REVIEWS 17 Wawrzynow, A. et al. (1995) EMBO J. 14,

1867-1877 18 Parsell, D. A., Kowal, A. S. and Lindquist, S.

(1994) J. Biol. Chem. 269, 4480-4487 19 Singh, S. K. and Maurizi, M. R. (1994) J. Biol.

Chem. 269, 29537-29545 20 Seol, J. H. et al. (1995) J. Biol. Chem. 270,

8087-8092 21 Kessel, M. et al. (1995) J. Mol. Biol. 250,

587-594 22 Kim, S., Willison, K. R. and Horwich, A. L.

(1994) Trends Biochem. Sci. 19, 543-548 23 Sanchez, Y. et al. (1992) EMBO J. 11,

2357-2364 24 Lindquist, S. et al. (1995) The Role of

Hspl04 in Stress Tolerance and [PSI+] Propagation in Saccharomyces cerevisiae (Vol. LX), pp. 451-460, Cold Spring Harbor Laboratory Press

25 Hubel, A. et al. (1995) Mol. Biochem. Parasitol. 70, 107-118

26 Schirmer, E. C., Lindquist, S. and Vierling, E. (1994) Plant Cell 6, 1899-1909

27 Lee, Y-R. J., Nagao, R. T. and Key, J. L. (1994)

Plant Cell 6, 1889-1897 28 Skowyra, D., Georgopoulos, C. and Zylicz, M.

(1990) Cell 62, 939-944 29 Parsell, D. A. eta/. (1994) Nature 372,

475-478 30 Vogel, J. L., Parsell, D. A. and Lindquist, S.

(1995) Curr. Biol. 5, 306-317 31 Tobias, J. W. et al. (1991) Science 254,

1374-1377 32 Mhammedi, A. A. et al. (1994) Mol. Microbiol.

11, 1109-1116 33 Lehnherr, H. and Yarmolinsky, M. B. (1995)

Proc. Natl. Acad. Sci. U. S. A. 92, 3274-3277

34 Schweder, T. et al. (1996) J. Bacteriol. 178, 470-476

35 Inoue, I. and Rechsteiner, M. (1994) J. Biol. Chem. 269, 29241-29246

36 Baker, T. A. (1993) Curr. Opin. Genet. Dev. 3, 708-712

37 Levchenko, I., Luo, L. and Baker, T. A. (1995) Genes Dev. 9, 2399-2408

38 Kruklitis, R., Welty, D. J. and Nakai, H. (1996) EMBO J. 15, 935-944

TIBS 2 1 - AUGUST1996

39 Geuskens, V. et al. (1992) EMBO J. 11, 5121-5127

40 Wickner, R. B., Masison, D. C. and Edskes, H. K. (1995) Yeast 11, 1671-1685

41 Cox, B. (1994) Curr. Biol. 4, 744-748 42 Gething, M-J. and Sambrook, J. (1992) Nature

355, 33-45 43 Sanchez, Y..et al. (1993) J. Bacteriol. 175,

6484-6491 44 Shapiro, J. A. (1993) J. Bacteriol. 175,

2625-2631 45 Walker, J. E. et al. (1982) EMBO J. 1,

945-951 46 Engel, A. eta/. (1995) Science 269, 832-836 47 Peters, J-M. (1994) Trends Biochem. Sci. 19,

377-382 48 Sander, C. and Schneider, R. (1991) Protein

Struct. Funct. Genet. 9, 56-68 49 Swofford, D. L. (1991) PAUP: Phylogenetic

Analysis Using Parsimony (version 3.1), Illinois Natural History Survey

50 Maddison, W. P. and Maddison, D. R. (1992) MacClade- Analysis of Phylogeny and Character Evolution (version 3.0), Sinauer Associates

The SR protein family: pleiotropic functions in

pre.mRNA splicing

Juan Valc rcel and Michael R. Green A family of proteins with arginine-serine-rich domains has recently come into the limelight of studies on the mechanisms of constitutive and regu- lated pre-mRNA splicing. Implicated in an ever increasing variety of func- tions, these proteins act as driving forces during spliceosome assembly and also play decisive roles in alternative splice-site selection, suggesting that they are crucial players in the regulation of splicing during cell differ- entiation and development.

MOST MESSENGER RNAs in higher eukaryotes are synthesized as precur- sors that contain intervening sequences 0ntrons). Introns are removed in the cell nucleus and the flanking exons are spliced together to generate functional mRNAs (Fig. 1). Very often, the use of alternative splice-sites in the same pre- mRNA is regulated in a cell-type-specific

J. Valchrcel is at the Gene Expression Programme, European Molecular Biology Laboratory, Meyerhofstrasse 1, D-69117 Heidelberg, Germany; and M. R. Green is at the Howard Hughes Medical Institute, Program in Molecular Medicine, University of Massachusetts Medical Center, 373 Plantation St, Worcester, MA 01605, USA. Email: michael.green@ummed.edu

296

manner, allowing the synthesis of differ- ent polypeptides from the same gene.

