Defective importin β recognition and nuclear import of the sex-determining factor SRY are associated with XY sex-reversing mutations

  1. Vincent R. Harley*,†,
  2. Sharon Layfield‡,
  3. Claire L. Mitchell§,
  4. Jade K. Forwood,∥,
  5. Anna P. John,
  6. Lyndall J. Briggs,
  7. Sharon G. McDowall‡, and
  8. David A. Jans,∥,**
  1. *Prince Henry's Institute of Medical Research, Melbourne 3168, Victoria, Australia; ‡Howard Florey Institute, Parkville 3010, Victoria, Australia; §Cytopia Proprietary Ltd., P.O. Box 6492, St. Kilda Road Central, 3008, Victoria, Australia; John Curtin School of Medical Research, Canberra 2601, Australian Capital Territory, Australia; ∥Department of Biochemistry and Molecular Biology and **Australian Research Council Centre of Excellence for Biotechnology and Development, Monash University, Clayton 3152, Victoria, Australia

  1. Edited by Barbara J. Meyer, University of California, Berkeley, CA (received for review January 8, 2003)

 
Next Section

Abstract

The architectural transcription factor SRY (sex-determining region of the Y chromosome) plays a key role in sex determination as indicated by the fact that mutations in SRY are responsible for XY gonadal dysgenesis in humans. Although many SRY mutations reduce DNA-binding/bending activity, it is not clear how SRY mutations that do not affect interaction with DNA contribute to disease. The SRY high-mobility group domain harbors two nuclear localization signals (NLSs), and here we examine SRY from four XY females with missense mutations in these signals. In all cases, mutant SRY protein is partly localized to the cytoplasm, whereas wild-type SRY is strictly nuclear. Each NLS can independently direct nuclear transport of a carrier protein in vitro and in vivo, with mutations in either affecting the rate and extent of nuclear accumulation. The N-terminal NLS function is independent of the conventional NLS-binding importins (IMPs) and requires unidentified cytoplasmic transport factors, whereas the C-terminal NLS is recognized by IMPβ. The SRY-R133W mutant shows reduced IMPβ binding as a direct consequence of the sex-reversing C-terminal NLS mutation. Of the N-terminal NLS mutants examined, SRY-R62G unexpectedly shows a marked reduction in IMPβ binding, whereas SRY-R75N and SRY-R76P show normal IMPβ binding, suggesting defects in the IMP-independent pathway. We conclude that SRY normally requires the two distinct NLS-dependent nuclear import pathways to reach sufficient levels in the nucleus for sex determination. This study documents cases of human disease being explained, at a molecular level, by the impaired ability of a protein to accumulate in the nucleus.

Mammalian sex is determined by a dominant gene on the Y chromosome known as SRY (sex-determining region of the Y chromosome) (1). In XY humans, mutations in SRY cause male-to-female sex reversal (2). These “XY female” patients have complete gonadal dysgenesis, which results in female genitalia with no testis differentiation (3). In the presence of Sry, the supporting cells of the indifferent gonad become testicular Sertoli cells; without Sry, the supporting cells become ovarian follicular cells (4).

SRY contains a conserved 80-aa DNA-binding domain, the high-mobility group (HMG) box. The majority of human SRY mutations reside in the HMG box, highlighting its importance for proper SRY function during testis formation. No in vivo targets for SRY have been defined to date, and thus the mechanism by which SRY initiates testis differentiation remains to be elucidated. Both weak repression and weak activation by human SRY have been reported in model transactivation assays in cultured cells (5–8). Many sex-reversing SRY mutations affect DNA-binding/bending activities of the mutant SRY proteins in vitro. However, in several cases, it is not clear how the sequence alteration in XY females contributes to disease. For example, SRY-F109S, SRY-M64T, SRY-M67V, SRY-S18N, and SRY-type HMG box (SOX)9-A20V mutations do not affect DNA-binding/bending activities (9, 10, 12, ††).

How SRY is transported into the nucleus is only partly defined. SRY contains two distinct nuclear localization signals (NLSs) at either end of the HMG box (Fig. 1a). Either NLS is sufficient to target the large heterologous protein β-galactosidase (β-gal) to the nucleus (13). Although the C-terminal NLS (cNLS) is recognized by the nuclear import receptor protein importin (IMP)β1 (14), little is known regarding the nuclear import pathway mediated by the N-terminal NLS (nNLS); this NLS is not recognized by IMPα or -β (14). Both NLSs are highly conserved in SRY among mammals (Fig. 1a) and are believed to be required for complete nuclear localization (13, 15).

Fig. 1.
View larger version:
Fig. 1.

