Genomic organization, chromosomal localization, and transcriptionalvariants of the murine pactolus gene

Rebecca L. Margraf, Yiyou Chen, Sean Garrison, Janis J. Weis, John H. Weis

The Division of Cell Biology and Immunology, Department of Pathology, University of Utah School of Medicine, 50 N. Medical Drive,Salt Lake City, Utah 84132, USA

Received: 13 April 1999 / Accepted: 18 June 1999

Abstract. The Pactolus gene encodes a cell surface protein simi-lar to that of theb integrin subunit family. The murine Pactolusgene is comprised of 16 exons that encompass 24 kb of the ge-nome. The genomic organization of the Pactolus gene is verysimilar to that described for the humanb2 integrin gene (anddeduced for murineb2 integrin), including a separate exon con-taining only 58 untranslated sequences. The Pactolus gene wasmapped to a terminal region of murine Chromosome (Chr) 16,distinct from the previously mapped site of theb2 integrin gene onmurine Chr 10. The Pactolus gene encodes three distinct tran-scripts via alternative splicing. The Pactolus A transcript encodesthe full-length protein including transmembrane and cytoplasmicdomains, while the Pactolus B transcript truncates translation be-fore reaching these membrane-anchoring sequences. A newly dis-covered form, Pactolus C, is found in neonatal samples (along withPactolus A) and would also encode a prematurely terminated pro-tein. This form is derived from an alternative splicing event thatskips part of exon 11, deletes exon 12, and uses an alternativeacceptor site upstream of exon 13. The formation of the PactolusB and C forms is thus governed by a complex alternative splicingmechanism that is affected by the developmental status of theanimal.

Introduction

Pactolus was first defined as a differentially expressed gene prod-uct present in murine bone marrow mast cells derived in stem cellfactor (SCF2) but lacking in the analogous cells derived in IL-3(Chen et al. 1998). Sequence analysis of the full-length PactoluscDNA indicated it contained regions of high homology to mem-bers of theb integrin subunit family, particularlyb2 integrin.Transcript analysis suggested that bone marrow possessed thehighest level of Pactolus expression. Of interest was the findingthat the Pactolus gene, in adult bone marrow, produced two dis-tinct transcripts. One of these transcripts (dubbed Pactolus A)would be expected to encode a protein very similar to that of theb integrins with C-terminal transmembrane and cytoplasmic do-mains, while the other transcript (dubbed Pactolus B) would en-code a truncated protein lacking these C-terminal domains. Biotinlabeling of cell surface proteins followed by immunoprecipitationwith an antibody specific for cytoplasmic domain residues con-

firmed the expression of the full-length Pactolus protein on thesurface of bone marrow cells (Chen et al. 1998).

The importance of the integrin family in bone marrow devel-opment, inflammation, targeted cell homing, and general develop-ment is evident in the phenotype of natural or targeted mutationsin such genes. For example, spontaneous mutations within theb2integrin give rise to a disease known as Lymphocyte AdhesionDeficiency in which the migration of neutrophils to sites of infec-tion is abrogated, resulting in chronic and life-threatening bacterialinfections (Springer 1990). Mouse knockout studies on such chainsasb7 anda4 have demonstrated an absence of lymphocytes withinmucosal tissues (Wagner et al. 1996), or developmental defects inthe heart and placenta (Yang et al. 1995), respectively. RAG-1-deficient blastocysts reconstituted witha4-deficient ES cells giverise to viable animals that lack such morphological sites of immu-nity as the intestinal Peyer’s patches (Arroyo et al. 1996). Thus, thefamily of integrin chains plays many roles in the functional devel-opment of the animal.

The homology between Pactolus and theb2 integrin subunit(Wilson et al. 1989) is evident for the first 631 N-terminal residuesof the Pactolus protein (63% identity). However, two regions arequite dissimilar. First, the amino acids from residue 631 to thec-terminus of the Pactolus protein (residue 738) are quite distinctfrom any otherb integrin subunit. Second, the Pactolus proteinlacks residues presumably important for the functional folding ofthe MIDAS domain (metal ion dependent adhesion sequence). TheMIDAS domain has been demonstrated to be critical fora andbintegrin chain pairing, and subsequent ligand recognition (Bajt andLoftus 1994; Michishita et al. 1993; Rieu et al. 1994; Ueda et al.1994; Wardlaw et al. 1990). The Pactolus protein is altered at thissite by losing the strict conservation of the DxSxS region (replac-ing it with GxSxS) and lacking approximately 28 amino acidswithin that region compared with the otherb integrins. Thus, whilePactolus possesses many characteristic features of theb integrins,it is distinct enough such that the protein may be more appropri-ately referred to asb integrin-like instead of simply referring to itas ab integrin subunit.

