The current paradigm proposes that the innate immune systems of invertebrates are much more complex than previously thought. The highly diverse 185/333 gene family in the purple sea urchin encodes a family of closely related proteins of varying length and sequence composition. Subsets of small phagocytes and polygonal cells express 185/333 proteins with localization on the surface of the small phagocytes and within perinuclear vesicles in both cell types. In short-term cultures, coelomocytes form small aggregates that progress to syncytia that are thought to be equivalent to encapsulation in vivo. These aggregates were found to be enriched for 185/333-positive (185/333+) small phagocytes. In response to lipopolysaccharide challenge, coelomocytes transiently increased, including frequencies of both 185/333+ and 185/333-negative (185/333–) small phagocytes and 185/333+ polygonal cells. The 185/333 proteins were present in a broad array of sizes, most of which were larger than that predicted from the cDNAs. Recombinant 185/333 proteins expressed in bacteria and insect cells were also larger than expected, suggesting that the proteins dimerize and multimerize. The diversity of the 185/333 proteins, their expression in response to immune challenge, and their cellular localization suggests this protein family and the small phagocytes have an important immunological role in the sea urchin.
Invertebrate immunology has recently undergone a major paradigm shift away from the notion that invertebrates have simple immune systems that recognize broad groups of pathogens and lack the capacity for precise discrimination of individual pathogenic species (Klein, 1997; Medzhitov and Janeway, Jr, 2002). Current theories state that invertebrates have complex immune responses with high levels of immune diversity (Litman et al., 2001; Flajnik, 2004; Flajnik and Du Pasquier, 2004; Litman et al., 2005; Smith et al., 2006). Recent evidence indicates that immune responses of some invertebrates may be capable of highly discriminatory immune recognition (Little and Kraaijeveld, 2004; Kurtz and Armitage, 2006; Sadd and Schmid-Hempel, 2006). Molecular characterization of immune functions in jawless vertebrates, invertebrates and higher plants have uncovered the presence of large and/or polymorphic gene families that are associated with defense against pathogens (Hurlburt et al., 2001; Litman et al., 2001; Flajnik, 2004; Flajnik and Du Pasquier, 2004; Loker et al., 2004; Pancer et al., 2004; Litman et al., 2005; Cooper and Alder, 2006; Jones and Dangl, 2006; Stout et al., 2008). Examples include defense resistance (R) genes in higher plants (Bittner-Eddy et al., 2000), variable-region-containing chitin-binding proteins (VCBPs) in protochordates (Litman et al., 2001; Cannon et al., 2002; Litman et al., 2005), and fibrinogen-related proteins (FREPs) in snails (Zhang et al., 2004). Diversity of R proteins and FREPs may result from gene conversion, among other mechanisms (Parniske and Jones, 1999; Meyers et al., 2003; Zhang et al., 2004), whereas the diversity of VCBPs appears to be based on a large and polymorphic gene pool (Cannon et al., 2004). Although Down's syndrome cell adhesion molecules (DSCAMs) in insects are encoded by a single-copy gene, combinatorial diversity evident in DSCAM proteins is the result of extensive alternative splicing (Watson et al., 2005; Dong et al., 2006). There are also single copy genes or very small gene families that have large numbers of alleles, which provide high levels of diversity. Examples include the penaiedin genes in shrimp that encode antimicrobial peptides (Cuthbertson et al., 2004; O'Leary and Gross, 2005), and Fu/HC and its ligand fester, that mediate colony rejection or fusion in Botryllus schlosseri (De Tomaso et al., 2005; Nyholm et al., 2006).
Genes expressed in coelomocytes of the purple sea urchin, Strongylocentrotus purpuratus (Stimpson), have been identified from expressed sequence tags (ESTs) (Smith et al., 1996; Nair et al., 2005) and have been deduced from genome annotation studies (Hibino et al., 2006; Sodergren et al., 2006). A number of large gene families, including those encoding Toll-like receptors, NOD-like receptors (Smith et al., 2006; Hibino et al., 2006) and scavenger receptors (Pancer, 2000) have been identified in the sea urchin genome. Another large gene family, called 185/333, has been estimated to have about 50 members that are highly expressed in response to immune challenge (Rast et al., 2000; Nair et al., 2005; Terwilliger et al., 2006; Terwilliger et al., 2007). The 185/333 transcripts make up more than 60% of the ESTs identified after challenge with lipopolysaccharide (LPS), and over 6% of the clones in a bacterially activated coelomocyte cDNA library. A variety of different 185/333 genes are expressed in response to challenge with bacteria, LPS, β-1,3-glucan and double stranded RNA. When 185/333 cDNAs are aligned optimally, the insertion of large gaps is required, which defines 25 blocks of sequence called elements (Terwilliger et al., 2006). Sequence diversity is present among members of each type of element, which is a result of small insertions or deletions (indels) and single nucleotide polymorphisms (SNPs). The deduced 185/333 proteins have a typical signal peptide, an N-terminal glycine-rich region, a C-terminal histidine-rich region, numerous large tandem and interspersed repeats, patches of histidines, regions of acidic amino acids, conserved glycosylation sites and an RGD motif (Terwilliger et al., 2006). They have no cysteines, no recognizable transmembrane region, no conserved GPI linkage motif, no clear predictions for folding, and are not similar to any known protein.