(a) Sequence conservation of the NLSs of SRY HMG domains. The nNLSs and cNLSs are highly conserved with basic residues (in bold) almost invariant between species. Arrows indicate the position of SRY mutations in four patients with XY gonadal dysgenesis; each substitution was of conserved arginines (R62G, R75N, R76P, and R133W). (b) Immunostaining of wild-type and mutant SRY proteins. COS-7 cells were transiently transfected with plasmids encoding wild-type or mutant FLAG epitope-tagged SRY. Mutant proteins identified in XY gonadal dysgenesis patients demonstrated partly cytoplasmic staining, whereas wild-type SRY showed strictly nuclear staining. Images were derived by using CLSM at ×60 magnification 2 days posttransfection. (c) Extent of nuclear accumulation of SRY mutants over a 3-day period. CLSM images were analyzed to determine the Fn/c. Results (mean ± SEM, for n > 22) relative to SRYwt (Fn/c of 15) are shown for SRYR62G, SRYR133W, and SRYR62G/R133W compared with SRYwt in one series (Left) and R75N and R76P to SRYwt in another (Right). Significant differences between wild type and mutants are indicated by * (P < 0.05), ** (P < 0.01), and *** (P < 0.001).


It has been hypothesized that 46,XY gonadal dysgenesis in patients carrying mutations causing the amino acid substitutions within either of the NLSs of SRY could arise through reduced efficiency of the mutant proteins to localize in the nucleus, thereby reducing their nuclear role in DNA binding. The present study tests this theory by elucidating the underlying biochemical defects in four such patients (16, 17) and documents cases of human disease being explained, in molecular terms, by the impaired ability of a protein to accumulate in the nucleus. This finding is thus in contrast to diseases involving atrophin-1 and Huntingtin in neurodegenerative disorders that derive from increased protein nuclear accumulation (18).

Previous SectionNext Section

Materials and Methods

Patient Reports. All four patients have 46,XY gonadal dysgenesis and a nucleotide transversion changing a single amino acid. R62G (16) had streak gonads, gonadoblastoma, a de novo mutation. R75N (17) had streak gonads, no tumors, and a normal brother suggesting a de novo mutation. For R76P (19), no further clinical details are known. Two unrelated families (sporadic cases) carrying the R133W mutation have been described. In one family, the father was not studied, and the affected had streak gonads (lacking ovarian tissue) and bilateral gonadoblastoma (16). In the second, a de novo mutation, streak gonads, was reported (20).

Plasmid Construction. For expression of full-length SRY, a forward primer including a Kozak consensus sequence, a FLAG epitope, and reverse primer amplified a PCR fragment cloned into pcDNA3. For expression of SRY HMG domains, pT7-7 SRY HMG plasmid was used (21). Mutations in expression plasmids were introduced by using GeneEditor (Promega) and confirmed by DNA sequencing.

Preparation of SRY HMG Domains. Wild-type and mutant SRY HMG domains were expressed in Escherichia coli strain BL21 (DE3) (10) and purified by FPLC as described (22).

Cell Culture. COS-7 and hepatoma tissue culture (HTC) cells were cultured as described (23).

Transient Cell Transfections. Lipofectamine–DNA complexes were formed and incubated with COS-7 cells according to manufacturer instructions (GIBCO/BRL).

Immunocytochemistry. Immunocytochemistry with a mouse anti-FLAG antibody was carried out essentially as described (22). Cells were imaged by confocal laser scanning microscopy (CLSM, Bio-Rad MRC-500), and image analysis was performed by using NIH IMAGE 1.60 public-domain software (24).

Preparation of SRY NLS-β-Gal Proteins and GST-IMPs. Oligonucleotides encoding the nNLSs and cNLSs of SRY were annealed and ligated into SmaI-digested β-gal fusion protein expression vector pPR2. Fusion protein was induced with 1 mM isopropyl β-D-thiogalactoside, and protein was purified by affinity chromatography and labeled with 5-iodoacetamidofluorescein as described (23). Mouse IMPα-GST and IMPβ-GST were expressed as described (24). Their functionality in terms of mediating nuclear import has been demonstrated (25, 26).

Nuclear-Transport Assays. Analysis of nuclear import kinetics at the single-cell level by using mechanically perforated (in vitro) or microinjected (in vivo) HTC cells, in conjunction with CLSM, was as described (23). NLS-dependent nuclear protein import can be reconstituted in the former system through the addition of exogenous cytosolic extract and an ATP-regenerating system (23, 24). Image analysis and curve fitting were as described (23, 24).

ELISA. An ELISA-based binding assay was used to examine the binding affinities of IMPα and β-GST fusion proteins and the SRY NLS-containing peptides (nNLS, qdrvKRpmnafivwsRdqRRKmagc; cNLS, hReKypnyKyRpRRKaKmlgc) or HMG domains, or mutant derivatives thereof, as described (23, 24). The simian virus 40 large-tumor antigen NLS peptide (large-tumor antigen amino acids 111–132 cgpgsddeaaanaqhaapPKKKRKVgy) was used as a control.

Electrophoretic Mobility-Shift Assays (EMSAs). The oligonucleotide used containing the SRY consensus-binding site (underlined) comprised the upper strand sequence GGG TTA ACG TAA ACA ATA AAT CTG GTA GA. Complementary oligonucleotides were annealed and end-filled by using superscript reverse transcriptase in the presence of [α-32P]dCTP. EMSAs were performed as described by using either 35S-labeled in vitro-translated full-length SRY or recombinant HMG domains (10). For competition experiments, 4 ng of SRY HMG box was incubated with varying amounts of either IMPα and/or IMPβ before the addition of 20 fmol of SRY DNA consensus probe.