Multiple discovery Six years ago, it was shown that a

single purified polypeptide, named ASF, when added to standard in vitro splic- ing reactions, could change the relative use of two competing 5' splice-sites of an SV40 early pre-mRNA 1. Sequence analy- sis revealed that ASF had also been iso- lated as an activity required for consti- tutive pre-mRNA splicing, and named SF2 (Ref. 2). Thus, ASF/SF2 is a general splicing factor that can switch between alternative splice-sites when present in excess, an observation subsequently confirmed in vivo 3'4. An implication of

�9 1996, Elsevier Science Ltd

this dual discovery was that physiologi- cal variations in the concentration of general splicing factors can regulate alternative splicing.

ASF/SF2 contains an amino-terminal RJ~lA-binding domain composed of ribo- nucleoprotein (Pd~IP-CS) motifs, which are 70-90-amino acid modules com- monly found in polypeptides that recog- nize Rd~IA in a sequence-specific manner, and which have a variety of functions in RNA metabolism (for review, see Ref. 5). ~F/SF2 also contains a carbo~- terminal region rich in arginine-serine dipeptides (Fig. 2). Similar modular organization and activities were also found in the polypeptide SC35, which had been identified as a component of splicing complexes 6.

At about the same time, a set of pro- teins was identified by a monocional antibody that recognizes components of transcriptionally active sites both in Xenopus lampbrush and in Drosophila pol~ene chromosomes 7. This group of proteins included ~F/SF2, SC35 and four additional polypeptides of 20, 40, 55 and 75 Id)a 8. All six proteins were co-purified from various sources by a simple, two- step salt precipitation procedure and were collectively named SR proteins. They share a similar domain organiz- ation (Fig. 2) and the ability to modu- late 5' splice-site choice. In addition, SR proteins can complement cytoplasmic S100 extracts, which lack all six poly- peptides and therefore cannot support splicing reactions.

SR proteins belong to a larger family of polypeptides with 'alternating arginine' domains, which include snRNP-associated

PII: S0968-0004(96) 10039-6

REVIEWS TIBS 2 1 - A U G U S T 1 9 9 6

(e.g. U1 70K) and non-snRNP associated (e.g. U2A~ splicing factors, splicing regulators [e.g. su(wa), Tra and Tra-2] and an increasing number of previously unidentified spliceosomal components 9 (for a recent comprehensive review, s e e

Ref. 10). The term SR proteins, however, is usually reserved for the six polypep- tides mentioned above and illustrated in Fig. 2, with a few more recent ad- ditions that share most of their charac- teristic features.

The structural and functional simi- larities among SR proteins suggests that they could perform redundant roles. De- letion or mutation of one particular SR member in Drosophila, however, prevents normal development n,]2, indicating dis- tinct functions in vivo. Although SR pro- teins are conserved across metazoa and have a ubiquitous tissue distribution, cell-type differences in their relative abundance have been observed n,13,14. Combined with recent findings of dif- ferences in activity among family mem- bers 13,]5-]7, these observations suggest that each cell type might have a distinct pattern of relative SR protein concen- trations, which might define alternative splicing decisions. Consistent with this hypothesis, overproduction of a particu- lar SR member in Drosophila causes multiple developmental abnormalities, which could result from aberrant pre- mP, NA splicing regulation ]8.

Functions of the structural domains SR proteins have a modular structure

consisting of an RNA-binding domain and an arginine--serine-rich (RS) region (Fig. 2). The amino-terminal RNA-binding domain is essential for all the known ac- tivities of SR proteins, both in vitro and in vivo 4,~9,2~ This region consists of one or two repeats of the aforementioned RNP-CS motif. In those SR proteins con- taining two repeats, the amino acid se- quence of the carboxy-terminal motif is less well conserved. This 'degenerate' RNP-CS, however, also contributes to define the overall RNA-binding affinity and specificity of the protein ]9,2].

