The Phosphorylation of Bovine DNase I Asn-linked Oligosaccharides Is Dependent on Specific Lysine and Arginine Residues*

(Received for publication, May 5, 1997)

Atsushi Nishikawa Dagger , Walter Gregory , John Frenz §, Jerry Cacia § and Stuart Kornfeld

From the Department of Medicine, Washington University School of Medicine, St. Louis, Missouri 63110 and § Genentech, South San Francisco, California 94080

ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
REFERENCES


ABSTRACT

The secretory glycoprotein DNase I acquires mannose 6-phosphate moieties on its Asn-linked oligosaccharides, indicating that it is a substrate for UDP-GlcNAc:lysosomal enzyme N-acetylglucosamine-1-phosphotransferase (phosphotransferase) (Cacia, J., Quan, C., and Frenz, J. (1995) Glycobiology 4, 99). Phosphotransferase recognizes a conformation-dependent protein determinant that is present in lysosomal hydrolases, but absent in most secretory glycoproteins. To identify the amino acid residues of DNase I that are required for interaction with phosphotransferase, wild-type and mutant forms of bovine DNase I were expressed in COS-1 cells and the extent of oligosaccharide phosphorylation determined. Phosphorylation of DNase I oligosaccharides decreased from 12.6% to 2.3% when Lys-50, Lys-124, and Arg-27 were mutated to alanines, indicating that these residues are required for the basal level of phosphorylation. Mutation of lysines at other positions did not impair phosphorylation, demonstrating the selectivity of this process. When Arg-27 was replaced with a lysine, phosphorylation increased to 54%, showing that phosphotransferase prefers lysine residues to arginines. Mutation of Asn-74 to a lysine also increased phosphorylation to 50.3%, and the double mutant (R27K/N74K) was phosphorylated 79%, equivalent to the values obtained with lysosomal hydrolases. Interestingly, Lys-27 and Lys-74 caused selective phosphorylation of the neighboring Asn-linked oligosaccharide. Finally, mutation of Lys-117 to an alanine stimulated phosphorylation, demonstrating that some residues may be negative regulators of this process. We conclude that selected lysine and arginine residues on the surface of DNase I constitute the major elements of the phosphotransferase recognition domain present on this secretory glycoprotein.


INTRODUCTION

In many cell types, the sorting of newly synthesized acid hydrolases from secreted proteins is mediated by the phosphomannosyl recognition system (1, 2). The specificity of this pathway is determined by the enzyme UDP-GlcNAc:lysosomal enzyme N-acetylglucosamine-1-phosphotransferase (phosphotransferase)1 which recognizes a conformation-dependent protein determinant that is present in lysosomal hydrolases but absent in most secretory glycoproteins (3). Using a variety of approaches, evidence has been obtained that the recognition determinant involves a broad surface patch that includes critical lysine residues (4-10). The interaction of phosphotransferase with its acid hydrolase substrates results in the transfer of GlcNAc-P to mannose residues on the Asn-linked high mannose oligosaccharides of the lysosomal hydrolases. The GlcNAc residues are then removed by N-acetylglucosamine-1-phosphodiester alpha -N-acetylglucosaminidase to generate phosphomannosyl residues which mediate binding of the hydrolases to mannose 6-phosphate (Man-6-P) receptors present in the Golgi. These complexes are subsequently translocated via clathrin-coated vesicles to endosomes where the hydrolases are discharged for packaging into lysosomes.

While phosphotransferase acts primarily on lysosomal hydrolases, a few secretory glycoproteins acquire Man-6-P moieties (11-14). One of these is human DNase I which is secreted by the pancreas and the salivary gland (14). Since DNase I is a relatively small protein which has been cloned from several species and its crystal structure has been determined (15-17), we reasoned that this would be an excellent model for studying how secretory glycoproteins acquire Man-6-P residues.

In this study we have focused on bovine DNase I which undergoes limited mannose phosphorylation when expressed in COS-1 cells. We show that this low level of phosphorylation is dependent on selected lysine and arginine residues. Interestingly, substitution of a different lysine with an alanine enhances the extent of phosphorylation, indicating that this particular lysine residue impairs phosphorylation. Most strikingly, insertion of two lysines that are present in the mouse sequence but absent in the bovine sequence results in a dramatic increase in the extent of phosphorylation. Furthermore, each of these critical lysines stimulates phosphorylation of a different Asn-linked oligosaccharide.


EXPERIMENTAL PROCEDURES

Materials

COS-1 cells were obtained from ATCC. [2-3H]Mannose was purchased from NEN Life Sciences Products Inc. QAE-Sephadex, concanavalin A-Sepharose, and the pSVK3 expression vector were from Pharmacia Biotech Inc. Recombinant endoglycosidase H fused to maltose-binding protein and Protein A-agarose beads were from New England Biolabs and Repligen, respectively. Bovine calf serum, Lipofectin, and Opti-MEM were from Life Technologies, Inc. Other reagents were purchased from Sigma.

