in the regulation of larval attachment and metamorphosis in H. elegans and B. amphitrite (Wang and Qian, 2010;He et al., 2012) are possibly involved in larval settlement and metamorphosis of these animals. Furthermore, Zhang et al. (2013) demonstrated that MKK3 (the upstream kinase of p38MAPK) activates the function of p38MAPK, thereby influencing the attachment of B. amphitrite larvae. Nitric oxide (NO) and cyclic GMP (cGMP) signaling molecules have been shown to act as regulators of the initiation of metamorphosis in ascidians and in an echinoid Leise et al., 2001).  demonstrated that the NO/cGMP system acts as a negative regulator during metamorphosis of the sea urchin Lytechinus pictus. Zhang et al. (2012a) documented the role of NO and cGMP during larval attachment of B. amphitrite. Recently Ueda and Degnan (2014) reported that NO facilitates the induction of metamorphosis while inhibition of nitric oxide synthase (NOS) reduces rates of metamorphosis in the abalone Haliotis asinine. It appears that nitric oxide signaling can contribute to stress-related responses (up-regulation of HSP-90) and regulate the metamorphic transitions. The role of calcium in signal transduction pathways during larval metamorphosis has been well documented in the polychaete worm Phragmatopoma californica (Ilan et al., 1993). Chen et al. (2012) examined the role of calmodulin (CaM) in larval attachment and metamorphosis of H. elegans Wong et al. (2012) found that the Wnt signaling pathway appears to be important for morphogenesis in B. neritina. Based on the findings above, we anticipate that MAPK signaling is important for larval attachment in B. amphitrite and H. elegans. Wnt signaling typically influences larval metamorphosis of B. neritina. The importance of Ca2+/CaM signaling in the development of polychaetes is of significant interest. Heyland and Moroz (2006) showed that larval attachment and metamorphosis in marine invertebrates is mediated by common metamorphic events. If this is true, one may expect that larval attachment and metamorphosis are regulated by a common set of proteins in multiple marine invertebrates. To identify common protein expression signatures (PESs) during larval attachment and metamorphosis of B. amphitrite, P. vexillosa, and B. neritina, Chandramouli et al. (2012b) found 11 commonly expressed proteins involved in cytoskeleton, metabolism, and stress responses. A similar set of common proteins were differentially regulated in multiple marine invertebrates. Table 2 lists the commonly regulated proteins and their expression patterns. In general, changes in the expression patterns of cytoskeletal proteins (tubulin, actin, and myosin) were observed in all of the species studied so far, indicating the importance of cytoskeletal dynamics in maintaining the architecture and contractility of larvae. Similarly, proteins related to energy metabolism (ATP synthase, disulfide isomerase, vitellogenin, enolase-phosphatase, and malate dehydrogenase) and the stress response (HSP-90, HSP-70, 14-3-3 protein, and calreticulin) were often differentially expressed during larval attachment and metamorphosis, indicating that these proteins may be synergistically regulated and may control events in the metamorphosis of larvae to juveniles. To identify possible common proteins that are the molecular targets for antifouling compounds, Thiyagarajan et al. (2009) showed that genistein, which is an inhibitor of metamorphosis (Zhou et al., 2009), down-regulated the expression of phosphoproteins in settling larvae of B. neritina. In contrast, this compound induced the expression of phosphoproteins in B. amphitrite. Another potential antifoulant, butenolide (Xu et al., 2010) sustained the expression of stress and energy metabolism proteins in settling larvae of B. neritina and B. amphitrite Zhang et al., 2010a). Dash et al. (2012) showed that polyethers affect proteins related to energy metabolism, oxidative stress and chaperones. Han et al. (2013) also found that meleagrin, a potential non-toxic antifoulant, affect the expression of cyprid major protein (CMP) and adhesive plaque matrix proteins (APMPs) and thereby interfere with the larval endocrine system and molting processes (Billinghurst et al., 2000). These studies indicated that proteins involved in oxidative stress, cytoskeletal dynamics, chaperones, apoptosis, and energy homeostasis are possible molecular targets of those antifouling compounds. In contrast, several other types of antifoulants act on ion channels, quorum sensing systems, neurotransmitters, genotoxic damage, and adhesive production (as reviewed in Qian et al., 2013), which indicates voltage dependent anion channels, acetylcholinesterase, phenoloxidase, chitinase, Ran GTPase activating protein, and NADH ubiquinone oxidoreductase are proposed as molecular targets. Because of housekeeping function and responsive to environmental changes, several PESs identified by proteomics studies may not be the direct targets but their differential expression may be an indirect physiological and biochemical response to various antifoulants.