RNA binding. Using iterative selection from a pool of random RNA sequences, different purine-rich-binding consensus sequences have been obtained for ASF/SF2 and SC35 (Ref. 21). The ASF/SF2 consensus is found in some 5' splice-sites and in particular types of exonic sequence, known as the purine- rich exon enhancers, which stimulate the use of weak splice-sites (see below). In fact, ASF/SF2 has been shown inde- pendently to bind to both of these

P r e - m R N A

5' splice site

exon 1 ~. .~ intron

Polypyrimidine tract

Branch 3' splice point site

A ~ U U U U U U ~

+- . exon2- -~

Pre -sp l i ceosome

Sp l i ceosome

D.. I~. z z rc

r O9

JUUU-I

/'/'-- U(,w . . . . . . . .

U4 snRNP

C Spl iced m R N A Lariat intron

A - - U U U U U U - -

( J Figure 1

Assembly of splicing complexes. The process of pre-mRNA splicing involves the elimination of intervening sequences (introns) in mRNA precursors and the ligation of their flanking regions (exons) to produce mature mRNAs. Four sequence elements are essential for the splicing process in higher eukaryotic pre-mRNAs: both the 5' and 3' splice-sites, the branch point (an adenosine residue that forms a 2'-5' covalent link with the excised 5' end of the intron, generating an RNA with a lariat configuration), and a polypyrimidine tract present be- tween the branch point and the 3' splice-site [represented by a poly(U) stretch]. Pre- spliceosome complexes are formed by interaction of the U1 small nuclear ribonucleo- protein particle (U1 snRNP) with the 5' splice-site and of U2 snRNP with the branch-point region. Catalytically active spliceosomes are subsequently formed by recruitment of a U4-U5-U6 tri-snRNP and concurrent destabilization of U1 snRNP binding. At this stage, U6 snRNA interacts with the 5' splice-site, which is therefore sequentially recognized dur- ing spliceosome assembly by Ul snRNP and U6 snRNP.

sequences 22-24. The SC35 consensus also resembles 5' splice-sites21o Little is known about Pd~IA sequences recog- nized by other SR proteins, and infor- mation is urgently needed to under- stand their functions.

Protein-protein interactions. Arginine- serine (RS) regions are essential for some, but not all, functions of SR proteins ]9,2~

These domains differ among SR proteins in their length, number of arginine-serine dipeptides and content of other amino acids (for review, see Ref. 5). Presently, it is not clear to what extent these domains are interchangeable. A variety of in vitro and in vivo techniques, including far-western blots, co-precipi- tation and yeast two-hybrid assays, have

297

REVIEWS TIBS 21 - AUGUST 1996

SRp75 v ~ [ i l l I l l l l I l l l I li I l l l l l l H I

SRp55 ~ . , [ I I I I I I I I I I I H i I I I l i l l e

SRp40 [| IIIII I I II I I I I I I I

SRp30a (ASF/SF2) [11 I I I IH I I I l l l ]

SRp30b (SC35) I I I I I I I I I I I I I I I I I I fl I I I I I I I

SRp20

[ I II l iD II IHI I ]

Key: RNP-CS domain -4,~m ~ �9 H~- RS domain

- ~ - ~t' RNP-CS domain i RS dipeptide -o- Gly-rich hinge

Figure 2 Domain structure of human SR proteins. These polypeptides contain an amino-terminal RNA-binding domain consisting of one RNP-CS motif (see text) with or without a second, less conserved RNP-CS (~P RNP-CS), and a carboxy-terminal arginine-serine rich (RS) domain of variable length and primary se- quence (each vertical bar represents an RS dipep- tide). A glycine-rich hinge connects these two domains in some of the SR proteins.

revealed that some RS regions can mediate protein-protein interactions 25,26. Partners for these interactions include: (1) other SR proteins; (2) other RS- containing splicing factors [e.g. U1 70K and the 35kDa subunit of U2AF]; and (3) RS-containing splicing regulators (e.g. Tra and Tra-2). Some RS domains can also influence RNA binding 19, pro- mote RNA-RNA annealing 27 or contain sequences that act as subcellular local- ization signals 28.

Splicing factor recruitment. The modu- lar organization of SR proteins suggests a basic mechanism for their function: their RNA-binding domain could recog- nize specific splicing signals and recruit (through interactions mediated by the RS domain) other splicing factors by protein-protein or RNA-protein inter- actions or by promoting RNA-RNA base pairing. This model is reminiscent of the modular organization and function of transcriptional activators, which similarly trigger the assembly of, or conformational rearrangements within another type of nucleoprotein assembly, the transcription pre-initiation complex.

Roles of SR proteins in spliceosome assembly

Five small nuclear r ibo- nucleoprotein particles (U snRNPs) and multiple pro- tein factors sequentially in- teract with the pre-mRNA to assemble a large spliceo- some complex, in which the splicing reaction occurs (Fig. 1). SR proteins have been implicated in almost every step of this assembly process (Table I).

