An Invertebrate Histocompatibility Complex

Corresponding author: Leo W. Buss, 165 Prospect St., Yale University, New Haven, CT 06520-8106., leo.buss{at}yale.edu (E-mail)

Communicating editor: D. CHARLESWORTH

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We have developed defined genetic lines of the hydroid Hydractinia symbiolongicarpus and confirmed earlier results showing that allorecognition is controlled by a single chromosomal region within these lines. In a large backcross population, we detected recombinants that display a fusibility phenotype distinct from typical fusion and rejection. We show that this transitory fusion phenotype segregates in a fashion expected of a single Mendelian trait, establishing that the chromosomal interval contains at least two genes that interact to determine fusibility. Using bulked segregant analysis, we have identified amplified fragment length polymorphisms (AFLP) cosegregating with fusibility, used these markers to independently confirm linkage of the two loci, and constructed a 3.4-cM map of an invertebrate histocompatibility complex.

THE hydroid Hydractinia has a long been a favored model in efforts to understand invertebrate allorecognition (VON HAUENSCHILD 1954 , VON HAUENSCHILD 1956 ; MULLER 1964 ; BUSS et al. 1984 ; MULLER et al. 1987 ; LANGE et al. 1989 , LANGE et al. 1992 ; BUSS and GROSBERG 1990 ; SHENK and BUSS 1991 ; GROSBERG et al. 1996 ; MOKADY and BUSS 1996 ). Colonies, which often encrust the shells of hermit crabs, grow by elongation and branching of stolons. When two or more larvae recruit to the same substratum, stolons of different colonies may eventually come into contact. Close proximity of an approaching stolon induces the production of a new stolon tip along the flank of the stolon being approached. If the two colonies are histocompatible, the two tips will first adhere, and then fuse, establishing functional gastrovascular continuity and a permanent genetic chimera. In contrast, tips of incompatible colonies fail to adhere to one another, but instead swell with the migration of nematocysts, which discharge to effect tissue damage.

The transmission genetics of allorecognition in Hydractinia were first addressed in the mid-1950s by von Hauenschild, who suggested a single-locus, codominant system (VON HAUENSCHILD 1954 , VON HAUENSCHILD 1956 ). While the vast majority of his data supported a single-locus model, segregation ratios in some cases proved problematic. MOKADY and BUSS 1996 inbred Hydractinia symbiolongicarpus to standardize the genetic background and subsequently applied a conventional incross/intercross/backcross analysis. Within these defined genetic lines, Hydractinia allorecognition segregated as a one-locus trait with codominant expression of alleles, such that two colonies fuse if they share at least one allele.

We here report the further development of these inbred and congenic lines and confirm, with an expanded population, that segregation of fusibility follows single-locus codominant Mendelian expectations. A fusibility phenotype first observed by VON HAUENSCHILD 1954 , VON HAUENSCHILD 1956 , in which interacting colonies initially fuse only to later separate, appeared at low frequencies in this cross. We report a series of crosses that establish that the transitory fusion (TF) phenotype itself behaves as a single Mendelian trait linked to the original allorecognition locus. Moreover, we show that this TF phenotype is readily detected in only one of the two conventional fusion assays. Finally, we use bulked segregant analysis to identify molecular markers linked to the interval and report the use of these markers to construct a genetic map of an invertebrate histocompatibility complex.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Animal cultures, crosses, and fusibility assays:
Culture methods have been described in detail elsewhere (BLACKSTONE and BUSS 1991 ). Briefly, colonies of H. symbiolongicarpus were grown on glass microscope slides and maintained at 20° ± 2° in 35-liter aquaria in recirculating artificial seawater (Reef Crystals) with daily 25% water changes. Animals were typically fed to repletion three times a week with 3- to 4-day-old nauplii of Artemia salina. Sexually mature colonies released gametes 50–70 min after exposure to light (BUNTING 1895 ; BALLARD 1942 ). Fertilized eggs develop into planula larvae within 36 hr at room temperature. Larvae were exposed to 53 mM CsCl for 3–4 hr to induce metamorphosis and competent planulae were placed on glass microscope slides where they completed their metamorphosis within 48 hr (MULLER 1973 ). Single-polyp colonies were fed daily, by hand when necessary, with newly hatched A. salina nauplii until they developed 3–5 polyps. Fusibility phenotypes were evaluated using two types of compatibility assays, the polyp and the colony fusibility tests. In the polyp fusibility test (LANGE et al. 1992 ), polyps from the two contesting colonies were excised, strung onto a human hair, and held with their cut ends opposed with agar blocks. After 2 hr, the agar blocks were removed. Fusibility was assayed on the following day. Compatible colonies developed continuous endodermal and ectodermal cell layers and a common gastric cavity; incompatible polyps, on the other hand, separated. In the colony fusibility test (MULLER 1964 ; IVKER 1972 ), an explant of stolonal mat containing 3–5 polyps from each of the colonies to be tested was placed 0.5 cm from one another and attached with a string onto a glass slide. Tissue attached to the glass within 24–48 hr and the string was then removed. Fusibility was assayed when colonies grew into contact with one another. Colonies were observed daily until they contacted and for at least the first 5 days thereafter. When observed to fuse, colonies were typically observed three times per week for an additional 2 weeks and once a week for an additional 3–4 weeks or until at least one member of the pair developed gametes. Experiments with all colonies displaying fusion results that deviated from conventional fusion and rejection responses (BUSS and SHENK 1990 ) were repeated at least three times.