Plasmids

The human DNase I cDNA (15) containing the signal sequence region was subcloned into the pSVK3 expression vector. Bovine DNase I cDNA lacking a signal sequence (18) was kindly provided by Dr. D. Suck (EMBL, Heidelberg). To prepare the bovine DNase I expression vector, an AflII restriction site was introduced into the human DNase I cDNA at the site where an AflII restriction site already existed in the bovine cDNA (amino acid residues 1-2 of mature sequence). This was accomplished using the sequential polymerase chain reaction procedure (19) and did not result in any change in the amino acid coding sequence. Next the portion of the cDNA encoding the mature human DNase I (minus the signal sequence) was excised (the AflII to HindIII site) and replaced by the same region of the bovine DNase I cDNA.

The expression vectors containing the various mutations in the bovine DNase I cDNA were also constructed using the sequential polymerase chain reaction procedure. All plasmids used for transfections were purified with the Qiagen plasmid kit.

COS-1 Cell Transfections

COS-1 cells were cultured in Dulbecco's modified Eagle's medium containing 10% calf serum. The cells were transfected with purified plasmid DNA using the Lipofectin method as follows: on the day before transfection, approximately 2 × 105 cells were plated into a 60-mm culture dish and incubated at 37 °C for 16 h in a CO2 incubator. Three micrograms of plasmid DNA were dissolved in 100 µl of Opti-MEM. Lipofectin solution was prepared by mixing 5 µg of Lipofectin and 100 µl of Opti-MEM. The DNA and Lipofectin solutions were gently mixed and incubated at room temperature for 15 min. The cells were washed with 2 ml of phosphate-buffered saline followed by 2 ml of Opti-MEM. Then 0.8 ml of Opti-MEM and the Lipofectin-DNA complex were added to the plate. After 24 h of incubation, the transfection reagents were replaced with 3 ml of culture medium. Following an additional 24-h incubation, the cells were washed twice with phosphate-buffered saline and incubated with 1.0 ml of Dulbecco's modified Eagle's medium containing 10% dialyzed bovine serum, 5 mM glucose, 10 mM NH4Cl, and 200 µCi of [2-3H]mannose for 4 h at 37 °C. Then 0.5 ml of Dulbecco's modified Eagle's medium containing 10% calf serum, 10 mM glucose, 10 mM mannose, and 10 mM NH4Cl was added to stop the uptake of [2-3H]mannose. The culture medium was harvested 4 h later.

Immunoprecipitation and Oligosaccharide Analysis

Rabbit anti-bovine DNase I serum was prepared by immunizing a rabbit with bovine DNase I purchased from Sigma and purified one additional step on a concanavalin A-Sepharose column. Rabbit anti-human DNase I serum was prepared using recombinant human DNase I as antigen (15).

The harvested culture media were incubated with 2 µl of antisera overnight with rotation at 4 °C. Then 100 µl of a 50% Protein A-agarose bead suspension was added. After 1 h of additional rotation, the bead suspension was transferred to a fresh tube. The beads were washed three times with 100 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1% Triton X-100, 5 mM EDTA, and 10 trypsin inhibitory units/ml aprotinin. The immunoprecipitated protein was eluted from the beads in 50 µl of 0.5% SDS, 10 mM beta -mercaptoethanol by heating at 78 °C for 10 min and rinsing the beads with 50 µl of water. Oligosaccharide analysis and the calculation of phosphorylation percentage were carried out according to the previously described methods (8).


RESULTS

Mutational Analysis of Lysine Residues in Bovine DNase I

The alignment of bovine, human, and mouse DNase I sequences reveals that they are very similar (Fig. 1). All four proteins have potential Asn-linked glycosylation sites at positions 18 and 106. Bovine DNase I has nine lysine residues (positions 2, 15, 50, 77, 88, 117, 124, 126, and 157). Of interest, there are four positions where mouse DNase I contains lysines but bovine and human DNase I do not (positions 27, 59, 74, and 261).


Fig. 1. Alignment of amino acid sequences of bovine, human, and mouse DNase I. The numbers of the residues refer to the sequence of the mature enzyme minus the signal sequence. N-Linked glycosylation sites are at positions 18 and 106.
[View Larger Version of this Image (42K GIF file)]

Since lysines have been shown to be important components of the phosphotransferase recognition site on acid hydrolases, the nine lysines on bovine DNase I were mutated individually to either an alanine or a glutamine and the effect on mannose phosphorylation determined. In the standard assay, plasmids encoding the various mutant proteins were transfected into COS-1 cells and the expressed proteins were labeled with [2-3H]mannose. The secreted molecules were immunoprecipitated and their [3H]mannose-labeled oligosaccharides were analyzed for the level of phosphorylation by ion exchange chromatography on QAE-Sephadex as described previously (8). Since the COS-1 cells secrete most of the DNase I regardless of the extent of phosphorylation, this material is representative of the total that is synthesized (data not shown). The data are summarized in Fig. 2. The wild-type bovine DNase I had 12.6% of its oligosaccharides phosphorylated, mostly with a single Man-6-P residue. Six of the lysine mutations had no significant effect on the extent of phosphorylation while two mutations (K50A and K124A) caused a modest decrease in phosphorylation (10 and 10.3%, respectively). When these two lysines were mutated in the same construct, inhibition of phosphorylation was more marked (7.7% phosphorylation). It is also of note that the K50A and K124A mutations caused a marked decrease in the formation of oligosaccharides with two Man-6-P residues. In contrast to these results, mutation of Lys-117 right-arrow alanine (K117A) produced a large stimulation of oligosaccharide phosphorylation (28.3%), with most of the oligosaccharides containing a single Man-6-P residue. Thus Lys-117 appears to impair phosphorylation of the oligosaccharides in DNase I. 


