Missense Mutations in hMLH1 and hMSH2 Are Associated with Exonic Splicing Enhancers

Abstract

There is a critical need to understand why missense mutations are deleterious. The deleterious effects of missense mutations are commonly attributed to their impact on primary amino acid sequence and protein structure. However, several recent studies have shown that some missense mutations are deleterious because they disturb cis-acting splicing elements—so-called “exonic splicing enhancers” (ESEs). It is not clear whether the ESE-related deleterious effects of missense mutations are common. We have evaluated colocalization of pathogenic missense mutations (found in affected individuals) with high-score ESE motifs in the human mismatch-repair genes hMSH2 and hMLH1. We found that pathogenic missense mutations in the hMSH2 and hMLH1 genes are located in ESE sites significantly more frequently than expected. Pathogenic missense mutations also tended to decrease ESE scores, thus leading to a higher propensity for splicing defects. In contrast, nonpathogenic missense mutations (polymorphisms found in unaffected individuals) and nonsense mutations are distributed randomly in relation to ESE sites. Comparison of the observed and expected frequencies of missense mutations in ESE sites shows that pathogenic effects of ⩾20% of mutations in hMSH2 result from disruption of ESE sites and disturbed splicing. Similarly, pathogenic effects of ⩾16% of missense mutations in the hMLH1 gene are ESE related. The colocalization of pathogenic missense mutations with ESE sites strongly suggests that their pathogenic effects are splicing related.

Missense mutations—nucleotide substitutions that change an amino acid in a protein—are among the most common types of mutations underlying inherited human diseases. The deleterious effects of missense mutations are usually attributed to their effects on protein function. However, recent studies of normal and alternative splicing suggest that the deleterious effects of nucleotide substitutions might, in fact, be splicing related when they are located in exonic splicing enhancers (ESEs) (Cartegni and Krainer 2002; Cartegni et al. 2002; Fackenthal et al. 2002; Moseley et al. 2002; Pollard et al. 2002). ESEs are discrete, degenerate motifs of 6–8 nts located inside exons (Liu et al. 1998; Blencowe 2000). The study of normal splicing suggests that most exons contain at least one functional ESE site (Blencowe 2000; Hastings and Krainer 2001; Cartegni et al. 2002). ESEs are target sequences for the family of conserved essential splicing factors—the serine- and arginine-rich (SR) proteins (Stojdl and Bell 1999; Graveley 2000; Hastings and Krainer 2001). ESEs play an important role in exon recognition. Nucleotide substitutions in ESEs can result in failure of SR proteins to bind to the ESE, which leads to failure of splisosome machinery to recognize the sequence as exonic and causes exon skipping (Ars et al. 2000; Cartegni et al. 2002; Fackenthal et al. 2002; Moseley et al. 2002). Each SR protein recognizes specific, albeit degenerate and partially redundant, sequence motifs. ESE motifs for four members of the SR family (SF2/ASF, SRp40, SRp55, and SC35) have been identified (Liu et al. 1998; Stojdl and Bell 1999; Graveley 2000; Liu et al. 2000). To identify the ESE motifs that are recognized by individual SR proteins, a PCR-based approach called “SELEX” (systematic evolution of ligands by exponential) enrichment was used. In this approach, a natural splicing enhancer in a minigene is replaced by short, random sequences derived from an oligonucleotide library. The generated pool of minigenes is transfected into cultured cells, and spliced mRNAs are amplified by RT-PCR and sequenced (Liu et al. 1998, 2000). On the basis of the frequencies of the individual nucleotides at each position, a score matrix for each nucleotide in each position was calculated. This score matrix can be used to predict SR protein–specific ESEs (ESEfinder).

We studied the association of pathogenic missense mutations (found in affected kindreds), nonpathogenic missense mutations (polymorphic mutations and sequence variants found in nonaffected individuals), nonsense mutations, and frameshifts (here restricted to 1- or 2-nt deletions and insertions) with ESE sites in hMSH2 and hMLH1, human mismatch repair genes that are related to human nonpolyposis colon cancer (HNPCC [MIM 114500]) (Peltomaki and Vasen 1997). We used published and our own data on pathogenic and nonpathogenic missense mutations in the hMSH2 and hMLH1 genes (tables A, B, C, D, and E [online only]). Only mutations found in independent families were used. We excluded multiple reported mutations found in the same family. The numbers of different types of mutations analyzed are shown in table 1.

Table A.

Pathogenic Missense Mutations Found in the hMSH2 Gene^[Note]