Development and characterization of inbred and congenic lines:
MOKADY and BUSS 1996 reported the production of a partially inbred line of H. symbiolongicarpus produced by six generations of brother-sister inbreeding, and an F₂ population derived from an outcross from this line to another wild-type individual. From this starting point, we developed two inbred lines and a congenic line. The two inbred lines (33 and 41) were produced by brother-sister inbreeding, whenever possible, to fix different alleles at the allorecognition (alr) locus identified by MOKADY and BUSS 1996 in two different genetic backgrounds (Fig 1). Breeders for each successive generation were chosen only if they fused within the same line and rejected the other line in a polyp fusibility assay. Fusibility within and between individuals of the terminal populations of both lines is reported elsewhere (CADAVID 2001 ) and conforms to Mendelian expectations. The 33 line is said to be homozygous for the f allele (alr-f/-f) and the 41 line homozygous for the r allele (alr-r/-r).

View larger version (34K): In this window In a new window Download PPT slide   Figure 1. Mating program used to generate the 41 and 33 inbred lines. The shaded area represents the work of MOKADY and BUSS 1996 and the nonshaded area is the continuation of that inbreeding program in this study. Circles and squares represent females and males, respectively. Horizontal lines represent crosses, thick horizontal lines are consanguineous crosses, and vertical lines represent descendants from such matings. Wild-type colonies (closed symbols) WT1 and WT2 are the founders of the 33 inbred line. Wild-type colony 045 was mated to the 33 line to initiate the 41 line.

A congenic line (431) was produced to maintain variation at alr within the genetic background of the 33 line (Fig 2). This line was generated by three successive backcross cycles to the 33 line, starting with the cross of an F₂ animal of the 41 line with an F₇ animal of the 33 line. Donor parents for each backcrossing cycle were chosen to fuse both inbred lines in the polyp assay, i.e., to be a heterozygote (alr-f/-r). Segregation of fusibility in a cross between two heterozygous, third-generation colonies from the congenic line, designated as 431 in Fig 2, are reported elsewhere (CADAVID 2001 ) and likewise conformed to Mendelian expectations.

View larger version (52K): In this window In a new window Download PPT slide   Figure 2. Mating program used to generate the congenic line. The congenic line was generated to introgress the allotypic determinant of the 41 line into the genetic background of the 33 line. The terminal cross, labeled 431, is a population. Symbols and shading as described for Fig 1.

Segregation of fusibility in a large backcross population:
All previous crosses with these inbred and congenic lines (MOKADY and BUSS 1996 ; CADAVID 2001 ) had involved population sizes of <100 individuals and hence were unsuitable for either detailed genetic mapping or the detection of closely linked loci. A large backcross population was generated by crossing a heterozygous alr-f/-r animal of the 431 line (431-63) to an alr-f/-f inbred animal of the 33 line (833-8). All 490 offspring generated in this cross were tested for fusibility against a homozygous alr-r/-r tester of the 41 line (colony 4117-2) using colony fusion assays. Of these, 127 colonies were also tested using the polyp fusion assay.

Bulked segregant analysis:
Molecular markers cosegregating with fusibility were identified by bulked segregant analysis (GIOVANNONI et al. 1991 ; MICHELMORE et al. 1991 ). For a trait known to be controlled in a single chromosomal interval, DNA pools are generated from multiple offspring of a single cross that do and do not display the trait in question. Amplified fragment length polymorphism (AFLP) fingerprints are generated for each pool and markers are detected by presence in one pool and absence in the other. Genomic DNA was extracted from individual colonies of H. symbiolongicarpus following the method of SHURE et al. 1983 . Two different sets of pools were used to generate markers linked to the allorecognition interval. A first set involved the two homozygous classes, i.e., 10 alr-f/f individuals vs. 10 alr-r/r individuals, from the 431 population (Fig 2). The second set involved animals from the backcross population (431-63 x 833-8) in which one pool contained 10 homozygotes (alr-f/f) and the other 10 heterozygotes (alr-f/r). Markers present in one pool and absent in the other were subsequently screened on five individuals from each pool to assay repeatability. If repeatable, the offspring of the backcross population were screened for the marker as detailed below.

