A Mutation in the Insulin Receptor That Impairs Proreceptor Processing but Not Insulin Binding*

Here we report the identification of a new mutation in the a-chain of the insulin receptor, changing Trp412 into Ser using DNA from consanguineous parents who gave birth to a child with leprechaunism. The mutant receptor was expressed stably in CHO and transiently in COS-1 cells. It was found that the Ser4I2 mutant is not cleaved into a- and P-subunits and remains as a 210-kDa proreceptor at an intracellular site. This property of the mutant receptor is in line with the observed decreased insulin binding to the parental fibroblasts. Cross-link-ing experiments show that the Ser412 proreceptor is able to bind insulin with an affinity comparable to that of the wild-type a-chain. Despite its capacity to bind insulin, the mutant receptor is not autophosphorylated. We postulate that the patient was homozygous for the Trp412 + Ser mutation and that the mutation was respon- sible for the leprechaun phenotype. This is the first de-scription of a transport-defective receptor with the mutation outside the tetrabasic processing site and a functional

Recently, we and others investigated the effect of a number of different mutations in the a-chain of the IR on receptor properties by transfection experiments. The examined mutations came from patients with leprechaunism or the Rabson-Mendenhall syndrome. The mutations were alike in that they resulted in insulin receptor precursors that were deficient in receptor processing. The transport to the cell surface of these proreceptors is inhibited (2-6). The transport and processing defect of these mutants resulted from an altered conformation of the precursor receptor, as judged from studies with monoclonal antibodies recognizing conformational epitopes and a loss of insulin binding to the proreceptor. This paper describes a new mutation in the a-chain of the IR, which is expected to be responsible for leprechaunism. As there was no material available from the deceased patient, material from both consanguineous parents was used for the experiments. The effect of the mutation on the properties of the IR was examined after overexpression of the mutant receptor in CHO and COS-1 cells.
Subject-Patient KG was a female born from a consanguineous marriage. The parents originated from Turkey. After an uncomplicated pregnancy, she was born at term with typical clinical symptoms of leprechaunism: extreme dystrophic appearance, low ear implant, hypertrichosis, genital hypertrophy, and lipodystrophy. From day 3, the proband suffered from general infections, and at the age of 24 days, she died from septic shock. Autopsy showed hepatomegaly and marked hypertrophy of the islets of Langerhans. No material from the patient was available. Skin biopsies and blood were obtained from the parents and used for the experiments.
Insulin Binding and DNA Sequencing-Determination of high afflnity insulin binding sites on cultured fibroblasts (7) and DNA sequencing (2) has been described previously.
Construction of Mutant Expression Vectors-Recombinant polymerase chain reaction was used to create the Ser412 mutant receptor from a cDNAclone encoding the wild-type human insulin receptor (WT-HIR). With two specific oligonucleotide sets, two cDNA fragments (A and B) were amplified, both containing the mutation. FragmentAwas amplified with primers 1057 and 1377 (AAGGTGTGCCACCTCCTAGA TTT-GCTCCAGTCCGAGAGCTGCCTTA). Fragment B with primers 1352 and 1661 (TAAGGCAGCTCTCGGACTGGAGCU, GGGCCTCTITG-TAGAACAGC). Primers 1352 and 1377 are complementary 26-mers  with the mutated nucleotide at position 1364 (indicated by bold italics).
Fragment A (1057-1377) and fragment B (1352-1661) were purified by low melting agarose (FMC) gel electrophoresis. They were isolated and denaturated together at 97 "C. Renaturation yielded the original fragments, and also the heteroduplexes A-B. In this mixture the heteroduplex was selectively reamplified with primers 1057 and 1661, thereby generating a mutated fragment of 604 nucleotides (1057-1661). This Ser412 fragment and vectors containing the wild-type IR were digested with BstEIIIPuuII and the restriction fragments were purified by low melting agarose gel electrophoresis. The BstEIIIPuuII fragment carrying the Ser412 mutation was ligated in the expression vectors. The resulting vectors were transfected in (2-600 bacteria. Colonies expressing the Ser412 mutant vector (Ser412-HIR) were identified by colony hybridization with a 32P-labeled mutant oligonucleotide. By DNA sequencing, the correct sequence of the Ser4I2-HIR vectors was reconfirmed. The COS-1 cells were grown to 4040% confluence, and the DNA mixture was added dropwise in 5 ml of Dulbecco's modified Eagle's medium. After 18 h at 37 "C in a 2 4 % CO, atmosphere, the precipitation was completed. The medium was then changed to standard Dulbecco's modified Eagle's medium, and the cells were cultured for another 56 h at 5% CO,.