Coelomocytes are the immune cells of the purple sea urchin. They are located in the coelomic fluid (CF) that fills the body cavity and they express the 185/333 genes. Coelomocytes are a mixed population of four cellular morphotypes termed phagocytes, colorless spherule cells, red spherule cells and vibratile cells (Holland et al., 1965; Johnson, 1969a; Chien et al., 1970; Terwilliger et al., 2006) (reviewed in Gross et al., 1999). The phagocytes are a complex set of cells with at least three different morphologies. Large phagocytes are the most common and when spread on a slide, two morphotypes can be differentiated, based on their cytoskeletal architecture (Henson et al., 1992; Edds, 1993; Henson et al., 1999). Discoidal cells (type 1) are disc-shaped with radially oriented actin filaments, whereas polygonal cells (type 2) are generally larger and have actin cables oriented with the long axes of this more angular cell. The third type of phagocyte, called small phagocytes (Gross et al., 2000), are smaller than the discoidal and polygonal cells and have a filopodial morphology that does not change upon spreading. This is in contrast to the larger phagocytes that readily change their morphology from lamellipodial to filopodial and back (Edds, 1977; Edds, 1979; Henson et al., 1999).
The expression patterns and sequence diversity of 185/333 cDNAs suggest the encoded proteins have some involvement in immune responses. We show here that rather than being secreted into the CF as predicted by the deduced signal sequence, 185/333 proteins are associated with the small phagocytes and polygonal cells, although these proteins are located in different cellular compartments for the two cell types. The percentage of 185/333-positive (185/333+) small phagocytes and polygonal cells increases in response to challenge with LPS. On gels, the 185/333 proteins are seen to have a broad array of sizes, most of which are larger than that predicted from the cDNA sequences. Although the function of this highly diverse and polymorphic family of proteins is not currently known, their expression in the small phagocytes and their unexpected variation in size provide important information on their possible involvement in host defense.
Small phagocytes and polygonal cells express 185/333 proteins
Conventional (Fig. 1) and confocal (Fig. 2) imaging of glass-adherent phagocyte populations, which were stained for actin, identified typical discoidal cells, polygonal cells and filopodial small phagocytes (Fig. 1A, Fig. 2B,E,H). The cytoskeleton of small phagocytes revealed an actin array that was less dense than the other phagocyte types, but was well ordered and often concentrated in filopodial core bundles (Fig. 2B,E,H). Preliminary western blots indicated that the 185/333 proteins were associated with the cellular rather than fluid fraction of the CF (not shown), which was not expected from the deduced signal sequence and lack of transmembrane region. Immunofluorescent labeling of the 185/333 proteins showed that a discrete subset of cells with the morphology of small phagocytes stained intensely for these proteins (Fig. 1B, Fig. 2A,D,G). The 185/333+ small phagocytes appeared distinct from either discoidal or polygonal phagocytes based on their overall morphology and actin filament cytoskeletal organization (Fig. 1A, Fig. 2B,E,H). Small phagocytes were almost always observed to have filopodial morphology with the most intense 185/333 staining along the plasma membrane (see below) and in perinuclear vesicles (Fig. 1B,D, Fig. 2G). Two morphological variants of small phagocytes were noted; some were slightly larger and more spread (Fig. 2B), but most were smaller with more extensive filopodia (Fig. 1A, Fig. 2E,H). Some of the polygonal cells expressed the 185/333 proteins, although their staining pattern was typically restricted to a collection of perinuclear vesicles (Fig. 1B,D) and was not associated with the plasma membrane. However, some polygonal cells showed 185/333 proteins widely distributed throughout the cytoplasm (Fig. 1E). The immunolocalization pattern of the 185/333 proteins in coelomocytes was similar in cells treated with either a combination of the three anti-185 antibodies or with each antibody alone (not shown).
To determine the distribution of the 185/333 proteins relative to microtubules, coelomocytes were double labeled for these proteins. The distribution of microtubules in 185/333+ small phagocytes was different from that observed for the discoidal or polygonal coelomocytes (Fig. 3). Microtubules in the 185/333+ cells appeared to associate into large bundles that were affiliated with filopodia in some cells (Fig. 3A) and were present in the cytoplasm in others (Fig. 3B). By contrast, microtubules in discoidal cells tended to be restricted to the perinuclear region, whereas microtubules in polygonal cells formed an extensive array that was spread throughout the cytoplasm (Fig. 3).
Although morphological and immunofluorescent analyses showed that most of the 185/333+ cells belonged to a subpopulation of small phagocytes, this cell type had only been reported once previously (Gross et al., 2000), and may only have been considered to be fragments of larger phagocytes by other investigators. Consequently, our observation of 185/333+ small phagocytes after fixation and processing for immunofluorescent imaging prompted us to confirm that these cells were present in living cell preparations and were not an artifact of fixation. Small, often highly filopodial live cells, which were morphologically similar to the fixed 185/333+ small phagocytes, were observed in live preparations (Fig. 4A,B). The overall morphology of these cells was such that they could easily be distinguished from both discoidal and polygonal coelomocytes. In an effort to correlate the observations on living cells with the results seen in the immunofluorescent labeling of fixed cells, live small phagocytes were imaged and then fixed by perfusion and immunolabeled. Results clearly indicated that the small, highly filopodial cells observed in the living state corresponded to the 185/333+ small phagocytes in fixed preparations (Fig. 4C-F).