Amplified fragment length polymorphisms:
AFLPs were performed as previously described (VOS et al. 1995 ) with minor modifications. A total of 1–400 ng of genomic DNA were double digested with 5 units of EcoRI and MseI (New England Biolabs, Beverly, MA) for 3–8 hr at 37° in the presence of 50 pmol of MseI-adaptor and 5 pmol of EcoRI-adaptor, 50 ng/µl BSA, 1 unit of T4 DNA ligase (Promega, Madison, WI) in a total volume of 10 µl. This digestion/ligation mix was subsequently diluted with 90 µl of 10 mM TE pH 8.0. Four microliters of the diluted mix were used as template for the preselective amplification reaction using primers with one selective nucleotide at the 3'-end (+1/+1). Three different combinations of selective nucleotides were used for this step: EcoRI-A/MseI-C, EcoRI-A/MseI-G, and EcoRI-T/MseI-G (EcoRI and MseI refer to the primer core sequence, 5'-gactgcgtaccaattc-3' and 5'-gatgagtcctgagtaa-3', respectively). The preselective amplification mix contained 10 mM Tris pH 8.0, 1.5 mM MgCl₂, 0.2 mM dNTPs, 0.5 units of Taq polymerase, 50 mM KCl, and 30 ng of each primer for a total volume of 20 µl. The amplification was carried out at 95° for 2 min followed by 20 cycles of 94° for 30 sec, 56° for 30 sec, and 72° for 1 min, and a final cycle of 60° for 30 min. The amplification product was diluted with 80 µl of TE pH 8.0, and 3 µl of it was used as template for the selective amplification reaction. For this reaction, primers with three selective nucleotides were used (+3/+3). The EcoRI primer was 5'-end labeled with one of three fluorescent dyes, FAM, HEX, or NED (ABI, Columbia, MD). PCR reactions were performed using 5 ng of labeled EcoRI primer and 30 ng of unlabeled MseI primer. The PCR profile consisted of a denaturing cycle of 95° for 2 min followed by a cycle of 94° for 30 sec, 65° for 30 sec, and 72° for 1 min. For the next 11 cycles the annealing temperature was decreased by 0.7° for each cycle and 24 additional cycles were made using this last annealing temperature. Samples were run on a 377 ABI automated sequencing machine for 3.5 hr and analyzed using the GeneScan software (Perkin-Elmer, Norwalk, CT). One hundred ninety-two and 64 different primer combinations (+3/+3 selective nucleotides) were used to screen the 431 population (Fig 2) and the large backcross population (431-63 x 833-8), respectively, where each primer combination amplified an average of 40 dominant AFLP loci.

Marker isolation, cloning, sequencing, and PCR typing:
Four AFLP markers that cosegregated with fusibility (F29, F18, F28, and R28) were cloned and sequenced. Briefly, 10 µl of diluted AFLP preselective amplification from each of the selected AFLP markers were used as template for 50-µl reactions. The reaction contained 30 ng of the corresponding +3/+3 AFLP primers with no fluorescent labeling and was amplified with the same profile used for AFLPs as described above. This reaction amplifies preferentially fragments having EcoRI adaptors at one end and MseI adaptors at the other (VOS et al. 1995 ). PCR products were run in a 2.5% NuSieve agarose gel in 0.5x TBE at 6 V/cm. Bands of the appropriate size were picked from the gel and reamplified using the same reaction conditions as that of the original amplification. Products were purified using the Qiaex II gel extraction kit (QIAGEN, Chatsworth, CA). AFLP markers were cloned into the pGEM vector (Promega) and 10–12 clones/marker were sequenced in both directions using the T7 and SP6 promoter primers in an ABI 377 automated sequencing machine.

Two of the four PCR products that we cloned (F29 and R28) proved to be heterogeneous as they were composed of two classes of DNA sequences. The other two PCR products (F18 and F28) yielded only one class of DNA sequences. Specific PCR primers were designed for each of the different sequences and used to amplify 10 alr-f/f and 10 alr-r/r individuals. Bona fide primers were expected to amplify fragments only in individuals from one homozygous group. From the two pairs of primers derived from each F29 and R28 AFLP marker, only one pair amplified fragments cosegregating with alr, and they were selected for further analysis. Primers derived from F18 and F28 AFLP markers amplified bands that cosegregated absolutely with alr. Primers thus selected had the following sequences: F29, 5'-cgccacctctcgcagctacc-3' and 5'-gtgcattccttcagattaag-3'; F18, 5'-aacagacgaaatgggaaatc-3' and 5'-agcataaaatacataccacc-3'; F28, 5'-ctcagtttagcgactttcac-3' and 5'-gctgactctgccggttcggt-3'; R28, 5'-ctcagtttagcgactttcac-3' and 5'-gctgaatagtctctgccggt-3'. These PCR markers were used to screen the large backcross population (431-63 x 833-8; with marker F18, n = 505 and marker R28, n = 311) and terminal individuals of both inbred lines (33 and 41). In addition to markers screened by PCR, two additional markers (markers R174 and R194) were screened in a sample of the large backcross population (n = 86 and 120, respectively) by AFLP reactions. Genotyping for all markers was repeated at least three times for any animal detected to be recombinant for any marker.

Mapping populations and linkage analysis:
Markers identified in the bulked segregant analysis were evaluated for departure from the single-locus Mendelian expectations in the backcross population using a ² test for goodness of fit. Linkage analysis was performed using MAPMAKER software (LANDER et al. 1987 ). Linkage between loci was established by two-point analysis using the log-likelihood approach where recombination fractions were estimated by maximum likelihood (LIU 1998 ). Loci were considered linked if the LOD score was 3.0, which represents a relative likelihood ratio of linkage to nonlinkage of 1000:1 (OTT 1991 ). Recombination fractions were converted to map distances using Kosambi's map function (KOSAMBI 1944 ). Locus ordering was done by multi-point analysis using maximum likelihood (LANDER et al. 1987 ). Relative confidence of map order was evaluated by a likelihood-ratio test (LIU 1998 ). Hydractinia haploid genome size has been estimated as 7.5 x 10⁸ bp by microspectrophotometry of Fuelgen-stained cells and C₀t curve analysis (M. A. SHENK, C. W. CUNNINGHAM and M. H. DICK, unpublished results).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Allorecognition in H. symbiolongicarpus is controlled in a single chromosomal interval:
The inbred and congenic lines generated in this study are presented in Fig 1 and Fig 2, respectively. The terminal generation of the 33 inbred line, the 41 inbred line, and the congenic line had an expected inbreeding coefficient of 0.82, 0.80, and 0.82, respectively. A large backcross population was generated from a cross between a terminal congenic line heterozygote (f/r, colony 431-63) and an eighth-generation 33 line inbred colony (f/f, colony 833-8). Fusibility was assayed against an r/r tester (colony 4117-2) using the colony fusibility assay. Of 490 colonies, 239 fused and 251 rejected the r/r tester, a result not significantly different from the expected 1:1 Mendelian ratio (² = 0.29, P = 0.59).