Circular Permutation Assays. Circularly permutated probes bearing the SRY-binding site were isolated and labeled as described (10). Probes (4–16 fmol) were mixed with crude E. coli extract containing SRY wild type (SRYwt), SRYR62G, or SRYR133W HMG domains in binding buffer (27). Products were resolved by electrophoresis, and bend parameters were determined as described (28).

Previous SectionNext Section

Results

Nuclear Accumulation of Wild-Type and Mutant SRY. Four patients with XY gonadal dysgenesis carry SRY mutations in either one of the SRY NLSs (Fig. 1a). To assess their effect on SRY nuclear import, plasmids encoding the full-length SRY and mutants were transiently expressed in COS-7 cells (Fig. 1b). Subcellular localization of SRY protein was determined 1–3 days after transfection by using CLSM and quantitated by using image analysis (Fig. 1c). Whereas wild-type SRY efficiently accumulated in the nucleus, all four mutants showed significant cytoplasmic staining after 2 days. A drastic reduction in nuclear accumulation was observed after 1 day for R62G, R75N, and the engineered mutant R62G/R133W with a mutation in both NLSs. Both R76P and R133W mutants showed milder defects with nuclear accumulation almost that of wild-type SRY by 3 days.

Nuclear Import Properties of SRY NLSs. Next we measured the individual nuclear-localizing ability of each SRY NLS. Recombinant proteins containing the SRY nNLS or cNLS fused to β-gal (SRY-nNLS-β-gal and SRY-cNLS-β-GAL) were affinity-purified, fluorescently labeled, and tested for their nuclear import ability either in vitro or in vivo by using established assays (24, 29). The maximal nuclear/cytoplasmic ratio (Fn/c) was ≈2 for both nNLS-β-gal and cNLS-β-gal, with half-maximal accumulation being achieved at ≈12.0 and 10.5 min, respectively, in vitro in the presence of cytosol (Fig. 2 a and b) and ≈1.2 and 1.7 min, respectively, in vivo (Fig. 2c). The SRY mutation-carrying R62G nNLS-β-gal and R133W cNLS-β-gal proteins, in contrast, were completely excluded from the nucleus (Fn/c < 0.3) either in the absence or presence of exogenous cytosol, indicating that the mutations abolished NLS function. Thus, each NLS of SRY can function independently, with NLS mutations directly reducing nuclear accumulation.

Fig. 2.
View larger version:
Fig. 2.

Kinetic parameters for nuclear uptake of β-gal fusion proteins in mechanically perforated (a and b) and microinjected (c) HTC cells. (a) Results shown are from a single typical experiment representative of a series of similar experiments where each data point represents six separate measurements of each Fn and Fc with background subtracted, where the SEM for the individual measurements was not more than 12% of the value of the mean. Data were fitted for the function Fn/c = Fn/cmax (1 - e - kt), where Fn/cmax is the maximal level of nuclear accumulation, k is the rate constant, and t is the time in min. The time at which nuclear accumulation is half-maximal (t 1/2) is 12 ± 3.7 min for wild-type (wt) SRY nNLS-β-gal in the presence of cytosol. Nuclear accumulation of a conventional IMPα/β-recognized large-tumor antigen-NLS-β-gal fusion protein in the same system (not shown) gave an Fn/cmax of >4 on the same day, with accumulation half-maximal at 6 min. In contrast, β-gal alone was excluded from the nucleus (Fn/cmax of ≈0.4, not shown). (b) The t 1/2 for wild-type SRY cNLS-β-gal is 10.5 ± 3.1 and 8.3 ± 1.7 min in the presence or absence of cytosol, respectively. CLSM images are shown for cells after 35 min of incubation. (c) In vivo studies of nuclear uptake of wild-type SRY nNLS-β-gal and cNLS-β-gal, compared with that of β-gal alone, in microinjected HTC cells. The t 1/2 values for wild-type SRY nNLS-β-gal and cNLS-β-gal nuclear accumulation were 1.2 and 1.7 min, respectively.


Conventional NLS-mediated nuclear protein import depends on cytosolic factors including IMPs and the monomeric GTP-binding protein Ran. To investigate whether the NLSs of SRY show this dependence, the nuclear import of nNLS-β-gal and cNLS-β-gal was assessed in vitro in the presence or absence of exogenously added cytosol. In the absence of cytosol, nuclear accumulation of nNLS-β-gal was prevented (Fig. 2a), whereas cNLS-β-gal still showed good accumulation (Fig. 2b). Although nuclear import mediated by the cNLS thus seems to be independent of cytosolic factors, it should be noted that residual IMPβ bound to the nuclear envelope can mediate nuclear import of proteins such as insulin-like growth factor binding protein 3 (IGFBP-3) in this assay system (30).

Recognition by IMPs. To characterize SRY nuclear import further, we tested whether SRY NLS peptides and HMG domain are recognized by IMPα, IMPβ, or the IMPα/β heterodimer using an ELISA-based assay (Table 1). Both NLS peptides were recognized poorly by IMPα; in contrast, the cNLS peptide was recognized by IMPβ with a high affinity (K d = 8 nM) similar to that of IMPα/β (K d = 5 nM), suggesting that SRY is recognized by IMPβ at a site distinct from that involved in interaction with IMPα. The nNLS peptide showed negligible binding by either IMPβ (K d = 140 nM) or IMPα/β (K d = 95 nM). Thus, the SRY nNLS does not seem to be recognized by IMPs, whereas the SRY cNLS is recognized by IMPβ rather than the conventional NLS-recognizing IMPα subunit.