To determine whether the 185/333 proteins were present on the surface of the small phagocytes, live coelomocytes were incubated with anti-185/333 antibodies prior to fixation and subsequent incubation with secondary antibodies. Confocal images indicated that 185/333 proteins were present on the extracellular face of the plasma membrane of small phagocytes (Fig. 5), but not the polygonal cells (not shown). The distribution was not uniform but appeared as focal points with spherical structures (Fig. 5A). The labeling on the filopodia showed discrete packets, or knobs, of intense 185/333 staining (Fig. 5B,C). It is noteworthy that the filopodia of living small phagocytes that were 185/333+ often had regularly spaced knobs that may be related to the distribution patterns of the 185/333 proteins (Fig. 5C). In these live-cell labeling experiments, anti-actin was included as a permeabilization control in which there was no indication of anti-actin staining in the cells. In other experiments a similar pattern of anti-185 labeling was generated in cells that were fixed in the glutaraldehyde prefix and the formaldehyde postfix in the absence of any detergent or methanol.
In short-term cultures, coelomocytes tend to form small aggregates that commonly progress to larger aggregates in which cells fuse to form syncytia (Johnson, 1969a; Johnson, 1969b; Chien et al., 1970; Bertheussen and Seljelid, 1978; Bertheussen, 1979). Coelomocyte syncytia that form in vitro have been considered to be the equivalent of in vivo encapsulation responses. Small aggregates were noted in our preparations (Fig. 6A), which appeared to be enriched for 185/333+ small phagocytes, possibly to the exclusion of other cell types (Fig. 6B). When coelomocytes were allowed to aggregate for longer periods of time, larger aggregates and syncytia were noted, always with very intense staining for 185/333 proteins (data not shown). Although this is the first observation of a correlation between small phagocytes and syncytia formation, which will require further testing, this result leads us to speculate that the small phagocytes may initiate the formation of syncytia and that the 185/333 proteins may somehow be involved in that process.
Coelomocytes respond to LPS challenge
The initial discovery of 185/333 cDNAs resulted from screening a coelomocyte library with a subtracted probe to identify transcripts that accumulated in response to challenge with LPS (Nair et al., 2005). The number of 185/333 clones in the LPS-activated cDNA library compared to those in the non-challenged library indicated a 76-fold increase (Nair et al., 2005), which was in agreement with analyses of sea urchins responding to marine bacteria (Rast et al., 2000) and other pathogen-associated molecular patterns (Terwilliger et al., 2007). However, it was not clear whether the increase in transcript content was due to increased gene expression in individual coelomocytes, to increased numbers of coelomocytes expressing 185/333 genes, or to both. To address this question, immunoquiescent sea urchins (n=4) were injected with LPS. In response, almost twice as many cells were present in the CF 12 hours after the challenge compared with pre-injection levels (P<0.05; 0 hours vs 12 hours post injection; Fig. 7A). After the initial increase, coelomocyte levels declined, and by 48 hours were similar to the pre-injection level (P>0.05; 0 hours vs 48 hours post injection). Frequencies of 185/333+ cells in response to LPS also increased (see supplementary material Fig. S1), but was slightly delayed relative to changes in total coelomocytes (Fig. 7B). The peak of 185/333+ cells occurred at 48 hours after LPS challenge, and was almost ten times greater than that prior to injection. These results indicated that the 185/333+ coelomocytes may be responsive to LPS challenge. Although there appears to be an increase in the 185/333 transcript content in coelomocytes (Nair et al., 2005), this change may be complicated by a corresponding increase in the number of 185/333+ cells.
Counts of the differential cell types were undertaken to characterize the types of 185/333+ cells and their frequencies. Of the total number of coelomocytes in immunoquiescent sea urchins (n=7) small phagocytes made up a variable percentage (13.14±10.5%) of which only 3.02±5.37% were 185/333+ (Fig. 7C and supplementary material Fig. S2). After challenge with LPS, the percentage of small phagocytes increased significantly (P<0.0005; pre- vs post-challenge) including the percentage of 185/333+ small phagocytes (P<0.05; pre- vs post-challenge). However, the percentage of polygonal cells in immunoquiescent sea urchins did not change in response to challenge with LPS, whereas the number of 185/333+ polygonal cells increased significantly (P<0.0005; pre- vs post challenge; Fig. 7C, and supplementary material Fig. S2). This suggested that the 185/333+ small phagocytes may be new cells that entered the CF, whereas the increase in 185/333+ polygonal cells may have been due to the induction 185/333 gene expression and protein production in existing cells. These results are in general agreement with the significant increases in the 185/333 transcripts observed in response to LPS and other immune challenges (Rast et al., 2000; Nair et al., 2005; Terwilliger et al., 2006; Terwilliger et al., 2007).