Nucleotide Position in ORF	ESE	Exon	Codon	Nucleotide	Nucleotide Change	Consequence	Type of Amino Acid Substitutiona	Geographic Origin	Family ID	Reference
4	0	1	2	4	G→A	Ala→Thr	C	None specified	…	Myriad et al.
182	1	1	61	182	A→C	Gln→Pro	C	Western Europe	…	Myriad et al.
308	1	2	103	308	A→G	Tyr→Cys	R	None specified	…	Myriad et al.
319	0	2	107	319	G→C	Ala→Prol	C	Czech Republic	PT523	A. Krepelova (unpublished)
380	0	3	127	380	A→G	Asn→Ser	C	None specified	…	Myriad et al.
380	0	3	127	380	A→G	Asn→Ser	C	Africa	…	Myriad et al.
380	0	3	127	380	A→G	Asn→Ser	C	Africa	…	Myriad et al.
380	0	3	127	380	A→G	Asn→Ser	C	None specified	…	Myriad et al.
380	0	3	127	380	A→G	Asn→Ser	C	None specified	…	Myriad et al.
435	1	3	145	435	T→G	Ile→Met	C	Czech Republic	Patient 338	A. Krepelova (unpublished)
435	1	3	145	435	T→G	Ile→Met	C	None specified	…	A. M. Deffenbaugh et al. (unpublished)
435	1	3	145	435	T→G	Ile→Met	C	Western European	…	Myriad et al.
446	0	3	149	446	G→A	Gly→Asp	R	None specified	…	A. M. Deffenbaugh et al. (unpublished)
505	1	3	169	505	A→G	Ile→Valb	C	Asia	…	A. M. Deffenbaugh et al. (unpublished)
593	1	3	198	593	GAA→GGA	Glu→Gly	R	Sweden	199	Unpublished
595	0	3	199	595	T→C	Cys→Arg	R	Hong Kong	[Turcot]	Yuen et al., Am J Pathol 153:1181–1188 (1998)
742	0	4	248	742	A→G	Lys→Glu	C	Western Europe	…	Myriad et al.
792	1	4	264	792	G→C	Gln→His	R	Western Europe	…	Myriad et al.
806	1	5	269	806	C→T	Ser→Leu	R	Latin America/Caribbean	…	A. M. Deffenbaugh et al. (unpublished)
815	1	5	272	815	C→T	Ala→Val	C	None specified	0108	MDACCc
913	1	5	305	913	G→A	Ala→Thr	C	Netherlands	NL-38	Wijnen et al. (unpublished)
965	1	6	322	965	G→Ab	Gly→Asp	C	Russia	1	Maliaka et al. Hum Genet 97:251–255 (1996)
965	1	6	322	965	G→Ab	Gly→Asp	C	Germany	186	Brieger et al. Clinical Chemistry 45: 1564–1567 (1999)
965	1	6	322	965	G→Ab	Gly→Asp	C	United States	0205	MDACCc
997	0	6	333	997	T→C	Cys→Arg	R	Nigeria	114-I-OL	A. de la Chapelle et al. (unpublished)
998	0	6	333	998	G→A	Cys→Tyr	R	Australia	IMS5	S. E. Bennett (unpublished)
1012	1	6	338	1012	G→A	Gly→Arg	R	None specified	…	A. M. Deffenbaugh et al. (unpublished)
1082	0	7	361	1082	A→G	Arg→Ser	R	Western Europe	…	Myriad et al.
1319	1	8	440	1319	T→C	Leu→Pro	R	Lithuania	O3	E. Avizienyte et al. (unpublished)
1508	0	9	503	1508	T→C	Leu→Pro	R	None specified	…	Myriad et al.
1516	1	10	506	1516	G→T	Asp→Tyr	R	Korea	SNU-YC13	Han et al., J. Natl. Cancer Inst. 88: 1317–1319 (1996)
1571	1	10	524	1571	G→C	Arg→Pro	R	None specified	Patient 2774 with ovarian cancer	Orth et al., Cold Spring Harbor Symp Quant Biol LIX: 349 (1994)
1600	1	10	534	1600	C→T	Arg→Cys	R	United States	3080	MDACCc
1774	1	12	592	1774	A→G	Met→Val	C	None specified	…	Myriad et al.
1807	0	12	603	1807	G→A	Asp→Asn	R	Finland	Patient 138	Salovaara et al., J Clin Oncol 18: 2193–2200 (2000)
1864	0	12	622	1864	C→A	Missense	R	Argentina	CRE001	Chialina et al. (unpublished)
1865	0	12	622	1865	C→T	Pro→Leu	R	New Zealand	J	Liu et al., Cancer Res. 54: 4590–4594 (1994)
1906	1	12	636	1906	G→Cb	Ala→Pro	C	France	CHI-S6351	B. Bressac-de Paillerets, unpublished
1906	1	12	636	1906	G→Cb	Ala→Pro	C	United States	54	A. de la Chapelle (unpublished)
1906	1	12	636	1906	G→C	Ala→Pro	C	Ashkenazi Jewish	MONO7 (y)	Yuan et al. J Med Genet 36:790–793 (1999)
1906	1	12	636	1906	G→C	Ala→Pro	C	None specified	…	Myriad et al.
2090	0	13	697	2090	G→T	Cys→Phe	R	Germany	56	Raedle et al. Gastroenterology 116:489A (1999)
2090	0	13	697	2090	G→T	Cys→Phe	R	Germany	62	Wehner et al., Hum Mut 10:241–244 (1997)
2164	1	13	722	2164	G→A	Val→Ile	R	United States	0024	MDACCc
2251	0	14	751	2251	G→A	Gly→Arg	R	France	BIL-IGR1924	B. Bressac-de Paillerets
2500	0	15	834	2500	G→A	Ala→Thr	C	Netherlands	NL57	J. T. Wijnen et al. (unpublished)
2714	1	16	905	2714	C→G	Thr→Arg	R	United Kingdom	C015	Froggatt et al., J. Med. Genet. 33:726–730 (1996)
2714	1	16	905	2714	C→T	Thr→Ile	R	Central/eastern Europe	…	Myriad et al.
2790	0	16	930	2790	A→Gb	Ile→Met	C	Spain	GE008	P. Hutter (unpublished) (1999)

Note.— The table enlists missense mutations from the HNPCC mutation database (accessed March 2003) and mutation data obtained by M.L.F. at The University of Texas M. D. Anderson Cancer Center (MDACC).

^aC = conservative; R = radical.

^bOf unknown pathogenicity.

^cThe University of Texas M. D. Anderson Cancer Center.

Table B.

Pathogenic Missense Mutations Found in the hMLH1 Gene^[Note]