Stable Expression of the Insulin Receptor (-Exon 11) in CHO Cells-The WT-HIR (-exon 11, kindly provided by Dr. A. Ullrich) and Ser412-HIR (-exon 11) vector were transfected in CHO cells using the calcium phosphate method, as described before (3). Single colonies resistant to 500 pglml G-418 were picked and brought to high expression by amplification on 400 n~ methotrexate.
Cross Analysis of receptor labeling was as described above. Autophosphorylation of Insulin Receptors-CHO cells were grown to confluence in 6-cm dishes, washed in phosphate-buffered saline, and lysed in 200 pl of lysis buffer. A total of 10 pl(60 pg of protein) lysate was used, and autophosphorylation was performed with and without 50 n~ insulin. In case of autophosphorylation on immune precipitates, the cells were grown to confluence and lysed in 300 pl of ice-cold lysis buffer. Nuclei were removed by centrifugation. The supernatant was diluted to 750 pl with IP buffer (20 m~ Tris-HC1, pH 7.6, 150 m~ NaCl, 0.1 m M EDTA, 0.1% Triton X-100, 0.1 m M phenylmethylsulfonyl fluoride). Three pl of monoclonal anti-IR (Amersham) was added. After incubation overnight, the immune complex was isolated by protein G-Sepharose (Pharmacia). Autophosphorylation was performed in 50 m M Hepes, pH 7.4, 0.1 m~ EDTA, 1% Triton X-100, 5 m~ MgCl,, 25 m o l of [3zP]ATP (60 Cibnmol). In case of insulin stimulation, 50 n~ insulin was present. Autophosphorylation was started by adding L3' P1ATP (final volume 15 111). Incubation was for 6 min, after which total protein was precipitated by adding 1 ml of 10% trichloroacetic acid. The mixture was centrifuged, and the pellet was washed with 70% ethanol. The protein pellet was dissolved in SDS sample buffer and analyzed by SDS-PAGE and autoradiography.

RESULTS
Insulin Binding to Fibroblasts-In most cases, fibroblasts from leprechaun patients show a loss of high affinity insulin binding sites (8). In relatives of the patients, who were heterozygous for mutations in the IR, we observed an approximately 70% loss of these sites. As no cells were available from the deceased proband KG, we examined fibroblasts from the parents of KG (641 and 642) for the number of high affinity insulin binding sites. In the binding experiment, we included fibroblasts from an individual (RD 330) who is heterozygous for a mutation in the IR. Homozygosity for the mutation results in leprechaunism and total loss of high affinity insulin binding to fibroblasts. This mutant is transport-defective when expressed in CHO cells (3). As control we used fibroblasts from healthy individuals. Insulin binding experiments with fibroblasts from RD 330,641, and 642 showed comparable binding values which were approximately 30% of the values seen in fibroblasts from 10 controls (Fig. 1). This observation suggests that in the deceased leprechaun KG, IR mutations were involved that are linked to a loss of high affinity insulin binding sites.
Sequence Analysis-To determine the presence of mutations in the IR, DNA was isolated from parental blood. The coding region of the IR gene was amplified, and the polymerase chain reaction product was sequenced. We detected a non-polymorphic mutation in exon 6 changing a tryptophan (TGG) into a serine (TCG) at position 412 (Fig. 2). Both parents are heterozygous for this Ser4" mutation. Insulin Binding to Cells Expressing the Ser4I2 IR-To examine the effect on the properties of the IR of the conversion of tryptophan to serine at position 412, we introduced this mutation in IR cDNA by recombinant polymerase chain reaction. Subsequently, CHO cells were transfected stably with an SV40 driven vector containing the sequence for WT-HIR and Ser4I2-HIR, respectively. This vector contains a dihydrofolate reductase gene, which enables amplification with methotrexate. Overexpression was obtained after selection with G-418 in the presence of increasing concentrations of methotrexate.