The similarity between Pactolus and theb integrins suggeststhat they may have evolved from a common precursor geneticelement. To investigate this possibility, we cloned the murine Pac-tolus gene, compared its intron/exon structure with that known forthe b2 integrin, and determined its chromosomal location. Whilethe Pactolus gene is most similar in structure to that of theb2integrin subunit, it does not share a common chromosomal locationwith that gene (MacDonald et al. 1991). In addition, further tran-script analysis of the Pactolus gene products has identified a thirdCorrespondence to:J.H. Weis

Mammalian Genome 10, 1075–1081 (1999).

© Springer-Verlag New York Inc. 1999

Incorporating Mouse Genome

Pactolus transcript (dubbed Pactolus C) present in neonatal liverthat is also predicted to encode a truncated form of the proteinsimilar to the Pactolus B product.

Materials and methods

Genomic library analysis.Phage containing the Pactolus gene wereisolated by screening a lambda dash (Stratagene) murine genomic librarycreated with 129/sv liver DNA with two Pactolus cDNA probes. Theprobes (a 58 probe 859 and a 38 probe 814: see Fig. 1) were made by PCRof Pactolus cDNA (Chen et al. 1998). Pactolus phage are named 58 Phageor 38 Phage to indicate which cDNA probe was used to isolate these phagefrom the library. Fourteen independent clones were isolated: all were spe-cific for the same gene, Pactolus. Each phage contained about 15 kb ofgenomic DNA.

Mapping and sequence of the Pactolus gene.Pactolus phage weregrown in and purified from LE392E. coli. The phage were ordered 58 to38 along the Pactolus gene by dot blot analysis of the phage DNA, utilizingfour Pactolus cDNA probes (listed as 901, 902, 881, and 837). To locatethe boundaries of the exons, 1mg of phage DNA was sequenced withPactolus primers (10 pmol total). This was done for both intron boundariesof each exon. Exact intron sizes were determined where the sequencesoverlapped. Other introns were estimated by PCR between exons with thesame primers as for sequencing. Placements of exons 15 and 16 are esti-mates based on exon 16’s inclusion of the terminalHindIII site. Exons #2and #4 were accurately placed on the restriction enzyme (RE) map byrestriction sites found in the sequences 58 and 38 of each exon, respectively.Sizes of introns #1, #4, and #14 were estimated after exon placement wasknown. In addition, subcloning of specific Pactolus genomic fragmentsinto plasmid vectors allowed for a confirmation of exon and restrictionenzyme site placement.

Transcript RT-PCR analysis.Neonatal mouse liver RNA and adultbone marrow RNA were analyzed for Pactolus transcripts with RT-PCR aspreviously published (Tan and Weis 1992). Primers designated for the PCRanalysis are shown in Fig. 1.

FISH analysis.Three overlapping Pactolus phage (58 Phage #1, 58 Phage#2, and 38 Phage #2) were used as independent probes of a murine chro-mosomal spread. The Pactolus probes were labeled by nick translation withSpectrumOrange dUTP (Vysis, Inc.), and the chromosomes were visual-ized with DAPI.

Oligonucleotides.Oligonucleotides used in this study are as follows(sequence written 58 to 38): #803-CAAAGTGGACTTCCAGCGGAC,#804-GATTCCACACTGGCTCCACATG, #807-GTCCAGCTG-GCAAGCAAACTG, #814-GCTGTTGGACAACCTACACTC, #815-GCACCTGCAGATGCCACACTC, #816-TGCTTGCCAGCTGGAC-C A C T G , # 8 1 7 - C C A G A G A G G C C C A T C A G G A A G , # 8 2 6 -T T G C T G A A T C C C A G G A A G A C , # 8 2 8 - G A G T C C G C A G -TCAGCTTCAG, #837-CGAGTGCGACAATGTCAACTG, #842-CATCACAGCTTCAGAGTGCT, #848-TAGTACTCGGAGCAGC-G A T G G , # 8 5 9 - G C T A T A A G G A G C T G A A G A A T C , # 8 8 1 -AGGGAAGCTGCCGGAAGGAC, #900-CCGCTCGAGACAGCT-GCCTCTGGTCCCTC, #901-CTGGGTACCAGACAGAACAC, #903-TCGGCCACAGGAGCAGTGGC and #1005-ACTGTAGGC-GTTCCTGATGAG.

