Intragenic Duplication and Divergence in the Spectrin Superfamily of

Intragenic Duplication and Divergence in the Spectrin Superfamily of Proteins

Graham H. Thomas,*? E. Claire Newbern,+ Carol C. Korte,* Mark A. Bales,*l Spencer V. Muse, *lc2 Andrew G. Clark, * and Daniel P. Kiehart$ *Department of Biology and TDepartment of Biochemistry and Molecular Biology, The Pennsylvania State University; and $Department of Cell Biology, Duke University Medical Center

Many structural, signaling, and adhesion molecules contain tandemly repeated amino acid motifs. The a-actinin/ spectrin/dystrophin superfamily of F-actin-crosslinking proteins contains an array of triple o-helical motifs (spectrin repeats). We present here the complete sequence of the novel P-spectrin isoform Pu,,,,,-spectrin (Pn). The sequence of PH supports the origin of (Y- and P-spectrins from a common ancestor, and we present a novel model for the origin of the spectrins from a homodimeric actin-crosslinking precursor. The pattern of similarity between the spectrin repeat units indicates that they have evolved by a series of nested, nonuniform duplications. Furthermore, the spectrins and dystrophins clearly have common ancestry, yet the repeat unit is of a different length in each family. Together, these observations suggest a dynamic period of increase in repeat number accompanied by homogenization within each array by concerted evolution. However, today, there is greater similarity of homologous repeats between species than there is across repeats within species, suggesting that concerted evolution ceased some time before the arthropod/vertebrate split. We propose a two-phase model for the evolution of the spectrin repeat arrays in which an initial phase of concerted evolution is subsequently retarded as each new protein becomes constrained to a specific length and the repeats diverge at the DNA level. This evolutionary model has general applicability to the origins of the many other proteins that have tandemly repeated motifs.

Introduction

The spectrin-based membrane skeleton is a ubiquitous cellular structure that has been implicated in a number of fundamental cellular processes. These include the maintenance of cell shape and integrity, the generation and/or maintenance of cell polarity, and the formation of cell adhesion complexes (Bennett and Gilli- gan 1993; Drubin and Nelson 1996). The utilization of different spectrin isoforms is probably a key step in cre- ating biochemical differences between different membrane domains. In vertebrates, several spectrin isoforms are known. Diversity at the protein level is encoded by multiple genes, by alternatively spliced mRNAs from individual genes, and by the combinatorial association of (x- and P-spectrins (Luna and Hitt 1992; Bennett and Gilligan 1993). This diversity of spectrin isoforms and the fact that these genes are often developmentally regulated is consistent with the view that variability in the function of the membrane skeleton in diverse tissues is due, at least in part, to the particular spectrin isoform that is expressed.

Spectrins are elongated proteins that form tetramers from two (x and two l3 subunits. Such tetramers crosslink F-actin via actin-binding domains on the P-spectrin subunits. The bulk of the mass and the ropelike nature of the spectrins comes from a short -106-amino-acid re-

l Present address: Department of Cell Biology and Anatomy, Col- legc of Medicine, University of Arizona.

* Present address: Division of Biological Sciences, University of Missouri.

Key words: a-actinin, spectrin, dystrophin, membrane skeleton, concerted evolution, repetitive DNA.

Address for correspondence and reprints: Graham H. Thomas, De- partment of Biology The Pennsylvania State University, 208 Erwin W. Mueller Laboratory, University Park, Pennsylvania 16802. E-mail: [email protected].

Mol. Biol. Evol. 14(12):1285-1295. 1997 0 1997 by the Society for Molecular Biology and Evolution. ISSN: 0737-4038

peat (spectrin repeat) that forms a triple o-helical structure (Speicher and Marchesi 1984; Yan et al. 1993). Spectrin repeats are not identical, but follow consensus rules for forming the three cx-helices that fold to form a short triple a-helical unit (Speicher and Marchesi 1984; Yan et al. 1993), and specific regions of the repeat array have evolved specific functions. These include binding to the protein ankyrin (Kennedy et al. 1991), an adapter between some integral membrane proteins and the spectrin network (Bennett 1992), and the ar/@spectrin di- merization nucleation site (Veil and Branton 1994).

Drosophila has three known spectrin genes encoding an o-spectrin and two isoforms of P-spectrin. One l3 isoform has the conventional @spectrin-like structure consisting of an amino-terminal (N-terminal) actin-binding domain followed by 17 spectrin repeats and a car- boxy-terminal (C-terminal) nonrepetitive domain containing a pleckstrin homology (PH) domain (Byers et al. 1992; Musacchio et al. 1993). The second l3 isoform, P neavy-spectrin (Pu), is unusually large for a P-spectrin (470 kDa) but dimerizes with the same ar-spectrin as the conventional l3 isoform (Dubreuil et al. 1990) and typ- ically has a polarized distribution in epithelial cells (Thomas and Kiehart 1994; Lee et al. 1997). Thus, in Drosophila, the combinatorial association of two p- spectrin isoforms with one o-spectrin to produce two different membrane skeletons provides a simple model system to investigate the differences between spectrin isoforms in establishing and/or maintaining membrane subdomains with specific characteristics.

Analysis of the amino acid sequences of a-actinin, spectrins, and dystrophin proteins has suggested that all three protein families arose from a single common ancestral protein that was a-actinin-like (Dubreuil 199 1; Byers et al. 1992). Specifically, cx-actinin has an N-terminus resembling that of P-spectrins and dystrophins (and that of other actin-crosslinking proteins [Dubreuil

1285

Dow

nloaded from https://academ

ic.oup.com/m

be/article/14/12/1285/1272991 by guest on 09 Decem

ber 2021

1286 Thomas et al.

a-actinin

dystrophin

I ..t

0$-spectrin

C&j-spectrin

**

key - actin binding domain mm EF-hand containing domain

0 spectrin repeats V SH3 domain

5/ cc/p head-to-head interaction site e PH domain

FIG. I.-Schematic representation of the members of the spectrin superfamily of proteins in their multimeric actin-crosslinking conformation. In each multimer, the amino terminus is to the left on the top and to the right on the bottom. Repeats with specific functions or differing lengths are not identified. o-Actinin and dystrophin are shown as homodimers; a/p and CX/& spectrins are heterotetramers. &, dimerizes with the same o-spectrin as the p isoform (Dubreuil et al. 1990), and its sequence is reported in this paper.

1991; Hartwig and Kwiatkowski 19911; fig. l), a short repeated a-helical motif common to the whole family, and Ca2+-binding EF-hands at the C-terminus related to those of a-spectrins and dystrophins. The spectrin repeats are reiterated a distinct number of times in each protein, resulting in a characteristic actin-crosslinking distance. cx-Actinin has 4 repeats, B-spectrin 17, a-spectrin 20, dystrophin 24, and Bn 30 (fig. 1). a-Actinin forms a homodimer, while 01- and B-spectrins form (x2/ B, (or 02/BH2) heterotetramers to crosslink F-actin. Based on homology, dystrophin may well have been a homodimer in the past, while the native oligomeric state of dystrophin today remains controversial (Rybakova, Amann, and Ervasti 1996). These similarities have led to the speculation that the lengths we see today arose by intragenic duplication of repeats in an cx-actinin-like ancestor and that (x- and B-spectrins are derived from a longer homodimeric precursor (Dubreuil et al. 1989; Byers et al. 1992).

