Determination of gene organization in individual haplotypes by analyzing single DNA fragments from single spermatozoa

Abstract

To determine human Ig heavy chain variable region (V_H) gene segment organization on individual homologous chromosomes, an efficient approach has been developed. Single spermatozoa were used as subjects for the study. Upon sperm lysis, V_H regions in each sperm were randomly sheared into fragments by the random Brownian force. The fragments were separated from each other by aliquoting the lysate into a certain number of tubes. The gene segments in the V_H1 and V_H4 families in each tube were identified by denaturing gradient gel electrophoresis after PCR amplification. The polymorphic V_H sequences were used to determine the parental origins of the analyzed sperm. V_H segment organization in the parental haplotypes was determined by aligning the overlapping fragments from the spermatozoa with the corresponding haplotypes. Based on this comparison between the resulting haplotype maps and the composite map reported previously, the V_H region on chromosome 14 could be subdivided into four portions. The numbers and compositions of the V_H gene segments differ considerably among the maps in two portions, but are highly conserved in the other two. The data also indicate that the V_H region on chromosome 15 may contain a large duplicated block with copy number varying among haplotypes. The approach used in the present study may be used to construct high-resolution haplotype maps without molecular cloning.

For many human multigene families, determination of gene organization on individual homologous chromosomes is a challenging issue for several reasons: (i) the diploidy of the human genome, (ii) the genes sharing a high degree of sequence identity in each family, (iii) variation in DNA sequences and gene compositions among the haplotypes, (iv) different chromosomal locations, (v) occupation of large chromosomal regions, and (vi) recent duplications. In this paper, we show that all of these issues can be addressed by analyzing single DNA fragments from the haploid genomes of individual spermatozoa.

Two human Ig heavy chain variable region (V_H) families, V_H1 and V_H4, were used for the analysis. Human V_H gene segments are located on three chromosomes, 14q32, 15q11, and 16p11 (1–6). However, only about half of the V_H segments on chromosome 14 are functional (7, 8). Each haploid human genome contains ≈120 distinct V_H gene segments (5, 7, 8). The V_H segments are classified into seven families (V_H1–V_H7) with segments sharing >80% sequence identity in each family (9–15). Extensive polymorphisms for the V_H region on chromosome 14 have been reported (16–22). A composite map for this region was constructed recently (8). Several studies showed that the number and composition of the V_H gene segments in some portions of the V_H region on chromosome 14 varied among the haplotypes (23–30). However, the extent of variation in the entire region remains unclear because no haplotype maps for this region have been constructed. The V_H segment organization on the other two chromosomes has not been determined.

MATERIALS AND METHODS

Preparation of Single V_H Segment-Containing DNA Fragments from Individual Spermatozoa.

The procedure by Lien et al. (31) was modified and used for single sperm preparation. Briefly, sperm (≈500) were suspended in 200 μl of 0.5% melted low melting point agarose in H₂O at 37°C. About 100 μl of the suspension was pipetted onto a microscopic slide (at a 45° angle) across the width and was allowed to flow to form a thin layer. After 10 min at room temperature, single sperm were scraped up with a syringe needle (27G½) and each was placed into a 0.5-ml microtube containing 4 μl of H₂O. The tubes were incubated at 72°C for 1 min to melt the agarose. Each sperm was lysed by incubating at 37°C for 15 min in a 5-μl solution containing 10 mM Tris⋅HCl (pH 8.3), 10 mM EDTA, 0.1% SDS, 40 mM DTT, and 50 μg/ml of proteinase K. After adding 27 μl of H₂O and flushing through a pipette tip once, each lysate was subdivided into a desired number of tubes (eight or 12 tubes for 10 sperm analyzed in the early stage of the study, and 16 tubes for 37 sperm analyzed later). To inactivate proteinase K, the tubes were incubated at 85°C for 10 min after adding a drop of mineral oil.

PCR Amplification.