View this table:
Table 1. IMP binding of normal and mutant SRY NLS peptides and HMG domains

The SRY HMG domain is recognized with higher affinity by IMPβ (K d = 3.1 nM) than is the cNLS peptide alone (K d = 8 nM), suggesting that regions outside the cNLS influence binding. Similar to the cNLS, the HMG domain was recognized poorly by IMPα (K d = 112 nM), although IMPβ recognition was slightly increased in the presence of IMPα (K d = 2.4 nM). Thus, SRY joins a less common but growing list of arginine-rich NLS-containing proteins recognized by IMPβ rather than IMPα (14, 22, 29, 31–33) that include the Rev (RqaRRnRRRRwR) from HIV type 1 (32, 34), the Rex protein of human T cell leukemia virus type 1 (mpKtRRRpRRsqRKRppt) (33), and the AP-1 family transcription factor CREB (26).

The cNLS peptide carrying the R133W mutation showed reduced IMPβ binding affinity (K d = 12 nM) when compared with the wild-type cNLS peptide (K d = 8 nM). Similarly, SRY HMG domain carrying the R133W substitution showed a reduction in IMPβ binding (K d = 5.2 nM) when compared with wild-type HMG domain (K d = 3.1 nM), which suggests that the R133W mutation reduces IMPβ binding by directly affecting the ability of the cNLS to be recognized by IMPβ. SRY HMG domain carrying the nNLS substitution R62G showed a large reduction (5-fold) in IMPβ binding (K d = 16 nM) as well as reduced maximal binding (2-fold), which was unexpected given that the wild-type nNLS peptide was poorly recognized by IMPβ and thus suggests indirect effects possibly arising from the close proximity of the NLSs in solution (see Fig. 3c Top).

Fig. 3.
View larger version:
Fig. 3.

(a) DNA binding of SRYwt, SRYR62G SRYR133W, and SRYR62G/ R133W. (Upper) An autoradiograph of in vitro-translated 35S-labeled SRYwt, SRYR62G and SRYR133W proteins analyzed by 12% SDS/PAGE, showing a single 30-kDa band. Luc, luciferase (negative control). (Lower) A typical single EMSA experiment demonstrating SRYwt, SRYR62G, and SRYR133W (860 pg) binding to the SRY DNA-binding consensus site AACAAT. Luciferase (lane 1) is included as a negative control. Products of the binding assay were analyzed on a 6% native polyacrylamide gel in 0.5× TBE buffer (90 mM Tris/64.6 mM boric acid/2.5 mM EDTA, pH 8.3), dried, and autoradiographed for 5 h at -70°C. SRYwt and SRY-R133W proteins form DNA complexes (arrow), but SRY-R62G does not. (b) DNA bending by wild-type and R62G and R133W mutant SRY HMG domains. Circularly permuted probes (a to g) were labeled with 32P and incubated with ≈100 ng of R62G and 4 ng of R133W and wild-type HMG domains. Bend angles were estimated as described (27). (c Top). The HMG box of SRY is an 80-aa DNA-binding domain, comprising three α-helices (numbered in red) that form an L conformation. Note that the DNA (green) is severely bent and unwound after binding (35). The nNLS is formed by the extended chain at the N-terminal end of the HMG box and part of helix 1 and is predicted to be in close proximity to the cNLS. The lower three panels show close-up views of arginine side chains R62, R75, and R76 (in light blue) substituted in cases of XY sex reversal. Although R62 and R75 contact DNA, R76 does not. (d) EMSA analysis of SRY binding to DNA in competition with IMPα/β. SRY HMG box (40 ng) was incubated with 0, 20, 40, 60, 80, or 100 ng of IMPα/β (lanes 2–7). Protein–DNA complexes (arrow) were resolved from free DNA on nondenaturing 6% polyacrylamide Tris/borate/EDTA gels.


DNA-Binding and -Bending Activity of Mutant SRYs. To establish whether import-defective SRY mutants also show altered DNA-binding/bending properties, the mutant SRY proteins were incubated with double-stranded oligonucleotides bearing the high-affinity DNA-binding site (TAAACAATAA) and complexes analyzed by EMSA. R62G and R133W mutant SRY proteins were stably translated and expressed at comparable levels to wild-type SRY (Fig. 3a Upper), which formed detectable complexes with DNA as did the R133W mutant, which bound with near wild-type affinity (Fig. 3a Lower; ref. 35). In contrast, the R62G mutant failed to form detectable complexes with DNA (Fig. 3) even at up to 25-fold (0.25 μM) higher concentrations (Fig. 3b). NMR studies indicate that R62 forms hydrophobic contacts and a salt bridge with the sugar-phosphate backbone of the DNA (Fig. 3c, second from the top; ref. 36); the R62G mutation is presumed to disrupt these contacts. Comparable studies with the other SRY mutants indicated that R75N failed to form detectable complexes with the DNA probe, whereas R76P showed wild-type DNA-binding activity (12). Consistent with this, the NMR data indicate that R75 interacts with the DNA backbone at one end of the DNA-binding site (Fig. 3c, second from the bottom), and R76 makes no contact with the DNA or polypeptide chain (Fig. 3c Bottom). We also determined the bend angles induced after binding of the SRY HMG domains to the DNA site AACAAT by using a circular permutation assay to be 64° for wild type and R133W but 44° for the R62G mutant (Fig. 3b). In summary, of four import-defective mutants, DNA recognition is intact in R76P and R133W, whereas DNA recognition is altered in R62G and R75N. Thus, the only detectable change in SRY function effected by the R76P and R133W mutations is impaired nuclear accumulation ability.