We report here the complete sequence of Bn and extend the evolutionary analysis of the o-actinin/spectrin/dystrophin superfamily of proteins. Moreover, we provide direct evidence for the hypotheses that CY- and B-spectrins had a common ancestor by analyzing the Src homology 3 (SH3) motif, which is found in both Bn and cx-spectrin. In addition, we present a novel model for the evolutionary origin of these two proteins from their most recent common ancestor.

Spectrin repeat arrays presumably arose through intragenic duplication events from some archetypal repeat. Such repetitive DNA sequences are expected to exhibit concerted evolution; they should tend to remain homo- geneous in sequence due to the action of sequence exchanges between repeats resulting from gene conversion

and unequal crossing over events (Zimmer et al. 1980). However, sequences of the spectrin repeats today have diverged, and we have previously shown that there has been no sequence exchange among the repeats in the arrays since the divergence of arthropods and vertebrates some 500 MYA (Muse, Clark, and Thomas 1997). Here, we test the hypothesis that concerted evolution occurred prior to this date by analyzing the relationships between the o-spectrin, B-spectrin, and Bn repeat sequences. We find evidence for a dynamic phase of concerted evolution during which the evolution of the repeated array was dominated by the behavior of the underlying repeated DNA. This dynamic phase gave way to a second stable phase prior to the arthropod/vertebrate divergence some 500 MYA. This two-phase evolutionary model has general applicability to the many other proteins containing tandemly repeated protein motifs.

Materials and Methods

The Bn mRNA is 13 kb in size (Dubreuil et al. 1990). Approximately 7 kb of this has been reported in previous publications (Dubreuil et al. 1990, Thomas and Kiehart 1994; GenBank accession numbers X53992 and UO7629). Using the C-terminal clone, pBH12 (Thomas and Kiehart 1994), as a probe to screen an embryonic cDNA library (Brown and Kafatos 1988), we recovered a larger cDNA clone, p3F1, which contains an addition- al 3 kb of coding sequence. The remaining sequences were isolated by PCR using the genomic clone X3-8 as a template and primers from flanking cDNA fragments. The 5 ’ untranslated sequences were isolated using 5 ‘- RACE PCR (Frohman 1990; Life Technologies, Gaith- ersburg, Md.). Nested deletions of clones (Erase-a-Base, Promega, Madison, Wis.) were sequenced using the Se- quenase kit (Amersham Life Sciences Inc., Arlington Heights, Ill.) or by automated cycle-sequencing in core facilities for DNA sequencing at the University of North Carolina (Chapel Hill, N.C.) and The Pennsylvania State University (University Park, Pa.). Synthetic oligonucleo- tides were used as necessary to obtain complete cover- age. Sequences derived from cloned PCR fragments were verified by sequencing a single strand on the genomic DNA template clones. The total length of the Bn mRNA was found to be 12,952 bp (with one small in- tron of 182 bp identified by sequence analysis in the genomic PCR fragment) and predicts a protein product of 4,097 amino acids (M, = 47 1,7 17).

Accession numbers for the sequences used in the evolutionary analysis presented here are given in table 1. Neighbor-joining trees were constructed using MEGA (Kumar, Tamwa, and Nei 1993) with Poisson correction to amino acid distance.

Results and Discussion The Complete Sequence of Bn

We have completed the sequence of the atypical B- spectrin isoform Bn (fig. 1). Bn is distinguished from other B-spectrin isoforms in that it contains an SH3 domain (Musacchio, Wilmanns, and Saraste 1994), 30 rather than 17 spectrin repeats, and a larger C-terminal

Dow


ic.oup.com/m


ber 2021

Concerted Evolution in the Spectrin Superfamily 1287

Table 1 Database Accession Numbers and References for Sequences Used in this Paper

Sequence

Spectrins Drosophila CY ............... Drosophila p ............... Drosophila PH .............. Human a1. ................. Human B&I ............... Human a11 ................. Human B&311 .............. Mouse &#I. ............... Mouse &@I1 ...............

Chicken OL ................... Other proteins

Dictyostelium a-actinin ....... Drosophila a-actinin ......... Chicken cr-actinin. ........... Human a-actinin ............ Chicken dystrophin .......... Human dystrophin ........... Mouse dystrophin. ........... Dictyostelium ABPl20. ....... Human ABP280. ............ Schistosoma fimbrin. ......... Chicken fimbrin ............. Dictyostelium fimbrin. ........

Accession Number Citation

M26400 Dubreuil et al. (1989) M92288 Byers et al. (1992) X53992, U07629, AF022656 Dubreuil et al. (1990); Thomas and Kiehart (1994); this paper M61877, JO5244 Sahr et al. (1990) 505500, M37884 Winkelman et al. (199Oa, 1990b) JO5243 Moon and McMahon (1990) M96803 Hu, Masayo, and Bennett (1992) S66283 Bloom, Birkenmeier, and Barker (1993) M74773, S59200, U73 17 1 Ma et al. (1993); Muse, Clark, and Thomas (1997) x14519, x13701 Wasenius et al. (1989)

YO0689 Noegel, Witke, and M. (1987) x5 1753 Fyrberg et al. (1990) M74143 Baron et al. (1991) M95 178, M3 1300 Youssoufian, McAfee, and Kwiatkowski (1990) Xl3369 Lemaire, Heilig, and Mandel (1988) M18533, M17154, Ml8026 Koenig, Monaco, and Kunkel (1988) M68859 Bies et al. (1992) x15430 Noegel et al. (1989) X53416 Gorlin et al. (1990) L33405 Saber et al. (1996) X52562 de Arruda et al. ( 1990) L36202 Prassler et al. (1997)

domain containing the ubiquitous pleckstrin homology (PH) domain (Musacchio et al. 1993) (fig. 2). Through- out this paper, we use the segment nomenclature described in Byers et al. (1992) to describe regions of the spectrin molecules discussed. Thus, conventional p- spectrins have segments 1-19 (segment 1 is the actin- binding domain; segments 2-17 are 16 spectrin repeats; segment 18 is the part-repeat which participates in the cx@ head-to-head interaction; and segment 19 is the C- terminal domain; fig. 2B), cx-spectrins have segments O- 22 (segment 0 is the part-repeat which participates in the o/p head-to-head interaction; segments l-9 and 1 l- 21 are 20 spectrin repeats; segment 10 is an SH3 domain embedded in the spectrin repeat of segment 9; and segment 22 is the Ca2+-binding C-terminal domain). Pn has segments l-33 (segment 1 is the actin-binding domain; segments 2-6 and 8-31 are 29 spectrin repeats; segment 7 is an SH3 domain embedded in the spectrin repeat of segment 6; segment 32 is a part-repeat predicted to par- ticipate in the CX/~~ head-to-head interaction; and segment 33 is the C-terminal domain; fig. 2). A brief description of the sequence features found in the Bn sequence is given in the legend of figure 2.