Family-specific primers were designed for the V_H1 and V_H4 families according to the sequences conserved among the sequences in each family. The primers used for the first PCR step were V1M2 (5′-CCTCAGTGAAGGTCTCCTGCAAG-3′) and V1M4 (5′-cgccgcccccgccccCTGCTCAGCTCCATGTAGGCTGTG-3′) for the V_H1 family, and V4M1 (5′-aatgagctccACACATTTCCTTAAA-TTC -3′) and V4M4 (5′-cgccgcccccgccccGAAGGCTTCACCAGTCCTG-3′) for the V_H4 sequences. In the second step, V_HGC (5′- cgcccgccgcgccccgcgcccgtcccgccgcccccgcccc-3′), as part of the GC clamp, was used for both families to replace V1M4 and V4M4 and two internal (nested) primers V1M3 (5′-GGACAAGGGCTTGAGTGGATGGGA-3′) and V4M2 (5′-GAACATGAAACACCTGTGGTTCT-3′) were used to replace V1M2 and V4M1 for the two families, respectively. Letters in lowercase in the primer sequences represent nongenomic sequences used either for enhancing amplification efficiency or for attaching the GC clamp to facilitate denaturing gradient gel electrophoresis (DGGE) separation. All PCR amplifications were performed with a DNA Thermal Cycler 480 (Perkin–Elmer). In the first step, each sample contained 1× PCR buffer (100 mM Tris⋅HCl, pH 8.3/50 mM KCl/1.5 MgCl₂/100 μg/ml gelatin), the four dNTPs each at 100 μM, primers each at 50 nM for the V_H4 family and 20 nM for the V_H1, and 0.5 units of the Taq DNA polymerase. The final volume was 25 μl. Each PCR cycle consisted of a denaturation step at 95°C for 30 sec and an extension step at 72°C for 30 sec. The annealing step was at 50°C for 5 min for the first three cycles, 55°C for 3 min for seven cycles, and 60°C for 2 min for the last 30 cycles. In the second PCR step, the V_H1 and V_H4 segments were amplified separately with 1.5-μl aliquots from the first-step products. Each sample contained 1× PCR buffer, the four dNTPs each at 100 μM, the corresponding primers each at 0.2 μM, and 0.5 units of the enzyme in a final volume of 25 μl. Each PCR cycle was 95°C for 30 sec for denaturation and 72°C for 30 sec for extension. The annealing step was 55°C for 1 min for the first three cycles and 60°C for 1 min for 25 cycles. An additional PCR cycle was performed at 95°C for 2 min, 60°C for 1 min, and 72°C for 10 min after adding 5 μl of solution containing 0.5 units of the Taq polymerase and adjusting the concentration of each primer to 0.5 μM to minimize DNA heteroduplexes that may form DGGE bands.

DGGE.

The DGGE gels were prepared according to the manufacturer of the DGGE apparatus (C.B.S. Scientific, Del Mar, CA). Ten percent polyacrylamide was used for all gels. The denaturing gradients were 52–68% and 55–73% (100% denaturing strength defined as 7 M urea and 40% formamide) for the V_H1 and V_H4 families, respectively. Electrophoresis was performed at 60°C at 113 v for 15 h. Some V_H1 bands were resolved with 50–65% denaturing range.

RESULTS

Identification of the V_H Gene Segments in Single DNA Fragments from Individual Sperm.

The procedure for identifying the V_H gene segments in single DNA fragments from individual sperm is diagramed in Fig. 1. Single sperm cells were prepared from a healthy donor and lysed. Because the V_H regions are relatively large (1,100 kb for the one on chromosome 14, ref. 8), these regions are unavoidably sheared by the random Brownian force when released from sperm. To separate the V_H segment-containing fragments from each other, each sperm lysate was aliquoted into 16 tubes (into eight or 12 for a few analyzed at the early stage of the study). The V_H1 and V_H4 sequences in each tube were amplified to analyzable amounts with the two-round PCR procedure described in Materials and Methods. In the final PCR products, the sequences flanked by the primers were ≈83 bp for V_H1 and ≈145 bp for V_H4.

Figure 1.

Diagram for identification of V_H gene segments in single DNA fragments from a single sperm. Only the V_H region on chromosome 14 is shown. V_H segments of two different families are shown as squares with different backgrounds and fillings. Family-specific PCR primers are represented by black and white arrows. The two vertical squares at the bottom represent DGGE gel strips with bands shown as bars.

