Vox Sanguinis (2015) 108, 403–409 © 2015 International Society of Blood Transfusion DOI: 10.1111/vox.12236
ORIGINAL PAPER
Dissecting alternative splicing in the formation of Miltenberger glycophorin subtype III (GYP.Mur) K. Hsu, C.-C. Yao, Y.-C. Lin, C.-L. Chang & T.-Y. Lee Mackay Memorial Hospital Transfusion Medicine Laboratory & Blood Bank, Tamsui, Taiwan
Background and Objectives Miltenberger subtype III (Mi.III, GP.Mur) is one of the most important red cell phenotypes in the fields of transfusion in South-East Asia. GP.Mur is believed to evolve from homologous gene recombination events between glycophorin A (GYPA) and glycophorin B (GYPB). GYP.Mur differs from GYPB in only seven nucleotides dispersed near the region of 30 exon 3 of GYP.Mur. The goal of this study was to dissect how these nucleotide variants affected splicing of exon 3. Materials and Methods We first designed two minigene constructs: one containing GYP.Mur from exon 2 to exon 4 and the other containing GYPB in the same region. To test how these nucleotide variations between GYP.Mur and GYPB affected the splicing, a repertoire of the GYP.Mur-like minigene constructs with different point mutations were created. These minigene variants were evaluated for their abilities to induce splicing of exon 3 using a heterologous expression system. Results (1) GYP.Mur minigene expressed exons 2, 3 and 4, whereas GYPB minigene expressed only exon 2 and exon 4. (2) The single nucleotide alteration at the position of the 50 splice site of glycophorin intron 3 reversed the splicing decision. (3) The nucleotide variations between GYP.Mur and GYPB other than that at the 50 splice site showed very little or no effect on splicing of exon 3. Received: 8 April 2014, revised 14 November 2014, accepted 18 November 2014, published online 6 March 2015
Conclusion Splicing of the glycophorin B-A-B hybrids (GYP.Mur and GYP.BUN) and unsplicing of GYPB follow the GU-AG rule strictly. Key words: alternative splicing, exonic splicing enhancer, glycophorin B, GYP.Mur, Miltenberger type III, minigene.
Introduction Miltenberger red cell phenotypes in the MNS blood group system are generally rare, except Miltenberger type III (Mi.III, GP.Mur), with 2–10% occurrence frequencies in several South-East Asian (SEA) countries [1–4]. The epitopes on GP.Mur protein, such as ‘Mia’ and Mur, could elicit strong allo-immune responses and are associated with intravascular haemolytic transfusion reactions and haemolytic diseases of foetus and newborn (HDFN) [5–7]. Therefore, GP.Mur is considered one of the most important RBC phenotypes in transfusion medicine in SEA, and Correspondence: Kate Hsu, PhD, 45 Min-Sheng Rd, Research Building 616, Tamsui 251, Taiwan E-mail: [email protected]
‘Mia’+ red cells have been included in routine antibody screening panels for ~20 years in Taiwan [8]. The rise of Miltenberger phenotypes is believed to evolve from gene crossover events among glycophorin A (GYPA), glycophorin B (GYPB) and glycophorin E (GYPE) [9]. Resolved by Huang et al. [10] in 1991, GYP.Mur, the Mi.III-specific gene encoding GP.Mur, bears the structure of glycophorin B-A-B hybrid. The genomic sequence of GYP.Mur is almost identical to that of GYPB, except for seven nucleotides dispersed in or near the 30 end of exon 3 (Fig. 1) [11]. The protein sequences of GP.Mur and glycophorin B (GPB) are almost identical, except for that GP.Mur, like glycophorin A (GPA), expresses its exon 3-encoded residues and GPB does not. It was predicted by sequence comparison that an emergence of the 50 splice site in this region due to gene crossover may enable
403
404 K. Hsu et al.
Fig. 1 GYP.Mur (JN_201202) and GYP.Bun alignment to GYPA (NG_007470.3) and GYPB (JN_201201) in the region of exons 2–4. Pseudoexon 3 of GYPB is labelled with dark grey colour; exon 3 of GYPA is with light grey colour. The first portion of exon 3 of GYP.Mur is coloured with dark grey, and the rest of exon 3 is with light grey, as GYP.Mur is a hybrid form of GYPB and GYPA in this region. GYP.Bun shares a similar glycophorin B-A-B hybrid structure as GYP.Mur. The 30 and 50 splice sites adjacent to exon 3 are boxed. The nucleotide differences between GYPB and GYP.Mur are marked with asterisks [11].
splicing of exon 3 in GYP.Mur. The corresponding sequence in GYPB lacking a functional 50 splice site is not subjected to pre-mRNA splicing and is thus termed ‘pseudoexon 30 . In this report, we created ‘minigenes’ of GYP.Mur and GYPB and tested their splicing or unsplicing of exon 3 using a heterologous expression system. Our experimental results indicate that conservation of the 50 splice site immediately following exon 3 is essential for the splicing event to take place. On the other hand, the other nucleotide variations between GYP.Mur and GYPB appear to have very small or no effects on splicing of exon 3.
Materials and methods Construction of minigene plasmids and mutant forms The genomic sequence of GYP.Bun (Miltenberger antigen subtype VI, Mi.VI) was determined from a serologically identified GP.Bun+ blood sample. Our GYP.Bun sequence (KM655769) is identical to the previously published GYP.Bun sequence (M60710.1) in exon 3 [10], except few sporadic variations between the two sequences in their intron regions. The genomic fragments of GYPB
(Accession number JN_201201), GYP.Mur (JN_201202) and GYP.Bun from exon 2 to exon 4 were subcloned into pN2-EGFP (Clontech, Mountain, CA, USA). The constructed minigene-GFP fusion plasmids for GYPB, GYP.Mur and GYP.Bun were named pGYPBe2e4, pMURe2e4 and pBUNe2e4, respectively. Selected single point mutations in GYPBe2e4 and MURe2e4 were created using QuikChange Site-Directed Mutagenesis kit (Staratagene, La Jolla, CA, USA).