5' splice-site recognition. A model has been proposed for how ASF/SF2 can pro- mote the first of these inter- actions: binding of U1 snRNP to the 5' splice-site (Fig. 3a). It has been shown that the ASF/SF2 RNA-binding domain can recognize 5' splice-sites or nearby sequences 21,22, and that its RS domain con- tacts the RS domain of the Ul-specific 70kDa polypep- tide 25,26. Although it has not been demonstrated that these two interactions can occur simultaneously, ASF/SF2 might act as a bridge between U1 snRNP and the 5' splice- site, thereby facilitating di- rect base-pairing between the splice-site and the 5' end

of U1 snRNA. Recognition of 5' splice- sites by U1 snRNA could, in turn, in- crease the binding affinity and speci- ficity of ASF/SF2 for these sequences 29.

Surprisingly, inactivation of the 5' end of U1 snRNA with antisense oligonucleo-. tides 3~ or even U1 snRNP depletion 31, can be compensated by an excess of SR proteins. These observations indicate that SR proteins can also help to define 5' splice-sites through another U1 snRNP-independent mechanism. Recent data revealed that recognition of the 5' splice-site by U6 snRNA (Fig. 1) becomes rate-limiting in U1 snRNP- depleted extracts 32, suggesting that this could be the step facilitated by SR pro- teins under these conditions. Alterna- tively, SR proteins could facilitate an earlier event that is normally triggered by U1 snRNP, e.g. U2 snRNP binding (see below), which ceases to be rate- limiting in the presence of SR proteins 32. While SC35 alone can fulfill this function 3~ it is not known whether other SR proteins share this activity.

Communication between splice-sites. The first functional relationship between

splice-site partners is established in a complex containing U1 snRNP bound to the 5' splice-site, and the splicing factor U2AF bound to the polypyrimidine tract associated with 3' splice-sites. U2AF consists of a 65 kDa subnnit that specifi- cally recognizes polypyrimidine tracts, and a 35 kDa subunit that does not bind to the pre-mRNA, but is tethered to the polypyrimidine tract through its inter- action with the 65 kDa subunit. The SR protein SC35 has been proposed to establish a physical link between both splice-sites through a network of~ domain-mediated protein-protein inter- actions. In this model, SC35 interacts with ASF/SF2 (and/or U1 70K) bound to the 5' splice-site, and with the 35kDa subunit of U2AF associated to the 3' splice-site 25 (Fig. 3b). Although in ap- parent contradiction to the observation that the 35 kDa subunit of U2AF is dis- pensable for splicing in vitro 33, the model is particularly attractive because it suggests a mechanism whereby two splice-sites become committed to splic- ing together. Indeed, pre-incubation of some pre-mRNAs with SC35 alone is sufficient to give them a kinetic advan- tage when challenged with an excess of competitor substrate 16, the operational definition of 'commitment'. Interest- ingly, a different pre-mRNA can be simi- larly committed by pre-incubation with ASF/SF2. These surprising results indi- cate that individual SR proteins can selectively associate with different pre- mRNAs and promote events that facili- tate the assembly of early splicing complexes 34, an idea with important im- plications for splice-site selection. It is unclear, however, whether these results are applicable to the mechanisms defin- ing splice-site partners in vivo.

Later steps in spliceosome assembly. The next step in spliceosome assembly is the stable binding of U2 snRNP to the pre-mRNA branch-point region, which requires ATP and several auxiliary fac- tors, including U1 snRNP. Because the interaction is absent in $100 extracts, and is established upon addition of SR proteins, these factors directly or indi- rectly stimulate U2 snRNP recruitment, even in the absence of functional U1 snRNP or a 5' splice-site 35.

Pre-spliceosome complexes contain- ing U1 and U2 snRNP are converted to mature spliceosomes by the addition of a U4-Ub-U5 tri-snRNP (Fig. 1), and recent results indicate that SR proteins also promote this step 36.

By contrast to the roles of SR pro- teins in early spliceosome assembly, no

298

REVIEWS TIBS 2 1 - AUGUST1996

detailed molecular model can be drawn for how SR proteins stimulate those later events. Because many of them involve pre-mPd~lA-snRNA and snRNA-snRNA base-pairing interactions, it has been suggested 35 that SR proteins can act as 'matchmakers' to provide adequate RNA conformations for base pairing. Indeed, formation of base-paired regions between U2 and U6 snRNAs, which is an impor- tant feature of the pre-spliceosome to spliceosome transition, has been shown to be stimulated by SR proteins 35,37.