Because the lengths of the amplified V_H sequences from each family were either the same or very similar, these sequences could not be separated by regular gel electrophoresis and were resolved by DGGE (32–36). DGGE was designed to separate the DNA fragments differing by as little as 1 bp. Because most V_H-amplified sequences differed from each other by >1 bp, these sequences should be readily separated from each other by DGGE. The V_H1 and V_H4 sequences amplified from the diploid genome of the sperm donor were used as the molecular markers in DGGE. The DGGE bands for the V_H sequences amplified from the tubes with V_H segment-containing fragments from one sperm are shown in Fig. 2. The DGGE bands were numbered based on their positions on the DGGE gels (from top to bottom). The V_H loci represented by the DGGE bands were named with prefixes 1- and 4-, respectively, followed by the band numbers.

Figure 2.

DGGE separation and chromosomal location of the V_H sequences amplified from single DNA fragments generated from a single sperm. Results for the V_H1 and V_H4 families are shown separately. The chromosomal origin of the V_H sequences in each lane is indicated by the chromosome number after SC for a sperm chromosome and HC for a human chromosome in the somatic cell hybrids. M, V_H sequences amplified from the sperm donor’s genomic DNA and used as molecular markers. Lanes 1–10, V_H sequences amplified from single DNA fragments and resolved by DGGE. V_H1 and V_H4 sequences in the lanes sharing the same numbers were those in the same tubes before the two families were amplified separately in the second round of PCR. Some bands such as 1–6 and 1–5b could not be separated well in the gel shown here but were well resolved in a gel with a slightly changed denaturing range. A band amplified from the Chinese hamster genome background in lanes HC14 and HC15 is indicated by an arrow (lane HC16 was from a human-mouse hybrid). Results from agarose gel (1%) electrophoresis for the same set of sperm samples are aligned with the DGGE results for illustrating the size uniformity of the amplified V_H sequences.

Determination of the V_H Gene Segment Organization in Individual Haplotypes.

V_H segment-containing fragments from 47 single sperm were analyzed with the procedure described above. The results were used to determine the V_H gene segment organization in individual haplotypes through three steps.

Determining the chromosomal locations of the V_H gene segments. To determine the V_H segment organization in individual haplotypes, it is necessary to learn the chromosomal locations of these segments. Human-rodent somatic cell hybrids containing single human chromosomes 14, 15, or 16, respectively, in mapping panel 2 of the National Institute of General Medical Sciences repository were used as standards for the analysis. By comparing the DGGE bands (Fig. 2, lanes M) from the sperm donor’s diploid genome with those from the somatic hybrids (Fig. 2, lanes HC14, HC15, and HC16), the chromosomal locations of most V_H sequences in the sperm donor were determined. To confirm that comigration of the DGGE bands from the two sources was because of their sequences identity rather than comigration of different sequences, the DGGE bands were excised from the gel, reamplified with PCR, and subjected to sequence analysis. The results showed that all comigrating bands from these two sources contained identical sequences. With the known chromosomal locations, the V_H sequences amplified from the sperm donor’s genome were used as molecular markers to determine the chromosomal locations of the V_H sequences identified from the sperm samples.

Several V_H sequences detected from the sperm donor’s genomic DNA were not detected from the somatic hybrids. The chromosomal locations of these sequences were readily determined because in the analysis of single V_H segment-containing fragments from single sperm these sequences frequently were codetected with certain sequences whose chromosomal locations were known through the somatic hybrid assay.

Determination of the V_H segment compositions in the parental haplotypes. The parental haplotypes for chromosome 14 were named as haplotypes 4 and 7, respectively, based on the initial observation that two allelic V_H4 sequences, 4–4 and 4–7, segregated among the sperm analyzed. Of the 47 sperm analyzed, 28 contained haplotype 4, and 19 contained haplotype 7. As shown in Fig. 3 (Upper), 17 V_H segments (nine V_H1 and eight V_H4) were detected from sperm with haplotype 4, and 21 (13 V_H1 and eight V_H4) from those with haplotype 7. Eight V_H segments (Fig. 3, 1–12L, 1–3, 1–12R, 4–11, 4–4, 4–17, 1–14, and 1–17) were specific for haplotype 4, and 12 V_H segments (Fig. 3, 1–6, 4–7, 4–12, 1–7, 1–11, 1–7b, 1–9b, 1–1, 1–13, 4–15, 1–5b, and 1–18) were for haplotype 7. These haplotype-specific sequences were, in turn, used to confirm the haplotypes in the sperm analyzed.