Spicing experiments with minigenes The day prior to transient transfection, HEK-293 cells were seeded in a 12-well cell culture plate at a density of 2 9 105 cells/well and maintained in DMEM supplemented with FBS and antibiotics. Minigene plasmids were individually transfected using T-Pro transfection reagent II (T-Pro Biotechnology, Taipei, Taiwan) for 4 h. Minigene splicing events in these cells were usually studied at 26– 44 h post-transfection. Splicing of (pseudo)exon 3 in these minigenes was assessed by the lengths of mRNA transcribed between exon 2 and exon 4, followed by sequencing validation. Total RNA was extracted from transfected HEK-293 cells using a total RNA mini kit (Viogene-BioTek, Taipei, © 2015 International Society of Blood Transfusion Vox Sanguinis (2015) 108, 403–409
Alternative splicing for GYP.Mur (Mi.III) 405
Taiwan) and reverse transcribed to cDNA with ReverTra Ace (TOYOBO, Osaka, Japan). PCRs were performing with the two pairs of primers (GB18064F & GFP151R; b-actin F & b-actin R). PCR products were then resolved on a 2% agarose gel. The relative quantities of RT-PCR products were analysed using Image J (NIH). Expression of GYPBe2e4 minigene completely skipped pseudoexon 3 and resulted in a 308-bp DNA fragment (lacking exon 3) from RT-PCR; MURe2e4 minigene yielded complete splicing of exon 3 and resulted in a 401-bp DNA fragment (containing exon 3). To determine the impacts of nucleotide variations on splicing of (pseudo)exon 3, the intensity of each PCR band was compared quantitatively with pGYPBe2e4 (set 0%) and pMURe2e4 (set 100%). The splicing experiments for each minigene construct were performed independently for 3–5 times.
Prediction of splicing factors To test how sequence variations could affect pre-mRNA splicing, the sequences of GYPA, GYPB and GYP.Mur were analysed using Human Splicing Finder, an online prediction program for splicing (http://www.umd.be/HSF/) [12].
Results The nucleotide sequences and protein sequences of many Miltenberger glycophorin hybrid variants were delineated in early 1990s [9]. It was predicted at that time that both glycophorin B-A-B hybrids (Mi.III & Mi.VI) express exon 3 like GYPA, because in their sequences, the 50 splice site immediately following exon 3 is retained [10]. There are
Fig. 2 An outline for the minigene experiments. The genomic sequences of GYPB and GYP.Mur between exon 2 and exon 4 were constructed into pN2-EGFP to create two minigene-GFP fusion plasmids – pGYPBe2e4 and pGYPMURe2e4 (pMURe2e4). These minigene-GFP fusion plasmids were then transfected into cultured cells, and their expressions were assessed by the appearance of green fluorescence. RT-PCR experiments ensued to test whether mRNA transcribed from these minigenes expressed exon 3 or not. A pair of the PCR primers were designed to amplify the sequence from exon 2 to exon 4. So from the sizes of the PCR products, whether exon 3 was spliced or skipped could be determined.
© 2015 International Society of Blood Transfusion Vox Sanguinis (2015) 108, 403–409
7 nucleotide differences between GYPB and GYP.Mur on or near exon 3 [11]. One of the nucleotide differences is located right at the 50 splice site of intron 3. GYPB lacks this splice site and is thus disqualified from splicing of exon 3 (Fig. 1). However, is this single nucleotide difference at the 50 splice site between GYPB and GYP.Mur [Mi.III] solely responsible for the expression of exon 3? Could nucleotide variations other than the one at the 50 splice site also influence the decision of exon 3 splicing? In this study, we designed minigenes of GYPB and GYP.Mur and examined their splicing patterns using a heterologous expression system (Fig. 2). As GYPB differs from GYP.Mur in or near (pseudo)exon 3, their genomic sequences from exon 2 to exon 4 were individually subcloned into the pN2-EGFP vector to create minigene-GFP fusion sequences. These two minigenes (GYPBe2e4 and MURe2e4) differ only in seven nucleotides (marked * in the sequence alignment in Fig. 1), as previously described [11]. To assess the splicing patterns of these minigenes, they were each transiently transfected into HEK-293 cells, and their mRNA expressions were evaluated by RT-PCR. Transfected HEK-293 cells expressed green fluorescence from reporter GFP (Fig. 3a), the N-terminus of which is fused with a fragment of glycophorin B (or GP.Mur). We observed these GFP fusions with GP.Mur/GPB fragments to be primarily localized in a punctate distribution intracellularly (Fig. 3a). Total RNA of these transfected cells was extracted for RT-PCR; cDNA was amplified between exon 2 and exon 4, and their PCR products were verified by DNA sequencing. As expected, GYPBe2e4 minigene, like glycophorin B, did not undergo splicing of exon 3 and only expressed exon 2 and exon 4 (represented by a band of 308 bp). In
406 K. Hsu et al.
(a)
(b)
(c)
contrast, MURe2e4 minigene, like GYP.Mur, expressed exons 2, 3 and 4 (represented by a band of 401 bp) (Fig. 3b,c). We also created a minigene for GYP.Bun (Mi.VI), another glycophorin B-A-B hybrid structurally very similar to GYP.Mur. The BUNe2e4 minigene also expressed exons 2, 3 and 4, as MURe2e4. We did not observe any exon 3-spliced form from the expression of GYPBe2e4 minigene, or any residual exon 3-unspliced form from MURe2e4 (Fig. 3c). Thus, the splicing or skipping of exon 3 entirely depends on the minigene sequences themselves and appears independent of the choice of the expression system. To test whether the different splicing patterns of exon 3 were primarily due to the single nucleotide difference at the 50 splice site of intron 3 between GYPBe2e4 and MURe2e4 (Fig. 3b), two plasmids containing single point mutations at the 50 splice site were created: pGYPBe2e4905T>G and pMURe2e4-905G>T. The point mutation for the former supplied a 50 splice site for GYPB, whereas the point mutation for the latter disrupted the 50 splice site of GYP.Mur to make it resemble GYPB. Our experiments showed that the single mutation (G>T) at the 50 splice site of intron 3 of GYP.Mur prevented much splicing of exon 3 (Fig. 4). A reversal change at the same position on GYPB (T>G) enabled splicing of ‘pseudoexon 30 in GYPB, as shown in Fig. 4. These experimental data support that a single nucleotide difference at the 50 splice site of intron 3 primarily determined whether exon 3 should be spliced or skipped. Since the decision to splice or skip exon 3 for GYP.Mur/GYPB solely depends on their coding sequences (Fig. 3), do the other nucleotide differences between GYP.Mur and GYPB affect the decision of exon 3 splicing
Fig. 3 Pseudoexon 3 of GYPBe2e4 minigene was not spliced, whereas exon 3 of MURe2e4 and BUNe2e4 minigenes were spliced and ultimately expressed. (a) The expressions of GYPBe2e4 and MURe2e4 minigenes were validated by reporter GFP that emitted green fluorescence in transfected HEK-293 cells. (b) Splicing or skipping of exon 3 in GYPA, GYPB, GYP.Mur and GYP.Bun was shown. Light grey colour and dark grey colour indicate the sequences of GYPA and GYPB, respectively. A2/ B2 refers to exon 2 of GYPA/GYPB. Similarly, A3/B3 refers to exon 3 and A4/B4 refers to exon 4. The critical sequences at the 50 and 30 splice sites (GT & AG) are specified. (c) The RTPCR results showed that exon 3 was spliced and expressed with MURe2e4 and BUNe2e4 minigenes and not with GYPBe2e4 minigene.