185/333 proteins have a broad array of molecular masses
A broad range of molecular masses has been deduced from full-length cDNAs (Terwilliger et al., 2006; Terwilliger et al., 2007) and ranged from 4.1 to 54.6 kDa, with most of the cDNAs (378 of 381 clones) encoding proteins ranging in size from 29.8 to 42.3 kDa. Protein sizes have also been deduced from truncated sequences (Terwilliger et al., 2007) and were slightly smaller, ranging from 1.5 to 38.9 kDa, with most being 14.7 kDa (281 of 309 clones). Although a few of the sequences from these two data sets may have been identical, the diversity of the mRNAs is, in part, based on the size of the 185/333 gene family, which has been estimated to have 100±20 alleles per individual (Terwilliger et al., 2006; Terwilliger et al., 2007) (K. Buckley, S. Munshaw, T. Kepler and L.C.S., unpublished). Variations in the molecular masses of these proteins would result from variations in element constitution in the mRNAs, in addition to any post-translational modifications. This prediction was indeed observed on western blots, which showed a wide range of 185/333 protein bands in all animals sampled (n=9). However, the range was expanded relative to what was expected with most bands over 60 kDa and some over 200 kDa (Fig. 8A,B). In addition to the wide range of protein sizes within individuals, size ranges differed among individual sea urchins (Fig. 8A). Although the range of 185/333 protein sizes was much larger than expected suggesting non-specific binding, the pre-immune sera failed to show any reactivity to sea urchin proteins (supplementary material Fig. S3). When western blots were analyzed with individual antisera, the complexity of the bands was not reduced.
When individual sea urchins (n=9) were challenged with LPS and coelomocytes were collected before challenge and at various times after challenge, the 185/333 protein content in the cell lysates varied substantially. Results for animal 7 showed a significant increase in the amount of 185/333 protein expression in coelomocytes between 1 and 6 hours after challenge (Fig. 8B). When individual sea urchins were sampled at 6, 24, 48 and 96 hours after immune challenge, the patterns and size ranges of bands did not appear to change (data not shown). The results presented here, in agreement with previously observed increases in 185/333 gene expression in challenged coelomocytes (Nair et al., 2005; Terwilliger et al., 2006; Terwilliger et al., 2007), indicated that the increase in the amount of the 185/333 protein was most likely a result of increases in the numbers of 185/333+ cells plus an induction of 185/333 gene expression.
Expression of recombinant 185/333 proteins Expression in bacteria
To simplify the 185/333 protein analysis, we employed a bacterial expression system to analyze a single recombinant protein. Inserts from several full-length cDNAs (Sp0032, Sp0164 and Sp0313) were expressed in E. coli (M15 strain). Because the Sp0164 and Sp0313 constructs were highly toxic, incubation and induction temperature was reduced to 25°C. Induction of Sp0032 resulted in slow bacterial growth and western blot analysis of cell lysates showed that the recombinant Sp0032 protein was expressed (Fig. 9A, lanes 1, 2). However, the band was ∼60 kDa, consistent with a dimer, rather than the expected size of 31.5 kDa. To determine whether the Sp0032 protein had aggregated as a result of being packaged into inclusion bodies, both pellets and supernatants from bacterial lysates were analyzed by western blot. Results showed that the Sp0032 protein was located in the cytoplasm and was mostly absent from the insoluble fraction (Fig. 9A, lanes 4, 5), indicating that the large size of the recombinant Sp0032 protein was not due to aggregation in inclusion bodies.
Expression in insect S2 cells
A recombinant 185/333 protein was also expressed in Drosophila S2 cells. Sp0296 was chosen because it was the most common message identified by Terwilliger et al. (Terwilliger et al., 2006). The element patterns for the deduced proteins encoded by Sp0296 and Sp0032 (expressed in bacteria) are E2 and E1, respectively. These patterns have the same 16 elements, differing only at nine amino acids, and Sp0032 has two additional elements near the middle, which encode 37 amino acids (Terwilliger et al., 2006, see Figure 1 within). After 1 or 2 days of induction, the recombinant protein appeared as monomers (∼31 kDa), dimers (∼66 kDa) and possibly tetramers (∼120 kDa; Fig. 9B). Several additional bands were observed between 45 kDa and 66 kDa, which may be evidence of either interaction of the recombinant protein with insect proteins, breakdown products of larger recombinant Sp0296 multimers, post-translational modifications, or the expression of truncated Sp0296 proteins (Terwilliger et al., 2007). However, the presence of 185/333 protein bands of ∼60 kDa in both the prokaryotic and eukaryotic expression systems strongly suggests homodimerization.
The first, most definitive report of sea urchin coelomocytes delineated four morphologically distinct cell types and described them as phagocytes, vibratile cells, and red and colorless spherule cells (Johnson, 1969a; Chien et al., 1970). The functions of most of these cell types are only vaguely understood (reviewed by Gross et al., 1999; Smith et al., 2006), and that for the colorless spherule cells are entirely unknown. The vibratile cells may be involved in initiating clot formation upon release of their granules (Johnson, 1969b; Bertheussen and Seljelid, 1978). Red spherule cells produce echinochrome, a red naphthoquinone molecule, which is released in the presence of bacteria (Johnson, 1969b) and has antibacterial properties against both Gram-positive and Gram-negative bacteria (Service and Wardlaw, 1984; Gerardi et al., 1990; Haug et al., 2002). Red spherule cells in hanging drop cultures move towards bacteria and line up in palisades forming a wall to block bacterial spread (Johnson, 1969b). Similarly, in vivo, they migrate towards wounded or infected tissue and form a dark rim of live tissue at the edge of damage or infection (Heatfield and Travis, 1975; Coffaro and Hinegardner, 1977; Höbaus, 1979).