Alternative fusibility phenotypes appear at low frequencies:
Of the 239 colonies that fused in the colony fusion assay, 226 remained fused for the duration of the observation period. The remaining 13 individuals displayed one of three novel fusibility phenotypes. Each of these assays was repeated a minimum of three times.

Of the 13 colonies, 10 displayed a transitory fusion phenotype resembling that originally described by VON HAUENSCHILD 1954 , VON HAUENSCHILD 1956 (Fig 3). In these cases, colonies initially adhere to one another upon contact of ectodermal cells. Within 2–4 hr postcontact a common endodermal gastrovascular system was established within which transport of food particles from one colony to the other was observed. In these initial stages the reaction is indistinguishable from a permanent fusion. After 12–24 hr postcontact, a gray band appeared at the point of initial contact, which subsequently spread to form a line spanning the original contact zone. The emergence of the gray line was accompanied by occlusion of the once-fused endodermal canals. Transport between colonies became increasingly disrupted and 12–24 hr after the first appearance of the gray line, colonies separated from one another. From this point on, the fusion phenotype was indistinguishable from a rejection reaction, except at the extreme growing edges of the contact zone, which, upon contact, displayed the same time course and phenomenology described above.

View larger version (90K): In this window In a new window Download PPT slide   Figure 3. Temporal progression of the TF phenotype. (A) Functional gastrovascular continuity is established between the colonies. (B) A gray line of dense tissue starts to develop at the contact margin while transport becomes increasingly disrupted. (C) A gray line of necrotic tissue separates the colonies. (D) In 12–24 hr postcontact, colonies have separated from one another other, except at the extreme margins of the contact zone. Colonies in B and C are the same fusion tests at different time points; A and D were chosen from other fusion tests to capture earlier and later stages, respectively. Bar, 0.5 cm.

The remaining three colonies displayed phenotypes that appeared to involve modifications in the time course and progression of transitory fusion. Specifically, colonies BC129 and BC393 displayed typical fusion for 7–21 days following initial contact, after which a line of occluded gastrovascular canals appeared similar to that observed in TF. However, unlike TF, this line did not demarcate a zone of subsequent separation, but rather the line disappeared over a period of several days and the colonies were thereafter indistinguishable from normal fusions. The final phenotype, displayed by colony BC375, initially fused and, like transitory fusion, eventually separated. The time course of this event differed from TF, with the first appearance of gastrovascular occlusion occurring 3–4 days postcontact and complete separation of the two colonies occurring after 8–14 days.

Transmission genetics of transitory fusion reveals a second allorecognition locus:
The appearance of TF at a low frequency in a population otherwise segregating in a monofactorial fashion suggested a system of linked loci. We hypothesized that colonies from the large backcross population displaying TF were recombinants between two linked loci, that is, colonies that share an allele at one of the two loci, but that share no alleles at the other locus.

We therefore designated the original alr locus described by MOKADY and BUSS 1996 as alr1 and hypothesized the existence of a second locus alr2. The 33 and 41 inbred lines are expected to be homozygous at both loci where alr1 has the previously described alleles f and r, while alr2 is given alleles and ß. Thus, the 33 line is predicted to have the genotype alr1-f, alr2-/alr1-f, alr2-, hereafter f/f; the 41 line is predicted to have the genotype alr1-r, alr2-ß/alr1-r, alr2-ß, hereafter rß/rß. Likewise, a colony fusing both inbred lines is predicted to be a double heterozygote with genotype alr1-f, alr2-/alr1-r, alr2-ß, hereafter f/rß. If we assume that alleles of the second locus exhibit codominance, then colonies that displayed TF against the 33 line and fused with the 41 line would be predicted to have the recombinant genotype rß/r (or rß/fß), whereas colonies that displayed TF against the 41 line and fused with the 33 line would be expected to possess the recombinant genotype f/fß (or f/r).

Two types of crosses were performed to test these predictions. The first mating involved two individuals from the 431 population, one that displayed TF against the 41 line and fused with the 33 line (colony 431-85, hypothesized genotype f/fß) and the other that fused with the 33 line and rejected the 41 line (colony 431-72, hypothesized genotype f/f). The second mating also involved two individuals from the 431 population, one that displayed TF against the 33 line and fused with the 41 line (431-55, hypothesized genotype rß/r), and the other that fused with the 41 line and rejected the 33 line (431-2, hypothesized genotype rß/rß).

Results appear in Table 1. All offspring of the first mating (f/fß x f/f) fused with the f/f tester (833-8; n = 13). Fusibility against the rß/rß tester (4117-2), on the other hand, segregated for rejection and TF in proportions indistinguishable from the 1:1 Mendelian ratio (n = 27, ² = 0.93, P = 0.34). The second mating was hypothesized to be the reciprocal of the first (rß/r x rß/rß). All tested offspring fused with the rß/rß colony (n = 25) whereas rejection and TF segregated in a 1:1 ratio in tests against the f/f tester (n = 24, ² = 0.17, P = 0.68). These results are consistent with a two-locus model wherein colonies fuse if they share one (or more) allele(s) at both fusibility loci, reject if they share no alleles at either locus, and display TF if they share only one allele at either fusibility locus.