Insulin binding studies showed in WT-HIR CHO cells a marked increase of binding compared to untransfected CHO cells. The number of high affinity sites in WT-HIR CHO cells ranges between 2 x lo6 and 1 x 106/cell in various individual clones. In the parental CHO cells and the Ser4I2-HIR CHO cells, p-chains were detected in the Ser412-HIR CHO cells (Fig. 3A), even after overexposure. The same result was obtained using a monoclonal antibody (RPN 538, Amersham) against the IR, which recognizes an epitope on the a-chain that is sensitive to conformational changes (Fig. 3B). These results suggest that the Ser412 mutant is not converted into a-and P-subunits. Western Blot Analysis of COS-1 Cells-The properties of the Ser4I2-HIR were also examined in COS-1 cells. The W-HIR and the Ser412-HIR were expressed transiently using the SR-aderived expression vector. T w o days after transfection, the IR was visualized by Western blotting using a polyclonal antibody against a peptide epitope on the P-chain of the IR. In case of the WT-HIR expression, next to the proreceptor a t 210 kDa the P-chain (95 kDa) was also present. Expression of the Ser412-HIR showed, as in CHO cells, only the 210-kDa band. In the untransfected COS-1 cells, no insulin receptor expression was seen (Fig. 4).
Enzymatic Deglycosylation-To determine whether the 210-kDa band actually represents the proreceptor, immune precipitates from the WT-HIR CHO and the Ser412-HIR CHO cells were subjected to enzymatic deglycosylation with endoglycosidase H and neuraminidase. After endoglycosidase H treatment, the 210-kDa band shifted to a 180-kDa band in both cell lines.
Use of neuraminidase showed no shift of the 210-kDa bands (Fig. 5). This behavior is characteristic for the proreceptor of the IR, as it lacks sialic acid residues (91, and underscores that the 210-kDa band seen in the immune precipitates is indeed the proreceptor. Cell Surface Iodination in CHO Cells-To investigate whether the mutant proreceptor is transported to the cell surface, the various CHO cell lines were subjected to cell surface radioiodination using the lactoperoxidase-catalyzed procedure. After cell lysis, immune precipitation, SDS-PAGE and autoradiography, cells expressing the WT receptor showed clearly the a-chain at 135 kDa and the P-chain a t 95 kDa, whereas in the mutant receptor neither the IR chains nor the proreceptor could be detected. It should be noted that the nonimmune serum precipitates a labeled protein, migrating at a position of approximately 85 kDa (Fig. 6).

Cross-linking of 125Z-Znsulin to the Receptor in Cell
Lysates from CHO Cells-As the proreceptor did not reach the cell surface, the binding defect to intact cells seems a secondary effect. To examine the ability of the mutant proreceptor to bind insulin, cell lysates of transfected CHO cells were prepared and used for insulin binding studies. It was found that whole cell lysates of CHO cells showed high binding values, which could not be competed with an excess of nonradioactive insulin. To circumvent this problem of non-specific binding, an indirect assay was used to determine the amount of insulin bound to receptors in whole cell lysates. The method is based on covalent cross-linking of 1251-insulin to the IR. Cell lysates from CHO cells expressing WT-HIR and Ser412-HIR were incubated over-  RPN 538 (lane 1 and 3) or nonimmune serum (lanes 2 and 4). diography. In the WT-HIR CHO cells, the a-subunit (135 kDa) and also (in small amounts) the proreceptor (210 kDa) were labeled. In the Ser412-HIR CHO cells, only the proreceptor was seen. Radioactive insulin was competed away by addition of excess of nonradioactive insulin (Fig. 7A). Thus, the Ser412 mutant proreceptor can be cross-linked to insulin. The experiments were also done with various concentrations of insulin (0.5,1,5, and lo-' M) and labeled (pro)receptor was quantitated by densitometry. The yield of labeled (pro)receptor depends on the amount of IR-insulin complex formed and is an indication for the KD of insulin binding. The observed dependence of crosslinking on the insulin concentrations is similar for the WT a-subunit and the Ser412 proreceptor. From the concentration dependence, it is estimated that the half-maximal labeling in both situations is reached at 2-4 n~ '251-insulin (Fig. 7B). In addition, the dissociation rate of the IR-insulin complex was found to be similar in both situations. This was examined by incubating the cell lysates with 1251-insulin for 18 h, followed by addition of 1 p~ nonradioactive insulin. Subsequently, at 5-min intervals, DSS was added and the labeling of the a-subunit in case of the WT IR and the proreceptor in case of the S e F 2 mutant was determined (Fig. a). Densitometry of the (prolreceptor bands is represented as percentage of the value measured a t t = 0. The decline is comparable in both cell lines (Fig.   8B). The t, for dissociation is approximately 19 min.