Results

Isolation and characterization of the murine Pactolus gene.Amouse genomic library was screened in two successive hybridiza-

Fig. 1. Organization of the murine Pactolus gene.A. The structure of themurine Pactolus gene was determined by analysis of a series of genomiclambda phage contigs. Three representative overlapping phage are denotedas 58 Phage #1, #2, and 38 Phage #2. The distance from the first to the lastexon is approximately 24 kb. Exons 15 and 16, denoted with the *, whichindicates that their position was approximated from the Southern data.

Restriction sites are denoted: E,EcoR1; B,BamH1; G,BgII; H, HindIII. B.The coding region (shaded area) and untranslated regions (white) of thefull-length Pactolus cDNA are shown in relation to the positions of theexons. In addition, the position of probes described in Materials and meth-ods is shown, as are the relative locations of PCR primers used in theRT-PCR analysis of Pactolus transcripts (see Fig. 5).

R.L. Margraf et al.: Pactolus gene structure1076

tions with cDNA sequences specific for the 58 and 38 regions of thecoding sequence of Pactolus (Chen et al. 1998). This screeningidentified 14 total distinct phage, each possessing approximately15 kb of genomic insert. Overlap analysis indicated all 14 werespecific for Pactolus and did not represent any other related genesequence. A subset of phage selected from the overlap analysis

were utilized for sequencing of the Pactolus gene. As shown inFig. 1, panel A, together the three overlapping genomic phageencompass the entire genomic sequence of the Pactolus gene,which is approximately 24 kb in size. The Pactolus gene is com-posed of 16 distinct exons, as defined by both PCR probe hybrid-ization and sequence analysis. The various cDNA probes used in

Fig. 2. Sequence analysis of the Pactolus intron/exon borders.A. The sequenceof the 16 exons of murine Pactolus are shown in order, 58 to 38. The first exonis untranslated, while the second exon contains the initiating ATG, which is inbold and underlined. The exon sequences that are shown in the middle dark boxrepresent only part of the sequence, with the exact number of base pairs withineach exon noted in the parentheses. The numbering above the exons indicatestheir position within the cDNA sequence. Flanking intron sequences are shownin the lightly shaded boxes with the donor sites on the right (with the conservedGT sequence) and the acceptor sequences on the left (possessing the conserved[C/T]AG sequence). The sizes of the introns are shown on the right: they areeither exact determinations based upon sequence analysis or approximate val-ues (denoted with *) based on PCR. ** indicates intron size approximated afterlocalization of the exons. The phase of coding is denoted under the donor sitesof the introns.Panel B. The first 750 bp 58 of the transcription start site forPactolus is shown. The +1 under the nucleotide G indicates the first nucleotideof the mature mRNA. A number of consensus site sequences for transcriptionfactors were located with MatInspector (http://www.gsf.de/cgi-bin/matsearch2.pl) (Quandt et al. 1995). Only those sites with the highest degree ofconservation are noted.

R.L. Margraf et al.: Pactolus gene structure 1077

this and following analyses of Pactolus are shown in Fig. 1,panel B.

The specificity of the Pactolus cDNA probes, and thus theuniqueness of the phage clones, was confirmed by probing mousegenomic DNA with the three Pactolus cDNA probes. All three ofthe probes identified genomic fragments similar in size to thosepredicted from the Pactolus genomic phage clones (data notshown). Of note, probe 881, which was derived from the Cys-richregion of the Pactolus sequence and has considerable sequencehomology to theb2 integrin, also faintly detected fragments pre-sumed to be derived from the murineb2 integrin gene (data notshown).