Exon	ORF Position	ESE	Exonic Position	Initial Nucleotide	Mutated Nucleotide	Initial Amino Acid	Mutated Amino Acid	Type of Amino Acid Substitutiona	Family ID	Reference
1	69	0	69	A	T	Glu	Asp	C	…	Myriad et al.
1	73	0	73	A	T	Ile	Phe	R	166	W. Weber and M. Rodriguez-Bigas (unpublished)
1	74	0	74	T	C	Ile	Thr	R	40547	G. Norbury et al. (unpublished)
1	83	1	83	C	T	Pro	Leu	R	52	Wehner et al., Hum Mut 10:241–244 (1997)
1	85	0	85	G	T	Ala	Ser	C	…	Myriad et al.
1	104	0	104	T	G	Met	Arg	C	3	Tannergard et al., Cancer Res 55:6092–6096 (1995)
1	116	0	116	G	A	Cys	Ter	R	MDACC3014	MDACC
2	184	1	68	C	A	Gln	Lys	C	498	Wehner et al., Hum Mut 10:241–244 (1997)
2	191	1	75	A	G	Asm	Ser	C	2104	J. T. Wijnen et al. (unpublished)
2	199	1	83	G	A	Gly	Arg	C	MDACC3011	MDACC
2	199	1	83	G	A	Gly	Arg	C	MDACC3202	MDACC
2	199	1	83	G	A	Gly	Arg	C	304	Tannergard et al., Cancer Res 55:6092–6096 (1995)
2	199	1	83	G	A	Gly	Arg	C	1652	Tannergard et al., Cancer Res 55:6092–6096 (1995)
2	199	1	83	G	A	Gly	Arg	C	68	Heinimann et al., Cancer 85 (12):2512–2518 (1999)
2	199	1	83	G	A	Gly	Arg	C	JPN-28	Sasaki et al., Hum Mut 9:164 (1977}
2	199	1	83	G	A	Gly	Arg	C	MFI	Herfarth et al., Genes Chrom Cancer 18:42 (1997)
2	199	1	83	G	A	Gly	Arg	C	VSO15	Hutter et al., Int J Cancer 78:680–684 (1998)
2	200	1	84	G	A	Gly	Glu	R	MDACC413075	MDACC
2	200	1	84	G	A	Gly	Glu	R	MDACC380548	MDACC
2	203	0	87	T	A	Ile	Asn	R	18	Tannergard et al., Cancer Res 55:6092–6096 (1995)
3	230	0	23	G	A	Cys	Tyr	R	IMS2	S. E. Bennet (unpublished)
3	250	0	43	A	G	Lys	Glu	R	171	Lamberti et al., Gut 44:839–843 (1999)
3	277	0	70	A	G	Ser	Gly	R	1	Quaresima et al., Hum Mut 12; 433 (1998)
3	299	1	92	G	C	Arg	Pro	C	…	Myriad et al.
3	304	1	97	G	A	Glu	Lys	R	61
3	306	1	99	G	T	Glu	Asp	R	…	Myriad et al.
4	320	1	14	T	G	Ile	Arg	R	28/51/67	Nyström-Lahti et al., Hum Mol Genet 5:763–769 (1996)
4	350	1	44	C	T	Thr	Met	R	…	A. M. Deffenbaugh et al. (unpublished)
4	350	1	44	C	T	Thr	Met	R	431	Krpelova (unpublished)
4	350	1	44	C	T	Thr	Met	R	S.B.-2575	Stachow-Kurzawski (unpublished)
4	350	1	44	C	T	Thr	Met	R	26	Raedle et al. Gastro 116:A-489 (1999)
4	350	1	44	C	T	Thr	Met	R	LG	Liu et al., Nature Med 2:169–174 (1996)
4	350	1	44	C	T	Thr	Met	R	h83	D. J. Bunyan (unpublished)
4	350	1	44	C	T	Thr	Met	R	BES-SG624	B. Bressac-de Paillerets (unpublished)
4	350	1	44	C	T	Thr	Met	R	84	A. de la Chapelle (unpublished)
4	350	1	44	C	T	Thr	Met	R	5	Maliaka et al., Hum Genet 97:251–255 (1996)
4	350	1	44	C	T	Thr	Met	R	434	Buerstedde et al., J Med Genet 32:909–912 (1995)
4	350	1	44	C	G	Thr	Arg	C	…	Myriad et al.
5	382	1	3	G	C	Ala	Pro	C	338	V. Pensotti et al. (unpublished)
6	479	1	25	C	T	Ala	Val	R	…	A. M. Deffenbaugh et al. (unpublished)
7	554	0	9	T	G	Val	Gly	R	OZ-1	A. M. Deffenbaugh et al. (unpublished)
7	554	0	9	T	G	Val	Gly	R	…	Kohonen-Corish et al., Am J Hum Genet 59:818–824 (1996)
7	577	0	32	T	C	Ser	Pro	C	NA-G.C.	P. Izzo et al. (unpublished)
8	595	0	7	G	C	Glu	Gln	R	34	T. Liu et al. (unpublished)
8	637	0	49	G	A	Val	Met	C	39356	Myriad et al.
8	637	0	49	G	A	Val	Met	C	…	Myriad et al.
8	637	0	49	G	A	Val	Met	C	…	G. Norbury et al. (unpublished)
8	637	0	49	G	A	Val	Met	C	…	Myriad et al.
8	649	0	61	C	T	Arg	Cys	R	9	Miyaki et al., J Mol Med 73:515–520 (1995)
8	649	0	61	C	T	Arg	Cys	R	SNU-H1006	Han et al., J Natl Cancer Inst 88:1317–1319 (1996)
8	677	1	89	G	A	Arg	Gln	R	…	A. M. Deffenbaugh et al. (unpublished)
8	677	1	89	G	A	Arg	Gln	C	6	Maliaka et al., Hum Genet 97:251–255 (1996)
8	677	1	89	G	A	Arg	Gln	R	NLH-20	Wijnen et al., Am J Hum Genet 58:300–307 (1996)
9	731	1	55	G	A	Gly	Asp	R	311	V. Pensotti et al. (unpublished)
10	791	0	1	A	G	Arg	Asp	R	333633	G. Norbury et al. (unpublished)
10	793	1	3	C	T	Arg	CyS	R	198	P. Hutter (unpublished) (1999)
10	793	1	3	C	T	Arg	CyS	R	VD0004	T. Liu et al. (unpublished)
10	794	1	4	G	A	Arg	His	C	A-PD1	Viel et al., Genes Chromosomes Cancer 18:8–18 (1997)
10	803	0	13	A	G	Glu	Gly	R	80	Liu et al., Clin Genet 53:131–135 (1998)
10	814	0	24	T	G	Leo	Val	R	…	Myriad et al.
10	842	1	52	C	T	Ala	Val	R	…	A. M. Deffenbaugh et al. (unpublished)
11	977	1	94	T	C	Val	Ala	R	3273	Myriad et al.
11	977	1	94	T	C	Val	Ala	R	1515	Liu et al., Nature Med 2:169–174 (1996)
11	977	1	94	T	C	Val	Ala	R	…	Buerstedde et al., J Med Genet 32:909–912 (1995)
11	986	1	103	A	C	His	Pro	C	96	Lamberti et al., Gut 44:839–843 (1999)
12	1166	1	129	G	A	Arg	Gln	C	…	A. M. Deffenbaugh et al.
12	1217	0	180	G	A	Ser	Asn	C	209	Cunningham et al., Am J Hum Genet 69:780–790 (2001)
12	1321	1	284	G	A	Ala	Thr	R	466	Cunningham et al., Am J Hum Genet 69:780–790 (2001)
12	1321	1	284	G	A	Ala	Thr	R	…	A. Krepelova (unpublished)
13	1421	0	13	G	A	Arg	Gln	R	DES-SG407	B. Bressac-de Paillerets (unpublished)
13	1474	1	66	G	A	Ala	Thr	R	A4T	Möslein et al., Hum Mol Genet 5:1245–1252 (1996)
13	1517	0	109	T	C	Val	Ala	R	3118	Liu et al., Nature Med 2:169–174 (1996)
13	1517	0	109	T	C	Val	Ala	R	MDACC387858	MDACC
13	1517	0	109	T	C	Val	Ala	R	…	A. M. Deffenbaugh et al. (unpublished)
14	1569	0	12	G	T	Glu	Asp	R	…	Myriad et al.
14	1625	1	68	A	T	Gln	Leu	R	SNUH-H2	Han et al., Hum Mol Genet 4:237–242 (1995)
14	1646	0	89	T	C	Leo	Pro	R	SNU-H14	Han et al., J. Natl. Cancer Inst. 88:1317–1319 (1996)
14	1652	1	95	A	T	Asp	Thr	R	VS012	Hutter et al., Int J Cancer 78:680–684 (1998)
15	1693	0	27	A	T	Ile	Phe	R	VD001	Hutter et al., Int J Cancer 78:680–684 (1998)
15	1721	1	55	T	C	Leo	Pro	R	SNUH-H1	Han et al., Hum Mol Genet 4:237–242 (1995)
16	1733	0	2	A	G	Glu	Gly	R	21	A. de la Chapelle et al. (unpublished)
16	1744	0	13	C	G	Leu	Val	C	JPN-1	Han et al., Hum Mol Genet 4:237–242 (1995)
16	1771	0	40	A	G	Glu	Gly	R	END0003	Hutter et al., Int J Cancer 78:680–684 (1998)
16	1853	0	122	A	C	Lys	Thr	R	US-6	T. Caldes (unpublished)
17	1958	1	62	T	G	Leu	Arg	R	…	Han et al., Hum Mol Genet 14:237–242 (1995)
17	1961	1	65	C	T	Pro	Leu	R	38469	G. Norbury et al. (unpublished)
17	1963	1	67	A	G	Ile	Val	C	…	Myriad et al.
17	1976	1	80	G	C	Arg	Pro	C	NL56	Nyström-Lahti et al., Hum Mol Genet 5:763–769 (1996)
17	1976	1	80	G	C	Arg	Pro	C	7	A. de la Chapelle et al. (unpublished) (also mut. MLH1 exon 16, codon 618)
17	1976	1	80	G	A	Arg	Gla	C	51	J. T. Wijnen et al. (unpublished)
17	1988	1	92	A	G	Glu	Gly	R	CHI-SG277	B. Bressac-de Paillerets (unpublished)
18	2027	0	38	T	G	Leu	Arg	R	…	Myriad et al.
18	2041	0	52	G	A	Ala	Thr	R	CO10	Kurzawska and G. Kurzawski (unpublished)
18	2041	0	52	G	A	Ala	Thr	R	B.T.-1881	Froggatt et al., J. Med. Genet.33: 726–730 (1996)
19	2146	1	44	G	A	Val	Meth	C	GE111	Myriad et al.
19	2146	1	44	G	A	Val	Meth	C	…	Hutter et al., Int J Cancer 78:680–684 (1998)
19	2152	0	50	C	T	His	Tyr	R	CHO-SG632	D. J. Bunyan (unpublished)