Because the NLSs were implicated with roles in both nuclear localization and DNA binding, the preformed wild-type SRY–DNA complex was incubated with increasing amounts of IMPα/β to test whether competition or ternary complex formation occurs. EMSA analysis showed that SRY HMG domain–DNA complex formation is effectively inhibited by substoichiometric amounts of IMPα/β (Fig. 3d). Thus DNA- and IMP-binding activities of SRY are mutually exclusive but possibly linked events, similar to observations for the yeast transcriptional activator GAL4 (37).

Previous SectionNext Section

Discussion

Altered in vitro DNA binding and bending of SRY from XY females imply that the activities of SRY in sex determination relate to its ability to interact with specific DNA sequences and thereby modulate transcription (38, 39). However, in cases where DNA recognition and bending by SRY is not affected by sex-reversing mutations, it is reasonable to postulate that some other essential function of SRY that contributes to its transcriptional modulatory role has been altered. On the basis of the similarity of the nNLS and cNLS sequences to known NLSs, we predicted that mutations in these sequences may affect nuclear import. Sudbeck and Scherer (13) fused the first 154 aa of SRY, which includes the nNLS, to β-gal and showed that the SRY-R62G mutant localized qualitatively to the nucleus in transfected COS-7 cells using fluorescence microscopy. The quantitative CLSM analysis here shows that the SRYR62G mutation does affect nuclear localization quite markedly. In addition, we demonstrate that SRY from three other XY females also shows reduced nuclear localization. Wild-type SRY, although only 26 kDa, shows exclusive nuclear staining, supporting signal-mediated rather than diffusional import mechanisms. Even when both NLSs are mutated at arginine residues, some nuclear localization still occurs, indicating that these NLSs are still active but at a reduced efficiency (data not shown). Thus, both intact NLSs seem to be necessary to ensure complete nuclear localization, a process central to the sex-determination function of SRY (refs. 13 and 15; see also below).

Both the SRY nNLS and cNLS sequences were functional in targeting β-gal to the nucleus in vivo and in vitro when compared with a large-tumor antigen-NLS-β-gal protein (24). Nuclear accumulation of the SRY nNLS-β-gal protein occurred in the presence of exogenous cytosol and was inhibited in the absence of cytosol, indicating that cytosolic factors are required for nuclear import mediated by the SRY nNLS. For SRY cNLS, nuclear uptake occurred in the absence of cytosol. Because the cNLS is known to mediate nuclear import through the action of IMPβ, the observed lack of a requirement for cytosol is attributable to residual IMPβ bound to the nuclear envelope, which has been shown for a GFP-SRY HMG box (14) as well as the IMPβ-mediated nuclear import substrate IGFBP-3 (30).

The SRY nNLS is functional but not recognized by IMPs, implying that SRY utilizes a novel pathway. Both R75N and R76P nNLS mutants bind IMPβ normally, suggesting that the observed nuclear accumulation defects are attributable to this novel pathway (Fig. 4). A Ran-independent nuclear import pathway that depends on calmodulin has been reported (40), and intriguingly in this context, SRY has been shown to bind to calmodulin in a Ca2+-dependent manner in vitro via its nNLS (41). Significantly, the basic helix–loop–helix and basic ZIP transcription factors (42) and p21Cip1 (43) also bind to calmodulin in vivo, and calmodulin antagonists decrease their nuclear accumulation. Clearly, these observations raise the possibility that the function of the nNLS may be regulated by calmodulin.

Fig. 4.
View larger version:
Fig. 4.

A model for nuclear import of SRY from normal males and XY females. The distinct NLSs of SRY use different import pathways. The cNLS of SRY is recognized by IMPβ, which docks the transport complex at the nuclear pore complex (NPC) followed by translocation through it. After nuclear entry of the complex, RanGTP binds to IMPβ to trigger release of SRY; DNA binding by SRY may also facilitate the release process. The SRY nNLS can mediate nuclear import through a novel pathway not utilizing conventional nuclear import factors such as IMPs but an unidentified transport factor, TF, possibly calcium-calmodulin. In the XY females carrying the R75N and R76P mutations, the nNLS-dependent nuclear import pathway is probably nonfunctional, but the cNLS pathway is most likely active because both mutants bind IMPβ normally. In the XY female carrying the cNLS mutation, R133W, SRY binds IMPβ with reduced affinity due to the direct effect of the mutation; the nNLS pathway is probably functional. In the XY female carrying the R62G nNLS mutation, reduced IMPβ binding may imply that both pathways are impaired, with the residual R62G protein that is translocated to the nucleus unable to bind DNA efficiently (similar to the R75N mutation), or bend DNA to the correct angle to establish the correct architecture in the chromatin.