Figure 3.

Maps for the V_H1 and V_H4 segments on chromosomes 14 and 15. The alignment of the maps for chromosome 14 is read from left to right and then from top to bottom, and is subdivided into four portions by vertical bars (see text). Loci in the dashed-lined boxes are aligned putatively because sequence data in the map by Cook et al. (8) are not available. The superscript letters indicate the numbers of base pairs different between the allelic sequences within the amplified regions. The unresolved pairs, 4–9 and 4–1 in both haplotypes 4 and 7, 1–17 and 1–14, and 1–16 and 1–4 in haplotype 4 are placed in the alignment according to the order information in the map by Cook et al. (8). 1–9b and 1–1 in haplotype 7 could not be amplified by the family-specific primers and were placed on the map by analyzing additional sperm with both gene segment-specific and family-specific primers. V_H segments that are closely located on chromosome 15 are boxed. The CEs on the right of the maps may be oriented differently as indicated by the observation that segments 1–23 and 1–5 were codetected with 1–10 and 4–2 from a sperm with haplotype 2 but with 1–19 and 1–9 from a sperm with haplotype 3. The sequences with no counterpart sequences in the maps and the accession numbers of their counterpart sequences in the GenBank are: 1–3, Z12305 (10); 1–5, Z29632 (5); 1–8, Z29632 (5); 1–9, Z17390 (5); 1–10, Z29596 (5); 1–12, Z12312 (10); 1–12b, AF030491; 1–19, Z29631 (5); 1–20, AF030492; 1–22, Z29633 (5); 1–23, L25542 (6); 4–2, X92231 (52); and 4–3, U23548.

Because the two parental haplotypes of chromosome 15 contained either two or three copies of a duplicated V_H segment-containing block, these haplotypes were called haplotypes 2 and 3 (Fig. 3, Lower). Of the 47 sperm analyzed, 17 apparently contained haplotype 2, and 28 were with haplotype 3. The remaining two sperm seemed to be recombinants. Thirteen V_H segments (11 V_H1 and two V_H4) were detected from haplotype 2 and 17 (14 V_H1 and three V_H4) from haplotype 3.

The sequence, 1–2 (GenBank accession no. AF030490), was the only V_H1 sequence detected from chromosome 16. The copy number of this sequence detected from each sperm ranged from zero to three. Because of the small number of the detected V_H1 sequences and the fact that there is no V_H4 sequence on chromosome 16 (4–6), no further analysis was performed for the V_H segments on this chromosome.

Determination of the V_H segment organization in individual haplotypes. By treating the V_H segments as restriction enzyme sites, the order of the V_H segments in each haplotype was determined by a procedure similar to restriction enzyme mapping. Practically, each haplotype was deduced by aligning the overlapping V_H segment-containing fragments from the sperm cells with the corresponding haplotype. The deduced order of the V_H segments in haplotype 7 and the aligned V_H segment-containing fragments from the 19 sperm is shown in Fig. 4. A small fraction (eight of 90 for haplotype 4 and five of 79 for haplotype 7) of tubes may have received two V_H segment-containing fragments in each. This assumption is consistent with the expectations based on the Poisson distribution. These tubes were easily identified because if they were not considered as the tubes containing two fragments in each, the gene order deduced based on the information from most tubes would have to be seriously disrupted. The orders of 4–9 and 4–1 in both haplotypes 4 and 7, 1–17 and 1–14, and 1–16 and 1–4 in haplotype 4 could not be resolved because of either their relatively close physical locations (for example, the distance between 1–16 and 1–4 is only ≈4 kb; ref. 7), or the low detection rate (for 4–1).