Fig. 4 Splicing of exon 3 followed the GU-AG rule. These RT-PCR results showed that exon 3 was spliced and expressed from GYPBe2e4-905T>G, a single mutant of GYPBe2e4 minigene that contains a functional 50 splice site in intron 3. In contrast, exon 3 was unspliced with MURe2e4905G>T, a single mutant of MURe2e4 minigene with a disrupted 50 splice site in intron 3.
at all? For this, we employed an online website – Human Splicing Finder (HSF) – to find out whether the other sequence variations between GYP.Mur and GYPB might have a role in promoting or suppressing splicing of exon 3 [12]. From the prediction by HSF, GYP.Mur primarily differs from GYPB with more motifs for exon splicing enhancers (ESEs) (Fig. S1). We did not find any major differences in the prediction for exon splicing suppressors (ESSs) in GYP.Mur, GYPB and GYPA by HSF (Fig. S2). Additionally, GYPA and GYP.Mur share identical sequences in 30 exon 3 and thus have the same prediction for splicing. GYPA and GYP.Mur do differ in 50 exon 3; however, their splicing prediction score (scaled in octamers strength by HSF) appeared the same in 50 exon 3 (Figs S1 and S2). HSF includes three well-studied algorithms for ESE prediction: ESE finder [13, 14], RESCUE-ESE [15], and Zhang and Chasin’s criteria for PESE-octamers [16]. The three ESE prediction methods all identified more ESE © 2015 International Society of Blood Transfusion Vox Sanguinis (2015) 108, 403–409
Alternative splicing for GYP.Mur (Mi.III) 407
signals present in GYP.Mur than in GYPB in 30 exons, where their octamers strengths (splicing scales) differed the most. The ESE finder program further specified that these additional ESE signals found in GYP.Mur are binding motifs for SF2/ASF (Fig. S1). Five SF2/ASF binding motifs were predicted in 30 exon 3 of GYP.Mur, compared to only two SF2/ASF motifs predicted in the same region of GYPB. ASF (alternative splicing factor-1), also known as SF2 (pre-mRNA splicing factor 2), is a serine/arginine (S/R)-rich splicing factor required for pre-mRNA splicing [17, 18]. Others have shown that ASF/SF2 supports splicing in a concentration-dependent fashion [19]. Could more ASF/SF2 motifs in GYP.Mur exon 3 support splicing of exon 3? To test this, we knocked down endogenous ASF by siRNA (Fig. S3A) and repeated the minigene experiments. We, however, did not find the splicing of MURe2e4 to be affected by ASF knock-down (Fig. S3B). These data suggest either that ESEs are not the critical determinants for splicing of glycophorin exon 3, or that the current ESE/ESS prediction programs are still not robust enough. To further explore how the nucleotide variations between GYPB and GYP.Mur, other than that located on
the 50 splice site of intron 3, could affect splicing of exon 3, three additional point mutants based on pMURe2e4 were created: MURe2e4-AAA, MURe2e4-CC and MURe2e4-791A>T (Fig. 5a). For the first two mutants (MURe2e4-AAA and MURe2e4-CC), nucleotide variants between GYP.Mur and GYPB were selectively reverted to resemble GYPB. The other mutant MURe2e4-791A>T has an A to T mutation at the predicted branch point in intron 2. The expressions of these point mutant minigenes were compared to the expression of MURe2e4 minigene, which was set 100% and indicative of complete exon 3 splicing. The expression of GBe2e4 was set 0%, indicative of complete exon 3 skipping. MURe2e4-905G>T (the GYP.Mur minigene with a disabled 50 splice site of intron 3) and GYPBe2e4-905T>G (the GYPB minigene with an enabled 50 splice site) were included in this experiment for comparison with the other point mutants of MURe2e4. With a single point mutation at the 50 splice site in pMURe2e4-905G>T, the expression of GP.Mur exon 3 dropped from 100% to 18 – 10% (Fig. 5b). In pMURe2e4-AAA, three nucleotides were reverted to the corresponding GYPB sequence (g.[895G>A; 898G>A; 902C>A]), and these changes lowered the expression of
(a)
(b)
Fig. 5 Selective single mutations in MURe2e4 minigene affected splicing and expression of exon 3 differentially. (a) Different symbols on the sequence alignment refer to different mutant constructs. Filled inverse triangle, filled diamond and hollow circle represented MURe2e4-based minigenes with different point mutations – MURe2e4-791A>T, MURe2e4-CC and MURe2e4-AAA – respectively. The positions of these point mutations are numbered starting from exon 2 of GYP.Mur and listed as shown. (b) The expressions of minigene point mutants in HEK-293 cells were assessed by RT-PCR and quantitatively analysed using Image J. The PCR band intensity at 401 bp (with exon 3 expression) was individually normalized from each experimental set. The expression levels of exon 3 from these MURe2e4 point mutants were then compared to that from MURe2e4 minigene (set 100%). The expression levels from GYPBe2e4 were set 0% for relative quantification. The relative expression of exon 3 (in percentile) from GYPBe2e4 and MURe2e4 mutant minigenes were calculated accordingly and summarized here from five independent experiments. The relative expressions of exon 3 from BUNe2e4 minigene were summarized from three independent experiments. Shown mean – SEM.
© 2015 International Society of Blood Transfusion Vox Sanguinis (2015) 108, 403–409
408 K. Hsu et al.
exon 3 from 100% to 83 – 11% (Fig. 5b). We also created a double point mutant MURe2e4-CC (g.[875G>C; 884A>C]), which disrupts a ESE motif predicted by the RESCUE-ESE program [15]. In MURe2e4-CC, the two nucleotide variants were reverted to resemble GYPB. This double mutant MURe2e4-CC reduced the expression of exon 3 from 100% to 80 – 10% (Fig. 5b). The minigene BUNe2e4, which resembles MURe2e4-CC at g.[875G>C], also only reduced the expression of exon 3 from 100% to 90 – 7% (Fig. 5b). Additionally, we created single nucleotide change at the predicted branch point of intron 2 of GYP.Mur (MURe2e4-791A>T). This branch point mutant did not affect splicing of exon 3 at all (Fig. 5b). In sum, only the point mutations at the 50 splice site play a determining role in splicing of glycophorin exon 3; all the other mutations tested showed minor or no statistically significant effects on the splicing of exon 3.