Consistent with the idea that DNA and IMPβ binding are linked inextricably, specific binding of SRY to DNA was shown to be blocked by IMPβ here (see also ref. 14). In terms of nuclear architecture, the promoters of active genes are known to be in close proximity to the nuclear pores such that DNA in active chromatin may play an important role in effecting release of IMPβ from SRY after transport through the nuclear pore. Analogous to GAL4 (37), DNA binding and IMPβ binding by SRY seem to be mutually exclusive events (14), indicating that the NLSs of SRY may be regulated in part by DNA binding. Because the promoters of active genes are close to nuclear pore complexes in terms of nuclear architecture (11), it seems possible that after transport through the nuclear pore, DNA binding in conjunction with RanGTP may effect release of IMPβ from SRY as has been hypothesized for GAL4 (37).

SRY is expressed in the primordial Sertoli cells in the developing genital ridge at week 7 of gestation. Although there is no appropriate cell-culture system currently available to test our observations directly, we think that the behavior of SRY characterized in the molecular and cellular systems used here are likely to represent SRY behavior in primordial Sertoli cells. The molecular consequences of the XY sex-reversal mutations analyzed here are summarized in Fig. 4 and Table 2, indicating the central importance of the nNLS- and cNLS-mediated nuclear import pathways. The results here represent the characterization of naturally occurring clinical mutations in NLSs that impair nuclear import by reducing IMP recognition. In analogous fashion to the mutations here, nuclear-transport defects could explain the deleterious effects of many other SRY/SOX mutations, in particular mutations that show normal DNA recognition (9, 12, ††). That this class of mutation almost certainly will include mutations that map outside the cNLS and nNLS sequences is supported by our observation that the SRY HMG box binds IMPβ better than the cNLS alone binds IMPβ, indicating that regions outside the cNLS influence binding, and the nNLS SRY mutation R62G significantly reduces IMPβ binding. In addition, the SOX9 mutation A158T, which lies outside the NLSs, isolated from an XY female patient with campomelic dysplasia, causes a significant reduction in nuclear import of SOX9 (22).

View this table:
Table 2. Summary of activities of wild-type and mutant SRY

In the case of the R62G mutant, sex reversal is attributable in part to insufficient import of the SRY protein into the nucleus of primordial Sertoli cells at the time of testis determination. The nNLS of SRY utilizes an IMP-independent pathway; the fact that the R62G substitution in the nNLS also affects IMPβ binding to the cNLS suggests that this mutation, which also affects DNA binding and bending, has dramatic effects on SRY structure such that the reduced amount of protein that is imported into the nucleus cannot bind and bend DNA. In contrast, the R76P and R133W mutations do not significantly impair DNA binding/bending, and in the case of R133W, IMPβ recognition is specifically affected, implying a critical role for Arg-133 in IMPβ binding. The sex-reversed phenotype resulting from the R76P and R133W mutations thus would seem to be the direct result of reduced levels of SRY protein in the nucleus, which supports a threshold activity of SRY in the nucleus and the notion that XY sex reversal can occur if this threshold is not reached. It should be noted that these experiments do not rule out the possibility that unidentified SRY activities may also be defective in R76P and R133W. Indeed, compound defects affecting protein structure, nuclear import, and transcriptional activation occur in SOX9 from campomelic dysplasia patients (22).

Previous SectionNext Section

Acknowledgments

We thank Anthea Chai and Helena Sim for technical assistance. This work was supported by National Health and Medical Research Council (Australia) Grants 983001 and 198713 (to V.R.H.) and 143790 and 143710 (to D.A.J.).

Previous SectionNext Section

Footnotes

  • ↵ † To whom correspondence should be addressed at: Prince Henry's Institute of Medical Research, Monash Medical Centre, 246 Clayton Road, Clayton 3152, Victoria, Australia. E-mail: vincent.harley{at}med.monash.edu.au.

  • This paper was submitted directly (Track II) to the PNAS office.

  • Abbreviations: SRY, sex-determining region of the Y chromosome; HMG, high-mobility group; NLS, nuclear localization sequence; β-gal, β-galactosidase; cNLS, C-terminal NLS; nNLS, N-terminal NLS; IMP, importin; HTC, hepatoma tissue culture; CLSM, confocal laser scanning microscopy, EMSA, electrophoretic mobility-shift assay; SRYwt, SRY wild type; SOX, SRY-type HMG box.

  • ↵ †† Tho, S. P., Zhang, Y. Y., Hines, R. S., Hansen, K. A., Khan, I. & McDonough, P. G. (1998) J. Soc. Gynecol. Investig. Suppl., 5, 138 (abstr.).