The coding regions within the genomic phage were determinedby sequencing the phage DNA or plasmid subclones. As shown inFig. 2, panel A, those sequences contained within the mature Pac-tolus transcript are derived from 16 exons of varying size. The firstintron contains only 58 untranslated sequence. The initiating ATGis present within the second exon, which encodes the predictedsignal sequence. The MIDAS-like domain region is includedwithin exons 5 and 6. Exon 15 encodes the transmembrane domainand a region of the cytoplasmic sequence. Exon 16 encodes theremaining cytoplasmic sequence as well as the full 38 untranslatedregion. The sizes of the exons and introns as well as the splicingphase are shown in Fig. 2, panel A.

The 58 end of the Pactolus transcript maps to the first exon,which is located approximately 6.8 kb upstream of the secondexon. The original Pactolus cDNA sequence contained the full

58 untranslated region in that RACE analysis of the 58 regionof Pactolus transcripts confirmed the guanine residue (GC-TATAA. . . .) as thefirst base of the mature transcript (data notshown). As shown in Fig. 2, panel B, the putative promoter se-quence of the Pactolus gene upstream of the first exon is unusualin that it does not appear to possess a TATA box to specify tran-scription initiation. However, at base −30 is a TAA sequence thatmay serve to help load the transcriptional apparatus. A number ofpossible transcriptional control sequences were found in this pu-tative promoter region. These include an ELK-1 site at −46, twoAP-1 sites at −56 and −80, an AP-4 site at −58, a C/EBP site at−117, and an E47 site at −140. Additionally, a strong consensussite for NF-kB binding is present at −175, multiple putative Ikarossites, and two GATA sequences. Most noteworthy of these lattersites is the GATA site at −560. The consensus sequence for thebinding of GATA family members is (A/T)GATA(A/G) (Evansand Felsenfeld 1989; Tsai et al. 1989) which matches the Pactolussequence of TGATAG. While there are a number of GATA tran-scription family members, GATA-1 might be suspected to play acentral role in Pactolus expression. GATA-1 is expressed only inhematopoietic derived cells including erythroid, megakaryocytic,mast cell and eosinophil lines, as well as in multipotential pro-genitors (Orkin 1992). Therefore, the expression pattern ofGATA-1 is consistent with the expression profile of Pactolus(Chen et al. 1998, data not shown). Also of interest are the putativeNF-kB and Ikaros sites. NF-kB regulation of Pactolus expressionwould suggest its level of transcription can be induced with vari-

Fig. 3. Intron/exon homology between murinePactolus and murine integrin subunitb2. Thecoding sequences of Pactolus andb2 integrin areshown from exons 2 to 16. The full intron/exonmap of the murineb2 integrin gene has not beenpublished; exons are inferred from comparison withthe intron/exon junctions of the humanb2 integringene (Weitzman et al. 1991; Wilson et al. 1993).Sequences have been adjusted to give the greatestdegree of alignment. The gapped sequence ofPactolus in exon 6 represents the sequences withinthe MIDAS domain that are lacking in the Pactolusgene (compared with other members of this genefamily).

R.L. Margraf et al.: Pactolus gene structure1078

ous immunological or inflammatory stimuli (May and Ghosh1998). Ikaros is a transcription factor preferentially expressed inlymphoid lineages (Kim et al. 1999; although it is also found innon-lymphoid cells such as mast cells: Y. Chen and J.H. Weisunpublished data). Regulation via Ikaros would be consistent withthe observed bone marrow-specific expression pattern seen forPactolus.

The Pactolus promoter region is similar to that of theb2 in-tegrin subunit in that the first intron ofb2 integrin is also quitelarge, the first exon does not encode protein sequences, and tran-scription initiation does not appear to make use of a TATA boxsequence (Rosmarin et al. 1992; Weitzman et al. 1991). The simi-larity of the Pactolus gene to theb2 integrin gene is also strikingwith regards to exon domain boundaries. When the splice sitesbetween Pactolus andb2 integrin are aligned, the two utilize vir-tually identical junctions (Fig. 3). This conservation breaks downwhere there is little homology between Pactolus andb2 integrin, atexons 15 and 16. The gap shown in the coding region of exon 6represents a sequence that is lacking in the MIDAS-like domain ofPactolus compared with the otherb integrin subunits. Thus, theabsence of a portion of the MIDAS region from the Pactolustranscript is not a result of alternative splicing, but instead repre-sents an apparent loss within the genomic sequence of Pactolus.