A major feature found in conventional P-spectrin isoforms is a binding site for the protein ankyrin (Ben- nett 1992), which links the spectrin membrane skeleton to a variety of integral membrane proteins. Recent evidence from immunofluorescent stainings has suggested that ankyrin colocalizes with P-spectrin but not with Pn, raising the possibility that Pn does not bind to ankyrin (Lee et al. 1997). Conventional P-spectrins bind to ankyrin in a region that overlaps segments 15/16 (Kennedy et al. 1991). Two regions of the Pn protein are poten- tially homologous to this region. Since PH and P-spectrins appear to be colinear over the initial 17 spectrin repeats (see fig. 7) one candidate region is Pn segments

16/17 (see fig. 2B). The second possible region is Pn segments 29/30, since this is part of a duplicated region originating from a region that included segments 16/17 (see below and fig. 7). While there is extremely good conservation of segments 15 and 16 in all conventional P-spectrins, Bn is quite divergent at both the segment 16/17 and 29/30 regions (fig. 3), suggesting that the primary structure of this region is not capable of folding to form an ankyrin-binding site.

The Evolutionary Relationship of Bn to other P-Spectrins

The N-terminal actin-binding domain of PH is also found in P-spectrin, a-actinin, and dystrophin (fig. 1) as well as other actin-crosslinking proteins (Dubreuil 1991; Hartwig and Kwiatkowski 1991). The proportion of identical amino acids in the actin-binding domain of the most divergent pair (human Po-spectrin vs. human dystrophin) is 0.484, indicating that the actin-binding domain is sufficiently conserved that interprotein distance estimates can be made reliably. Pairwise comparison of the actin-binding domain of PH with those of these other actin-crosslinking proteins suggests that PH defines a new and distinct class of P-spectrins that diverged from a-actinin at about the same time as conventional P-spectrins and dystrophins (fig. 4). The pairwise distances among the Drosophila genes’ actin-binding domains, expressed as the number of amino acid substitutions per site are: cx-actinin-P-spectrin, 0.671 + 0.060; P-spectrin-pa, 0.594 + 0.055; and o-actinin+n, 0.811 + 0.068. Examination of the bootstrap confidence in the nodes of the neighbor-joining tree constructed from the actin-binding domains shows that none of the three groupings ((cx-actinin, P-spectrin), Pn), (a-actinin, (p- spectrin, Pn)), or ((a-actinin, Pn), g-spectrin), is found to be significantly more likely, reflecting the ancient or-

Dow


ic.oup.com/m


ber 2021

1288 Thomas et al.

A 6 #AA-*AA PHeovy-swtr 1 n

01 l-100 MTQRDGlIKFENERIKTL~EERLHlOKKTFTKWNNSFLIKAKNEVEDLFTDLADGIKLLKLLEllSSEKLGKPNSGRNRVHKlENVNKSLAFLHTKVRLE 01 101-200 SlGAEDlVDGNPRLlLGLIWTl lLRFOIOEIEIDVDEENESSEKRSAKDALLLWCORKTHGYPGVNlTDFTNSWRSGLGFNALlHSHRPDLFEYSTlVNS 01 201-274 KNSNLDNLNHAFDTAANELGIPSLLDAEDIDSARPDEKSILTYVASYYHTFARNKNEOKSGKRIANIVG~LNDA

02 275-394

03 395-497 04 498-603 05 604-709 06 710-783

07 784-883 07 884-944

06 945-976 08 977-1078 09 1079-1184 10 1185-1289 11 1290-1395 12 1396-1499 13 1500-1605 14 1606-1711 15 1712-1817 16 1818-1923 17 1924-2028 18 2029-2134 19 2135-2240 20 2241-2350 21 2351-2458 22 2459-2565 23 2566-2672 24 2673-2776 25 2777-2881 26 2882-2987 27 2988-3093 28 3094-3199 29 3200-3305 30 3306-3411 31 3412-3517 32 3518-3591

33 3592-3691 33 3692-3791 33 3792-3891 33 3892-3991 33 3992-4097

B

~00000000000~~~ ~00d00000000_ 1 2 3 15 16 1 23 68 16 17 oooooooooo~g.2 29 30

FIG. 2.-The complete amino acid sequence of PHeavy- spectrin (Pn). A, The protein is broken down into segments according to the nomenclature of Byers et al. (1992). #S-segment number; #AA#AA-amino acid range shown on corresponding line. The repetitive segments (2- 6, 8-32) are aligned relative to one another (except for the eight-amino-acid extension at the C-terminus of segment 2, which is wrapped for convenience). Since insertions tend to occur between the a-helices, the appearance of three columns reflects the locations of helices a, b, and c (left to right). The alignment of the PH repeats shown here is that used to provide evidence of block duplications in figure 7. Briefly, the unique (nonrepeat) features present in the fin sequence are: segment l-conserved actin-binding domain. Pn binds actin (Dubreuil et al. 1990); segment 2-probable dimer nucleation site resembling that of P-spectrins (Dubreuil et al. 1990; Veil and Branton 1994); segment 7-Src homology 3 (SH3) domain inserted between the b and c helices of segment 6 (Dubreuil et al. 1990; Musacchio, Wilmanns, and Saraste 1994); segment 20- an unusual extension of four amino acids between helices a and b of unknown significance; segment 32-spectrin repeat truncated after the b helix and probably forming the head-to-head interaction site with a-spectrin during tetramer formation (Tse et al. 1990; Speicher et al. 1993); Segment 33-nonrepetitive C-terminal segment containing a pleckstrin homology (PH) domain and lysine-rich terminus characteristic of all nonerythroid P-spectrins (Musacchio et al. 1993; Thomas and Kiehart 1994; Veil and Branton 1996) . Also found within this domain is a short polyglutamine stretch (opa sequence; Wharton et al. 1985). Where these features comprise less than the entire segment they are underlined. The sequences surrounding the PH domain in the C-terminus are unique to PH. B, Schematic illustration of the differences between conventional p- spectrins (left) and Pn (right). Numbering illustrates the segment nomenclature. Domains are shaded as in figure 1. Also shaded are segments 2 and 3, indicating the location of the dimer nucleation site on the two proteins (Veil and Branton 1994); segments 15/16 on P-spectrin, indicating the known ankyrin-binding site on these isoforms (Kennedy et al. 1991); and the homologous repeats in Pn (16/17 and 29/30), which we do not believe to bind to ankyrin (see text for discussion). Note that, due to the segment nomenclature and the presence of the SH3 domain as segment 7 in PHI the repeat array numbering is increased by one relative to the number in P-spectrin after segment 6 (i.e., p l-6 = PH 1-6, but p 7-17 = Pn 8-18).