Previous Section
 

References

  1. ↵
    Sinclair, A. H., Berta, P., Palmer, M. S., Hawkins, J. R., Griffiths, B. L., Smith, M. J., Foster, J. W., Frischauf, A. M., Lovell-Badge, R. & Goodfellow, P. N. (1990) Nature 346 , 240-244.pmid:1695712
    CrossRefMedline
  2. ↵
    Berta, P., Hawkins, J. R., Sinclair, A. H., Taylor, A., Griffiths, B. L., Goodfellow, P. N. & Fellous, M. (1990) Nature 348 , 448-450.pmid:2247149
    CrossRefMedline
  3. ↵
    Hawkins, J. R. (1994) Hum. Mol. Genet. 3 , 1463-1467.pmid:7849739
    Abstract
  4. ↵
    Capel, B. (2000) Mech. Dev. 92 , 89-103.pmid:10704890
    CrossRefMedlineISI
  5. ↵
    Desclozeaux, M., Poulat, F., de Santa Barbara, P., Capony, J. P., Turowski, P., Jay, P., Mejean, C., Moniot, B., Boizet, B. & Berta, P. (1998) J. Biol. Chem. 273 , 7988-7995.pmid:9525897
    Abstract/FREE Full Text
  6. McElreavey, K., Vilain, E., Abbas, N., Herskowitz, I. & Fellous, M. (1993) Proc. Natl. Acad. Sci. USA 90 , 3368-3372.pmid:8475082
    Abstract/FREE Full Text
  7. Haqq, C. M., King, C. Y., Ukiyama, E., Falsafi, S., Haqq, T. N., Donahoe, P. K. & Weiss, M. A. (1994) Science 266 , 1494-1500.pmid:7985018
    Abstract/FREE Full Text
  8. ↵
    Cohen, D. R., Sinclair, A. H. & McGovern, J. D. (1994) Proc. Natl. Acad. Sci. USA 91 , 4372-4376.pmid:8183916
    Abstract/FREE Full Text
  9. ↵
    Jager, R. J., Harley, V. R., Pfeiffer, R. A., Goodfellow, P. N. & Scherer, G. (1992) Hum. Genet. 90 , 350-355.pmid:1483689
    MedlineISI
  10. ↵
    McDowall, S., Argentaro, A., Ranganathan, S., Weller, P., Mertin, S., Mansour, S., Tolmie, J. & Harley, V. (1999) J. Biol. Chem. 274 , 24023-24030.pmid:10446171
    Abstract/FREE Full Text
  11. ↵
    Grasser, K. D. & Feix, G. (1991) Nucleic Acids Res. 19 , 2573-2577.pmid:2041733
    Abstract/FREE Full Text
  12. ↵
    Mitchell, C. & Harley, V. R. (2002) Mol. Genet. Metab. 77 , 217-225.pmid:12409269
    CrossRefMedline
  13. ↵
    Sudbeck, P. & Scherer, G. (1997) J. Biol. Chem. 272 , 27848-27852.pmid:9346931
    Abstract/FREE Full Text
  14. ↵
    Forwood, J. K., Harley, V. & Jans, D. A. (2001) J. Biol. Chem. 276 , 46575-46582.pmid:11535586
    Abstract/FREE Full Text
  15. ↵
    Poulat, F., Girard, F., Chevron, M. P., Goze, C., Rebillard, X., Calas, B., Lamb, N. & Berta, P. (1995) J. Cell Biol. 128 , 737-748.pmid:7876301
    Abstract/FREE Full Text
  16. ↵
    Affara, N. A., Chalmers, I. J. & Ferguson-Smith, M. A. (1993) Hum. Mol. Genet. 2 , 785-789.pmid:8353496
    Abstract/FREE Full Text
  17. ↵
    Battiloro, E., Angeletti, B., Tozzi, M. C., Bruni, L., Tondini, S., Vignetti, P., Verna, R. & D'Ambrosio, E. (1997) Hum. Genet. 100 , 585-587.pmid:9341876
    CrossRefMedline
  18. ↵
    Ross, C. A., Wood, J. D., Schilling, G., Peters, M. F., Nucifora, F. C., Jr., Cooper, J. K., Sharp, A. H., Margolis, R. L. & Borchelt, D. R. (1999) Philos. Trans. R. Soc. London B 354 , 1005-1011.pmid:10434299
    CrossRefMedlineISI
  19. ↵
    Zhi, L., Zheng-gang, X. & Guanghui, X. (1996) Chinese Med. J. Genet. 12 , 258-261.
  20. ↵
    Veitia, R., Ion, A., Barbaux, S., Jobling, M. A., Souleyreau, N., Ennis, K., Ostrer, H., Tosi, M., Meo, T., Chibani, J., et al. (1997) Hum. Genet. 99 , 648-652.pmid:9150734
    CrossRefMedlineISI
  21. ↵
    Mertin, S., McDowall, S. G. & Harley, V. R. (1999) Nucleic Acids Res. 27 , 1359-1364.pmid:9973626
    Abstract/FREE Full Text
  22. ↵
    Preiss, S., Argentaro, A., Clayton, A., John, A., Jans, D. A., Ogata, T., Nagai, T., Barroso, I., Schafer, A. J. & Harley, V. R. (2001) J. Biol. Chem. 276 , 27864-27872.pmid:11323423