Transitory fusion is not detected in the polyp fusion assay:
Polyp fusibility was assayed for 127 of the 490 colonies tested in the backcross population. In addition, 18 colonies were tested for polyp fusibility in one of the crosses segregating for transitory fusion (431-72 x 431-85). In all cases, the results of the polyp fusion and colony fusion were identical. Colonies displaying TF in the colony assay remained fused in the 24-hr polyp assay (n = 8 from the backcross population and n = 9 from the TF cross). Close examination of polyp fusions involving colonies that displayed TF in colony assays showed the ectoderm of fusion zone between two polyps to be granular, rather than the smoothly joined epithelia of a standard polyp assay. This effect, however, was subtle and would not be easily detected in an assay that is scored as the appearance of two separated polyps vs. a single double-headed polyp.

Genetic map of the Hydractinia allorecognition complex:
Bulked segregant analysis was performed using the 431 population and the large backcross population. Fig 4 shows a typical AFLP marker cosegregating with fusibility. The 431 population yielded four dominant AFLP markers and the backcross pools an additional two markers. In the 431 population, markers F29 (EcoRI-ACC/MseI-CTA), F18 (EcoRI-AAC/MseI-CAC), and F28 (EcoRI-ACC/MseI-CAT) cosegregated with the f haplotype and marker R28 (EcoRI-ACC/MseI-CAT) with the rß haplotype. In the large backcross population, marker 174 (247 bp, primer combination EcoRI-TCT/MseI-CTA) and marker 194 (74 bp, primer combination EcoRI-TCC/MseI-CAA) were found to cosegregate with the rß haplotype.

View larger version (44K): In this window In a new window Download PPT slide   Figure 4. Bulked segregant analysis used to generate AFLP markers. The figure shows AFLP fingerprints generated with the primer combination EcoRI-ACC/MseI-CTA. The analysis was performed with the GenScan software where peaks represent AFLP fragments. Fragment sizes are shown in the x-axis and fluorescence intensities in the y-axis. (Top) Pools of 10 f/f and rß/rß double homozygotes, respectively. Note that the 100-nucleotide fragment (arrow) is present in the f/f pool and absent in the rß/rß pool. Similarly, the 104-nucleotide fragment (arrow) is unique for the rß/rß r pool. (Bottom) Amplification of three individual colonies of each nonrecombinant ARC genotype.

Of these, four AFLP products were subsequently cloned and sequenced, and specific primers were synthesized. Primers for marker F29 amplified a 62-bp product. F18 primers coamplified two allelic products of 386 bp (R18) and 249 bp (F18), which segregated with the rß haplotype and with the f haplotype, respectively. Sequence comparison of these allelic forms showed that they differed by an indel of 137 bp at position 103 (data not shown). Primers for marker F28 amplified a 60-bp product identical to that amplified by R28 primers (65-bp product) except for a 5-bp indel at position 5. F28 cosegregated with the f haplotype and the 65 bp of R28 cosegregated with the rß haplotype, so these markers were considered alleles of the same locus.

The large backcross population was used to map the allorecognition complex (ARC)-containing chromosomal interval. In this population, the only segregating parental haplotype is rß as it is derived from an f/f x f/rß cross. All four markers (R18, R28, 174, and 194) segregated as expected for a dominant Mendelian locus with no significant deviation from a 1:1 ratio of presence to absence (Table 2). Multi-point analysis with these four markers and the two fusibility loci produced a map with ARC at 3.4 cM, such that alr1 is flanked by markers 194 and 18 at 0.9 and 2.1 cM, respectively, and alr2 is flanked by marker R28 at 0.2 cM (Fig 5). With the population size surveyed, we are unable to discriminate between marker 174 and alr2. The log-likelihood value for this map was –185.0 and the next most likely map had a value of –186.8.

View larger version (5K): In this window In a new window Download PPT slide   Figure 5. Linkage map of Hydractinia ARC. Genetic distances are given in centimorgans.

A total of 13 animals showed recombination in the ARC region. Eleven of these recombinant animals had a crossing over between alr1 and alr2 and, of these, 10 were animals that had previously been detected as the colonies displaying TF against the rß/rß (4117-2) tester. The remaining colony, BC375, also displayed TF, albeit with an altered temporal progression as described above. Colony BC 493 proved recombinant for only the 194 marker, but was fusibility tested as f/f and hence was likely a double recombinant. Finally, recall that two colonies (BC129 and BC 393) displayed a unique phenotype. Neither proved to be recombinant in the marker-delineated interval.