Chromosomal location of the Pactolus gene.The high level ofhomology between Pactolus and theb2 integrin suggests theywere derived from a common ancestor gene. Thus, it was of in-terest to determine whether the two genes were linked within thegenome. FISH analysis was performed with three different Pacto-lus phage clones, displayed in Fig. 1, to probe mouse chromosomespreads (Fig. 4). All three clones mapped the Pactolus gene to theterminus of Chr 16, band C4. This is a different location frommurineb2 integrin, which maps to murine Chr 10 (MacDonald etal. 1991). However, both the Pactolus andb2 integrin regionsrepresent sequences syntenic to human Chr 21 (MacDonald et al.1991). This raises the possibility that the two genes were linkedduring mammalian evolution but separated during rodent speciation.

Identification of the Pactolus C transcript.As previously de-scribed, bone marrow is the tissue in the adult mouse with thehighest expression level of Pactolus (Chen et al. 1998). Transcriptanalysis of mature marrow identified two distinct forms of Pacto-lus which are derived from alternative splicing. One of these tran-scripts encodes a transmembrane form (full length) of Pactolus(Pactolus A), while the second encodes a smaller, truncated form(Pactolus B). The Pactolus B form appears to be unstable and tohave a very short half life (S. Garrison and J.H. Weis, unpublisheddata). Since the fetal and neonatal liver is a site of hematopoieticcell development, we screened the liver for Pactolus transcripts. Incomparison with adult liver, the neonatal liver did possess signifi-cant quantities of Pactolus transcripts (data not shown). We thenscreened neonatal liver cDNA samples, via RT-PCR, with a vari-ety of oligonucleotides (Fig. 1B) which would, in the entirety ofthe overlaps, include the entire Pactolus coding sequence. Asshown in Fig. 5, virtually the same level of Pactolus transcriptionis evident in neonatal liver as adult marrow (based upon actinequivalence). While the majority of the bands between the samplesare the same size (indicating identical transcript products), aunique band of 400 bp was evident in the neonatal liver samplewith the oligo set 842 and 848. This band was excised from the gel,reamplified, cloned, and sequenced. As shown in Fig. 6, the se-quence from this insert matched that of exon 11 and exon 13, withan apparent loss of 37 bp from exon 11 and an insertion of 43bp

58 of exon 13. Exon 12 was not present in this sequence. Thisproduct, dubbed Pactolus C, would result in a protein with alteredreading frame from the full length protein, thus creating a trun-cated product (Fig. 6). From the apparent instability of the PactolusB product, the Pactolus C protein would also be expected to beshort lived.

Discussion

The Pactolus gene is unusual in specifying three distinct tran-scripts, two of which would result in truncated products. As shown

Fig. 4. Chromosomal localization of the murine Pactolus gene. Fluores-cent in situ hybridization (FISH) was performed with Pactolus phageclones 58 Phage #1, #2, and 38 Phage #2, shown in Fig. 1A. The two FISHpanels are the same probed karyotype, the top panel is a DAPI stain,followed by a chromosomal banding pattern in the lower panel. The twored dots show the position of hybridization of the Pactolus probes. Whileonly the data for 58 Phage #1 are shown, all three probes did hybridize tothe terminal region of murine Chr 16 at band C4 (arrows).

R.L. Margraf et al.: Pactolus gene structure 1079

in Fig. 7, panel A, the three Pactolus proteins predicted from the A,B, and C transcripts would all contain the same extracellular do-mains encoded in the first 10 exons. At exon 11, the C formutilizes a distinct donor sequence not used by the A or B forms,

and skips over exon 12 to an acceptor site in intron 12, 43 bpupstream of exon 13. The B form splices as the A form until exon13, where the B form utilizes a distinct acceptor site present 37 bpinto exon 13 (Fig. 7, panel B). Both B and C transcripts produceproducts whose reading frames vary, owing to the alternative splic-ing of the primary transcripts, from each other as well as the nativePactolus sequence. The alternative splice variants would be ex-pected to result in truncated proteins since both the B and C formsgenerate stop codons within the exon 13 sequence.