Figure 4.

Diagram of the V_H segment-containing fragments in the 19 sperm with haplotype 7. The deduced order of the V_H segments in this haplotype is shown at the top. Sperm numbers are on the left. V_H1 (green) and V_H4 (red) segments are represented by squares. Undetected segments are left blank at the corresponding positions. Codetected but apparently nonadjacent fragments are connected with brackets. Unresolved segments, 4–1 and 4–9, because of the low detection rate of 4–1 are boxed.

The V_H Gene Segment Organization in the Haplotypes of Chromosome 14.

Based on the order of the V_H segments and sequence comparison, the two haplotype maps for chromosome 14 generated in the present study were aligned with the composite map by Cook et al. (8) (Fig. 3, Upper). The orders of the V_H segments in these maps are all consistent, indicating that the method used in the current study is highly reliable. The alignment for the three maps could be subdivided into four portions (indicated by the vertical bars): (i) the most J_H-distal portion (upper left). This portion is highly polymorphic in V_H segment composition. Of the seven loci, five (1–12b, 1–6/1–69, 1–3/1–68, 1–12, and 4–11/4–61) contained null alleles. It should be pointed out that two V_H1 segments, YAC7 and DP-10, both were placed to the 1–69 locus in the map by Cook et al. (8). These two sequences differ by 7 bp within 294 bp. Our sequence data indicate that 1–6 in haplotype 7 is identical to YAC7 and 1–12 in haplotype 4 is identical to DP-10. Our map data also showed that 1–12 is between 1–3 and 1–21 in haplotype 4. Therefore, 1–6 and 1–12 are likely two different loci as shown in Fig. 3 rather than two allelic sequences of one locus. Sasso et al. (30) showed that the V_H1 gene segment, 1–69 (V_H26), was present in a very complicated fashion. It may have been duplicated into two loci. This observation is consistent with our results, (ii) the upper-right portion of the alignment. This portion is surprisingly conserved. Not only are the V_H segment number in all three maps the same, but the nucleotide sequences of the eight segments are all identical in the amplified regions, except for a 1-bp difference between 4–7 and its allelic counterparts, (iii) the lower left portion. This portion features large insertion/deletion polymorphisms. No V_H segment in this portion was detected from haplotype 4. For the four V_H4 segments, 4–30.4, 4–30.2, 4–30.1, and 4–28 on the map of Cook et al. (8), only the counterparts for 4–30.2 were detected in haplotype 7. V_H4 segments in this portion also were shown to be polymorphic in the previous studies (8, 24–26), (iv) the most J_H-proximal portion (lower right). This portion includes three V_H segments with no variation in V_H segment composition although allelic differences at the nucleotide sequences level were detected at each locus.

The size of the V_H region covered by the composite map by Cook et al. (8) was estimated to be ≈1,100 kb. Based on this estimate, the differences between the V_H segment numbers in the maps, and the average distance between adjacent V_H segments in the map by Cook et al., the V_H region is estimated to be ≈800 kb in haplotype 4 and ≈970 kb in haplotype 7. Because a large portion of the composite map was generated with diploid materials (7), the estimated size by Cook et al. may be greater than the actual sizes of the V_H region in the haplotypes. In other words, the estimated sizes of haplotypes 4 and 7 could be closer to the actual sizes. The mean numbers of V_H segment-containing fragments detected from each sperm were 3.21 and 4.16 for haplotypes 4 and 7, respectively. Therefore, the average size of these fragments could be 230–250 kb with a range from several kb to >1,000 kb (Fig. 4). Based on the previous map information (8), the distances between the genes resolved in the current study ranged from 9 kb (between 4–15 and 1–5b) to 115 kb (between 1–15 and 1–16).

The V_H Gene Segment Organization in the Haplotypes of Chromosome 15.