Roles of SR proteins in splice-site selection Choice of 5' splice-site. One of the

characteristic activities of SR proteins is their ability to regulate 5' splice- site choice: higher concentrations of ~F/SF2 or SC35 proteins promote the use of sites proximal to the 3' splice- site ]-4, while SRp40 and SRp55 promote the use of distal sites 17. It has been pro- posed that ~F/SF2 affects 5' splice-site selection by increasing the occupancy by U1 snP-d~lP of all the competing sites 38. An intrinsic tendency to splice closest neighbours bound by splicing components would then favor proximal site selection. This hypothesis is in agreement with the model for ~F/SF2 assistance in 5' splice-site recognition described above (Fig. 3a). Atso consist- ent with the idea of modulation of U1 snRNP binding by SR proteins is the observation that SRp40 and SRp55 increase the association of U1 snRNP with distal sites 17. The model, however, neither explains why the RS domain of ?d~F/SF2, while essential for promoting U1 snPd~iP binding, is not required for this activit~ ,I9.2~ nor why the U1 snPd~lP- associated polypeptide U1A, while dis- pensable for formation of a complex between U1 snPd~lE ASF/SF2, and the 5' splice-site 29,35, is necessary for 5' splice- site switching 35.

An alternative hypothesis is that SR proteins mediate 5' splice-site selection after U1 snRNP binding has taken place 17,35, for example at the time of 5' splice-site recognition by U6 snRNA (see above and Fig. 1). Indeed, this step has been shown to act as a proof- reading activity during 5' splice-site definition in yeast. Conceivably, 5' splice- site selection by modulation of U6 snPdNA binding would require a different set of molecular contacts, which do not involve the dispensable RS domain of SR proteins.

SR proteins bind to exon enhancers. Another important mode of splicing regulation in which SR proteins seem to

Table I. Activities of SR proteins

Binding to pre-mRNA Refs 5' splice sites 21, 22 Exon enhancers 23, 24, 41, a

Promote U snRNP-U snRNP and U snRNP-pre-mRNA interactions U1 snRNP binding to 5' splice sites 17, 26 U6 snRNP binding to 5' splice sites 32 U2 snRNP binding to 3' splice sites 35 U1-U2 interactions at 3' splice sites b U2-U6 interactions 35, 37 U4-U5-U6 binding to pre-mRNA 36 Promote splicing of 3' splice sites in trans c

Complementation of splicing-deficient extracts Complementation of $100 cytoplasmic extracts 1, 2, 14, 15,

19, 20, d-i Complementation of U1 snRNP-depleted or U1 snRNP-inactivated extracts 30, 31 Complementation of Cap-binding protein-depleted extracts h

Regulation of splice-site choice Modulation of 5' splice site choice in vitro and in vivo 1-4, 17, d-f,

i ,k Activation of cryptic 5' splice sites 30, 35 Regulation of exon skipping vs inclusion in vitro and in vivo 3, I Commitment of pre-mRNA substrates to splicing 16 Stimulation of the use of 3' or 5' splice sites through exon enhancers 23, 24, 34, 41,

43, a, j -o Splicing inhibition through negative regulatory signals p

Effects of overexpression in vivo Alternative splicing modulation 3, 4 Splicing inhibition 4, q

aXu, R., Teng, J. and Cooper, T. A. (1993) Mol. Cell. Biol. 13, 3660-3674; bFu, X-D. and Maniatis, T. (1992) Proc. Natl. Acad. Sci. U. S. A. 89, 1725-1729; CBruzic, J. P. and Maniatis, T. (1995) Proc. Natl. Acad. Sci. U. S. A. 92, 7056-7059; dKrainer, A. R., Conway, G. C. and Kozak, D. (1990) Cell 62, 35-42; eGe, H., Zuo, P. and Manley, J. L. (1991) Cell 66, 373-382; fFu, X. D., Mayeda, A., Maniatis, T. and Krainer, A. R. (1992) Proc. Natl. Acad. Sci. U. S. A. 89, 11224--11228; gMayeda, A., Zahler, A. M., Krainer, A. R. and Roth, M. B. (1992) Proc. Natl. Acad. Sci. U. S. A. 89, 1301-1304; hLewis, J. and Mattaj, L., pers. commun.; ~Zahler, A. M., Neugebauer, K. M., Stolk, J. A. and Roth, M. B. (1993) Mol. Cell. Biol. 13, 4023-4028; JCavaloc, Y. et al. (1994) EMBO J. 14, 4336-4349; kHimmelspach, M. et al. (1995) RNA 1, 794-806; JMayeda, A., Helfman, D. M. and Krainer, A. R. (1993) Mol. Cell. Biol. 13, 2993-3001; mDirksen, W. E, Hampson, R. K., Sun, Q. and Rottman, E M. (1994) .L Biol. Chem. 269, 64314436; ~ M. et al. (1994) Nucleic Acids Res. 22, 1018-1022; ~ M. B., Bryan, J, Cooper, T. A. and Berget, S. M. (1995) Mol. Cell. Biol. 15, 3979-3988; PMcNally, L. M. and McNally, M. T. (1996) J. Virol. 70, 1163-1172; qRomac, J. M-J. and Keene, J. D. Genes Dev. 9, 1400-1410.