Autophosphorylation of Mutant Receptors-Cell lysates were also used to examine the ability of the Ser412 proreceptor to undergo insulin-stimulated autophosphorylation. For that, cell lysates from WT-HIR CHO cells and Ser412-HIR CHO cells were incubated with or without 50 n~ insulin in the presence of PAGE and autoradiography (Fig. 9). The lysates from the WT-HIR CHO cells showed a strong signal at the position of the P-chain (95 kDa) after insulin stimulation, comparable to that seen when partially purified IRs were used. The Ser412-HIR CHO cells did not show any induction at the position of the proreceptor or at the 95-kDa band @-chain). We also performed this experiment on insulin receptors purified by immune precipitation. Similar results were obtained compared to total cell lysates (not shown). DISCUSSION A number of IR mutations found in patients with leprechaunism or Rabson-Mendenhall syndrome interfere with transport of the IR to the cell surface (2,3,10). The patients are usually homozygous or compound heterozygous for the mutations. Cultured cells from these patients show, in general, a loss of high affinity insulin binding sites on the cell surface (5,6). Individuals who are heterozygous for such a mutation, like the parents of these patients, show a decrease in high affinity in-  (lunes 2,4, and 6 ) or absence of 50 nM insulin (lunes 1,3, and 5). Lunes 1 and 2, cell lysate from W-HIR CHO cells; lunes 3 and 4, cell lysates from Ser4I2-HIR CHO cells; lunes 5 and 6, glycoprotein fraction of W-HIR CHO cells. Positions of the proreceptor (210 kDa) and P-chain are indicated.
sulin binding sites to approximately 30% of control fibroblasts. This is due to a partially dominant effect of the mutant receptor (2, 11).
We detected a new mutation (Ser for Trp at amino acid position 412) in the insulin receptor gene in DNA from both parents of a patient who exhibited phenotypical features of leprechaunism. As the parents are heterozygous, we postulate that the deceased child was homozygous for this serine (TCG) for tryptophan (TGG) substitution. We cannot confirm this assumption, because no material from the proband is available.
Fibroblasts from the parents show a marked reduction of the number of high affinity insulin binding sites. A similar situation is observed in fibroblasts from heterozygous relatives of leprechaun Geldermalsen (Pro233) and leprechaun H (Argl). As the and the Arg' IR mutants are impaired in receptor processing and transport (2, 3), the same defect could be expected in case of the Ser412 mutant. We examined the properties of the Ser412 mutant receptor, expressed in CHO and COS-1 cells. In CHO cells the insulin receptor expression vector lacked exon 11, whereas the vector used for COS-1 cells had exon 11. In both cases the synthesized Ser4I2-HIR appears as a proreceptor that is not transported to the cell surface. We observe that the Ser412 proreceptor (-exon 11) expressed in CHO cells has an affinity for insulin similar as the WT-HIR. In addition, it seems that the WT proreceptor (-exon ll), which is present in small amounts in the CHO cells expressing WT-HIR (cf Fig.  7, lane 11, has a similar affinity for insulin as the WT-a-chain. This is inferred from the observation that the ratio of aftinitylabeled WT-a and WT-proreceptor (Fig. 8 A ) is similar to the biosynthetic levels of processed and unprocessed receptor (Fig.  7). Thus, we observe no marked decrease in affinity of both WT and Ser412 proreceptors for insulin. Other groups have shown that proreceptors resulting from mutations in the tetrabasic cleavage site have lower affinities for insulin as processed a-chains (12,13). However, it has been pointed out (13) that the affinity of the proreceptor for insulin depends on the cell type in which the proreceptor is expressed, which may explain the similar affinities of processed and unprocessed receptors for insulin that we observed. These authors also provide evidence that the presence of exon 11 in the proreceptor has no major effect on insulin binding.