Note.— The table enlists missense mutations from the HNPCC mutation database (accessed March 2003) and mutation data obtained by M.L.F. at MDACC.

^aC = conservative; R = radical.

Table C.

Database on hMSH2: Intragenic Polymorphisms and Sequence Variants^[Note]

Exon	Codon	Nucleotide Change	Consequence	ESE	Allele Frequency	Geographic Origin	Family ID	Reference
2	110	A→G at 329	Lys→Arg	0	.02	United States	HNPCC kindred	W. Weber and M. Rodriguez-Bigas (unpublished)
2	113	G→A at 339	Lys→Lys	0	…	Czech Republic	Patient 435	A. Krepelova (unpublished)
2	113	AAG→AAA	Lys→Lys	0	.0076		…	Liu et al. Clin Genet 53:131–135 (1998)
3	133	C→T at 399	Asp→Asp	0	…	African American	Patient 31	Johnson (unpublished)
3	153	C→T at 459	Ser→Ser	1	…	None specified	…	Möslein et al., Hum Mol Genet 5:1245–1252 (1996)
3	127	A→G at 380	Asn→Ser	0	.17		Blood donors	A. de la Chapelle et al. (unpublished)
6	322	G→A at 965	Gly→Asp	1	…	Czech Republic	Patients 338, 376, 461, 468, 491	A. Krepelova (unpublished)
6	328	GCC→GCT	Ala→Ala	0	…	France	Gue-SG618	A. Lindblom et al. (unpublished) (1999)
10	521	T→C at 1563	Tyr→Tyr	0	…	Czech Republic	Patient 424	A. Krepelova (unpublished)
11	556	T→C at 1666	Leu→Leu	0	.01	Germany	70	Wehner et al., Hum Mut 10:241–244 (1997)
11	579	A→G at 1737	Lys→Lys	0	.01	Netherlands	NL22	Wijnen et al., Am J Hum Genet 56:1060–1066 (1995)
11	585	T→C at 1755	Ser→Ser	1	.005	Finland	Finnish kindreds	Wahlberg et al. 74:134–137 (1997)
12	596	A→G at 1787	Asn→Ser	0	<.004	Italy	R-MD4	Viel et al., Genes Chrom Cancer 18:8–18 (1997)
13	713	G→C at 2139	Gly→Gly	1	.04	Finland	Finnish kindreds	Nyström-Lahti et al., Hum Mol Genet 5:763–769 (1996)
13	718	A→G at 2154	Gln→Gln	0	…	Lithuania	03	E. Avizienyte et al. (unpublished)

Note.— The table enlists missense mutations from the HNPCC mutation database (accessed March 2003).

Table D.

Database on hMLH1: Intragenic Polymorphisms and Sequence Variants^[Note]

Exon	Codon	Nucleotide Change	Consequence	ESE	Allele Frequency	Geographic Origin	Family ID	Reference
2	66	C→T at 198	Thr→Thr	0	.01	Germany	7	Wehner et al., Hum Mut 10:241–244 (1997)
8	219	A→G at 655	Ile→Val	1	.33	Italy	Italian kindreds	A. Piepoli (unpublished)
8	219	A→C at 655	Ile→Leu	1	.31	None specified	…	Möslein et al., Hum Mol Genet 5:1245–1252 (1996)
12	406	G→A at 1217	Ser→Asn	0	.015	Finland	Finnish patients with colorectal cancer	Wu et al., Genes Chrom Cancer 4:269–278 (1997)
16	618	AA→GC at 1852, 1853	Lys→Ala	0	.001	None specified	…	A. M. Deffenbaugh et al. (unpublished)
17	653	G→T at 1959	Leu→Leu	1	…	United Kingdom	36902	G. Norbury et al. (unpublished) (1999)
18	676	C→T at 2028	Leu→Leu	0	…	Czech Republic	Patient 386	A. Krepelova (unpublished)
19	718	C→T at 2152	His→Tyr	0	.144	African American	…	Kowalski et al., Genes Chromosomes Cancer 18:219–227 (1997)

Note.— The table enlists missense mutations from the HNPCC mutation database (accessed March 2003).

Table E.