have a key role is the function of exon enhancers. These sequences stimulate the use of sub-optimal splice-sites 39, and are often purine-rich 4~ At least some of the SR proteins bind directly to the en- hancer 23,24, and each specific enhancer sequence seems to be recognized by a distinct subset of SR proteins 34,4].

Do the activities of SR proteins at en- hancers resemble their functions at 5' splice-sites? Two lines of evidence sug- gest that this might be the case. First, U1 snRNP has been found in complexes assembled on some enhancers 34,39. Sec- ond, downstream 5' splice-sites have also been shown to stimulate the use of upstream weak 3' splice-sites.

How does the binding of SR proteins to an enhancer sequence promote the use of nearby splice-sites? Both purine- rich exon enhancers and downstream 5' splice-sites have been shown to increase U2AF binding to the poly- pyrimidine tract associated with 3' splice-sites 34,42,43. These data are also consistent with the observation that an increase in the polypyrimidine content

of some regulated 3' splice-sites makes them enhancer-independent. Because the 35 kDa subunit of U2AF can estab- lish protein-protein interactions with certain SR proteins 25, it has been pro- posed that this subunit acts as a bridge between SR proteins and the 65kDa subunit, thereby facilitating the cooper- ative binding of SR proteins to the en- hancer and of the 65 kDa subunit to the polypyrimidine tract 44 (Fig. 3c). Thus, bridging splice-site partners across an intron (see previous section) and en- hancing splice-site recognition from an exon could operate through a similar network of protein-protein interactions (Fig. 3b,c).

The function of other types of exon enhancers also involves SR proteins. In the Drosophila gene doublesex, six re- peats of 13 nucleotides are required for splicing of a weak upstream 3' splice- site, an important regulatory step in the control of sexual differentiation. Two splicing regulators, Tra and Tra-2, associate with the repeats and help to assemble a complex containing SR

299

REVIEWS TIBS 21 - AUGUST 1996

(a)

) (c)

( Exon

I enhancer

Exon enhancer

Figure 3 RS domain-mediated protein-protein interactions among SR proteins can explain some of their activities in spliceosome assembly. Protein-protein interactions that have been experimentally documented, are represented by double arrows. The pre-mRNA is represented as in Fig. 1. (a) Model of ASF/SF2 assisting U1 snRNP binding to 5' splice-sites 21,22,25,26. U1 snRNP consists of an RNA molecule (U1 snRNA), eight core proteins common to all snRNPs (brown circles) and three Ul-specific polypeptides (U1 7OK, UIA and UlC). ASF/SF2 recognizes the 5' splice-site or nearby sequences, and an interaction between its RS region and the RS region of U1 70K could facilitate the recruit- ment of U1 snRNP to the 5' splice-site. The 5' end of U1 snRNA can then form base pairs (represented by vertical black lines) with the 5' splice-site region to further stabilize the complex. (b) A network of interactions commits splice-site partners 25. Protein-protein interactions can link U1 snRNP-ASF/SF2 complexes assembled on the 5' splice-site (a), with U2AF bound to the polypyrimidine tract [represented by a poly(U) stretch]. U2AF is a heterodimer: its 65 kDa subunit binds directly to the polypyrimidine tract and the 35 kDa subunit can establish RS domain-mediated contacts with SR proteins. (c) A network of interactions in the function of exon enhancers. Blue arrows represent the equi- librium for binding of U2AF to a weak polypyrimidine tract, and of an SR protein to an exon enhancer (represented by a grey box). Direct or in- direct interactions between these proteins results in cooperative stabilization of their association with the pre-mRNA, thereby promoting the use of the 3' splice-site 34,44.

proteins 4s. Both Tra and Tra-2 contain regions, which interact with the

regions of SR proteins 2s. Embedded within the repeats is a purine-rich el- ement that is sufficient to promote splicing in the absence of Tra and Tra-2 when located close to the 3' splice-site. Thus, the function of Tra and Tra-2 could be to promote the function of a weak splicing enhancer that cannot effi- ciently recruit SR proteins and activate 3' splice-sites at a distance.