Dow


ic.oup.com/m


ber 2021

Ankyrin bind. c Fly &., 16/17 Flyp 29/30 Fly @75/16 Hum pG 15/16 M0u/3~15/16 Hum pR 1 S/16 Moup, 15/16

ons

Concerted Evolution in the Spectrin Superfai nily 1289

10 20 30 40 50 60

70 80 90 100 110 120 130 Ankyrinbind.cons.TLQR:HTAFEHDLD~LGVOVDDLD~VAARLDAAY~GDKADA*DNdEDEVLAAWD~LLDACAGRR: Fly pH 16/17 Flyf3 29/30 Fly f?l5/16 Hum pG 15/16 Mou pG 15/16 HumpR 15/16 Moup, 15/16

FIG. 3.-Alignment of ankyrin-binding regions in conventional P-spectrins with the homologous regions of PH. Abbreviations: Ankyrin bind. cons.-a majority consensus derived from the five conserved ankyrin-binding regions of the conventional P-spectrins; Fly---Drosophila melunogaster; Hum----Homo sapiens; Mou--Mus musculus. The segment numbers encompassing the sequences shown are indicated. Differences from this consensus in each line are shaded and illustrate the divergent nature of the Pi_, regions. The amino acids analyzed are: &-1902-2024 and 3278-3407 (segments 16/17 and 29/30 indicated in fig. 2); fly P-spectrin-1777-1906; human Po-spectrin-1776-1905, mouse Po-spectrin-17761905; human P,-spectrin-1768-1897; mouse P,-spectrin-1759-1888 (all segments 15/16). The l3n segments form a clear outgroup (verified by 100% bootstrap confidence in the relevant node of a neighbor-joining tree).

igin of l3u (fig. 4). Further corroborating this observation is the recent identification of a P,-like spectrin in Cae- norhabditis ekgans (J. Austin, personal communication) suggesting that this type of l3 chain is indeed old and will prove to be widespread. The coincident appearance of distinct lineages for cr-actinin, P-spectrin, /3n, and dystrophin suggests that closely spaced gene duplications gave rise to the precursors of these four gene fam-

A Common Origin for OL- and P-Spectrins? The Relationship of PH to cx-Spectrin

Since the N-terminus of p-spectrin resembles the N- terminus of cx-actinin and dystrophin, and the C-terminus of cx-spectrin resembles the C-terminus of these same two homodimeric proteins (fig. l), it has been postulated that there was a homodimeric spectrin precursor that gave rise to both the (x- and @spectrin lineages (Byers et al. 1992). The sequence of l3n is consistent with this suggestion. In addition to the actin-binding segment 1, & has a PH domain at its C-terminus (fig. 2), indicating that it shared a common ancestor with the @spectrin lineage. It also contains an SH3 domain (Dubreuil et al. 1990), heretofore a characteristic of ol-spectrins, supporting the idea that Pn shared a common ancestor with a-spectrin. The Pn SH3 domain tends to cluster with the cx-spectrin class of SH3 domains in trees, but SH3 domains give poor bootstrap values at many points in such analyses. However, it is possible to align these sequences based on the crystal structures of several such domains. The alignment presented in figure 5 adds the &_, SH3 domain to that of the eight major classes of SH3 domain aligned by Musacchio, Wilmanns, and Saraste (1994) in this way. The most conserved residues of the PH SH3 domain best match those of the cx-spectrin phylogenetic signature and suggest that the origin of the PH SH3 domain is the same as that of the cx-spectrins. This supports a common origin for (x- and @spectrins, while divergence, particularly in the RT-loop, indicates that they probably have different binding speci- ficity today (Musacchio, Wihnanns, and Saraste 1994).

ilies.

0.1

FIG. 4.-The relationship of the PH actin-binding domain to that in other actin-binding proteins. Shown is a neighbor-joining tree based on gamma-corrected proportions of amino acid differences (scale bar indicates number of substitutions per site). Numbers at the nodes of the tree are bootstrap confidence levels in percent. Spectrins are labeled as in figure 3. Abbreviations: Dct--Dictyostelium discoideum; Chk- chicken; Mou-mouse; Sch-Schistosomu munsoni; cwA-cx-actinin; Dy-dystrophin; Fim-fimbrin.

A Model for the Origin of (x- and /3-Spectrins from a Common Precursor

An obvious problem in the evolution of the spectrins from a common ancestor derived from an cx-actinin-like molecule is the transition from a homodimeric to

Dow


ic.oup.com/m


ber 2021

1290 Thomas et al.

la Hum KSRC GGVTTFVALYDYES....... RTETDLSFKKGERLQIVNNTE............... GDWWLAHSL.... STGOTGYIPSNYVAPSDS 1 b Mou LCK LQDNLVIALHSYEP....... SHDGDLGFEKGEQLRILEQS................ GEWWKAQSL.... TTGOEGFIPFNFVAKANS 1 c Fly ABL DDPQLFVALYDFQA....... GGENQLSLKKGEQVRILSYNK.............. SGEWCEAHSD.....SGNVGWVPSNYVTPLNS 2 Fly DRK PEEMLVQALYDFVP....... QESGELDFRRGDVITVTDRSD............... ENWWNGEIG...... NRKGIFPATYVTPYHS 3 Fly DLGl KRSLYVRALFDYDPNRDDGLPSRGLPFKHGDILHVTNASD................. DEWWQARRVLGDNEDEOIGIVPSKRRWERKM 4 Hum PIPS MPQRTVKALYDYKA....... KRSDELSFCRGALIHNVSKEP............... GGWWKGDYG..... TRIQQYFPSNYVEDIST 5 Yeast ABPl KENPWATAEYDYDA....... AEDNELTFVENDKIINIEFVD...............DDWWLGELE.... KDGSKGLFPSNYVSLGN. 6 BovpBSB PEGFOYRALYPFRR....... ERPEDLELLPGDVLVVSRAALOALGVAEGNERCPOSVGWMPGLNE....RTRORGDFPGTYVEFLGP 7 Yeast CC25 RPIGIVVAAYDFNYPIKKD..SSSQLLSVOQGETIYILNKNS............... SGWWDGLVIDDSNGKVNRGWFPONFGRPLRD 8 Fly a-SPEC TGKECVVALYDYTE....... KSPREVSMKKGDVLTLLNSNN...............KDWWKVEVN......DROGFVPAAYIKKIDA l&+pectrin TETKMLPHVKSLFP.......FEGQGMKMDKGEVMLLKSKTN...............DDWWCVRKD.....NGVEGFVPANYVREVEP Spectrin cons l * . . K k dD cvrkd ge A

FIG. 5.--Alignment of the Pn SH3 domain with representatives of the major classes of SH3 domain. Pn is the only P-spectrin with an SH3 domain. This alignment adds the Pn SH3 domain to that of Musacchio, Wilmanns, and Saraste (1994), which is based on several SH3 crystal structures (see text for discussion). Spectrin cons-p, residues matching the o-spectrin SH3 signature derived by Musacchio, Wilmanns, and Saraste are indicated-uppercase indicates a Pn residue matching an invariant position, lowercase indicates a PH residue matching a variable position, and asterisks indicate a failure to match.

a heterotetrameric actin crosslinker. Today, during tetramer formation, the head-to-head interaction between CX- and P-spectrins occurs by reconstruction of a complete spectrin repeat where two a-helices are provided by the final l3 chain repeat, and the third is provided by the N-terminal end of the o-chain (Tse et al. 1990; Speicher et al. 1993). This could arise by gene duplication followed by the precise loss of the N- and C- termini from OL- and P-spectrins, respectively, and the acquisition of a PH domain in the p lineage. This is a complex scenario. It could also arise via a simple “gene cleavage” event; however, this would not explain why all modern P-spectrin genes encode a C-terminal segment containing a PH domain beyond this site, nor would cx-spectrin necessarily be expressed immediately after such an event, since it would be separated from the promoter and other regulatory sequences that previously drove its expression.