The order of the V_H gene segments and the map alignment for the haplotypes 2 and 3 of chromosome 15 are shown Fig. 3 (Lower). More than one copy was detected for the four V_H segments, 1–10, 4–2, 1–9 (or 1–8), and 1–19, from most single sperm. These segments must be closely located because they were codetected very frequently (106 of 119). Because these duplicated blocks are the major components on the map for chromosome 15, they are called the core elements (CEs). The two haplotypes of chromosome 15, haplotypes 2 and 3, were named based on the number of CEs in these haplotypes. Other than the CEs, two copies of 1–20 and one copy of 1–5, 1–22, and 1–23 were detected from both haplotypes. Segments 1–5 and 1–23 also may be very closely located because they were codetected in 30 of 31 sperm from which both segments were detected.

The two V_H segments 1–8 and 1–9 differing by 1 bp can be used to distinguish the CEs. CEs containing 1–8 are called CE-8 and those contained 1–9 are called CE-9. As shown in Fig. 3 (Lower), haplotype 2 contained only two CE-8s, and haplotype 3 contained one CE-8 and two CE-9s. The haplotypes in two sperm could not be placed to any of these parental types. One of these sperm contained three CE-8s, and the other contained two CE-8s and one CE of unknown type because experimentally neither 1–8 nor 1–9 was detected from this CE. These sperm may have resulted either from duplication of a CE-8 in haplotype 2 or from genetic recombination between the two different parental haplotypes. In any case, the genetic event responsible for the occurrence of these recombinant sperm should be on the order of 2/47 = 0.04.

The means numbers of the V_H segment-containing fragments detected from chromosome 15 were 4.18 and 5.07 per sperm for haplotypes 2 and 3, respectively. Based on these numbers and the estimated size of the V_H segment-containing fragments from chromosome 14, the sizes of the V_H regions on chromosome 15 were estimated as ≈1,000 kb for haplotype 2 and 1,200 kb for haplotype 3. Because fewer V_H segments were detected from this chromosome, the density of the V_H segments on chromosome 15 could be lower than those on chromosome 14. Other than 1–12b and 1–20, all V_H segments on chromosome 15 identified in the present study were reported previously (5, 6). V_H segment duplication on this chromosome was implicated by analyzing the D_H segments (6). However, no analysis on V_H segment organization on this chromosome has been reported previously. Difficulty caused by the large distances between the V_H loci may be one of the reasons.

DISCUSSION

Determination of gene organization by sperm analysis has many advantages over the conventional methods. Because each sperm contains only one set of chromosomes, the complications involved in using diploid materials are removed. With the molecular cloning method, the sizes of the cloned DNA fragments are restricted by the upper and lower limits of the vector capacity. With sperm analysis, it is possible to generate DNA fragments of several hundred to >1,000 kb so that the order of the genes separated by relatively large distances can be determined and a large amount of work involved in cloning can be avoided. On the other hand, any closely located duplicated genes also may be resolved if the force applied to shearing sperm DNA is strong enough, or if the sperm DNA is digested with one or more restriction enzymes that do not cut the amplified sequences. Our group has developed a high-capacity multiplex PCR protocol (37) with which a large number of single-copy sequences can be amplified simultaneously by PCR and resolved either by regular gel electrophoresis or by DGGE (unpublished data). By combining this method with sperm analysis, the organization of a large number of single-copy sequences also could be examined. With the conventional approaches, resolving recently duplicated genes is a serious issue because the duplicates are, or almost are, identical. With sperm analysis, although it may require thousands of tubes to separate the copies from each other for the genes coding for ribosomal RNAs, for many duplicated genes with relatively small copy numbers, their copy numbers could be readily estimated as long as each sperm lysate is aliquoted into a sufficient number of tubes.

When a large number of genes with different chromosomal locations are studied, the analysis could be very complex. This complexity can be significantly reduced by first determining the chromosomal locations of the genes by using the sequences amplified from the human-rodent somatic hybrids containing single human chromosomes as molecular markers as shown in the present study. With known chromosomal locations of the genes, the number of tubes to which each sperm lysate is aliquoted to can be significantly reduced because genes on different chromosomes can be analyzed separately even if they are codetected from the same tubes.