Many of the activities of SR proteins in splicing regulation are antagonized

300

by other factors, the best characterized being hnPd~lP A1 ~efs 3, 46). However, no molecular model for how this antag- onism operates is available at present.

Role of phosphorylation The monoclonal antibody that al-

lowed the initial purification of the SR protein family recognizes a phospho- epitope in the ~ region 7, indicating that SR proteins are phosphorylated in vivo. ~though a number of kinases have been identified that phosphoryl- ate SR proteins in vitro ('Fable II), the

functions of these kinases in vivo are still unclear.

Recent studies have found that cycles of protein phosphorylation-dephos- phorylation occur during pre-mPd~IA splicing. Protein phosphatase inhibitors have been found to prevent catalytic ac- tivation of assembled spliceosomes 47, whereas an excess of protein phospha- tases inhibits spliceosome assembl~ 8. SR proteins are strong target candidates to mediate these effects 48. In this model, their phosphorylation would be required for spliceosome assembly, while their

REVIEWS TIBS 2 1 - A U G U S T 1 9 9 6

Table II. SR protein kinases

Name Features Substrates tested

Unknown a U1 snRNP-associated U1 7OK, ASF/SF2 SRPK1 b Cell-cycle regulated, causes redistribution of SR ASF/SF2, SC35

proteins in vivo and splicing inhibition in vitro Contains RS domains necessary for interaction

with SR proteins; overproduction causes redistribution of SR proteins in vivo

Specific inhibitors change the phosphorylation state of SR proteins in vivc

CIk/Sty r ASF/SF2

Topoisomerase I ~ ASF/SF2

aWoppmann, A. et al. (1993) Nucleic Acids Res. 21, 2815-2822; bGui, J. F., Lane, W. S. and Fu, X-D. (1994) Nature 369, 678-682; ~ K. T. et al. (1996) EMBO J. 15, 265-275; ~Rossi, F. et al. (1996) Nature 381, 80-83.

dephosphorylation would be necessary for the spliceosome to undergo catalysis. Phosphorylation of other RS-containing polypeptides, however, can also under- lie these effects, as has been shown for the U1 70K protein 49.

Remaining questions An issue with important phy]ogenetic

and mechanistic implications is whether SR proteins are present in all types of organisms that splice their mRNA pre- cursors, particularly in yeast. The fact that the yeast homologues of some RS- containing mammalian splicing factors do not contain RS domains suggests mechanistic differences between the two systems.

Although it is clear that RS domains can mediate protein-protein contacts, the biophysical basis, determinants of specificity and role of serine phosphoryl- ation in these interactions are not under- stood. Also unknown is whether phos- phorylation of SR proteins is entwined in signal transduction networks control- ling splicing, similar to those controlling transcription.

Another crucial question is to verify whether a tissue-specific configuration of relative SR protein concentrations acts as a code that dictates coordinated changes in multiple splicing patterns observed in vivo. If this is the case, we need to understand first how this code is deciphered by the general splicing machinery, and second how expression of the SR genes is regulated to achieve cell-specific SR protein configurations.

Answers to these central questions are likely to be elucidated in the next few years and will shed light on the workings of the splicing machinery.

Acknowledgements We thank F. Gebauer, R. Va|c~rcel and

members of the Green lab for comments on the manuscript. J. V. was supported by fellowships from EMBO and Spanish Ministerio de EducaciOn y Ciencia. Work in M. R. G.'s lab is supported by a grant from the NIH. M. R. G. is an Investigator of the Howard Hughes Medical Institute. We apologize to the many colleagues whose work has not been cited because of space constraints.

References 1 Ge, H. and Manley, J. L. (1990) Cell 62, 25-34 2 Krainer, A. R., Mayeda, A., Kozak, D. and

Binns, G. (1991) Cell 66,383-394 3 C~ceres, J. F., Stamm, S., Helfman, D. M. and

Krainer, A. R. (1994) Science 265, 1706-1709 4 Wang, J. and Manley, J. L. (1995) RNA 1, 335-346 5 Birney, E., Kumar, S. and Krainer, A. R. (1993)

Nucleic Acids Res. 21, 5803-5816 6 Fu, X-D. and Maniatis, T. (1992) Science 256,

535-538 7 Roth, M. B., Zahler, A. M. and Stolk, J. A.

(1991) J. Cell Biol. 115,587-596 8 Zahler, A. M., Lane, W. S., Stolk, J. A. and

Roth, M. B. (1992) Genes Dev. 6, 837-847 9 Neugebauer, K. M., Stolk, J. A. and Roth, M. B.

(1995) J. Cell. Biol. 129, 899-908 10 Fu, X-D. (1995) RNA 1, 663-680 11 Ring, H. Z. and Lis, J. T. (1994) Mol. Cell. Biol.