This analysis of the Pactolus gene has provided a variety ofinformation. We have been able to compare the structure of thePactolus gene with that of the other reportedb integrin genes. Aspreviously reported, the nucleotide and deduced amino acid se-quence of Pactolus is closely related to that ofb2 integrin (Chenet al. 1998). Both genes appear to have non-TATA box promoterregions and possess virtually identical exon/intron junctions. Thesedata thus suggest that these two genes are derived from a commonprecursor sequence. In contrast, however, Pactolus andb2 integrinreside at two distinct chromosomal locations. It will be of interestto determine whether the humanb2 integrin and Pactolus genesare linked, or whether their chromosomal separation occurred priorto rodent/primate differentiation.

The organization of the Pactolus gene has also shed under-standing on the nature of the Pactolus splice variants. These vari-ants are not rare products: the ratio of the Pactolus B to the A formis about 10 to 1 in adult mouse marrow, while that of C to A in theneonatal liver sample is approximately 1 to 2. It is interesting tonote that the production of such altered transcripts centers upon therecognition of splice sites within a confined region of the gene(i.e., exons 11, 12, and 13). The structure of the B and C variantssuggests the production of secreted, albeit unstable, proteins whilethe A form would be held within the membrane. Although thesedata help to explainhow such variants are made, they do notaddresswhy they are produced. Experimentation to address thatquestion is under way.

Acknowledgments.This research was supported by National Institutes ofHealth (NIH) grants AI-42032 and AI-24158 (J.H. Weis), an award fromthe American Lung Association (J.H. Weis), NIH grants AI-3223 andAR-43521 (J.J. Weis), and funds from the Center of Excellence in Hema-tology, grant DK-49219. This work was also supported by the HuntsmanCancer Institute and the National Cancer Institute grant 5 P30 CA-42014.

Fig. 6. Sequence and alignment of the Pactolus C transcript. The nucleo-tide sequence derived from the Pactolus C PCR fragment (PacC) shown inFig. 5 is compared with the analogous sequence from the full-length Pac-tolus transcript (PacA:underlined) from exon 11 to exon 13. Amino acidsencoded by the Pactolus A sequence are denoted above the PacA sequence,while the unique coding sequences of the PacC sequence are shown belowthe PacC sequence. Where there is coding unity between the PacA and PacC sequence, only the PacA amino acids are shown. The Pactolus C se-quence splices out of exon 11 early using an alternative donor site (con-served GT), skips exon 12, and utilizes an alternative acceptor site withinintron 12. The alternative splicing changes the reading frame for the Pac-tolus C product, resulting in a premature termination of the protein, de-noted with *.

Fig. 5. Identification of the Pactolus C transcript.cDNA derived from RNA obtained from neonatal(5-day-old) liver (N) and adult bone marrow (A)was equalized for actin transcript content and thenamplified via PCR with series of oligonucleotidespositioned at intervals over the full codingsequence of Pactolus. The position of theoligonucleotides is shown in Fig. 1, panel B.Molecular weight markers are shown on the left.The alternative spliced Pactolus C product is shownas a 400-bp band within the N lane of oligo’s842/848 (denoted by the box) and is denoted withthe arrow on the right and an * by the specificband.

R.L. Margraf et al.: Pactolus gene structure1080

References

Arroyo AG, Yang JT, Rayburn H, Hynes RO (1996) Differential require-ments for alpha4 integrins during fetal and adult hematopoiesis. Cell 85,997–1008

Bajt ML, Loftus JC (1994) Mutation of a ligand binding domain of beta 3integrin. Integral role of oxygenated residues in alpha IIb beta 3 (GPIIb-IIIa) receptor function. J Biol Chem 269, 20913–20919

Chen Y, Garrison S, Weis JJ, Weis JH (1998) Identification of pactolus, anintegrin beta subunit-like cell-surface protein preferentially expressed bycells of the bone marrow. J Biol Chem 273, 8711–8718

Evans T, Felsenfeld G (1989) The erythroid-specific transcription factorEryf1: a new finger protein. Cell 58, 877–885

Kim J, Sif S, Jones B, Jackson A, Koipally J et al. (1999) Ikaros DNA-binding proteins direct formation of chromatin remodeling complexes inlymphocytes. Immunity 10, 345–355

MacDonald G, Chu ML, Cox DR (1991) Fine structure physical mappingof the region of mouse chromosome 10 homologous to human chromo-some 21. Genomics 11, 317–323

May MJ, Ghosh S (1998) Signal transduction through NF-kappa B. Im-munol Today 19, 80–88