    Abstract/FREE Full Text
  23. ↵
    Jans, D. A., Ackermann, M. J., Bischoff, J. R., Beach, D. H. & Peters, R. (1991) J. Cell Biol. 115 , 1203-1212.pmid:1659575
    Abstract/FREE Full Text
  24. ↵
    Hubner, S., Xiao, C. Y. & Jans, D. A. (1997) J. Biol. Chem. 272 , 17191-17195.pmid:9202041
    Abstract/FREE Full Text
  25. ↵
    Hu, W. & Jans, D. A. (1999) J. Biol. Chem. 274 , 15820-15827.pmid:10336485
    Abstract/FREE Full Text
  26. ↵
    Forwood, J. K., Lam, M. H. & Jans, D. A. (2001) Biochemistry 40 , 5208-5217.pmid:11318643
    Medline
  27. ↵
    van de Wetering, M., Oosterwegel, M., Dooijes, D. & Clevers, H. (1991) EMBO J. 10 , 123-132.pmid:1989880
    MedlineISI
  28. ↵
    Thompson, J. F. & Landy, A. (1988) Nucleic Acids Res. 16 , 9687-9705.pmid:2972993
    Abstract/FREE Full Text
  29. ↵
    Lam, M. H., Briggs, L. J., Hu, W., Martin, T. J., Gillespie, M. T. & Jans, D. A. (1999) J. Biol. Chem. 274 , 7391-7398.pmid:10066803
    Abstract/FREE Full Text
  30. ↵
    Schedlich, L. J., Le Page, S. L., Firth, S. M., Briggs, L. J., Jans, D. A. & Baxter, R. C. (2000) J. Biol. Chem. 275 , 23462-23470.pmid:10811646
    Abstract/FREE Full Text
  31. ↵
    Tiganis, T., Flint, A. J., Adam, S. A. & Tonks, N. K. (1997) J. Biol. Chem. 272 , 21548-21557.pmid:9261175
    Abstract/FREE Full Text
  32. ↵
    Truant, R. & Cullen, B. R. (1999) Mol. Cell. Biol. 19 , 1210-1217.pmid:9891055
    Abstract/FREE Full Text
  33. ↵
    Palmeri, D. & Malim, M. H. (1999) Mol. Cell. Biol. 19 , 1218-1225.pmid:9891056
    Abstract/FREE Full Text
  34. ↵
    Henderson, B. R. & Percipalle, P. (1997) J. Mol. Biol. 274 , 693-707.pmid:9405152
    CrossRefMedlineISI
  35. ↵
    Li, B., Zhang, W., Chan, G., Jancso-Radek, A., Liu, S. & Weiss, M. A. (2001) J. Biol. Chem. 276 , 46480-46484.pmid:11641389
    Abstract/FREE Full Text
  36. ↵
    Werner, M. H., Bianchi, M. E., Gronenborn, A. M. & Clore, G. M. (1995) Biochemistry 34 , 11998-12004.pmid:7547937
    CrossRefMedline
  37. ↵
    Chan, C. K., Hubner, S., Hu, W. & Jans, D. A. (1998) Gene Ther. 5 , 1204-1212.pmid:9930321
    CrossRefMedlineISI
  38. ↵
    Harley, V. R., Jackson, D. I., Hextall, P. J., Hawkins, J. R., Berkovitz, G. D., Sockanathan, S., Lovell-Badge, R. & Goodfellow, P. N. (1992) Science 255 , 453-456.pmid:1734522
    Abstract/FREE Full Text
  39. ↵
    Pontiggia, A., Rimini, R., Harley, V. R., Goodfellow, P. N., Lovell-Badge, R. & Bianchi, M. E. (1994) EMBO J. 13 , 6115-6124.pmid:7813448
    MedlineISI
  40. ↵
    Sweitzer, T. D. & Hanover, J. A. (1996) Proc. Natl. Acad. Sci. USA 93 , 14574-14579.pmid:8962094
    Abstract/FREE Full Text
  41. ↵
    Harley, V. R., Lovell-Badge, R., Goodfellow, P. N. & Hextall, P. J. (1996) FEBS Lett. 391 , 24-28.pmid:8706923
    CrossRefMedlineISI
  42. ↵
    Corneliussen, B., Holm, M., Waltersson, Y., Onions, J., Hallberg, B., Thornell, A. & Grundstrom, T. (1994) Nature 368 , 760-764.pmid:8152489
    CrossRefMedline
  43. ↵
    Taules, M., Rodriguez-Vilarrupla, A., Rius, E., Estanyol, J. M., Casanovas, O., Sacks, D. B., Perez-Paya, E., Bachs, O. & Agell, N. (1999) J. Biol. Chem. 274 , 24445-24448.pmid:10455103
    Abstract/FREE Full Text
« Previous | Next Article »Table of Contents

This Article

  1. Published online before print May 22, 2003, doi: 10.1073/pnas.1137864100 PNAS June 10, 2003 vol. 100 no. 12 7045-7050
  1. AbstractFree
  2. Figures Only
  3. » Full Text
  4. Full Text (PDF)

Classifications

Social Bookmarking