Discussion By constructing and experimenting with GYPB/GYP.Mur/ GYP.Bun minigenes that include exon 2 to exon 4, we identified the key determinant for exon 3 splicing to be at the 50 splice site of intron 3. These minigenes alone were sufficient for pre-mRNA splicing and expression in a typical mammalian HEK-293 expression system (Fig. 3). Thus, whether splicing of (pseudo)exon 3 of GYPB/GYP.Mur/GYP.Bun should take place or not depends entirely on the coding sequence. Most of the genes encoding for Miltenberger glycophorin hybrids were resolved about 20 years ago [9, 20]. GYP.Mur and GYP.Bun both have the glycophorin B-A-B gene structure and incorporate a sequence for the 50 splice site of intron 3 from GYPA [10]. By comparison, the corresponding sequence of GYPB lacks the 50 splice site. Because of the single nucleotide variation at the 50 splice site of intron 3, exon 3 was spliced with GYP.Mur and GYP.Bun, but not with GYPB (Fig. 3). Exchange of this particular nucleotide between GYP.Mur and GYPB (GYP.Mur: 905G>T; GYPB: 905T>G) critically reverse the
splicing of exon 3 (Fig. 4). So splicing of exon 3 in GYP.Mur follows the GU-AG rule, as previously predicted [10]. We also tested the other nucleotide variations between GYP.Mur and GYPB, which likely affect the numbers and distributions of ESE binding motifs (Fig. S1 and S5). Compared to MURe2e4-905G>T with disruption of the 50 splice site, all the other nucleotide variations resulted in slight or no significant impacts on the splicing of glycophorin exon 3 (Fig. 5). This could be due to that the ESE motifs in exon 3 of GYPB, despite fewer than in GYP.Mur, might be quite effective in facilitating pre-mRNA splicing (Fig. S1). From the HSF online program that predicts splicing events (Figs S1 and S2), ASF was predicted to interact differentially with pre-mRNA in exon 3 and affect the splicing efficiency. However, effective knock-down of endogenous ASF in HEK-293 cells did not alter the outcome of exon 3 splicing (Fig. S3). Conceivably, the splicing machinery was designed to be redundant functionally. This may allow expression tolerance for the numerous glycophorin variants in the MNS blood group system.
Acknowledgements We thank staff at Taiwan Mackay Memorial Hospital Blood Bank for serology support. This work was supported by grants from Taiwan Ministry of Science & Technology (MOST 103-2320-B-195-001-MY3) and National Health Research Institute (NHRI-EX10110122SI), and a Mackay Memorial Hospital Intramural Grant (MMH103-25).
Authors’ contribution K.H., C.C. and Y.L. designed the experiments and co-wrote the manuscript; Y.L., C.C., T.L. and C.Y. carried out the experiments; and C.Y., C.C. and T.L. performed data analyses.
References 1 Broadberry RE, Lin M: The incidence and significance of anti-”Mia” in Taiwan. Transfusion 1994; 34:349–352 2 Lin M, Broadberry RE: Immunohematology in Taiwan. Transfus Med Rev 1998; 12:56–72 3 Hsu K, Lin YC, Chao HP, et al.: Assessing the frequencies of GP.Mur (Mi.III) in several Southeast Asian
populations by PCR typing. Transfus Apher Sci 2013; 49:370–371 4 Palacajornsuk P, Nathalang O, Tantimavanich S, et al.: Detection of MNS hybrid molecules in the Thai population using PCR-SSP technique. Transfus Med 2007; 17:169–174 5 Lin M, Broadberry RE: An intravascular hemolytic transfusion reaction due
to anti-’Mi(a)’ in Taiwan. Vox Sang 1994; 67:320 6 Heathcote DJ, Carroll TE, Flower RL: Sixty years of antibodies to MNS system hybrid glycophorins: what have we learned? Transfus Med Rev 2011; 25:111–124 7 Tippett P, Reid ME, Poole J, et al.: The Miltenberger subsystem: is it obsoles-
© 2015 International Society of Blood Transfusion Vox Sanguinis (2015) 108, 403–409
Alternative splicing for GYP.Mur (Mi.III) 409
8
9
10
11
cent? Transfus Med Rev 1992; 6:170– 182 Broadberry RE, Lin M: The distribution of the MiIII (Gp.Mur) phenotype among the population of Taiwan. Transfus Med 1996; 6:145–148 Blumenfeld OO, Huang CH: Molecular genetics of the glycophorin gene family, the antigens for MNSs blood groups: multiple gene rearrangements and modulation of splice site usage result in extensive diversification. Hum Mutat 1995; 6:199–209 Huang CH, Blumenfeld OO: Molecular genetics of human erythrocyte MiIII and MiVI glycophorins. Use of a pseudoexon in construction of two delta-alpha-delta hybrid genes resulting in antigenic diversification. J Biol Chem 1991; 266: 7248–7255 Hsu K, Lin YC, Chang YC, et al.: A direct blood polymerase chain reaction
12
13
14
15
approach for the determination of GP.Mur (Mi.III) and other Hil+ Miltenberger glycophorin variants. Transfusion 2013; 53:962–971 Desmet FO, Hamroun D, Lalande M, et al.: Human Splicing Finder: an online bioinformatics tool to predict splicing signals. Nucleic Acids Res 2009; 37:e67 Cartegni L, Wang J, Zhu Z, et al.: ESEfinder: a web resource to identify exonic splicing enhancers. Nucleic Acids Res 2003; 31:3568–3571 Smith PJ, Zhang C, Wang J, et al.: An increased specificity score matrix for the prediction of SF2/ASF-specific exonic splicing enhancers. Hum Mol Genet 2006; 15:2490–2508 Fairbrother WG, Yeo GW, Yeh R, et al.: RESCUE-ESE identifies candidate exonic splicing enhancers in vertebrate exons. Nucleic Acids Res 2004; 32:W187–W190
16 Zhang XH, Kangsamaksin T, Chao MS, et al.: Exon inclusion is dependent on predictable exonic splicing enhancers. Mol Cell Biol 2005; 25:7323–7332 17 Kim DJ, Oh B, Kim YY: Splicing factor ASF/SF2 and transcription factor PPAR-gamma cooperate to directly regulate transcription of uncoupling protein-3. Biochem Biophys Res Commun 2009; 378:877– 882 18 Li X, Manley JL: Inactivation of the SR protein splicing factor ASF/SF2 results in genomic instability. Cell 2005; 122:365–378 19 Zuo P, Manley JL: Functional domains of the human splicing factor ASF/SF2. EMBO J 1993; 12:4727–4737 20 Huang CH, Johe KK, Seifter S, et al.: Biochemistry and molecular biology of MNSs blood group antigens. Baillieres Clin Haematol 1991; 4:821–848
Supporting Information Additional Supporting Information may be found in the online version of this article: Figures 1 & 2 Prediction of splicing signals in regions near exon 3 of GYPA, GYPB, and GYP.Mur, by an online bioinformatics program—the Human Splicing Finder (version 2.4.1). Figure 3 Knockdown of ASF did not affect splicing of MURe2e4.