Marker-based confirmation of classical segregation analysis:
Just as the markers provide an independent confirmation of the association of recombination with TF, the markers are expected to confirm the predicted genotypes of both the inbred lines and the parents of the crosses used to test Mendelian expectations of TF (Table 3). Recall that the parents of the TF crosses, colonies 431-85 and 431-55, displayed TF against the rß/rß and f/f testers, respectively. Marker R18 was present in 431-85 whereas marker F18 appeared in 431-55, indicating that indeed these two colonies had a recombination between alr1 and alr2. Marker 18 showed complete cosegregation with the TF phenotype in the offspring of these colonies, as expected (Table 3). Finally, as predicted, animals from the 33 and 41 inbred lines were found to be homozygous for alleles F18 and R18, respectively.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The major histocompatibility complex (MHC) of vertebrates was first discovered by study of allorecognition reactions in congenic lines of mice. Initially described as single locus (SNELL et al. 1976 ), the MHC is now known to represent a complex of >120 genes (MHC SEQUENCING CONSORTIUM 1999 ). Colonial invertebrates have been known for almost a century to display histocompatibility reactions, yet no invertebrate histocompatibility complex has been described up to this time. Allorecognition in the ascidian Botryllus scholesseri and the hydroid H. symbiolongicarpus, however, has been reported to exhibit single-locus segregation patterns (OKA and WATANABE 1957 ; SCOFIELD et al. 1982 ; MOKADY and BUSS 1996 ; DE TOMASO et al. 1998 ). In both cases, these findings have been based on crosses of relatively small population sizes; hence existing data have been insufficient to discriminate between a single locus and a complex mapping to a single chromosomal region. Using the classical techniques that proved successful in mice a half century ago, we report that the development of lines of the colonial hydroid H. symbiolongicarpus are congenic for fusibility and confirm our earlier report that fusibility is controlled by a single chromosomal interval in these lines. As was the case for the MHC, we here report that the chromosomal interval controlling invertebrate allorecognition is complex.

We have mapped a 3.4-cM interval with molecular and genetic markers and demonstrated that no less than two distinct allorecognition loci, alr1 and alr2, map within this interval. Our results suggest a dose-dependent interaction of these two loci in controlling fusibility. Specifically, our results are consistent with permanent fusion occurring when colonies share one or more alleles at both loci, rejection if colonies share no alleles at either locus, and transitory fusion occurring if colonies share an allele at only one of the two loci. It bears emphasizing, however, that we have yet to observe all allelic combinations, and further refinement of fusion rules may be necessary as the gene complex is further characterized. Indeed, the fact that two additional fusion phenotypes appear in this cross, represented by colonies BC 375 and BC 129/BC 393, suggests that the region will prove gene rich and that a genetic dissection of this interval will prove informative.

Our findings suggest caution in the use of methods heretofore conventional in the study of fusibility in this animal. First, we have found that animals that display TF in the colony assay remain fused in the polyp assay. Note that the polyp assay is terminated after 24 hr and that, in the colony assay using small explants, the transitory fusion phenotype is typically an event that takes a day to appear and at least another day to complete. It would be of interest to maintain polyp fusions for an extended duration and to further characterize the chimeric epithelial junctions to determine whether transitory fusion effects can be more readily detected in this otherwise useful assay. The failure of TF to readily manifest itself in the polyp assay may explain why this phenotype was not encountered in either the development of these lines or MOKADY and BUSS' (1996) work; both these studies determined fusibility by the polyp fusion assay. A second caution pertains to the treatment of transitory fusion in the colony fusion assay. Transitory fusion phenotype will be scored as a fusion if tests are observed one day after contact, but as a rejection 2–3 days after contact. Studies that employ 7-day observation windows (GROSBERG et al. 1996 ) would be expected to confound TF with rejection at appreciable frequencies.

Our finding that allorecognition is controlled by a single chromosomal interval recalls the early work of VON HAUENSCHILD 1954 , VON HAUENSCHILD 1956 . von Hauenschild performed F₁ crosses of four different wild-type colonies of H. echinata and a series of F₂ crosses among full and half-sibs. He produced a strong argument for single-locus segregation on the basis of a scheme whereby different alleles were assigned different numerical values and the outcome of fusibility reactions was predicted by the sum of these values. Transitory fusion was interpreted as a phenomenon that appeared when the sums of allelic values did not differ by a threshold required for fusion or rejection. While the vast majority of his data were accommodated under this interpretation, the results of some crosses could not be explained. DU PASQUIER 1974 reanalyzed von Hauenschild's data and noted that von Hauenschild's results were better explained by a single-locus system with a simple haplotype effect (as found here). Of particular interest, is Du Pasquier's observation that some, but not all, of the deviations from single Mendelian expectations in von Hauenschild's data could be explained by "suggest[ing] the existence of subloci in this genetic region," specifically, by postulating that one of von Hauenschild animals was recombinant (DU PASQUIER 1974 ).

In contrast to our findings, and the preponderance of those of von Hauenschild and Du Pasquier, are the suggestions of GROSBERG et al. 1996 , GROSBERG et al. 1997 . These investigators crossed five wild-type pairs of H. symbiolongicarpus and tested fusibility among full and half-sib colonies. From these data, the authors predicted the number of independently segregating loci that would be required to accommodate their findings under specific assumptions regarding dosage and the number of alleles at the hypothesized loci, ultimately deriving an estimate that some five to seven unlinked loci would be required to generate their findings. GROSBERG et al. 1996 , GROSBERG et al. 1997 , however, failed to consider the possibility of a system of linked loci. The departure of GROSBERG et al. 1996 , GROSBERG et al. 1997 findings from those presented here may reflect methodological difficulties (see above) or a strong effect of genetic background on fusibility (or both).