Alternatively, the transformation from a homodi- merit ancestor to the modern heterotetrameric spectrins could have occurred via a considerably simpler molecular mechanism. We propose that a reciprocal translocation event occurred involving a PH domain encoding gene and a site between the b and c helices of what we now know as P-spectrin segment number 18 (fig. 6). This event would fuse the PH-containing domain onto the @spectrin ancestor after helix b and result in an (Y- spectrin ancestor whose repeat array begins with helix c, both of which are features of modern spectrins. (x- Spectrin expression would be maintained from the promoter of the second gene, and this event would simul- taneously generate a prototype for the head-to-head interaction of CY- and @spectrin. The new heterotetrameric spectrin would therefore posses the ability to crosslink F-actin over the same distance as its homodimeric ancestor immediately after the translocation. Such a simple model, requiring only a single, very plausible event, is very attractive.

Evidence for Concerted Evolution in the Spectrin Family of Proteins

Intragenic duplications that result in proteins with tandemly repeated motifs presumably arise due to the dynamics of repeated DNA. Both unequal crossing over and replication slippage are mechanisms that allow repeated arrays of identical or closely related DNA se-

quences to expand and contract in a process expected to result in concerted evolution (Zimmer et al. 1980), whereby such tandemly repeated sequences tend to re- tain more similarity than expected. Loss and duplication of repeats during array length changes, and exchanges between repeats through unequal crossing over and gene conversion, respectively, would promote the spread of sequence variants across the array and maintain similarity among repeats. Gene conversion events will result in homogenization whether they involve mispairing of al- lelic repeat arrays or out-of-phase pairing of arrays in different genes. Mechanisms such as unequal crossing over presumably gave rise to the dramatic differences in repeat array length seen in the o-actinin/dystrophin/ spectrin superfamily of proteins and would be expected to result in concerted evolution within these repeat arrays; however, the lengths of the four proteins in the spectrin superfamily have remained fixed since the divergence of arthropods and vertebrates some 500 MYA. In addition, our previous analysis has shown that, since this time, each of the spectrin repeats has been evolving independently with little or no sequence exchange between them (Muse, Clark, and Thomas 1997), suggesting that concerted evolution has been arrested. Whether the degree of sequence divergence among repeats is a cause or effect of the arrest is not possible to determine rigorously.

Traces of events that are consistent with intragenic duplication followed by sequence divergence can be seen in these proteins today. From comparisons of (x- and @spectrin amino acid sequences, it has previously been suggested that there is an eight-repeat supramotif detectable in both (Y- and @spectrins (Byers et al. 1992). Such block duplications result from unequal crossing over in a tandem array. Our comparisons among repeats in the spectrin family reveal a pattern of sequence divergence among repeats consistent with a hierarchical set of duplication events (fig. 7). The repeats in both CX- spectrin and P-spectrin cluster in a way that strongly supports a block duplication offset by eight repeats, although subsequent exchange events have made this somewhat less distinct in the case of @spectrin (fig. 7A- 25). An alignment of the l3u repeats reveals two separate duplications offset by 8 repeats, which are in turn offset by 13 repeats (fig. 7F and G). This suggests the hy-

Dow


ic.oup.com/m


ber 2021


A homodimeric ancestor alp-spectrin tetramer

mm

I

homodimeric ancestor

gene encoding PH domain

gene encoding p-spectrin ancestor

C p side a side

* * PH domaln

a-spectrin ancestor abc

___I

FIG. 6.-A one-step model for the generation of the cx- and P-spectrin lineages that accounts for all of the general structural characteristics of spectrins today. A, Schematic illustration of the change in structure from a homodimeric ancestor to the heterotetrameric combination of cx- and P-spectrins. The spectrin repeat involved in the reciprocal translocation illustrated in this figure and described in the text is shaded black; all other domains are labeled as in figure 1. B, Proposed mechanism for the transition from homodimer to heterotetramer. A reciprocal translocation event is postulated to have occurred between the coding regions of the homodimeric precursor and a PH domain containing protein. The specific location of this event would have been between the b and c ar-helices of a spectrin repeat in the gene encoding the homodimer and N-terminal to the PH domain in the second gene (double-headed arrow). In one event, this would generate (1) a P-spectrin ancestor that contains a C-terminal PH domain; (2) (Y- and P-spectrin ancestors that are capable of interacting in a head-to-head fashion (C) while maintaining the F-actin-crosslinking activity at the same length as the homodimeric ancestor; and (3) an a-precursor that is immediately expressed through the acquisition of the promoter and regulatory sequences of the PH domain gene. C, Illustration of the head-to-head interaction that results from this event between the cx- and P-spectrin ancestors. In the homodimeric protein, the a, b, and c helices fold and interact to form a triple a-helical unit. After the translocation, this interaction now forms the basis for the new head-to-head binding site.

pothesis that the increase in length of the PH repeat array from the 17 seen in P-spectrin to 30 may have occurred in one unequal crossover event offset by 13 repeats.

A second indication that concerted evolution was once operating in the spectrin repeat arrays of this protein superfamily comes from the existence of arraywide differences in the sizes of these repeats in different proteins. Dystrophin repeats are distinct from spectrin repeats in that they tend to average 109 amino acids in length instead of 106 (Koenig, Monaco, and Kunkel 1988). This observation can be explained by the pro- pensity for sequence variants to spread through a tandemly repeated DNA array by unequal exchange and conversion events. The ancestral gene for this superfamily could presumably have had either the 106- or 109- amino-acid-long repeat, with the other length variant spreading in the appropriate lineage by either neutral or selective mechanisms, although the slight heterogeneity of repeat lengths today makes it hard to infer precise ancestral lengths. The appearance of a repeat length variant across the whole array in spectrins or dystrophins combined with the evidence for block duplications in cx-spectrin, P-spectrin, and PH strongly suggests that processes leading to concerted evolution were once active in this superfamily of proteins.