Michishita M, Videm V, Arnaout MA (1993) A novel divalent cation-binding site in the A domain of the beta 2 integrin CR3 (CD11b/CD18)is essential for ligand binding. Cell 72, 857–867

Orkin SH (1992) GATA-binding transcription factors in hematopoieticcells. Blood 80, 575–581

Quandt K, Frech K, Karas H, Wingender E, Werner T (1995) MatInd andMatInspector: new fast and versatile tools for detection of consensusmatches in nucleotide sequence data. Nucleic Acids Res 23, 4878–4884

Rieu P, Ueda T, Haruta I, Sharma CP, Arnaout MA (1994) The A-domainof beta 2 integrin CR3 (CD11b/CD18) is a receptor for the hookworm-

derived neutrophil adhesion inhibitor NIF. J Cell Biol 127, 2081–2091

Rosmarin AG, Levy R, Tenen DG (1992) Cloning and analysis of theCD18 promoter. Blood 79, 2598–2604

Springer TA (1990) Adhesion receptors of the immune system. [Review][165 refs]. Nature 346, 425–434

Tan SS, Weis JH (1992) Development of a sensitive reverse transcriptasePCR assay, RT-RPCR, utilizing rapid cycle times. Genome Res 2, 137–143

Tsai SF, Martin DI, Zon LI, D’Andrea AD, Wong GG et al. (1989) Cloningof cDNA for the major DNA-binding protein of the erythroid lineagethrough expression in mammalian cells. Nature 339, 446–451

Ueda T, Rieu P, Brayer J, Arnaout MA (1994) Identification of the comple-ment iC3b binding site in the beta 2 integrin CR3 (CD11b/CD18). ProcNat Acad Sci USA 91, 10680–10684

Wagner N, Lohler J, Kunkel EJ, Ley K, Leung E et al. (1996) Critical rolefor beta7 integrins in formation of the gut-associated lymphoid tissue.Nature 382, 366–370

Wardlaw AJ, Hibbs ML, Stacker SA, Springer TA (1990) Distinct muta-tions in two patients with leukocyte adhesion deficiency and their func-tional correlates. J Exp Med 172, 335–345

Weitzman JB, Wells CE, Wright AH, Clark PA, Law SK (1991) The geneorganization of the human beta 2 integrin subunit (CD18). FEBS Lett294, 97–103

Wilson RW, O’Brien WE, Beaudet AL (1989) Nucleotide sequence of thecDNA from the mouse leukocyte adhesion protein CD18. Nucleic AcidsRes 17, 5397

Wilson RW, Ballantyne CM, Smith CW, Montgomery C, Bradley A et al.(1993) Gene targeting yields a CD18-mutant mouse for study of inflam-mation. J Immunol 151, 1571–1578

Yang JT, Rayburn H, Hynes RO (1995) Cell adhesion events mediated byalpha 4 integrins are essential in placental and cardiac development.Development 121, 549–560

Fig. 7. Alignment and generation of the threealternatively spliced forms of Pactolus.Panel A.The cDNAs of the full-length (Pactolus A) andtruncated Pactolus (Pactolus B and C) are showndivided into their 16 exons. The solid shaded areabetween the dashed lines is translated sequence,while the white area is untranslated. Theextracellular (E-C), transmembrane (T),cytoplasmic (C), and the MIDAS-like (MIDAS)domains are shown below the cDNA (based onhomology withb2 integrin). In Pactolus B thesequence of exon 13 is partially lost owing toalternative splicing. Pactolus C is shown with ashortened exon 11 sequence, no exon 12, and aninsertion of part of intron 12. Codon phases ofselected introns are labeled within the white circlesto show the alternative reading frames (comparedwith the full-length Pactolus A transcript) of thePactolus B and C products. Forms B and Ctruncate early within exon 13 owing to theiralternative reading frames, which possess multipletranslation termination codons. All three formsapparently utilize the same transcriptional start andstop sequences.Panel B.Diagram demonstratinghow alternative splicing gives rise to the threeforms of Pactolus transcripts. The only knownalternatively spliced regions of Pactolus, exons 11through 13, are shown. Exons are in bold, with thedeletions shaded. The intronic sequenceincorporated in the Pactolus C transcript is striped.

R.L. Margraf et al.: Pactolus gene structure 1081