The fact that allorecognition segregates to a single interval in our defined genetic lines does not preclude the occurrence of minor histocompatibility loci at other locations in the genome. Indeed, we emphasize that congenic lines were produced to homogenize the effect of genetic background and hence eliminate the effects of any modifying loci. The vertebrate analogy is again germane. Shortly after the discovery of the MHC, mouse geneticists began to identify a number of unlinked modifiers and minor histocompatibility loci (SNELL et al. 1976 ). As proved fruitful in mice, crosses between our defined lines and wild-type colonies may now be used to search for unlinked genes that contribute to allorecognition phenotypes.

Much of the interest in invertebrate allorecognition derives from the suggestion that these molecules may prove homologous to elements in the immune system of higher chordates (BURNET 1971 ; BUSS 1982 ; SCOFIELD et al. 1982 ; BUSS and GREEN 1985 ). Identification of the invertebrate allorecognition determinants and the genes encoding them, however, has proven difficult and the issue remains an open one. Previous attempts to identify them have focused on the search for plausible candidate genes by either structural similarities to their putative vertebrate counterparts or analogy to other recognition systems. Such efforts have been fruitless, and in retrospect, the problem seems to demand a strategy that makes no a priori assumptions as to the identity of the gene product. Positional cloning from linked markers is such a strategy. With the availability of maps spanning the relevant interval in Hydractinia, the tools are in hand to at last isolate the invertebrate allorecognition determinants.

	FOOTNOTES

¹ These authors contributed equally to this work.
² Present address: Department of Biology, University of New Mexico, Albuquerque, NM 87131.
³ Present address: Department of Anatomy and Neurobiology, University of California, Irvine, CA 92697-4040.

	ACKNOWLEDGMENTS

We thank Stephen Dellaporta for drawing our attention to bulked segregant analysis, for his comments on this manuscript, and for ongoing discussion of the work. Bill Rose and Rafael Rosengarten provided attentive and essential animal care, as well as technical assistance in genotyping. This work is supported by a grant to L.W.B. from the National Science Foundation (MCB-9817380). M.M. was supported by a grant from the National Institutes of Health (RO1-GM38148).

Manuscript received January 28, 2003; Accepted for publication January 22, 2004.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BALLARD, W. W., 1942  The mechanisms of synchronous spawning in Hydractinia and Perannia.. Biol. Bull. 82:329-339.[Abstract/Free Full Text]

BLACKSTONE, N. W. and L. W. BUSS, 1991  Shape variation in hydractiniid hydroids. Biol. Bull. 180:394-405.[Abstract]

BUNTING, M., 1895  The origins of sex-cells in Hydractinia and Podocoryne.. J. Morphol. 9:203-236.[CrossRef]

BURNET, M., 1971  "Self-recognition" in colonial marine forms and flowering plants in relation to the evolution of immunity. Nature 232:230-235.[CrossRef][Medline]

BUSS, L. W., 1982  Somatic cell parasitism and the evolution of somatic tissue compatibility. Proc. Natl. Acad. Sci. USA 79:5337-5341.[Abstract/Free Full Text]

BUSS, L. W. and D. R. GREEN, 1985  Histocompatibility in vertebrates: the relict hypothesis. Dev. Comp. Immunol. 9:191-201.[CrossRef][Medline]

BUSS, L. W. and R. K. GROSBERG, 1990  Morphogenetic basis for phenotypic differences in hydroid competitive behaviour. Nature 343:63-66.[CrossRef]

BUSS, L. W., and M. A. SHENK, 1990 Hydroid allorecognition regulates competition at both the level of the colony and at the level of the cell lineage, pp. 85–105 in Defense Molecules, edited by J. J. MARCHALONIS and C. REINISCH. Alan R. Liss, New York.

BUSS, L. W., C. S. MCFADDEN, and D. R. KEENE, 1984  Biology of hydractiniid hydroids. 2. Histocompatibility effector system/competitive mechanism mediated by nematocyst discharge. Biol. Bull. 167:139-158.[Abstract/Free Full Text]

CADAVID, L. F., 2001 Genetic characterization of the hydroid allorecognition complex. Ph.D. Thesis, Yale University, New Haven, CT.

DE TOMASO, A. W., Y. SAITO, K. J. ISHIZUKA, K. J. PALMERI, and I. L. WEISSMAN, 1998  Mapping the genome of a model protochordate. I. A low resolution genetic map encompassing the fusion/histocompatibility (Fu/HC) locus of Botryllus schlosseri. Genetics 149:277-287.[Abstract/Free Full Text]

DU PASQUIER, L., 1974  The genetic control of histocompatibility reactions: phylogenetic aspects. Arch. Biol. 85:91-103.

GIOVANNONI, J. J., R. A. WING, M. W. GANAL, and S. D. TANKSLEY, 1991  Isolation of molecular markers from specific chromosomal intervals using DNA pools from existing mapping populations. Nucleic Acids Res. 19:6553-6558.[Abstract/Free Full Text]

GROSBERG, R. K., D. R. LEVITAN, and B. B. CAMERON, 1996  Evolutionary genetics of the allorecognition in the colonial hydroid Hydractinia symbiolongicarpus.. Evolution 50:2221-2240.[CrossRef]

GROSBERG, R. K., M. W. HART, and D. R. LEVITAN, 1997  Is allorecognition specificity in Hydractinia symbiolongicarpus controlled by a single gene? Genetics 145:857-860.[Medline]

IVKER, F. B., 1972  A hierarchy of histo-incompatibility in Hydractinia echinata.. Biol. Bull. 143:162-174.[Abstract/Free Full Text]

KOSAMBI, D. D., 1944  The estimation of map distances from recombination values. Ann. Eugen. 12:172-175.