A Two-Phase Model for the Evolution of the c;u-Actinin/Dystrophin/Spectrin Superfamily of Proteins

Our analysis supports the idea that o-spectrin, p- spectrin, and PH had a common ancestor. We have also found evidence of concerted evolution in the cx-actinin/ spectrin/dystrophin superfamily of proteins, which indicates that the mechanisms necessary to derive both dystrophin and the spectrins from an cx-actinin-like precursor were active in this superfamily. Combining these two observations with our previous analysis indicating stability in the recent evolution of the spectrin repeat arrays (Muse, Clark, and Thomas 1997) suggests that the evolution of the superfamily can be divided into two phases: an initial dynamic phase, characterized by intragenic duplications and concerted evolution, followed by a stable phase that began before the arthropod/vertebrate split in which the number of repeats became constant and their sequences evolved independently (fig. 8). During the first phase gene duplications produced the cx-actinin, spectrin, and dystrophin lineages, and the evolution of the repeat regions was dominated by the behavior of the tandemly repeated DNA encoding them rather than selection at the protein level for individuality among the repeats. This

Dow


ic.oup.com/m


ber 2021

1292 Thomas et al.

A a-spectrin B a-spectrin

3

LHum E a-5

r-i-+ Chk a-l 4 100

98 Hum N a-l 4

:FIy a-14 0 05

-2 L

IHum E a-14

I

-

0.1

-11

A;:

D a-spectrin E P-spectrin

C f3-spectrin

----- -i8) 7

1-9 -4 ----- 410)

17

18

F &-spectrin G PH-spectrin

FIG. 7.-Molecular evolution of spectrins revealed by gene trees. A detailed description of this nomenclature is given in Materials and Methods. A, Portion of the neighbor-joining tree of a-spectrin repeats including the human erythrocyte (Hum E) and nonerythrocyte (Hum N) forms and the chicken (Chk) and Drosophila melunoguster (Fly) a-spectrins. Distances are based on amino acid identity. Notice that the subtrees for each segment have identical topologies, indicating that the most recent common ancestor of differently numbered segments occurred before the arthropod/vertebrate split. These trees are robust; the bootstrap value at the earliest branch point on the example shown is high at the equivalent nodes that join other segment pairs in both the IX- and P-spectrin trees shown in B and C. In panel B, these are (in rank order) 88%, 88%, 92%, 94%, 96%, lOO%, lOO%, lOO%, and 100%. In panel C, all but one of these nodes has a 100% bootstrap value, the exception being 92%. B, Neighbor-joining tree for all a-spectrin repeats without the details of branches to the individual species’ gene. A block duplication of segments [2-81 * [ 1 l-171 is supported by this tree (indicated schematically in 0). Dashed branches are a-spectrin segments from individual isoforms or species that did not cluster with the equivalent segments at the other branches, and are not to scale. C, Neighbor-joining tree (drawn and annotated as in B) for all of the individual P-spectrin repeats plus segments 2-5 and 8-18 of Pu, which appear to be colinear. Note that due to the segment numbering system, segments 8-18 of l3u = segments 7-17 of conventional @spectrins due to the insertion of the SH3 domain (Pu segment 7; see fig 2). Where a given l3u segment groups with its P-spectrin homologs, only the P-spectrin segment number is shown for simplicity. Where a l3u segment does not group with its @spectrin homolog, it is labeled with the Pu segment number in parentheses and is on a dashed branch that is not to scale (i.e., &, segments 5, 8, 10, 11, 14). A block duplication of segments [4, 6-101 w [12, 14-181 is supported by this tree (indicated schematically in L’). D, Schematic illustration of the block duplication in a-spectrin identified by the nesting of the branches in B. The segment blocks related by the duplication are shaded black and gray and are connected by the bracket with arrows. Horizontal arrows indicate the colinear relationship between the most closely related segments. Hatched segments were not included in the analysis. E, Schematic illustration of the block duplication in p- spectrins identified by the nesting of the branches in C. Annotation is as in D. F and G, Schematic illustration of the results of aligning all of the fin repeat segments (fig. 2). The branching pattern of the tree constructed from the l3u repeat similarities (not shown) supports the hypothesis that two successive block duplication events occurred. Two pairs of blocks offset by eight repeats are more closely related than all the other repeat comparisons: [5, 8-121 tj [14, 16-201 has an average pairwise distance (pd) of 0.7408 + 0.0570 and for [21-241 * [29-321, pd = 0.7431 ? 0.0469, while all the other pairwise comparisons yield a pd = 0.815 ? 0.0019. These may represent traces of the same block duplication that was itself subsequently duplicated by an unequal crossing over event with an offset of 13 repeats. This is supported by a pd of 0.7356 + 0.0517 for the blocks [8-12, 15-191 ti [21-25, 28-321, shown schematically in G. Annotation is as in D.

Dow


ic.oup.com/m


ber 2021


a-actinin-like ancestor

GD tQ+ a-actinin

spectrin/dystrophin stable lineage

unstable lineage

homodimeric ancestor

SH3 ‘7 dystrophin

a/p ancestor

c P/P” ancestor

0 GD P-spectrin

PmsPectrin + isoforms pH spectrin

\ \

-- PH isoforms? --) (e.g. TW260?)

repeat array length becomes stable due to selective constraints on crosslinking distance and/or divergence of

individual repeats as they acquire different functions

FIG. K-A model for the evolution of the spectrin superfamily of proteins. We propose two distinct phases for the evolution of these proteins. The first phase, following the vertical hatched arrow, is dominated by the behavior of the repetitive DNA encoding the spectrin repeats. The repeat array lengths of these proteins would be unstable; sequence variants could spread across an array, and domains such as the SH3 domain would be mobile within an array and prone to duplication or elimination in unequal crossing over events. The second phase, following the horizontal hatched arrow, is characterized by loss of this behavior due to divergence at the DNA level. Abbreviations: GD-gene duplication(s); RT-reciprocal translocation diagrammed in figure 6. The structures of the various intermediates postulated are indicated at key steps. The number of repeats in each of these proteins is necessarily uncertain due to the dynamic nature of the repeated DNA encoding them. The N-terminus is to the left and the domains are shaded as in figure 1.

gave rise to both the repetitive arrays and the different repeat lengths of the spectrin and dystrophin lineages. During a transition period, we hypothesize that the new genes began to acquire distinct F-actin-crosslinking functions and that there was selection for specific lengths of crosslinking in each class of proteins. Sub- sequent selection against shorter or longer proteins would disfavor tandem arrays that continued to under- go array length changes and would favor those which remained stable. The molecular mechanism by which this occurred would be a self-reinforcing process as follows. A combination of genetic drift and selection for appropriate crosslinking properties would drive divergence at the DNA level while retaining the basic structural requirements for a spectrin repeat at the amino acid level. Such sequence divergence would cause a progressive diminution in the likelihood of both in-

terallelic and intergenic misalignments, preventing ad- ditional duplications, conversion events, and concerted evolution, thereby fixing the proteins at the lengths we see today and initiating the second evolutionary phase. The current length of the spectrin subunits is evidently very stable, having remained unchanged since the divergence of the arthropod and vertebrate lineages. Since the actin-binding domains from different proteins show similar binding kinetics today (Winkelman and Forget 1993; Rybakova, Amann, and Ervasti 1996), their function appears to have changed little during evolution. Thus, actin-crosslinking distance and the flexibility of the tether composed of spectrin-like repeats would be expected to change as a consequence of changes in array length and repeat sequence length, respectively, providing a selective basis for length sta- bilization.