When autophosphorylation is considered, we observe no insulin-stimulated autophosphorylation using the Ser412 proreceptor. Insulin proreceptors have been shown to undergo insulin-stimulated autophosphorylation, albeit with a lower efficiency as WT receptors (12,13). Despite its ability to bind insulin, the Ser412 proreceptor does not show an increase in autophosphorylation when insulin is added, indicating that A Processing-defective IR Mutant with

Insulin Binding
Trp412 somehow is important to transduce the insulin binding signal to activation of the Tyr-kinase.
Enzymatic deglycosylation experiments, showing the lack of sialic acid, suggest that the proreceptor is already retained in the endoplasmatic reticulum or early Golgi. This observation implies that the folding of the mutant receptor is incorrect. Chaperonine proteins can interact with these misfolded proreceptors, resulting in defective transport. Previously described mutations, which prevented receptor transport were likely to affect receptor folding, like the Arg3l, Pro233, and (1) mutants. These mutations occurred near cysteine residues involved in disulfide bond formation. The Ser412 IR is not located in a cysteine-rich domain, and a protein prediction program showed no major effect on the probability of a-helix or P-sheet formation. An additional indication that the mutation does not affect the folding of the entire a-chain of the receptor comes from immune precipitation experiments with the monoclonal antibody RPN 538. This is an antibody directed against a conformational epitope on the a-chain. RPN 538 recognizes the WT proreceptor but not the transport-defective mutant receptors and Arg3l. However, this monoclonal antibody does recognize the Ser412 mutant.
Direct insulin binding experiments with the Ser4"-HIR in CHO cell lysates were inconclusive due to the high background binding in CHO cells. Indirect binding experiments by crosslinking proved that the proreceptor was still able to bind insulin with an affinity comparable to that of the wild-type IR. The suggested domains of the insulin receptor that determine the specificity are located at the amino acid region 1-68 (14-161, and the cysteine-rich domain between residue 205-310 (16)(17)(18) or a domain between residue 450-601 (19). The immunological and insulin binding data on the Ser412 proreceptor show that the overall folding of the a-chain is not affected by the mutation. The regions 1-68, 205-310, and, probably, 450-601 are important for correct insulin binding. Therefore, it seems that the properties of the receptor in the Ser4I2 mutant are only locally altered around position 412. Remarkably, this apparently minor change in structure is sufficient to inhibit intracellular IR transport completely. Although the conformation of the Ser412 mutant has not been altered grossly, insulin-stimulated autophosphorylation of the receptor is abolished. I t has been reported recently by Collier et al. (20) that a loss of four N-linked glycosylation sites in the a-chain by changing Asn into Gln creates a transport-defective receptor. We have no evidence that, in the case of the Ser412 receptor mutant, the glycosylation pattern is grossly altered, suggesting that the conformational change introduced by the conversion of Trp into Ser is directly responsible for the transport defect.
In conclusion, we have found a new mutation in the IR changing a tryptophan into a serine at position 412 in exon 6 of the IR. This is the first mutation associated with impairment of IR function due to defective transport of the proreceptor to the cell surface, without directly interfering with its insulin binding capacity. The proband was most likely homozygous for this mutation and presumably this caused the leprechaun phenotype.