ESE Motifs Found in the hMSH2 and hMLH1 Genes

	SF2/ASF ESE Motifs				SC35 ESE Motifs				SRp40 ESE Motifs				SRp55 ESE Motifs
	Position in ORF				Position in ORF				Position in ORF				Position in ORF
Gene	Left	Right	Motif	Score	Left	Right	Motif	Score	Left	Right	Motif	Score	Left	Right	Motif	Score
hMSH2	10	16	CAGCCGA	5.3	90	97	GACCACCA	4.5	67	73	TTTCAGG	5.0	8	13	TGCAGC	4.7
hMSH2	14	20	CGAAGGA	3.2	121	128	GACTTCTA	4.7	95	101	CCACAGT	3.6	39	44	CGCGGC	4.5
hMSH2	41	47	CGGCCGA	4.9	149	156	CGCTGCTG	3.3	111	117	CGACCGG	3.4	61	66	CGCTTC	3.0
hMSH2	44	50	CCGAGGT	3.7	203	210	GGCCGGCA	3.5	277	283	CTTCTGG	4.4	101	106	TGCGCC	3.5
hMSH2	96	102	CACAGTG	3.1	346	353	GATTGGTA	3.0	306	312	TTATAAG	3.1	129	134	TACGGC	4.7
hMSH2	134	140	CGCACGG	5.4	366	373	GGCTTCTC	3.2	368	374	CTTCTCC	3.2	136	141	CACGGC	3.9
hMSH2	159	165	CGCCCGG	4.4	369	376	TTCTCCTG	3.5	371	377	CTCCTGG	3.0	196	201	TACATG	3.2
hMSH2	164	170	GGGAGGT	3.4	456	463	GTCCGCAG	3.9	422	428	TGTCAGC	3.6	224	229	TGCAGA	4.3
hMSH2	179	185	CCCAGGG	3.5	469	476	GGCCAGAG	4.0	431	437	CCATTGG	3.3	236	241	TGCTTA	3.5
hMSH2	207	213	GGCAGGA	4.0	498	505	GGATTCCA	3.4	475	481	AGACAGG	4.5	255	260	TGAATC	3.4
hMSH2	459	465	CGCAGTT	3.1	520	527	GGACTGTG	3.8	507	513	ACAGAGG	3.6	284	289	TTCGTC	3.8
hMSH2	476	482	GACAGGT	4.1	531	538	ATTCCCTG	3.5	586	592	CCAAAGG	4.4	420	425	TATGTC	4.0
hMSH2	508	514	CAGAGGA	5.7	549	556	GTTCTCCA	3.4	601	607	TTACCCG	4.0	495	500	TGTGGA	3.4
hMSH2	587	593	CAAAGGA	3.7	584	591	GACCAAAG	3.3	614	620	AGACTGC	3.4	506	511	TACAGA	3.7
hMSH2	606	612	CGGAGGA	5.3	615	622	GACTGCTG	5.0	671	677	TCACAGA	3.9	561	566	TGAGGC	3.2
hMSH2	654	660	AAGAGGA	3.2	662	669	GAATTCTG	4.1	698	704	CCACAAA	3.0	758	763	TGAATA	3.0
hMSH2	672	678	CACAGAA	3.4	669	676	GATCACAG	4.0	700	706	ACAAAAG	3.1	801	806	TTCATC	3.2
hMSH2	754	760	CAGATGA	3.8	718	725	GACCTCAA	3.9	712	718	TATCAGG	3.0	912	917	TGCAGC	4.7
hMSH2	812	818	CTGCGGT	3.3	728	735	GGTTGTTG	3.0	720	726	CCTCAAC	3.5	989	994	TGAATA	3.0
hMSH2	842	848	CAGATGA	3.8	759	766	GAATAGTG	3.5	734	740	TGAAAGG	3.9	1107	1112	TGCAGA	4.3
hMSH2	871	877	CTGACTA	3.1	792	799	GGTTGCAG	3.7	758	764	TGAATAG	3.1	1115	1120	TGAGGC	3.2
hMSH2	1030	1036	CAGTGGA	3.3	803	810	CATCACTG	3.0	805	811	TCACTGT	3.7	1139	1144	TACTTC	3.2
hMSH2	1120	1126	CAGACTT	3.9	847	854	GATTCCAA	3.8	809	815	TGTCTGC	3.4	1142	1147	TTCGTC	3.8
hMSH2	1282	1288	CACCAGA	3.1	873	880	GACTACTT	3.1	872	878	TGACTAC	3.8	1494	1499	TGCAGC	4.7
hMSH2	1550	1556	CACAGTT	3.5	883	890	GACTTCAG	4.1	885	891	CTTCAGC	3.8	1570	1575	CGTGTA	3.4
hMSH2	1633	1639	CAGAAGA	3.8	906	913	GGATATTG	3.0	920	926	TCAGAGC	3.8	1598	1603	TTCGTA	3.5
hMSH2	1718	1724	CCCAGGA	4.3	942	949	GGGTTCTG	3.2	937	943	TTTCAGG	5.0	1762	1767	TATGTA	3.6
hMSH2	1826	1832	CTCACGT	5.1	957	964	TACCACTG	3.6	959	965	CCACTGG	5.7	1775	1780	TGCAGA	4.3
hMSH2	1984	1990	CAGATGT	3.5	964	971	GGCTCTCA	3.4	973	979	TCTCTGG	5.0	1830	1835	CGTGTC	3.8
hMSH2	2016	2022	GGGAGGT	3.4	971	978	AGTCTCTG	3.4	1006	1012	CCTCAAG	4.4	1855	1860	TATGTA	3.6
hMSH2	2099	2105	CAGAAGT	3.5	1148	1155	GATTCCCA	4.2	1029	1035	CCAGTGG	4.0	1905	1910	AGCATC	3.6
hMSH2	2206	2212	CTCAGGT	4.5	1311	1318	GACTCCTC	3.6	1047	1053	TCTCATG	3.3	1993	1998	CACATC	4.1
hMSH2	2253	2259	AAGAGGA	3.2	1515	1522	GGACCCTG	5.4	1066	1072	ATAGAGG	3.2	2091	2096	TGAGTC	4.0
hMSH2	2406	2412	CACAGCA	3.4	1539	1546	GGATTCCA	3.4	1084	1090	TTAGTGG	4.0	2112	2117	TGTGGA	3.4
hMSH2	2699	2705	CAGAAGA	3.8	1540	1547	GATTCCAG	4.3	1126	1132	TTACAAG	5.4	2119	2124	TGCATC	5.5
hMSH2					1638	1645	GAATGGTG	3.0	1186	1192	AGACAAG	3.9	2196	2201	TGCTTC	3.8
hMSH2					1722	1729	GGATGCCA	3.2	1201	1207	TTACAAG	5.4	2295	2300	TATATC	3.4
hMSH2					1849	1856	GTTCCATA	3.5	1222	1228	TATCAGG	3.0	2332	2337	TGCATG	3.8
hMSH2					2005	2012	GGCCCCAA	5.0	1252	1258	ATACAGG	5.0	2399	2404	TACATG	3.2
hMSH2					2064	2071	GGCCCAAA	3.7	1310	1316	TGACTCC	4.1	2401	2406	CATGTC	3.1
hMSH2					2078	2085	GTTTTGTG	3.3	1319	1325	TTACTGA	3.4	2406	2411	CACAGC	3.3
hMSH2					2106	2113	GTCCATTG	3.6	1338	1344	CTTCTCC	3.2	2490	2495	TGCAGA	4.3
hMSH2					2139	2146	GGCTGGTG	4.4	1348	1354	TTTCAGG	5.0	2707	2712	AACATC	3.0
hMSH2					2210	2217	GGTCTGCA	3.1	1458	1464	TGACTTG	3.2	2799	2804	TACGTG	3.8
hMSH2					2211	2218	GTCTGCAA	3.2	1544	1550	CCAGTGC	3.2
hMSH2					2282	2289	GGTTAGCA	3.2	1632	1638	CCAGAAG	3.7
hMSH2					2319	2326	GATTGGTG	3.5	1754	1760	CTTCAGG	4.6
hMSH2					2415	2422	CACCACTG	3.6	1771	1777	CCAATGC	3.3
hMSH2					2481	2488	GATTCATG	4.0	1815	1821	TGTCAGC	3.6
hMSH2					2657	2664	AGTTCCTG	3.9	1900	1906	TTAAAAG	3.9
hMSH2									1992	1998	CCACATC	3.1
hMSH2									2000	2006	TTACTGG	5.8
hMSH2									2145	2151	TGACAGT	3.1
hMSH2									2156	2162	TGAAAGG	3.9
hMSH2									2205	2211	CCTCAGG	4.9
hMSH2									2360	2366	TTACTGC	4.9
hMSH2									2398	2404	CTACATG	3.6
hMSH2									2405	2411	TCACAGC	5.5
hMSH2									2412	2418	ACTCACC	3.1
hMSH2									2414	2420	TCACCAC	3.2
hMSH2									2417	2423	CCACTGA	3.3
hMSH2									2443	2449	TATCAGG	3.0
hMSH2									2578	2584	TCGCAAG	3.1
hMSH2									2650	2656	ATTCAGG	3.9
hMSH2									2687	2693	TTACTGA	3.4
hMSH2									2698	2704	TCAGAAG	4.0