Dow


ic.oup.com/m


ber 2021

1294 Thomas et al.

Conclusions

There are many examples of proteins that contain repeated domains, including other actin-crosslinking proteins such as ABP120 and ABP280, which contain a repeated crossed P-sheet motif (Noegel et al. 1989; Du- breuil 1991) while sharing the same actin-binding motif as the spectrin superfamily, the ankyrin-repeats that are found in several cytoskeletal and signaling proteins, armadillo repeats of p-catenins and plakoglobins (e.g. Riggleman, Wieschaus, and Schedl 1989; Peifer and Weischaus 1990), and the numerous repetitive exodo- mains on adhesion and signaling molecules (e.g., EGF repeats of Notch [Wharton et al. 19951 and crumbs [Te- pass, Theres, and Knust 19901). It is likely that these proteins originated through intragenic duplication in a dynamic phase resembling that of the spectrins, while the specific selective pressures that led to their stabili- zation would obviously differ. Finally, we note that tenascin-X has a block of near-identical FnIII repeats (Bris- tow et al. 1993), suggesting that this region may still be actively exchanging information between repeats.

Acknowledgments

This work was funded by grants from the National Institutes of Health (NIH; GM-33830) and the Muscular Dystrophy Association (U.S.A.) to DPK., by the Amer- ican Heart Association (Pennsylvania Affiliate) and the NIH (GM-52506) to G.H.T., and by the U.S. National Science Foundation to A.G.C. We are grateful to Dr. Harold Erickson for critically reading this manuscript. We are also grateful to Dr. William Gelbart for providing lab space to G.H.T. for part of this work and to Dr. Kenneth Weiss for some of the automated sequencing runs.

LITERATURE CITED

BARON, M. D., M. D. DAVISON, P A. JONES, and D. R. CRITCH- LEY. 1991. The sequence of chick alpha-actinin reveals ho- mologies to spectrin and calmodulin. J. Biol. Chem. 262: 17623-17629.

BENNETT, V. 1992. Ankyrins: adapters between diverse plasma membrane proteins and the cytoplasm. J. Biol. Chem. 267: 87038706.

BENNETT, V., and D. M. GILLIGAN. 1993. The spectrin-based membrane skeleton and micron-scale organization of the plasma membrane. Annu. Rev. Cell Biol. 9:27-66.

BIES, R. D., S. E PHELPS, M. D. CORTEZ, R. ROBERTS, C. T CASKEY, and J. S. CHAMBERLAIN. 1992. Human and murine dystrophin mRNA transcripts are differentially expressed during skeletal muscle, heart, and brain development. Nu- cleic Acids Res. 20: 1725-1731.

BLOOM, M. L., C. S. BIRKENMEIER, and J. E. BARKER. 1993. Complete nucleotide sequence of the murine erythroid p- spectrin cDNA and tissue-specific expression in normal and jaundiced mice. Blood 82:2906-2914.

BRISTOW, J., M. K. TEE, S. E. GITELMAN, S. H. MELLON, and W. L. MILLER. 1993. Tenascin-X: a novel extracellular ma- trix protein encoded by the human XB gene overlapping p450c21B. J. Cell Biol. 122:265-278.

BROWN, N. H., and E C. KAFATOS. 1988. Functional cDNA libraries from Drosophila embryos. J. Mol. Biol. 203:425- 437.

BYERS, T J., E. BRANDIN, R. A. LUE, E. WINOGRAD, and D. BRANTON. 1992. The complete sequence of Drosophila beta-spectrin reveals supra-motifs comprising eight 106-residue segments. Proc. Natl. Acad. Sci. USA 89:6187-6191.

DE ARRUDA, M. V., S. WATSON, C. S. LIN, J. LEAVITT, and P MATSUDAIRA. 1990. Fimbrin is a homologue of the cyto- plasmic phosphoprotein plastin and has domains homologous with calmodulin and actin gelation proteins. J. Cell Biol. 111: 1069-1079.

DRUBIN, D. G., and W. J. NELSON. 1996. Origins of cell polarity. Cell 84:335-344.

DUBREUIL, R. R. 1991. Structure and evolution of the actin crosslinking proteins. Bioessays 13:219-226.

DUBREUIL, R. R., T. J. BYERS, A. L. SILLMAN, D. BAR-ZVI, L. S. B. GOLDSTEIN, and D. BRANTON. 1989. The complete sequence of Drosophila alpha-spectrin: conservation of structural domains between alpha-spectrins and alpha-actinin. J. Cell Biol. 109:2197-2205.

DUBREUIL, R. R., T J. BYERS, C. T STEWART, and D. I? KIE- HART. 1990. A B-spectrin isoform from Drosophila (&) is similar in size to vertebrate dystrophin. J. Cell Biol. 111: 1849-1858.

FROHMAN, M. A. 1990. RACE: rapid amplification of cDNA ends. Pp. 28-38 in M. A. INNIS, D. H. GELFAND, J. J. SNIN- SKY and T. J. WHITE, eds. PCR Protocols. Academic Press, San Diego.

FYRBERG, E., K. MELISSA, E. BALL, C. FYRBERG, and M. C. REEDY. 1990. Molecular genetics of Drosophila alpha-actinin: mutant alleles disrupt Z disc integrity and muscle insertions. J. Cell Biol. 110: 1999-2011.

GORLIN, J. B., R. YAMIN, S. EGAN, M. STEWART, T. I? STOSSEL, D. J. KWIATKOWSKI, and J. H. HARTWIG. 1990. Human en- dothelial actin-binding protein (ABP-280, nonmuscle filam- in): a molecular leaf spring. J. Cell Biol. 111:1089-l 105.

HARTWIG, J. H., and D. J. KWIATKOWSKI. 1991. Actin-binding proteins. Cur-r. Opin. Cell Biol. 3:87-97.

Hu, R.-J., W. MASAYO, and V. BENNETT. 1992. Characteriza- tion of human brain cDNA encoding the general isoform of B-spectrin. J. Biol. Chem. 267:18715-18722.

KENNEDY, S., S. WARREN, B. FORGET, and J. MORROW. 1991. Ankyrin binds to the 15th repetitive unit of erythroid and non-erythroid beta-spectrin. J. Cell Biol. 115:267-277.

KOENIG, M., A. l? MONACO, and L. M. KUNKEL. 1988. The complete sequence of dystrophin predicts a rod-shaped cytoskeletal protein. Cell 53:219-228.

KUMAR, S., K. TAMWA, and M. NEI. 1993. Molecular evolutionary genetic analysis. Version 1.02. Institute of Molec- ular Evolutionary Genetics, Pennsylvania State University, University Park.

LEE, J. K., E. BRANDIN, D. BRANTON, and L. S. B. GOLDSTEIN. 1997. a-Spectrin is required for ovarian follicle monolayer integrity in Drosophila melunoguster. Development 124: 353-362.

LEMAIRE, C., R. HEILIG, and J. L. MANDEL. 1988. Nucleotide sequence of chicken dystrophin cDNA. Nucleic Acids Res. 16:11815.

LUNA, E. J., and A. L. HITT. 1992. Cytoskeleton-plasma membrane interactions. Science 258:955-964.