A method called radiation hybrid (RH) mapping was first described by Cox et al. (38) and has been used by many laboratories for constructing physical maps. With this method, human cells containing chromosomes sheared by radiation are fused with rodent cells. Because closely located markers tend to retain together in the hybrid cells, the order of the markers can be determined by examining the codetection rates of the markers. This method is powerful when the markers are separated by large distances (several megabases). Compared with the RH method, our approach is used for the genes separated by smaller distances (from several kb to several hundred kb) and for determining gene organization in greater detail. Experimentally, because no viable cells are required, our experimental procedure is much simpler. With our method, the gene order in individual haplotypes can be determined.

Another method called “Happy Mapping” was used to order DNA sequences (27, 39, 40) by aliquoting sheared genomic DNA to one haploid equivalent per tube and by analyzing the codetection rate of the markers. Both Happy Mapping and the sperm analysis used in the present study are based on limiting dilution without involving molecular cloning for map construction. These two approaches are different in the following aspects: (i) Haplotype map construction. Because each sperm contains precisely one haploid genome, sperm analysis can be used to construct haplotype maps readily. A few polymorphic markers are needed only for the purposes of distinguishing the sperm of different parental origins and the recombinants from nonrecombinants. Happy Mapping starts with a DNA solution containing DNA fragments from many copies of two mixed haplotypes. With this approach, haplotype maps can be constructed only if most, if not all, DNA fragments contain one or more known polymorphic markers. Practically, identification of these markers requires DNA blocks larger than 1 kb at a density comparable to or higher than the marker density and determination of the sequences in these blocks. (ii) Background noise and map resolution. For both approaches, the basic information “units” are the DNA fragments containing the markers rather than the haploid genomes. Sperm analysis pursues separation of these basic units and therefore generates data with very low background noise, i.e., the fraction of tubes receiving >1 marker. With sperm analysis, the distribution of two markers in a lysate is completely independent from those in others. When the two markers are linked, it is impossible to generate any “background noise” because each lysate contains only one DNA fragment with the two markers. Background noise can be generated only when the two markers are separated (unlinked) in a lysate and can be expressed as 1/n², where n is the tube number for each sperm lysate, or the dilution factor. According to this correlation, the background noise generated from the unlinked markers decreases exponentially as the dilution factor increases linearly. It also indicates that the effect on reducing the fraction of informative tubes caused by increasing the dilution factor may not “cancel out” that caused by reducing background noise. In contrast, when two markers are aliquoted from a DNA solution with the Happy Mapping approach, the background noise may be generated from a number of possible combinations of the markers and of different copies of the markers by the Poisson distribution. The background noise also may be generated from the intermixing of the linked and unlinked markers. Our simulation analysis indicated that with such a complex distribution pattern, the background noise cannot be reduced significantly by increasing the dilution factor at least in a practical range (1 to 1/20 haploid genome equivalent per tube) while this can be achieved by sperm analysis. To achieve the same degree of statistic confidence for linkage detection, 40–50 times more samples are required by Happy Mapping if the DNA sample is aliquoted into one haploid genome equivalent per tube compared with sperm analysis when each sperm lysate is aliquoted into 16 tubes.

Two V_H4 segments on chromosome 14, 4–61 (4–11 in haplotype 4) and 4–59 (4–4 and 4–7 in the haplotypes) were found at the border between a diversified and a conserved portion (Fig. 3). This finding indicates that the relative locations of these two genes may vary considerably among the haplotypes because gene composition in the adjacent diversified portion may vary considerably. Results from our recent study (41) indicate that the 4–61 locus is highly polymorphic with at least four alleles including a null allele at a frequency of 23%. The presence or absence of the 4–61 segment also could affect the relative location of 4–51. On the other hand, these two segments were shown to be involved in generating anti-DNA autoantibodies in the patients with systemic lupus erythematosus (42–44). The usage of the V_H segments could be considerably affected by their organization. For example, when a frequently used V_H segment is absent, other V_H segment(s) may be used more frequently. With sperm analysis, it is possible to compare the V_H gene organization in the haplotypes of normal individuals with that of patients with autoimmune diseases. The study could provide important information about the effect of V_H segment organization on autoimmune diseases.

ABBREVIATIONS

DGGE

denaturing gradient gel electrophoresis

V_H

heavy chain variable region

CE

core element