The phagocytes constitute the largest percentage of the coelomocytes (Bertheussen and Seljelid, 1978; Smith et al., 2006) and were named based on their phagocytic activities (Johnson, 1969b; Bertheussen and Seljelid, 1978; Bertheussen, 1981; Edds, 1993). Large phagocytes are present in two forms: bladder amoebocytes that have lamellipodia protruding in all directions from the cell body, and cells with numerous long filopodia (Johnson, 1969b; Chien et al., 1970). These two morphological forms of phagocytes transition from one type to the other and back in response to changes in buffer osmolarity (Edds, 1977; Edds, 1980) (reviewed in Edds, 1985). Gross et al. (Gross et al., 2000) first described a third phagocyte type, which was based on small size and different morphology. Unlike the discoidal and polygonal phagocytes, small phagocytes do not generally spread into the lamellipodial morphology, but remain filopodial. The structural organization of the cytoskeleton, including the distribution pattern of microtubules, is different from that in the larger phagocytes. Furthermore, small aggregates of cells that may lead to syncytia formation (Johnson, 1969a; Johnson, 1969b; Chien et al., 1970; Bertheussen and Seljelid, 1978; Bertheussen, 1979), appear to be composed chiefly of 185/333+ small phagocytes, and these cells may be of central importance in initiating encapsulation. The observation that small phagocytes are the major coelomocyte type to express 185/333 proteins has defined them as a new cell type and has brought them to the forefront of investigations of sea urchin coelomocytes.
Although definitions of coelomocyte cell types have been based largely on morphology, more recently, sub-populations of discoidal and polygonal phagocytes have been characterized by the expression of SpC3, the sea urchin homolog of the complement component C3 (Gross et al., 2000). Similarly, there are subpopulations of phagocyte morphotypes that express 185/333 proteins, and, as a result of these observations, there may be several functional types of phagocytes. Future investigations are expected to delineate the relationships among the cellular morphotypes and putative functional differences that may be based on protein expression.
185/333 protein expression in phagocytes
One of the cellular compartments in which the 185/333 proteins are localized is the outer cell membrane, even though these proteins do not have a recognizable transmembrane region or GPI anchor site. The 185/333 proteins are likely secreted perhaps from subsets of both polygonal cells and small phagocytes, and then subsequently associate with the cell surface of the small phagocytes. The surface association could be based on interactions between the RGD motif on the 185/333 proteins and integrins present on the cell surface. Predictions from the sea urchin genome reveal a number of integrin gene models, including four β subunits and eight α subunits (Whittaker et al., 2006), of which one has been identified as a coelomocyte EST (Nair et al., 2005) (reviewed in Smith et al., 2006). The sea urchin complement homolog, SpC3, is localized in vesicles present throughout the cytoplasm of subsets of both large polygonal and smaller discoidal phagocytes (Gross et al., 2000). Unlike the distribution of the 185/333 proteins, most of the SpC3 is secreted from the phagocytes and is present in the fluid fraction of the CF, whereas most of the 185/333 proteins are associated with the cell fraction. We speculate that the 185/333 proteins are also secreted from both small phagocytes and polygonal cells and subsequently associate with the surface of the small phagocytes. Although the functions of the 185/333 proteins are not known, the increased frequency of 185/333+ coelomocytes in response to LPS is in agreement with increases in 185/333 gene expression in response to immune challenge (Rast et al., 2000; Nair et al., 2005; Terwilliger et al., 2006; Terwilliger et al., 2007) and implies an involvement in the sea urchin immune response.
The differential cell counts indicate that the percentage of total polygonal cells does not change in response to LPS challenge, while the percentage of 185/333+ polygonal cells increases. This suggests that some of the polygonal cells may be induced to express the 185/333 protein, although a steady state cell turnover of 185/333-negative (185/333–) to 185/333+ polygonal cells could also explain this result. However, there is an increase in both percentage of total small phagocytes and the percentage of 185/333+ small phagocytes in response to LPS challenge. This may be due to an increase in small phagocytes as a result of additional 185/333+ cells that enter the CF, but it may also be due to an induction of 185/333 protein expression by small phagocytes and/or surface binding of 185/333 proteins secreted by other cells. The change in 185/333 protein expression and the change in subpopulations of coelomocytes in response to an immune challenge indicates that this diverse protein family and this subset of coelomocytes are likely to be intimately involved in immune defense.
The 185/333 protein diversity
Significant nucleotide sequence diversity is present in the 185/333 cDNAs with a corresponding high level of diversity predicted for the amino acid sequences. This diversity is a result of SNPs, small indels, large gaps, and subsets of elements that are present in different mRNAs (Nair et al., 2005; Terwilliger et al., 2006; Terwilliger et al., 2007). A high level of protein diversity was also predicted from the size of the 185/333 gene family, which is estimated to be about 50±10 loci per genome, and is highly polymorphic among different individuals (Terwilliger et al., 2006) (K. Buckley, S. Munshaw, T. Kepler and L.C.S., unpublished). Not only do the 185/333 proteins show a broad range of sizes, but different suites of proteins are expressed by different individuals, illustrating the protein diversity both within individuals and among members of the population (Terwilliger et al., 2007) (N. M. Dheilly, S. V. Nair, L.C.S. and D.A.F., unpublished). Similar results have been found in other invertebrates that express proteins encoded by large gene families, including the lancelet Branchiostoma floridae (Litman et al., 2001; Litman et al., 2005) and the freshwater snail Biomphalaria glabrata (Stout et al., 2008).