hMSH2									2711	2717	TCACAAT	3.4
hMSH2									2722	2728	TTAAAAC	3.0
hMSH2									2731	2737	CTAAAAG	3.5
hMSH2									2780	2786	TTTCACG	4.7
hMSH2									2795	2801	TTACTAC	4.4
hMLH1	12	18	GGCAGGG	3.2	48	55	GAACCGCA	3.1	20	26	TTATTCG	3.1	23	28	TTCGGC	3.0
hMLH1	59	65	CGGCGGG	3.6	163	170	GGCCTGAA	3.3	40	46	ACAGTGG	3.3	52	57	CGCATC	4.7
hMLH1	132	138	CACAAGT	3.7	180	187	GATCCAAG	3.2	131	137	CCACAAG	5.4	57	62	CGCGGC	4.5
hMLH1	195	201	CACCGGG	4.3	219	226	GGATATTG	3.0	139	145	ATTCAAG	3.3	77	82	AGCGGC	3.4
hMLH1	392	398	CAGATGG	3.1	237	244	GTTCACTA	4.3	188	194	ACAATGG	3.5	289	294	TATGGC	3.2
hMLH1	478	484	GCCACGA	3.3	258	265	GTCCTTTG	3.3	231	237	TGAAAGG	3.9	303	308	TGAGGC	3.2
hMLH1	698	704	GTGAGGA	3.0	292	299	GGCTTTCG	3.1	239	245	TCACTAC	4.7	316	321	AGCATA	3.3
hMLH1	842	848	CAGCCTA	3.2	312	319	GGCCAGCA	4.0	242	248	CTACTAG	4.8	327	332	TGTGGC	3.8
hMLH1	922	928	CACCCCA	3.0	330	337	GGCTCATG	4.9	262	268	TTTGAGG	3.2	372	377	TGCATA	5.2
hMLH1	950	956	TGCACGA	3.2	440	447	GGACCCAG	4.4	281	287	TTTCTAC	3.3	376	381	TACAGA	3.7
hMLH1	1141	1147	CACCAGA	3.1	614	621	GGACACTA	4.5	287	293	CCTATGG	3.1	627	632	TGCCTC	3.8
hMLH1	1144	1150	CAGATGG	3.1	639	646	GGACAATA	3.2	344	350	TTACAAC	4.6	652	657	TCCATC	3.2
hMLH1	1238	1244	CAGAGGA	5.7	794	801	GTCTGGTA	3.3	387	393	TTACTCA	3.1	733	738	TACATA	4.6
hMLH1	1264	1270	GGCAGGG	3.2	869	876	CATTCCTG	3.3	404	410	TGAAAGC	3.1	766	771	TGCATC	5.5
hMLH1	1376	1382	CAGAGAA	3.2	895	902	AGTCCCCA	3.0	468	474	TTACAAC	4.6	781	786	TTCATC	3.2
hMLH1	1383	1389	GAGAGGA	4.2	896	903	GTCCCCAG	4.8	616	622	ACACTAC	3.6	840	845	TGCAGC	4.7
hMLH1	1485	1491	CCCCCGG	3.1	962	969	GCATCCTG	3.2	623	629	CCAATGC	3.3	877	882	TACCTC	3.2
hMLH1	1486	1492	CCCCGGA	3.3	998	1005	AGCTCCTG	4.5	629	635	CCTCAAC	3.5	906	911	TGTGGA	3.4
hMLH1	1507	1513	CTCACTA	3.3	1006	1013	GGCTCCAA	4.7	671	677	TTAGTCG	3.7	961	966	AGCATC	3.6
hMLH1	1561	1567	CTCCGGG	3.1	1060	1067	GGCCCCTC	4.2	725	731	TGAATGG	3.7	977	982	TGCAGC	4.7
hMLH1	1622	1628	CACAGCA	3.4	1104	1111	GTCTTCTA	4.4	740	746	CCAATGC	3.3	985	990	CACATC	4.1
hMLH1	1720	1726	CTCAGGT	4.5	1110	1117	TACTTCTG	3.1	775	781	TTACTCT	3.1	1027	1032	TACTTC	3.2
hMLH1	1793	1799	CAGAGGA	5.7	1131	1138	GGTCTATG	3.8	793	799	CGTCTGG	3.9	1101	1106	CTCGTC	3.0
hMLH1	1868	1874	CAGACTA	4.2	1149	1156	GGTTCGTA	4.3	862	868	ACACACC	4.1	1110	1115	TACTTC	3.2
hMLH1	1997	2003	GGGACGA	4.3	1159	1166	GATTCCCG	4.6	893	899	TCAGTCC	3.2	1151	1156	TTCGTA	3.5
hMLH1	2075	2081	CTGAGGA	4.6	1193	1200	AGCCTCTG	3.9	926	932	CCACAAA	3.0	1155	1160	TACAGA	3.7
hMLH1	2177	2183	CACACAT	3.7	1205	1212	AACCCCTG	4.3	942	948	TCACTTC	3.2	1190	1195	TGCAGC	4.7
hMLH1	2202	2208	CACAGAA	3.4	1220	1227	AGCCCCAG	3.7	1000	1006	CTCCTGG	3.0	1392	1397	TACTTC	3.2
hMLH1	2204	2210	CAGAAGA	3.8	1227	1234	GGCCATTG	4.3	1017	1023	CTCCAGG	3.3	1434	1439	TGTGGA	3.4
hMLH1	2260	2266	GAGAGGT	3.8	1298	1305	AACTCCCA	3.0	1029	1035	CTTCACC	3.5	1473	1478	TGCAGC	4.7
hMLH1					1305	1312	AGCCCCTG	4.7	1045	1051	CTACCAG	3.4	1520	1525	TGAGTC	4.0
hMLH1					1320	1327	GGCTGCCA	4.2	1055	1061	TTGCTGG	3.2	1574	1579	TGCATA	5.2
hMLH1					1347	1354	GGATACAA	3.3	1064	1070	CCTCTGG	4.7	1601	1606	TGAATC	3.4
hMLH1					1362	1369	GACTTCAG	4.1	1085	1091	CCACAAC	4.6	1622	1627	CACAGC	3.3
hMLH1					1453	1460	GATTCCCG	4.6	1088	1094	CAACAAG	3.1	1625	1630	AGCATC	3.6
hMLH1					1470	1477	GACTGCAG	3.9	1112	1118	CTTCTGG	4.4	1824	1829	TGAATA	3.0
hMLH1					1523	1530	GTCTCCAG	4.6	1140	1146	CCACCAG	3.7	1866	1871	TGCAGA	4.3
hMLH1					1558	1565	GTTCTCCG	3.8	1186	1192	TTTCTGC	3.9	1958	1963	TGCCTA	3.5
hMLH1					1717	1724	GTTCTCAG	3.5	1222	1228	CCCCAGG	3.6	2053	2058	TCCATC	3.2
hMLH1					1775	1782	GTCCAGAG	3.4	1235	1241	TCACAGA	3.9	2068	2073	TACATA	4.6
hMLH1					1790	1797	GGACAGAG	3.0	1237	1243	ACAGAGG	3.6	2145	2150	TGTGGA	3.4
hMLH1					1804	1811	GGTCCCAA	4.4	1256	1262	TTTCTAG	4.2	2186	2191	TGCCTC	3.8
hMLH1					1805	1812	GTCCCAAA	3.1	1259	1265	CTAGTGG	3.7	2222	2227	TGCAGC	4.7
hMLH1					1838	1845	AGTTTCTG	3.1	1301	1307	TCCCAGC	3.1
hMLH1					1980	1987	AGCCACTG	4.2	1332	1338	TCAGAGC	3.8
hMLH1					2095	2102	GGCCAGCA	4.0	1349	1355	ATACAAC	3.6
hMLH1					2128	2135	AACTCCTG	4.1	1354	1360	ACAAAGG	3.7
hMLH1					2141	2148	GGACTGTG	3.8	1459	1465	CGAAAGG	3.6
hMLH1					2187	2194	GCCTCCTA	3.7	1469	1475	TGACTGC	4.4
hMLH1					2242	2249	GATCTATA	3.0	1508	1514	TCACTAG	5.5
hMLH1									1525	1531	CTCCAGG	3.3
hMLH1									1560	1566	TCTCCGG	3.5
hMLH1									1581	1587	CCACTCC	4.6
hMLH1									1608	1614	TCAGTGG	4.4
hMLH1									1629	1635	TCAAACC	3.6
hMLH1									1647	1653	TCTCAAC	3.9
hMLH1									1680	1686	CTACCAG	3.4
hMLH1									1719	1725	TCTCAGG	5.3
hMLH1									1792	1798	ACAGAGG	3.6
hMLH1									1915	1921	TTACCCC	3.1
hMLH1									1929	1935	TGACAAC	4.1
hMLH1									1975	1981	CGACTAG	4.3
hMLH1									1982	1988	CCACTGA	3.3
hMLH1									2019	2025	TGAAAGC	3.1
hMLH1									2074	2080	TCTGAGG	3.6
hMLH1									2090	2096	TCTCAGG	5.3
hMLH1									2130	2136	CTCCTGG	3.0
hMLH1									2176	2182	TCACACA	3.6
hMLH1									2201	2207	TCACAGA	3.9
hMLH1									2259	2265	TGAGAGG	3.8