© 2015 International Society of Blood Transfusion Vox Sanguinis (2015) 108, 403–409
ORIGINAL PAPER
Dissecting alternative splicing in the formation of Miltenberger glycophorin subtype III (GYP.Mur) K. Hsu, C.-C. Yao, Y.-C. Lin, C.-L. Chang & T.-Y. Lee Mackay Memorial Hospital Transfusion Medicine Laboratory & Blood Bank, Tamsui, Taiwan
Background and Objectives Miltenberger subtype III (Mi.III, GP.Mur) is one of the most important red cell phenotypes in the fields of transfusion in South-East Asia. GP.Mur is believed to evolve from homologous gene recombination events between glycophorin A (GYPA) and glycophorin B (GYPB). GYP.Mur differs from GYPB in only seven nucleotides dispersed near the region of 30 exon 3 of GYP.Mur. The goal of this study was to dissect how these nucleotide variants affected splicing of exon 3. Materials and Methods We first designed two minigene constructs: one containing GYP.Mur from exon 2 to exon 4 and the other containing GYPB in the same region. To test how these nucleotide variations between GYP.Mur and GYPB affected the splicing, a repertoire of the GYP.Mur-like minigene constructs with different point mutations were created. These minigene variants were evaluated for their abilities to induce splicing of exon 3 using a heterologous expression system. Results (1) GYP.Mur minigene expressed exons 2, 3 and 4, whereas GYPB minigene expressed only exon 2 and exon 4. (2) The single nucleotide alteration at the position of the 50 splice site of glycophorin intron 3 reversed the splicing decision. (3) The nucleotide variations between GYP.Mur and GYPB other than that at the 50 splice site showed very little or no effect on splicing of exon 3. Received: 8 April 2014, revised 14 November 2014, accepted 18 November 2014, published online 6 March 2015
Conclusion Splicing of the glycophorin B-A-B hybrids (GYP.Mur and GYP.BUN) and unsplicing of GYPB follow the GU-AG rule strictly. Key words: alternative splicing, exonic splicing enhancer, glycophorin B, GYP.Mur, Miltenberger type III, minigene.
Introduction Miltenberger red cell phenotypes in the MNS blood group system are generally rare, except Miltenberger type III (Mi.III, GP.Mur), with 2–10% occurrence frequencies in several South-East Asian (SEA) countries [1–4]. The epitopes on GP.Mur protein, such as ‘Mia’ and Mur, could elicit strong allo-immune responses and are associated with intravascular haemolytic transfusion reactions and haemolytic diseases of foetus and newborn (HDFN) [5–7]. Therefore, GP.Mur is considered one of the most important RBC phenotypes in transfusion medicine in SEA, and Correspondence: Kate Hsu, PhD, 45 Min-Sheng Rd, Research Building 616, Tamsui 251, Taiwan E-mail: [email protected]
‘Mia’+ red cells have been included in routine antibody screening panels for ~20 years in Taiwan [8]. The rise of Miltenberger phenotypes is believed to evolve from gene crossover events among glycophorin A (GYPA), glycophorin B (GYPB) and glycophorin E (GYPE) [9]. Resolved by Huang et al. [10] in 1991, GYP.Mur, the Mi.III-specific gene encoding GP.Mur, bears the structure of glycophorin B-A-B hybrid. The genomic sequence of GYP.Mur is almost identical to that of GYPB, except for seven nucleotides dispersed in or near the 30 end of exon 3 (Fig. 1) [11]. The protein sequences of GP.Mur and glycophorin B (GPB) are almost identical, except for that GP.Mur, like glycophorin A (GPA), expresses its exon 3-encoded residues and GPB does not. It was predicted by sequence comparison that an emergence of the 50 splice site in this region due to gene crossover may enable
403
404 K. Hsu et al.
Fig. 1 GYP.Mur (JN_201202) and GYP.Bun alignment to GYPA (NG_007470.3) and GYPB (JN_201201) in the region of exons 2–4. Pseudoexon 3 of GYPB is labelled with dark grey colour; exon 3 of GYPA is with light grey colour. The first portion of exon 3 of GYP.Mur is coloured with dark grey, and the rest of exon 3 is with light grey, as GYP.Mur is a hybrid form of GYPB and GYPA in this region. GYP.Bun shares a similar glycophorin B-A-B hybrid structure as GYP.Mur. The 30 and 50 splice sites adjacent to exon 3 are boxed. The nucleotide differences between GYPB and GYP.Mur are marked with asterisks [11].
splicing of exon 3 in GYP.Mur. The corresponding sequence in GYPB lacking a functional 50 splice site is not subjected to pre-mRNA splicing and is thus termed ‘pseudoexon 30 . In this report, we created ‘minigenes’ of GYP.Mur and GYPB and tested their splicing or unsplicing of exon 3 using a heterologous expression system. Our experimental results indicate that conservation of the 50 splice site immediately following exon 3 is essential for the splicing event to take place. On the other hand, the other nucleotide variations between GYP.Mur and GYPB appear to have very small or no effects on splicing of exon 3.
Materials and methods Construction of minigene plasmids and mutant forms The genomic sequence of GYP.Bun (Miltenberger antigen subtype VI, Mi.VI) was determined from a serologically identified GP.Bun+ blood sample. Our GYP.Bun sequence (KM655769) is identical to the previously published GYP.Bun sequence (M60710.1) in exon 3 [10], except few sporadic variations between the two sequences in their intron regions. The genomic fragments of GYPB
(Accession number JN_201201), GYP.Mur (JN_201202) and GYP.Bun from exon 2 to exon 4 were subcloned into pN2-EGFP (Clontech, Mountain, CA, USA). The constructed minigene-GFP fusion plasmids for GYPB, GYP.Mur and GYP.Bun were named pGYPBe2e4, pMURe2e4 and pBUNe2e4, respectively. Selected single point mutations in GYPBe2e4 and MURe2e4 were created using QuikChange Site-Directed Mutagenesis kit (Staratagene, La Jolla, CA, USA).