The large size range of 185/333 proteins expressed in coelomocytes (30 to >200 kDa) indicates that these proteins either multimerize with each other or may bind to other sea urchin proteins. However, we have no evidence that the antisera are non-specific. The pre-immune sera did not recognize proteins in the CF (supplementary material Fig. S3) and did not recognize proteins in the bacterial lysate that corresponded with the 185/333 bands (Fig. 9, lane 3). The antisera did not recognize bacterial proteins when the recombinant 185/333 protein was not induced (Fig. 9, lane 1). Furthermore, preliminary results from western blots of two dimensional gels suggest that all three antisera are specific for 185/333 proteins (N. M. Dheilly, S. V. Nair, L.C.S. and D.A.F., unpublished). Mass spectrometric analysis of tryptic peptides from proteins that are recognized by anti-185/333 antisera identified mass charge ratios comparable with those of deduced 185/333 tryptic peptides. Proteomic analyses have found no evidence that the anti-185/333 antisera cross react with any other sea urchin proteins (D.A.R., unpublished). All of our available data are consistent with the antisera being highly specific for 185/333 proteins.
It is not known how the recombinant 185/333 proteins appear predominantly as dimers. The protein sequences do not indicate any obvious means of stable interaction, and they do not have cysteines to form disulfide bonds. Furthermore, it is not clear whether these proteins oligomerize in the sea urchin with other full-length or truncated forms to generate mixed aggregates, and/or if they interact with other proteins. There are proteins that are known to form stable dimers and oligomers in other systems that do not resolve to monomers under reducing conditions, including glycophorin A dimers (Lemmon et al., 1992), oligomers of α2 adrenergic receptors (Salahpour et al., 2003) and soluble N-ethylmaleimide-sensitive factor attachment receptors (SNAREs) (Weninger et al., 2003). SNAREs form a tight tetrahelix that functions in membrane fusion, and is resistant to denaturation. Although the structural predictions for the 185/333 proteins do not indicate alpha helices, the recombinant proteins form aggregates after freeze-thaw cycles associated with storage. These aggregates are of similar large size as observed in the western blots shown here, and do not dissociate in lysis buffer (Y.-O.K. and L.C.S., unpublished). It appears that the 185/333 proteins aggregate easily, and when present on the surface of small phagocytes, their functions may be to form oligomers between cells to initiate cell-cell interactions that leads to cell aggregation and syncytia formation.
The 185/333 gene system and the family of proteins that they encode is one of several examples of immune complexity that is present in organisms that do not have an adaptive immune system. For any host organism with a long generation time relative to its microbial pathogens, mechanisms for immune diversification are required to combat microbial diversification that is inherent with short generations. Advantageous changes in diversity on both sides of this struggle can be selected for and require a balance on both sides for the struggle to continue, i.e. for both host and pathogen to survive. Immune diversification in higher vertebrates is relatively well understood, whereas the possibility that this occurs in invertebrates has only recently come to the forefront. The mechanisms employed by invertebrates for immune diversification will be one of the questions to hold the attention of comparative immunologists in the future. Full sequence analysis of the complex 185/333 gene locus, which is composed of closely linked 185/333 genes (Buckley and Smith, 2007) (K. Buckley, S. Munshaw, T. Kepler and L.C.S., unpublished) is expected to provide some preliminary clues to the means by which sea urchins survive interactions with the microbes that are present in both the water and substrate in the marine habitat of the southern California Pacific kelp forest.
Materials and Methods
Polyclonal rabbit antisera were raised against synthetic peptides (Quality Controlled Biochemicals, Hopkington, MA) that corresponded to three elements that are present in all known 185/333 cDNAs (supplementary material Table S1). The peptide sequences were derived from the most common protein sequence deduced from several hundred cDNA sequences (Nair et al., 2005; Terwilliger et al., 2006; Terwilliger et al., 2007). Only those antisera for which the prebleeds did not cross react with CF proteins by western blot (supplementary material Fig. S1) were chosen for this study. Anti-185-66, anti-185-68, anti-185-71 antibodies were used equally in dilutions.
Purple sea urchins, Strongylocentrotus purpuratus, were supplied by Marinus Scientific Inc. (Long Beach, CA). Animals were maintained according to Gross et al. (Gross et al., 2000). Immunoquiescent animals were generated by long term housing (>8 months) without significant disturbance as described previously (Gross et al., 1999; Clow et al., 2000).
Sea urchins were immunologically activated by injections of lipopolysaccharide (LPS; Sigma-Aldrich, St Louis, MO; 1.0 μg/μl) in artificial coelomic fluid (aCF; 10 mM CaCl2, 14 mM KCl, 50 mM MgCl2, 398 mM NaCl, 1.7 mM Na2HCO3, 25 mM Na2SO4) such that each animal received 1 μg LPS/ml of CF as described by Smith et al. (Smith et al., 1992). The volume of CF in adult sea urchins range from 12 ml to as much as 40 ml for large specimens. Sea urchins received one or two injections at 0 hours and 24 hours. CF (100-200 μl) was withdrawn (see below) just prior to the first injection and at 6 hours, 12 hours, 24 hours, 48 hours, 72 hours, 96 hours and 192 hours after the last LPS challenge. Control animals received equivalent volumes of aCF without LPS.