CHALLENGES AND FUTURE DIRECTIONS SAMPLE PREPARATION
The complexity of protein samples has always been a major obstacle of proteomic study (Hondermarck, 2003;Speicher, 2007;Shen et al., 2008). Larvae and adults of many marine invertebrates are protected by shell (e.g., B. amphitrite), calcareous tubes (H. elegans), and mucilaginous polysaccharide substances (P. vexillosa and C. teleta). Cell culture and tissue isolation methods have not yet been established for these species. The direct use of whole larvae or juveniles for proteomic studies results in an enormous increase in sample complexity, complicating downstream sample preparation for mass spectrometry (MS) analysis. Highly abundant proteins can mask those of a low abundance, potentially preventing the detection of signaling and receptor proteins that may be crucial for larval attachment and metamorphic transitions. To overcome these problems, we need to improve protein extraction protocols. In addition, larval proteins of polychaetes can rapidly degraded shortly after tissue lysis. To provent degradation, protease inhibitors are required during multiple sample preparation steps (Zhang et al., 2010b;Chandramouli et al., 2011b). In addition, to minimize the possible protein degradation, it is better to freeze the samples in liquid nitrogen or be used immediately. The combined use of strong detergents (8M urea), homogenization, and sonication methods can also significantly improve protein recovery . It is also highly recommended that prior to 2-DE, larval protein samples shall be purified from other contaminants through the use of acetone precipitation and 2-DE clean-up kits. Multistep strategies in sample preparation, fractionation, and MS separation of peptides may facilitate the detection of these proteins (Millioni et al., 2011). Among these methods, immunoaffinity columns seems to be more suitable for the depletion of high-abundant cytoskeletal proteins (Molnár et al., 2011), after which low-abundant proteins may bind to the column, be eluted with specific buffers, and identified by 2-DE and/or MS. This procedure facilitates the identification of potential protein candidates or biomarkers. Alternately, the enrichment of specific subset of phosphoproteins or peptides from a complex protein sample can be an effective approach (Batalha et al., 2012). Cost-effective specific immune-affinity reagents, such as antibodies and metal, are the subjects for targeted proteomic studies, particularly for elucidating the protein phosphorylation dynamics in signaling pathways. However, the development of protocols for protein enrichment is always challenged by the functional and structural diversity of phosphorylated proteins. Furthermore, we need to maximize the amount of proteins to begin with in order to have sufficient protein for down-stream analysis.

DE NOVO SEQUENCING AND MS BLAST
In general, mass spectra are searched against databases using dedicated softwares, such as Mascot and SEQUEST, and X! Tandem that compare peaks in MS spectra with precomputed peptide fragments (Liska and Shevchenko, 2003). However, this approach is not efficient for identifying proteins that are not included in the database. In de novo sequencing, peptide sequences are obtained from mass spectra without the assistance of databases either manually or via computer programs (Ma and Johnson, 2012). Sequence homology search engines, such as MS driven BLAST (MS BLAST) have been successfully applied in marine proteomics studies. For example, Shevchenko et al. (2005) identified 48 proteins in larval venom of the Brazilian moth Cerodirphia speciosa. Waridel et al. (2007) employed multiple layered integrated search strategy that included MASCOT, de novo sequencing and MS BLAST sequence-similarity search to identify proteins in green alga Dunaliella salina. Sun et al. (2010) identified 65 proteins from golden apple snail P. canaliculata using de novo crossspecies method. We combined conventional and customized transcriptome databases search with MS-BLAST sequence similarity search and identified 21 proteins from N. arenaceodentata (Chandramouli et al., 2012a). The de novo sequencing and MS BLAST have shortcomings because they use partial peptide sequences for database search that are very short sequence tags that can be used to search a standard database, such as NCBI nr, SwissProt, Uniprot etc. to identify the homologous peptide.
Frontiers in Marine Science | Marine Molecular Biology and Ecology October 2014 | Volume 1 | Article 52 | 8

SPECIES-SPECIFIC GENOME DATABASES
The major limitation in conducting molecular studies at either the genome or proteome level is the lack of genome sequences (López, 2007;Tomanek, 2011;Slattery et al., 2012). For example, in our study we only managed to identify about 200 proteins in N. arenaceodentata (Chandramouli et al., 2013a). Several studies have addressed this problem by constructing transcriptome databases of marine species under investigation, which has remarkably improved the quality of proteomics dataset. As more genome information becomes available, proteomics will provide a wealth of information about molecular processes that are associated with larval behavior and morphology.

VALIDATION OF PROTEOMIC DATA
The current proteomic tools were established mainly for model organisms, they need to be selectively optimized for investigating larval proteome. The lack of commercially available speciesspecific antibodies makes it challenging to experimentally validate proteomics data in functional studies. To overcome this problem, He et al. (2012) generated B. amphitrite species-specific p38 antibodies and overexpressed them in E. coli. The authors demonstrated activation of Bar-p38 MAPK during larval settlement through co-immunoprecipitating p38 antibodies with interacting partner proteins. He et al. (2013) used the same method to generate antibodies for cement proteins, cp20k, to localize those proteins in barnacle cyprids. Zhang et al. (2012b) also used co-immunoprecipitation and SDS-PAGE-LC-MS/MS to identify the binding proteins of icocyanide in three marine invertebrates. Furthermore, there has been no matured method for bridging the gap between gene/protein expression and larval development and behavior in marine species. RNAi or gene knockouts methods may offer the solution but require substantial effort to develop these methods for marine species.

CONCLUSION
Over the last 7 years, gel-based and gel-free proteomics have been used to study molecular mechanisms of larval attachment and metamorphosis in several species of marine invertebrates.
Notably, proteomics studies have provided new tools to systematically document the plasticity and variability of the proteome in larval morphogenesis and behavior as well as the response of larvae to antifoulants and environmental stressors. These larval proteomic investigations have also provided substantial information of protein expression patterns, PTMs, and signaling pathways during larval attachment and metamorphosis, which improved our understanding of possible molecular mechanisms of larval attachment and metamorphosis.