Spicing experiments with minigenes The day prior to transient transfection, HEK-293 cells were seeded in a 12-well cell culture plate at a density of 2 9 105 cells/well and maintained in DMEM supplemented with FBS and antibiotics. Minigene plasmids were individually transfected using T-Pro transfection reagent II (T-Pro Biotechnology, Taipei, Taiwan) for 4 h. Minigene splicing events in these cells were usually studied at 26– 44 h post-transfection. Splicing of (pseudo)exon 3 in these minigenes was assessed by the lengths of mRNA transcribed between exon 2 and exon 4, followed by sequencing validation. Total RNA was extracted from transfected HEK-293 cells using a total RNA mini kit (Viogene-BioTek, Taipei, © 2015 International Society of Blood Transfusion Vox Sanguinis (2015) 108, 403–409
Alternative splicing for GYP.Mur (Mi.III) 405
Taiwan) and reverse transcribed to cDNA with ReverTra Ace (TOYOBO, Osaka, Japan). PCRs were performing with the two pairs of primers (GB18064F & GFP151R; b-actin F & b-actin R). PCR products were then resolved on a 2% agarose gel. The relative quantities of RT-PCR products were analysed using Image J (NIH). Expression of GYPBe2e4 minigene completely skipped pseudoexon 3 and resulted in a 308-bp DNA fragment (lacking exon 3) from RT-PCR; MURe2e4 minigene yielded complete splicing of exon 3 and resulted in a 401-bp DNA fragment (containing exon 3). To determine the impacts of nucleotide variations on splicing of (pseudo)exon 3, the intensity of each PCR band was compared quantitatively with pGYPBe2e4 (set 0%) and pMURe2e4 (set 100%). The splicing experiments for each minigene construct were performed independently for 3–5 times.
Prediction of splicing factors To test how sequence variations could affect pre-mRNA splicing, the sequences of GYPA, GYPB and GYP.Mur were analysed using Human Splicing Finder, an online prediction program for splicing (http://www.umd.be/HSF/) [12].
Results The nucleotide sequences and protein sequences of many Miltenberger glycophorin hybrid variants were delineated in early 1990s [9]. It was predicted at that time that both glycophorin B-A-B hybrids (Mi.III & Mi.VI) express exon 3 like GYPA, because in their sequences, the 50 splice site immediately following exon 3 is retained [10]. There are
Fig. 2 An outline for the minigene experiments. The genomic sequences of GYPB and GYP.Mur between exon 2 and exon 4 were constructed into pN2-EGFP to create two minigene-GFP fusion plasmids – pGYPBe2e4 and pGYPMURe2e4 (pMURe2e4). These minigene-GFP fusion plasmids were then transfected into cultured cells, and their expressions were assessed by the appearance of green fluorescence. RT-PCR experiments ensued to test whether mRNA transcribed from these minigenes expressed exon 3 or not. A pair of the PCR primers were designed to amplify the sequence from exon 2 to exon 4. So from the sizes of the PCR products, whether exon 3 was spliced or skipped could be determined.
© 2015 International Society of Blood Transfusion Vox Sanguinis (2015) 108, 403–409
7 nucleotide differences between GYPB and GYP.Mur on or near exon 3 [11]. One of the nucleotide differences is located right at the 50 splice site of intron 3. GYPB lacks this splice site and is thus disqualified from splicing of exon 3 (Fig. 1). However, is this single nucleotide difference at the 50 splice site between GYPB and GYP.Mur [Mi.III] solely responsible for the expression of exon 3? Could nucleotide variations other than the one at the 50 splice site also influence the decision of exon 3 splicing? In this study, we designed minigenes of GYPB and GYP.Mur and examined their splicing patterns using a heterologous expression system (Fig. 2). As GYPB differs from GYP.Mur in or near (pseudo)exon 3, their genomic sequences from exon 2 to exon 4 were individually subcloned into the pN2-EGFP vector to create minigene-GFP fusion sequences. These two minigenes (GYPBe2e4 and MURe2e4) differ only in seven nucleotides (marked * in the sequence alignment in Fig. 1), as previously described [11]. To assess the splicing patterns of these minigenes, they were each transiently transfected into HEK-293 cells, and their mRNA expressions were evaluated by RT-PCR. Transfected HEK-293 cells expressed green fluorescence from reporter GFP (Fig. 3a), the N-terminus of which is fused with a fragment of glycophorin B (or GP.Mur). We observed these GFP fusions with GP.Mur/GPB fragments to be primarily localized in a punctate distribution intracellularly (Fig. 3a). Total RNA of these transfected cells was extracted for RT-PCR; cDNA was amplified between exon 2 and exon 4, and their PCR products were verified by DNA sequencing. As expected, GYPBe2e4 minigene, like glycophorin B, did not undergo splicing of exon 3 and only expressed exon 2 and exon 4 (represented by a band of 308 bp). In
406 K. Hsu et al.
(a)
(b)
(c)
contrast, MURe2e4 minigene, like GYP.Mur, expressed exons 2, 3 and 4 (represented by a band of 401 bp) (Fig. 3b,c). We also created a minigene for GYP.Bun (Mi.VI), another glycophorin B-A-B hybrid structurally very similar to GYP.Mur. The BUNe2e4 minigene also expressed exons 2, 3 and 4, as MURe2e4. We did not observe any exon 3-spliced form from the expression of GYPBe2e4 minigene, or any residual exon 3-unspliced form from MURe2e4 (Fig. 3c). Thus, the splicing or skipping of exon 3 entirely depends on the minigene sequences themselves and appears independent of the choice of the expression system. To test whether the different splicing patterns of exon 3 were primarily due to the single nucleotide difference at the 50 splice site of intron 3 between GYPBe2e4 and MURe2e4 (Fig. 3b), two plasmids containing single point mutations at the 50 splice site were created: pGYPBe2e4905T>G and pMURe2e4-905G>T. The point mutation for the former supplied a 50 splice site for GYPB, whereas the point mutation for the latter disrupted the 50 splice site of GYP.Mur to make it resemble GYPB. Our experiments showed that the single mutation (G>T) at the 50 splice site of intron 3 of GYP.Mur prevented much splicing of exon 3 (Fig. 4). A reversal change at the same position on GYPB (T>G) enabled splicing of ‘pseudoexon 30 in GYPB, as shown in Fig. 4. These experimental data support that a single nucleotide difference at the 50 splice site of intron 3 primarily determined whether exon 3 should be spliced or skipped. Since the decision to splice or skip exon 3 for GYP.Mur/GYPB solely depends on their coding sequences (Fig. 3), do the other nucleotide differences between GYP.Mur and GYPB affect the decision of exon 3 splicing
Fig. 3 Pseudoexon 3 of GYPBe2e4 minigene was not spliced, whereas exon 3 of MURe2e4 and BUNe2e4 minigenes were spliced and ultimately expressed. (a) The expressions of GYPBe2e4 and MURe2e4 minigenes were validated by reporter GFP that emitted green fluorescence in transfected HEK-293 cells. (b) Splicing or skipping of exon 3 in GYPA, GYPB, GYP.Mur and GYP.Bun was shown. Light grey colour and dark grey colour indicate the sequences of GYPA and GYPB, respectively. A2/ B2 refers to exon 2 of GYPA/GYPB. Similarly, A3/B3 refers to exon 3 and A4/B4 refers to exon 4. The critical sequences at the 50 and 30 splice sites (GT & AG) are specified. (c) The RTPCR results showed that exon 3 was spliced and expressed with MURe2e4 and BUNe2e4 minigenes and not with GYPBe2e4 minigene.