Coelomocytes were withdrawn as described previously (Clow et al., 2000) into syringes pre-loaded with 0.3-0.5 ml ice-cold calcium- and magnesium-free sea water containing 30 mM EDTA and 50 mM imidazole pH 7.4 (CMFSW-EI) (Gross et al., 1999). Diluted CF was expelled into 2 ml microcentrifuge tubes on ice containing an additional 0.5 ml CMFSW-EI. Cells were counted and adjusted to 2×104 to 2×106 cells/ml using CMFSW-EI, depending on the subsequent protocol.
Cell suspensions (100 μl) were settled onto poly-L-lysine [0.1 mg/ml Tris-buffered saline (TBS) 0.2 M Tris pH 7.4, 0.2 M NaCl]-coated glass coverslips (Sigma-Aldrich) or slides (Polysciences, Warrington, PA, or Newcomer Supply, Middleton, WI) and incubated in a humid chamber at room temperature (RT) for 5 minutes. CMFSW-EI was replaced with 100 μl of coelomocyte culture medium (CCM) (Henson et al., 1999) and incubated for 30 minutes at RT in a humid chamber. A three part fixative was used for most experiments: (1) a prefix in 0.002% glutaraldehyde (Sigma-Aldrich) in CCM for 5 minutes, (2) a fix in 1% formaldehyde, 0.25% Triton X-100 in AC320 buffer (320 mM sucrose, 75 mM KCl, 2 mM MgCl2.6H2O, 20 mM EGTA, 20 mM Pipes pH 7) for 5 minutes, and (3) a postfix in cold (–20°C) methanol for 5-13 minutes. Coverslips were washed three times (5 minutes each) in phosphate-buffered saline (PBS; 0.15 M phosphate buffer pH 7.4, 0.2 M NaCl) and incubated in blocking solution [2% v/v normal goat serum (NGS) with (1%) or without bovine serum albumin (BSA) in PBS] for 30 minutes in a humid chamber. Coelomocytes were incubated with an equal mixture of anti-185-66, anti-185-68 and anti-185-71 antibodies (1:4000 to 1:6000 dilution in blocking solution) plus mouse monoclonal anti-actin antibody (MP Biomedicals, Solon, OH; 1:300 to 1:600 dilution) for 45 minutes at RT in a humid chamber. In some experiments, cells were incubated with only one of the anti-185 antibodies and in others the anti-actin antibody was replaced with mouse monoclonal anti-α-tubulin antibody (1:300 dilution; Sigma-Aldrich). Cells were washed as described above and incubated with a mixture of goat anti-rabbit immunoglobulins (GαR-Ig) conjugated to Alexa Fluor 568 (Pierce Biotechnology, Rockford, IL; 1:400 dilution in blocking solution), donkey anti-mouse Ig (DαM-Ig) conjugated to Alexa Fluor 488 (Pierce; 1:200 to 1:400 dilution) and Hoechst 33258 (Sigma-Aldrich; 1:1000 dilution) in blocking solution for 45 minutes at RT in a humid chamber. Negative controls included pre-immune sera as well as omission of the anti-185 antisera. Cells were washed as described above and mounted with ProLong Gold Antifade (Invitrogen, Carlsbad, CA).
Coelomocytes were collected as described above and settled onto glass coverslips that were either coated with poly-L-lysine or were untreated. Perfusion chambers were constructed on a microscope slide using coverslips mounted on coverslip fragments that functioned as shims. Live cells were imaged (described below) followed by fixation and incubation with antibodies as described above prior to re-imaging. All steps were accomplished in the perfusion chambers. In other experiments live coelomocytes were treated with anti-185 antibodies prior to glutaraldehyde prefixation and formaldehyde fixation (with no Triton or methanol postfixation) and secondary antibody incubation. In these experiments anti-actin was included as a control for membrane permeabilization during processing. Finally in other experiments cells were fixed in prefix and fix (minus Triton) and then labeled with anti-185 and anti-actin antibodies.
Fixed cells were inspected with an Axioplan fluorescence microscope (Zeiss, Oberkochen, Germany) or a MRC 1024 confocal laser scanning system (Bio-Rad Laboratories, Hercules, CA) attached to an IMT2-RFC inverted microscope (Olympus, Center Valley, PA). Live cells were viewed using an E600 microscope (Nikon, Melville, NY) fitted with a 60× NA 1.4 planapochromatic phase-contrast objective lens coupled to a CCD camera (Hitachi Ltd, Tokyo, Japan). Video images from the camera were subjected to real-time digital enhancement involving background subtraction, frame averaging and contrast enhancement using a Hamamatsu Argus 10 image processor (Hamamatsu Photonics, Hamamatsu City, Japan). For some experiments, a 40× NA 0.75 plan phase-contrast objective lens was used. The E600 microscope was also used to collect conventional epifluorescence images of stained cells. Laser scanning confocal images of fluorescently labeled cells were collected on a Fluoview 500 instrument (Olympus) employing argon and krypton lasers for the appropriate excitation wavelengths and a 60× NA 1.4 planapochromatic objective lens.