Table 1.

Types and Numbers of Mutations Analyzed

	No. of Mutations
Gene	Pathogenic Missense	Nonsense and Frameshift	Nonpathogenic Missense	Total Mutations
hMSH2	50	81	17	148
hMLH1	99	68	8	175

First, we searched the coding regions of the genes for the presence of ESE motifs with ESEfinder software. To reduce the number of false-positive results, we used a more-stringent-than-recommended threshold value of 3.0 for all four types of ESE motifs. Potential ESE motifs found in the hMSH2 and hMLH1 genes are listed in table E (online only). We excluded ESEs in exon/exon boundaries, accounted for overlap between different ESEs by counting as a single ESE any segment containing two or more ESEs, and estimated the percentage of sequence that consists of ESE motifs for each entire gene and each exon. This estimate provided us with the proportion of mutations expected to be in ESE motifs under the null hypothesis that assumes that there is no association between pathogenic missense mutations and exonic splicing enhancers.

Different nucleotide substitutions in mutation databases for the hMSH2 and hMLH1 genes differ with respect to how many times they are reported in the databases. Some of them are listed only once, whereas others are reported several times (e.g., the C→T transition at position 350 in the hMLH1 gene is reported 11 times). Multiply reported mutations in the mutation databases originate from different families. Counting each reported mutation, we found that missense mutations are colocalized with ESEs (for hMSH2, χ²=11.8, df=1, P<.001; for hMLH1, χ²=7.9, df=1, P<.01) (fig. 1a). Alternatively, we also counted each separate mutation only once, no matter how often it was listed in the database. Again, we found that, in both genes, deleterious missense mutations are located in ESEs more frequently than expected (for hMSH2, χ²=9.4, df=1, P<.001; for hMLH1, χ²=4.3, df=1, P<.05). Counting each mutation only once eliminates all possibility that families are related but probably leads to a downward bias. Nucleotide substitutions can occur at any position in a coding region of a gene; yet only deleterious mutations that disturb important functional sites would lead to cancer and thus have a chance of being detected by screening of cancer-affected families. The more deleterious the mutation, the higher the chance it has of causing disease and of being detected. Therefore, deleterious mutations are expected to be most frequent in mutation databases (Martin et al. 2002; Olivier et al. 2002). Counting multiple mutations only once leads to loss of information and may cause downward bias by reducing the number of observations and by eliminating variation in the number of mutations between different mutant sites.

Figure 1.

a, Frequency of ESE-associated mutations. Horizontal lines show the expected frequencies of mutations calculated as a proportion of the sequence occupied by ESE motifs. Differences between the expected and observed frequencies of pathogenic missense mutations (PM) were significant for both hMSH2 and hMLH1 (marked by an asterisk [*]). Nonpathogenic missense mutations (NP) show a trend to be underrepresented in ESEs. Nonsense and frameshift mutations (NF) did not show preferential association with ESE sites. Bars represent SEs of the frequencies. b, Positions of pathogenic missense mutations with respect to the 5′ and 3′ ends of an exon. The last 20 nts of exons had the highest frequency of mutations in ESEs: 83% of the mutations were in ESE sites (marked by a double asterisk [**]). The difference between the observed and expected numbers of mutations in that region was highly significant (χ²=33.8; df=1; P<.001). When the analysis was limited to short (average size, 80 nts) exons, the colocalization of the pathogenic missense mutations with ESE sites was even higher: 96% of the mutations were in ESEs.