Fig. 4 Splicing of exon 3 followed the GU-AG rule. These RT-PCR results showed that exon 3 was spliced and expressed from GYPBe2e4-905T>G, a single mutant of GYPBe2e4 minigene that contains a functional 50 splice site in intron 3. In contrast, exon 3 was unspliced with MURe2e4905G>T, a single mutant of MURe2e4 minigene with a disrupted 50 splice site in intron 3.
at all? For this, we employed an online website – Human Splicing Finder (HSF) – to find out whether the other sequence variations between GYP.Mur and GYPB might have a role in promoting or suppressing splicing of exon 3 [12]. From the prediction by HSF, GYP.Mur primarily differs from GYPB with more motifs for exon splicing enhancers (ESEs) (Fig. S1). We did not find any major differences in the prediction for exon splicing suppressors (ESSs) in GYP.Mur, GYPB and GYPA by HSF (Fig. S2). Additionally, GYPA and GYP.Mur share identical sequences in 30 exon 3 and thus have the same prediction for splicing. GYPA and GYP.Mur do differ in 50 exon 3; however, their splicing prediction score (scaled in octamers strength by HSF) appeared the same in 50 exon 3 (Figs S1 and S2). HSF includes three well-studied algorithms for ESE prediction: ESE finder [13, 14], RESCUE-ESE [15], and Zhang and Chasin’s criteria for PESE-octamers [16]. The three ESE prediction methods all identified more ESE © 2015 International Society of Blood Transfusion Vox Sanguinis (2015) 108, 403–409
Alternative splicing for GYP.Mur (Mi.III) 407
signals present in GYP.Mur than in GYPB in 30 exons, where their octamers strengths (splicing scales) differed the most. The ESE finder program further specified that these additional ESE signals found in GYP.Mur are binding motifs for SF2/ASF (Fig. S1). Five SF2/ASF binding motifs were predicted in 30 exon 3 of GYP.Mur, compared to only two SF2/ASF motifs predicted in the same region of GYPB. ASF (alternative splicing factor-1), also known as SF2 (pre-mRNA splicing factor 2), is a serine/arginine (S/R)-rich splicing factor required for pre-mRNA splicing [17, 18]. Others have shown that ASF/SF2 supports splicing in a concentration-dependent fashion [19]. Could more ASF/SF2 motifs in GYP.Mur exon 3 support splicing of exon 3? To test this, we knocked down endogenous ASF by siRNA (Fig. S3A) and repeated the minigene experiments. We, however, did not find the splicing of MURe2e4 to be affected by ASF knock-down (Fig. S3B). These data suggest either that ESEs are not the critical determinants for splicing of glycophorin exon 3, or that the current ESE/ESS prediction programs are still not robust enough. To further explore how the nucleotide variations between GYPB and GYP.Mur, other than that located on
the 50 splice site of intron 3, could affect splicing of exon 3, three additional point mutants based on pMURe2e4 were created: MURe2e4-AAA, MURe2e4-CC and MURe2e4-791A>T (Fig. 5a). For the first two mutants (MURe2e4-AAA and MURe2e4-CC), nucleotide variants between GYP.Mur and GYPB were selectively reverted to resemble GYPB. The other mutant MURe2e4-791A>T has an A to T mutation at the predicted branch point in intron 2. The expressions of these point mutant minigenes were compared to the expression of MURe2e4 minigene, which was set 100% and indicative of complete exon 3 splicing. The expression of GBe2e4 was set 0%, indicative of complete exon 3 skipping. MURe2e4-905G>T (the GYP.Mur minigene with a disabled 50 splice site of intron 3) and GYPBe2e4-905T>G (the GYPB minigene with an enabled 50 splice site) were included in this experiment for comparison with the other point mutants of MURe2e4. With a single point mutation at the 50 splice site in pMURe2e4-905G>T, the expression of GP.Mur exon 3 dropped from 100% to 18 – 10% (Fig. 5b). In pMURe2e4-AAA, three nucleotides were reverted to the corresponding GYPB sequence (g.[895G>A; 898G>A; 902C>A]), and these changes lowered the expression of
(a)
(b)
Fig. 5 Selective single mutations in MURe2e4 minigene affected splicing and expression of exon 3 differentially. (a) Different symbols on the sequence alignment refer to different mutant constructs. Filled inverse triangle, filled diamond and hollow circle represented MURe2e4-based minigenes with different point mutations – MURe2e4-791A>T, MURe2e4-CC and MURe2e4-AAA – respectively. The positions of these point mutations are numbered starting from exon 2 of GYP.Mur and listed as shown. (b) The expressions of minigene point mutants in HEK-293 cells were assessed by RT-PCR and quantitatively analysed using Image J. The PCR band intensity at 401 bp (with exon 3 expression) was individually normalized from each experimental set. The expression levels of exon 3 from these MURe2e4 point mutants were then compared to that from MURe2e4 minigene (set 100%). The expression levels from GYPBe2e4 were set 0% for relative quantification. The relative expression of exon 3 (in percentile) from GYPBe2e4 and MURe2e4 mutant minigenes were calculated accordingly and summarized here from five independent experiments. The relative expressions of exon 3 from BUNe2e4 minigene were summarized from three independent experiments. Shown mean – SEM.
© 2015 International Society of Blood Transfusion Vox Sanguinis (2015) 108, 403–409
408 K. Hsu et al.
exon 3 from 100% to 83 – 11% (Fig. 5b). We also created a double point mutant MURe2e4-CC (g.[875G>C; 884A>C]), which disrupts a ESE motif predicted by the RESCUE-ESE program [15]. In MURe2e4-CC, the two nucleotide variants were reverted to resemble GYPB. This double mutant MURe2e4-CC reduced the expression of exon 3 from 100% to 80 – 10% (Fig. 5b). The minigene BUNe2e4, which resembles MURe2e4-CC at g.[875G>C], also only reduced the expression of exon 3 from 100% to 90 – 7% (Fig. 5b). Additionally, we created single nucleotide change at the predicted branch point of intron 2 of GYP.Mur (MURe2e4-791A>T). This branch point mutant did not affect splicing of exon 3 at all (Fig. 5b). In sum, only the point mutations at the 50 splice site play a determining role in splicing of glycophorin exon 3; all the other mutations tested showed minor or no statistically significant effects on the splicing of exon 3.