Mean percentage and standard deviation were calculated for 185/333+ cells versus the total number of coelomocytes per microscope field (n>5 fields) using 40× magnification. The statistical significance of differences between mean values was determined by two-tailed Student's t-tests for populations with unequal variance using the SPSS software package (Microsoft, Sydney, Australia) or Microsoft Excel (Microsoft Corp., Redmond, WA).
Samples of pelleted coelomocytes were resuspended in 2×SDS lysis buffer (1:1) (Gross et al., 2000), resolved under reducing conditions by 12% SDS-PAGE, and transferred by electroblotting (HEP-1 Panther Semi Dry Electroblotter, Owl Separation Systems, Portsmouth, NH) onto polyvinylidene fluoride membrane (Immobilon-PSQ Transfer Membrane; Millipore Co., Bedford, MA) using transfer buffer (10% methanol, 20 mM Tris pH 8.8, 150 mM glycine, 0.05% SDS) for 1 hour at 100 mA per filter (7.5×10 cm). After transfer, membranes were stained with 0.1% PonceauS (Sigma-Aldrich) in 5% acetic acid and rinsed in 5% acetic acid to visualize sample bands and mark the positions of the standard bands. Membranes were washed in Tris-NaCl-Tween (TNT; 25 mM Tris pH 7.4, 0.5 M NaCl, 0.1% Tween 20) at RT for 10 minutes before rocking overnight in `blotto' (5% milk proteins in TNT) at RT. Blotto was replaced with an equal mixture of anti-185 antisera (1:10,000-1:15,000 dilution for each antisera in blotto) for 1.5 hours at RT with rocking, followed by three washes in TNT. Membranes were incubated with GαR-Ig conjugated with horseradish peroxidase (GαR-Ig-HRP; 1:10,000-1:30,000 dilution in blotto; Pierce) for 1.5 hours at RT with rocking. Membranes were washed in TNT followed by washes in TN buffer (TNT without Tween 20). Membranes were incubated briefly in SuperSignal West Pico Chemiluminescent Substrate System (Pierce) before exposure to X-OMAT AR X-ray film (Eastman Kodak, Rochester, NY).
Recombinant 185/333 protein expression
Degenerate primers F2.1 and R10 (supplementary material Table S2) were used to amplify the coding region of several 185/333 cDNAs: Sp0032 (GenBank acc. no. DQ183168), Sp0164 (GenBank acc. no. DQ183183), Sp0313 (GenBank acc. no. DQ183171). These primers omitted the leader and amplified elements 1-25 (Terwilliger et al., 2006). The amplicons were ligated into pQE-UA expression vector (Qiagen, Hilden, Germany), transfected into chemically competent M15 strain of E. coli and selected with ampicillin (100 μg/ml) and kanamycin (25 μg/ml). Overnight cultures were inoculated into fresh broth, grown at 25°C until the optical density (OD600) reached 0.6-0.7 and were induced with 1 mM isopropyl β-D-thiogalactoside (IPTG; Fisher Science, Atlanta, GA). Cells were harvested at various times after induction, resuspended in 5× SDS lysis buffer with a protease inhibitor cocktail mix (P2714, Sigma-Aldrich) and analyzed by western blot. Uninduced cultures were treated similarly.
To determine whether the recombinant 185/333 protein was localized within inclusion bodies, induced cultures were pelleted for 10 minutes at 4°C, resuspended in 10 mM Tris pH 7.4 and lysed by sonication using a probe sonicator attached to a Sonic Dismembrator (model 500, Fisher Science). Lysates were centrifuged at 1.9×104g for 15 minutes at 4°C, and the pellets and supernatants were analyzed by western blot.
Insect cell expression
The Drosophila expression system (DES) vector pMT/BiP/V5-His6 A (Invitrogen) was used to make an expression construct containing the coding region from Sp0296 (GenBank acc. no. DQ183121). The insert was amplified by PCR using forward and reverse primers (supplementary material Table S2), which generated a fragment encoding amino acid Q22 to the 3′ end of the open reading frame and included elements 1-25a (Terwilliger et al., 2006). The incorporation of SmaI and AgeI restriction sites onto the ends of the amplicon enabled directional cloning. After restriction digestion and ligation of the amplicon into the vector according to standard methods, the construct was transformed into E. coli TOP10' cells (Invitrogen) and correct ligation was verified by sequencing. The construct was transformed into Drosophila melanogaster S2 cells using Cellfectin reagent (Invitrogen) following the manufacturer's protocols. Propagation of the transformed cells and induction of recombinant 185/333 protein expression was performed following the DES protocols. Resultant media were analyzed for 185/333 protein by western blot.
The authors would like to thank Priya Moorjani who helped with 185/333 protein expression in E. coli. Katherine Buckley, Rebecca Easley and Lindsay Edwards provided editorial improvements on the final draft of the manuscript. This work was funded by a George Washington University Columbian College postdoctoral award to V.B., a National Institutes of Health (GM60925-02) award and a National Science Foundation instrumentation award (MRI 0320606) to J.H.H., an Australian Research Council Discovery Grant and funding from the Outside Studies Program from Macquarie University to D.A.R., and an award from the National Science Foundation (MCB-0424235) to L.C.S.