14, 7499-7506 12 Peng, X. and Mount, S. M. (1995) Mol. Cell.

Biol. 15, 6273-6282 13 Zahler, A. M., Neugebauer, K. M., Lane, W. S.

and Roth, M. B. (1993) Science 260,219-222 14 Screaton, G. R. et al. (1995) EMBO J. 14,

4336-4349

15 Kim, Y-J., Zuo, P., Manley, J. L. and Baker, B. S. (1992) Genes Dev. 6, 2569-2579

16 Fu, X-D. (1993) Nature 365, 82-85 17 Zahler, A. M. and Roth, M. B. (1995) Proc. Natl.

Acad. Sci. U. S. A. 92, 2642-2646 18 Kraus, M. E. and Lis, J. T. (1994) Mol. Cell. Biol.

14, 5360-5370 19 C&ceres, J. F. and Krainer, A. R. (1993) EMBO J.

12, 4715-4726 20 Zuo, P. and Manley, J. L. (1993) EMBO J. 12,

4727-4737 21 Tacke, R. and Manley, J. L. (1995) EMBO J. 14,

3540-3551 22 Zuo, P. and Manley, J. L. (1994) Prec. Natl.

Acad. Sci. U. S. A. 91, 3363-3367 23 Sun, Q. et al. (1993) Genes Dev. 7,

2598-2608 24 Lavigueur, A., La Branche, H., Kornblihtt, A. R.

and Chabot, B. (1993) Genes Dev. 7, 2405-2417

25 Wu, J. Y. and Maniatis, T. (1993) Cell 75, 1061-1070

26 Kohtz, J. D. et al. (1994) Nature 368, 119-124 27 Lee, C. G., Zamore, P. D., Green, M. R. and

Hurwitz, J. (1993) J. Biol. Chem. 268, 13472-13478

28 Li, H. and Bingham, P. M. (1991) Cell 67, 335-342

29 Jamison, S. F. et al. (1995) Nucleic Acids Res. 23, 3260-3267

30 Tam, W. Y. and Steitz, J. A. (1994) Genes Dev. 8, 2704-2717

31 Crispino, J. D., Blencowe, B. J. and Sharp, P. A. (1994) Science 265, 1866-1869

32 Crispino, J. D. and Sharp, P. A. (1995) Genes Dev. 9, 2314-2323

33 Zamore, P. D. and Green, M. R. (1991) EMBO J. 10, 207-214

34 Staknis, D. and Reed, R. (1994) Mol. Cell. Biol. 14, 7670-7682

35 Tam, W. Y. and Steitz, J. A. (1995) Proc. Natl. Acad. Sci. U. S. A. 92, 2504-2508

36 Roscigno, R. F. and Garcia-Blanco, M. A. (1995) RNA 1, 692-706

37 Wassarman, D. A. and Steitz, J. A. (1993) Proc. Natl. Acad. Sci. U. S. A. 90, 7139-7143

38 Eperon, I. C. et al. (1993) EMBO J. 12, 3607-3617

39 Watakabe, A., Tanaka, K. and Shimura, Y. (1993) Genes Dev. 7,407-418

40 Tanaka, K., Watakabe, A. and Shimura, Y. (1994) Mol. Cell. Biol. 14, 1347-1354

41 Ramchatesingh, J. et al. (1995) Mol. Cell. Biol. 15, 4898-4907

42 Hoffman, B. E. and Grabowski, P. J. (1992) Genes Dev. 6, 2554-2568

43 Wang, Z., Hoffman, H. M. and Grabowski, P. J. (1995) RNA 1, 335-346

44 Tian, M. and Maniatis, T. (1994) Genes Dev. 8, 1703-1712

45 Tian, M. and Maniatis, T. (1993) Cell 74, 105-114

46 Mayeda, A. and Krainer, A. R. (1992) Cell 68, 365-375

47 Mermoud, J. E., Cohen, P. and Lamond, A. I. (1992) Nucleic Acids Res. 20, 5263-5269

48 Mermoud, J. E., Cohen, P. T. W. and Lamond, A. I. (1994) EMBO J. 13, 5679-5688

49 Tazi, J. et al. (1993) Nature 363, 283-286

C o m p u t e r C o r n e r

Have you come across any applications (freeware or commercially available), CDs, servers or tips recently that might be of interest to other biochemists and molecular biologists? If so, why not let us know so that we can review them in Computer Cornet?.

Please contact: Jo McEn ty re

Tel: +44 1223 315961 Fax: +44 1223 464430

Emai l : tibs@elsevier.co.uk

301