There may be two possible explanations for missense mutations in hMSH2 and hMLH1 genes being preferentially located in ESE motifs: either missense mutations arise more frequently in ESEs, or mutations in ESEs are more pathogenic than mutations outside ESEs and therefore are more likely to be detected during screening of affected individuals. Our analysis favors the second explanation. If a missense mutation becomes pathogenic because it is located in an ESE, then one can expect pathogenic missense mutations to be located in ESE sites more frequently, compared with nonpathogenic ones. We found that 57% (28/49) of pathogenic missense mutations were located in ESEs of the hMSH2 gene versus 24% (4/17) of nonpathogenic mutations (χ²₁=5.66; P=.02; Fisher's exact test). For the hMLH1 gene, we also found that pathogenic missense mutations were more frequently located in ESE sites than nonpathogenic mutations—58% (57/99) versus 38% (3/8), respectively (χ²₁=1.21; P=.30; Fisher's exact test)—but there were too few nonpathogenic mutations to draw a meaningful conclusion.

If deleterious effects of nucleotide substitutions located inside ESE motifs result from disruption of splicing, then even missense mutations that are unlikely to affect protein structure (e.g., mutations that do not change the type of amino acid) will have a chance to be deleterious because they disturb ESEs and, therefore, splicing. This idea is supported by stratified analysis. We stratified missense mutations into “conservative” and “radical” (classifying them according to specifications of Dagan et al. [2002]), and we found that, in both genes, missense mutations located outside ESE sites tended to be “radical,” strongly affecting protein-structure mutations, whereas those located in ESE motifs are more likely to be “conservative” mutations that have no or slight effect on protein structure. For hMSH2 genes, the frequency of conservative missense mutations located in ESEs is 0.61±0.09, whereas the frequency of conservative missense mutations outside ESEs is 0.50±0.10. For the hMLH1 gene, we found the same trend: the frequency of conservative missense mutations in ESEs is 0.31±0.06, whereas the frequency of conservative mutations outside ESEs is 0.21±0.10. For both genes, the differences in the proportions of conservative missense mutations located inside and outside of ESE sites are not significant, even after we combine the data for both genes, probably because of the relatively low number of mutations in analysis.

The correlation of different types of mutations with ESE sites can be explained as follows: nonsense and frameshift mutations always produce truncated nonfunctional proteins and therefore always—no matter where they are located with respect to ESEs—are sufficiently damaging to cause disease. Thus, truncating mutations have a high chance of being detected by screening affected families. Missense mutations, especially those located outside important functional domains, may not change protein structure sufficiently to be pathogenic. However, if a nucleotide substitution occurs in a functional ESE site, it could disturb normal splicing and be sufficiently deleterious to cause disease. This could explain why affected individuals are enriched with missense mutations that are located in ESE sites and why polymorphisms found in unaffected individuals are not associated with ESEs and even show a trend not to be localized there.

Different single-nucleotide substitutions can change the ESE score in different directions: some substitutions increase the score, whereas others decrease it. If nucleotide substitutions located in ESE sites are deleterious because they disturb functional ESE sites, then such substitutions are expected mainly to decrease ESE scores. We compared the observed and expected proportion of score-decreasing missense mutations located inside ESE sites. The expected proportion of score-decreasing substitutions was calculated on the basis of all possible substitutions in the ESEs that lead to missense mutations. We found that ESE-located missense mutations reported in hMSH2 and hMLH1 mutation databases decrease ESE scores significantly more frequently than one would expect. For the hMSH2 gene, we found that the expected frequency of score-decreasing mutations is 0.77±0.03, whereas the observed frequency of score-decreasing mutations is 0.96±0.04. The differences are highly significant (χ²=6.5; df=1; P<.01). A similar result was obtained for the hMLH1 gene: the expected frequency of score-decreasing mutations is 0.78±0.02, whereas the observed frequency of score-decreasing mutations is 0.91±0.06 (χ²=5.9; df=1; P<.01).

The excess of pathogenic mutations in ESE sites compared with the expected frequency provides a minimal estimate of the proportion of missense mutations, the pathogenic effects of which are ESE related. As an upper limit for the estimate of the proportion of ESE-related mutations, one can suggest that all pathogenic missense mutations located in ESE sites are deleterious because they disturb functional splicing enhancers. This approach is likely to overestimate the proportion of ESE-related pathogenic mutations. First, not all ESE motifs are actual functional splicing enhancers (Cartegni 2002). Second, not all nucleotide substitutions in functional ESEs disturb their function (Cartegni and Krainer 2002; Fackenthal et al. 2002; Moseley et al. 2002; Pollard et al. 2002; ESEfinder). For the hMSH2 gene, the observed frequency of pathogenic missense mutations in ESEs is 55%, whereas the expected frequency is 36%. This means that 20%–55% of missense mutations in hMSH2 are pathogenic, because they affect ESE sites and therefore disturb normal splicing. A similar reasoning shows that the frequency of ESE-related mutations in the hMLH1 gene is 16%–58%.

Although an exon usually has several ESE motifs, the splicing machinery does not use most of them (Cartegni 2002; ESEfinder). The question is whether the functional ESE sites are distributed randomly. If functional ESEs are preferentially located in some specific regions within an exon, then the association of the pathogenic missense mutations with ESEs will be higher in that region. A study of the molecular mechanisms of splicing suggests that functional ESE sites occupy specific positions relative to the 5′ or 3′ ends of an exon (Blencowe 2000; Hastings and Krainer 2001; Cartegni et al. 2002). We compared the expected and observed frequencies of pathogenic missense mutations in four regions: the first 20 nts located near the 5′ end of an exon, nts 21–40 near the 5′ end of an exon, the first 20 nts near the 3′ end of an exon, and nts 21–40, starting from the 3′ exon region (fig 1b). Because the number of missense mutations located in these specific regions is relatively low, we combined the data on both the hMLH1 and hMSH2 genes. We found that >80% of pathogenic missense mutations that are located in the last 20 nts of exons (especially in short exons—80 nts, on average) strongly colocalize with ESEs (fig. 1b). This finding suggests that functional ESEs are preferentially located near the 3′ ends of exons. However, since we used mostly short exons (∼80 nts), it is noteworthy that, in fact, the functional ESEs are located 60–65 nts from the 5′ ends of exons.

Aside from the SR proteins that we have studied here, other classes of exonic enhancers exist; for example, those driven by hnRNP proteins (Chabot et al. 2003). In future research, studies of the frequency of mutations that affect other ESE sites would be of interest, once algorithms have been developed for evaluating the effect that mutations have on splicing related to these proteins. In summary, our analyses provide compelling evidence that many missense mutations associated with deleterious effects for hMLH1 and hMSH2 affect splicing. Of note, we found that conservative mutations that heretofore may not have had an obvious role in causing HNPCC may disrupt splicing. Further studies to evaluate the isoforms of mRNAs in individuals with missense mutations in ESE sites would help to confirm the role of missense mutations in disease causation and to provide clinical insight into the significance of these mutations.