Discussion By constructing and experimenting with GYPB/GYP.Mur/ GYP.Bun minigenes that include exon 2 to exon 4, we identified the key determinant for exon 3 splicing to be at the 50 splice site of intron 3. These minigenes alone were sufficient for pre-mRNA splicing and expression in a typical mammalian HEK-293 expression system (Fig. 3). Thus, whether splicing of (pseudo)exon 3 of GYPB/GYP.Mur/GYP.Bun should take place or not depends entirely on the coding sequence. Most of the genes encoding for Miltenberger glycophorin hybrids were resolved about 20 years ago [9, 20]. GYP.Mur and GYP.Bun both have the glycophorin B-A-B gene structure and incorporate a sequence for the 50 splice site of intron 3 from GYPA [10]. By comparison, the corresponding sequence of GYPB lacks the 50 splice site. Because of the single nucleotide variation at the 50 splice site of intron 3, exon 3 was spliced with GYP.Mur and GYP.Bun, but not with GYPB (Fig. 3). Exchange of this particular nucleotide between GYP.Mur and GYPB (GYP.Mur: 905G>T; GYPB: 905T>G) critically reverse the
splicing of exon 3 (Fig. 4). So splicing of exon 3 in GYP.Mur follows the GU-AG rule, as previously predicted [10]. We also tested the other nucleotide variations between GYP.Mur and GYPB, which likely affect the numbers and distributions of ESE binding motifs (Fig. S1 and S5). Compared to MURe2e4-905G>T with disruption of the 50 splice site, all the other nucleotide variations resulted in slight or no significant impacts on the splicing of glycophorin exon 3 (Fig. 5). This could be due to that the ESE motifs in exon 3 of GYPB, despite fewer than in GYP.Mur, might be quite effective in facilitating pre-mRNA splicing (Fig. S1). From the HSF online program that predicts splicing events (Figs S1 and S2), ASF was predicted to interact differentially with pre-mRNA in exon 3 and affect the splicing efficiency. However, effective knock-down of endogenous ASF in HEK-293 cells did not alter the outcome of exon 3 splicing (Fig. S3). Conceivably, the splicing machinery was designed to be redundant functionally. This may allow expression tolerance for the numerous glycophorin variants in the MNS blood group system.
Acknowledgements We thank staff at Taiwan Mackay Memorial Hospital Blood Bank for serology support. This work was supported by grants from Taiwan Ministry of Science & Technology (MOST 103-2320-B-195-001-MY3) and National Health Research Institute (NHRI-EX10110122SI), and a Mackay Memorial Hospital Intramural Grant (MMH103-25).
Authors’ contribution K.H., C.C. and Y.L. designed the experiments and co-wrote the manuscript; Y.L., C.C., T.L. and C.Y. carried out the experiments; and C.Y., C.C. and T.L. performed data analyses.
References 1 Broadberry RE, Lin M: The incidence and significance of anti-”Mia” in Taiwan. Transfusion 1994; 34:349–352 2 Lin M, Broadberry RE: Immunohematology in Taiwan. Transfus Med Rev 1998; 12:56–72 3 Hsu K, Lin YC, Chao HP, et al.: Assessing the frequencies of GP.Mur (Mi.III) in several Southeast Asian
populations by PCR typing. Transfus Apher Sci 2013; 49:370–371 4 Palacajornsuk P, Nathalang O, Tantimavanich S, et al.: Detection of MNS hybrid molecules in the Thai population using PCR-SSP technique. Transfus Med 2007; 17:169–174 5 Lin M, Broadberry RE: An intravascular hemolytic transfusion reaction due
to anti-’Mi(a)’ in Taiwan. Vox Sang 1994; 67:320 6 Heathcote DJ, Carroll TE, Flower RL: Sixty years of antibodies to MNS system hybrid glycophorins: what have we learned? Transfus Med Rev 2011; 25:111–124 7 Tippett P, Reid ME, Poole J, et al.: The Miltenberger subsystem: is it obsoles-
© 2015 International Society of Blood Transfusion Vox Sanguinis (2015) 108, 403–409
Alternative splicing for GYP.Mur (Mi.III) 409
8
9
10
11
cent? Transfus Med Rev 1992; 6:170– 182 Broadberry RE, Lin M: The distribution of the MiIII (Gp.Mur) phenotype among the population of Taiwan. Transfus Med 1996; 6:145–148 Blumenfeld OO, Huang CH: Molecular genetics of the glycophorin gene family, the antigens for MNSs blood groups: multiple gene rearrangements and modulation of splice site usage result in extensive diversification. Hum Mutat 1995; 6:199–209 Huang CH, Blumenfeld OO: Molecular genetics of human erythrocyte MiIII and MiVI glycophorins. Use of a pseudoexon in construction of two delta-alpha-delta hybrid genes resulting in antigenic diversification. J Biol Chem 1991; 266: 7248–7255 Hsu K, Lin YC, Chang YC, et al.: A direct blood polymerase chain reaction
12
13
14
15
approach for the determination of GP.Mur (Mi.III) and other Hil+ Miltenberger glycophorin variants. Transfusion 2013; 53:962–971 Desmet FO, Hamroun D, Lalande M, et al.: Human Splicing Finder: an online bioinformatics tool to predict splicing signals. Nucleic Acids Res 2009; 37:e67 Cartegni L, Wang J, Zhu Z, et al.: ESEfinder: a web resource to identify exonic splicing enhancers. Nucleic Acids Res 2003; 31:3568–3571 Smith PJ, Zhang C, Wang J, et al.: An increased specificity score matrix for the prediction of SF2/ASF-specific exonic splicing enhancers. Hum Mol Genet 2006; 15:2490–2508 Fairbrother WG, Yeo GW, Yeh R, et al.: RESCUE-ESE identifies candidate exonic splicing enhancers in vertebrate exons. Nucleic Acids Res 2004; 32:W187–W190
16 Zhang XH, Kangsamaksin T, Chao MS, et al.: Exon inclusion is dependent on predictable exonic splicing enhancers. Mol Cell Biol 2005; 25:7323–7332 17 Kim DJ, Oh B, Kim YY: Splicing factor ASF/SF2 and transcription factor PPAR-gamma cooperate to directly regulate transcription of uncoupling protein-3. Biochem Biophys Res Commun 2009; 378:877– 882 18 Li X, Manley JL: Inactivation of the SR protein splicing factor ASF/SF2 results in genomic instability. Cell 2005; 122:365–378 19 Zuo P, Manley JL: Functional domains of the human splicing factor ASF/SF2. EMBO J 1993; 12:4727–4737 20 Huang CH, Johe KK, Seifter S, et al.: Biochemistry and molecular biology of MNSs blood group antigens. Baillieres Clin Haematol 1991; 4:821–848
Supporting Information Additional Supporting Information may be found in the online version of this article: Figures 1 & 2 Prediction of splicing signals in regions near exon 3 of GYPA, GYPB, and GYP.Mur, by an online bioinformatics program—the Human Splicing Finder (version 2.4.1). Figure 3 Knockdown of ASF did not affect splicing of MURe2e4.
© 2015 International Society of Blood Transfusion Vox Sanguinis (2015) 108, 403–409