LANDER, E. S., P. GREEN, J. ABRAHAMSON, A. BARLOW, and J. M. DALY et al., 1987  MAPMAKER, an interactive computer package for constructing primary genetic linkage maps of experimental and natural populations. Genomics 1:174-181.[CrossRef][Medline]

LANGE, R., G. PLICKERT, and W. A. MÜLLER, 1989  Histocompatibility in a low invertebrate, Hydractinia echinata: analysis of the mechanism of rejection. J. Exp. Zool. 249:284-292.[CrossRef]

LANGE, R. G., M. H. DICK, and W. A. MÜLLER, 1992  Specificity and early ontogeny of historecognition in the hydroid Hydractinia.. J. Exp. Zool. 262:307-316.[CrossRef]

LIU, B. H., 1998 Statistical Genomics: Linkage Mapping and QTL Analysis. CRC Press, Boca Raton, FL.

Complete sequence and gene map of a human major histocompatibility complex. (1999) Nature 401:921-923.[CrossRef][Medline]

MICHELMORE, R. W., I. PARAN, and R. V. KESSELI, 1991  Identification of markers linked to disease resistance genes by bulked segregant analysis: a rapid method to detect markers in specific genomic regions by using segregating populations. Proc. Natl. Acad. Sci. USA 88:9828-9832.[Abstract/Free Full Text]

MOKADY, O. and L. W. BUSS, 1996  Transmission genetics of allorecognition in Hydractinia symbiolongicarpus (Cnidaria: Hydrozoa). Genetics 143:823-827.[Abstract]

MÜLLER, A. W., 1964  Experimentelle Untersuchungen über Stockentwicklung, pelypendifferenzierung und Sexualchimären bei Hydractinia echinata.. Wilhem Roux' Arch. 155:182-268.

MÜLLER, A. W., 1973  Induction of metamorphosis by bacteria and ions in the planulae of Hydractinia echinata: an approach to mode of action. Publ. Seto Marine Biol. Lab. 20:195-208.

MÜLLER, W. A., A. HAUCH, and G. PLICKERT, 1987  Morphogenetic factors in hydroids: I. Stolon tip activation and inhibition. J. Exp. Zool. 243:111-124.[CrossRef]

OKA, H. and H. WATANABE, 1957  Colony-specificity in compound ascidians as tested by fusion experiments. Proc. Jpn. Acad. 33:657-659.

OTT, J., 1991 Analysis of Human Genetic Linkage. The Johns Hopkins University Press, Baltimore.

SCOFIELD, V. L., J. M. SCHLUMPBERGER, L. A. WEST, and I. L. WEISSMAN, 1982  Protochordate allorecognition is controlled by a MHC-like gene system. Nature 295:499-502.[CrossRef][Medline]

SHENK, M. A. and L. W. BUSS, 1991  Ontogenetic changes in fusibility in the colonial hydroid Hydractinia symbiolongicarpus.. J. Exp. Zool. 257:80-86.[CrossRef]

SHURE, M., S. WESSLER, and N. FEDORFF, 1983  Molecular identification and isolation of the Waxy locus in maize. Cell 35:225-233.[CrossRef][Medline]

SNELL, G. D., J. DAUSSET and S. NATHENSON, 1976 Histocompatibility. Academic Press, New York.

VON HAUENSCHILD, C., 1954  Genetische und entwichlungphysiologische Untersuchungen über Intersexuälitat und Gewebeverträlichkeit bei Hydractinia echinata Flem. Wilhem Roux' Archiv. 147:1-41.[CrossRef]

VON HAUENSCHILD, C., 1956  Uber die vererbung einer geweberverträglichkeits eigenschaft bei dem hydroidpolypen Hydractini echinata.. Z. Naturforsch. 11b:132-138.

VOS, P., R. HOGERS, M. BLEEKER, M. REIJANS, and T. VAN DE LEE et al., 1995  AFLP: a new technique for DNA fingerprint. Nucleic Acids Res. 23:4407-4414.[Abstract/Free Full Text]

This article has been cited by other articles:

				A. E. Powell, M. L. Nicotra, M. A. Moreno, F. G. Lakkis, S. L. Dellaporta, and L. W. Buss Differential Effect of Allorecognition Loci on Phenotype in Hydractinia symbiolongicarpus (Cnidaria: Hydrozoa) Genetics, December 1, 2007; 177(4): 2101 - 2107. [Abstract] [Full Text] [PDF]

				M. Poudyal, S. Rosa, A. E. Powell, M. Moreno, S. L. Dellaporta, L. W. Buss, and F. G. Lakkis Embryonic chimerism does not induce tolerance in an invertebrate model organism PNAS, March 13, 2007; 104(11): 4559 - 4564. [Abstract] [Full Text] [PDF]

				L. F. Cadavid Self/non-self Discrimination in Basal Metazoa: Genetics of Allorecognition in the Hydroid Hydractinia Integr. Comp. Biol., August 1, 2005; 45(4): 623 - 630. [Abstract] [Full Text] [PDF]

				M. L. Nicotra and L. W. Buss A Test for Larval Kin Aggregations Biol. Bull., June 1, 2005; 208(3): 157 - 158. [Full Text] [PDF]