MA, Y., W. E. ZIMMER, B. M. RIEDERER, and S. R. GOODMAN. 1993. The complete amino acid sequence for brain l3 spectrin (p fodrin): relationship to globin sequences. Mol. Brain Res. 18:87-99.

MOON, R. T., and A. P MCMAHON. 1990. Generation of diversity in nonerythroid spectrins. J. Biol. Chem. 256:4427- 4433.

MUSACCHIO, A., T. GIBSON, I? RICE, J. THOMPSON, and M. SARASTE. 1993. The PH domain: a common piece in the

Dow


ic.oup.com/m


ber 2021


structural patchwork of signaling proteins. Trends Biochem. Sci. l&343-348.

MUSACCHIO, A., M. WILMANNS, and M. SARASTE. 1994. Struc- ture and function of the SH3 domain. Prog. Biophys. Mol. Biol. 61:283-297.

MUSE, S. V., A. G. CLARK, and G. H. THOMAS. 1997. Com- parisons of the nucleotide substitution process among repetitive segments of the (Y- and P-spectrin genes. J. Mol. Evol. 44:492-500.

NOEGEL, A., W. WITKE, and M. SCHLEICHER. 1987. Calcium- sensitive non-muscle alpha-actinin contains EF-hand structures and highly conserved regions. FEBS Lett. 221:391- 396.

NOEGEL, A. A., S. RAPP, E LOTTSPEICH, M. SCHLEICHER, and M. STEWART. 1989. The Dictyostelium gelation factor shares a putative actin-binding site with ol-actinins and dystrophin and also has a rod domain containing six lOO-residue motifs that appear to have a cross-p conformation. J. Cell Biol. 109:607-618.

PEIFER, M., and E. WEISCHAUS. 1990. The segment polarity gene armadillo encodes a functionally modular protein that is the Drosophila homolog of human plakoglobin. Cell 63: 1167-l 178.

PRASSLER, J., S. STOCKER, G. MARRIOTT, M. HEIDECKER, J. KELLERMAN, and G. GERISCH. 1997. Interaction of a Dic- tyostelium member of the plastin/fimbrin family with actin filaments and actin-myosin complexes. Mol. Biol. Cell 8: 83-95.

RIGGLEMAN, B., E. WIESCHAUS, and l? SCHEDL. 1989. Molec- ular analysis of the armadillo locus: uniformly distributed transcripts and a protein with novel internal repeats are associated with a Drosophila segment polarity gene. Genes Dev. 3:96-l 13.

RYBAKOVA, I. N., K. J. AMANN, and J. M. ERVASTI. 1996. A new model for the interaction of dystrophin with F-actin. J. Cell Biol. 135:661-672.

SABER, M. A., H. ABDEL-HEMID, M. ROMEIH, H. AHMED, and R. HARISON. 1996. GenBank Accession L33405, Release 98.0.

SAHR, K. E., I? LAURILA, L. KOTULA et al. (12 co-authors). 1990. The complete cDNA and polypeptide sequences of human erythroid a-spectrin. J. Biol. Chem. 265:4434-4443.

SPEICHER, D. W., T. M. DESILVA, K. D. SPEICHER, J. A. UR- SITTI, I? HEMBACH, and L. WEGLARZ. 1993. Location of the human red cell spectrin tetramer binding site and detection of a related “closed” hairpin loop dimer using proteolytic footprinting. J. Biol. Chem. 268:4227-4235.

SPEICHER, D. W., and V. T. MARCHESI. 1984. Erythrocyte spectrin is comprised of many homologous triple helical segments. Nature 311: 177-180.

TEPASS, U., C. THERES, and E. KNUST. 1990. crumbs encodes an EGF-like protein expressed on apical membranes of Drosophila epithelial cells and required for organization of epithelia. Cell 61:787-799.

THOMAS, G. H., and D. l? KIEHART. 1994. PHeavy-spectrin has a restricted tissue and sub cellular distribution during Dro- sophila embryogenesis. Development 120:2039-2050.

TSE, W. T., M.-C. LECOMTE, E E COSTA, M. GARBARZ, C. FEO, P BOIVIN, D. DHERMY, and B. G. FORGET. 1990. Point mu- tation in the P-spectrin gene associated with aI/ heredi- tary elliptocytosis. J. Clin. Invest. 86:909-916.

VEIL, A., and D. BRANTON. 1994. Interchain binding at the tail end of the Drosophila spectrin molecule. Proc. Natl. Acad. Sci. USA 91: 10839-10843.

-. 1996. Spectrin: on the path from structure to function. Curr. Opin. Cell Biol. 8:49-55.

WASENIUS, V.-M., M. SARASTE, I? SALVI?N, M. ER~~MAA, L. HOLM, and V.-I? LEHTO. 1989. Primary structure of the brain a-spectrin. J. Cell Biol. 108:79-93.

WHARTON, K. A., K. M. JOHANSEN, T. Xu, and S. ARTAVANIS- TSAKONAS. 1995. Nucleotide sequence from the neurogenic locus Notch implies a gene product that shares homology with proteins containing EGF-like repeats. Cell 43:567- 581.

WHARTON, K. A., B. YEDVOBNICK, V. G. FINNERTY, and S. ARTAVANIS-TSAKONAS. 1985. opa: a novel family of tran- scribed repeats shared by the Notch locus and other developmentally regulated loci in D. melunoguster. Cell 40:55- 62.

WINKELMAN, J. C., J.-G. CHANG, W. T. TSE, V. T. MARCHESI, and B. G. FORGET. 1990~. Full-length sequence of the cDNA for human erythroid P-spectrin. J. Biol. Chem. 265: 11827-l 1838.

WINKELMAN, J. C., E E COSTA, B. L. LINZIE, and B. G. FOR- GET. 1990b. p Spectrin in human skeletal muscle. J. Biol. Chem. 265:20449-20454.

WINKELMAN, J. C., and B. G. FORGET. 1993. Erythroid and nonerythroid spectrins. Blood 81:3 173-3 185.

YAN, Y., E. WINOGRAD, A. VEIL, T. CRONIN, S. C. HARRISON, and D. BRANTON. 1993. Crystal structure of the repetitive segments of spectrin. Science 262:2027-2030.

YOUSSOUFIAN, H., M. MCAFEE, and D. J. KWIATKOWSKI. 1990. Cloning and chromosomal localization of the human cytoskeletal alpha-actinin gene reveals linkage to the beta-spectrin gene. Am. J. Hum. Genet. 47:62-71.

ZIMMER, E. A., S. L. MARTIN, S. M. BEVERLEY, Y. W. KAN, and A. C. WILSON. 1980. Rapid duplication and loss of genes coding for the a-chains of hemoglobin. Proc. Natl. Acad. Sci. USA 77:2158-2162.

THOMAS H. EICKBUSH, reviewing editor

Accepted September 5, 1997

Dow


ic.oup.com/m


ber 2021

Date post:	09-Feb-2022
Category:	Documents
Upload:	others
View:	4 times
Download:	0 times

Intragenic Duplication and Divergence in the Spectrin Superfamily of

Documents