Fig. 2. Effect of lysine mutations on oligosaccharide phosphorylation. Constructs were prepared that either replaced lysines that are present in the bovine DNase I sequence with alanines (except for Lys-88 which was changed to the Gln) (Panel A) or inserted a lysine (or an alanine or arginine) at positions where lysines are lacking in the bovine sequence but present in the mouse sequence (Panel B). The various constructs were expressed in COS-1 cells and the extent of oligosaccharide phosphorylation determined.
[View Larger Version of this Image (21K GIF file)]

In the next set of experiments, lysine residues were placed at positions 27, 59, 74, and 261 since mouse DNase I has lysines at these locations (Fig. 1). The addition of a lysine at either position 27 (R27K) or position 74 (N74K) resulted in a 4-fold increase in oligosaccharide phosphorylation (53.9 and 50.3%, respectively) (Fig. 2). Furthermore, the increased phosphorylation in the R27K mutant was exclusively in oligosaccharides with one phosphate while the N74K mutation stimulated the formation of oligosaccharides with two phosphates. When Asn-74 was changed to an arginine, the increase in phosphorylation was only 2-fold, although all of the increase was in oligosaccharides with two phosphates. Replacement of Asn-74 with an alanine had no effect, demonstrating that this asparagine residue does not inhibit phosphorylation. Of note, substitution of Arg-27 with an alanine residue decreased the extent of phosphorylation to 9.4%. Thus at both positions 27 and 74, arginines have a modest stimulatory effect on phosphorylation, but are considerably less active than lysine residues at these positions. The location of the lysines was critical for the stimulatory effect. When the arginine at position 73 was replaced with a lysine (R73K), the oligosaccharide phosphorylation was only 19.6%, in sharp contrast to the effect of having a lysine at position 74. Similarly, a construct with a lysine at position 75 (S75K) was phosphorylated at the basal level of 12.2% (data not shown). When Arg-73 was mutated to Ala (R73A), there was a modest decrease in the basal phosphorylation to 10.2%. However, the R73A/N74K double mutant was phosphorylated as well as N74K single mutant (49.6 and 50.3%, respectively), demonstrating that the arginine at position 73 does not contribute in a major way to the level of oligosaccharide phosphorylation.

The placement of a lysine at position 59 (D59K) had only a small stimulatory effect on phosphorylation (16.5%). To prepare the final construct, threonine 260 was mutated to an arginine and two additional amino acids were added to give the mutation T260RKT. This construct was phosphorylated to the same extent as the wild-type bovine DNase I (12.6%), indicating that a lysine at position 261 does not enhance phosphorylation.

Effect of Multiple Mutations on Oligosaccharide Phosphorylation

The analysis of the various lysine mutants implicated lysines at positions 27 and 74 as being strongly stimulatory, lysines at positions 50 and 124 as being weakly stimulatory, and Lys-117 as impairing oligosaccharide phosphorylation. To determine whether these lysines acted independently or cooperatively, constructs containing various combinations of these residues were prepared and analyzed in the standard COS-1 assay system (Table I). Construct 1, which has the three stimulatory residues present in bovine DNase I mutated to alanines (R27A/K50A/K124A), was phosphorylated only 2.3%, indicating that these residues account for most of the basal phosphorylation that occurs with wild-type bovine DNase I. Construct 2, which has lysine residues at positions 27 and 74 in addition to the lysines at positions 50 and 124, was phosphorylated to a very large extent (79.2%). And construct 3, which contains the lysines at the four positive positions as well as having the inhibitory lysine at position 117 replaced with an alanine, was almost completely phosphorylated (94.9%), with 44% of the oligosaccharides containing one Man-6-P residue and 51% having two Man-6-P residues. By comparison, the lysosomal protease cathepsin D is phosphorylated about 60% in the COS-1 cell expression system whereas the secretory protein glycopepsinogen is phosphorylated less than 0.1% (data not shown).

Table I. Effect of combinations of lysine mutations on oligosaccharide phosphorylation

COS-1 cells transfected with the various constructs were labeled with [2-3H]mannose and the secreted DNase I was immunoprecipitated from the media. The labeled oligosaccharides were analyzed to determine the percentage of phosphorylated Asn-linked oligosaccharides. The results given for constructs 1-3 are the average of two to three separate experiments.

Construct Residues
% Oligosaccharide phosphorylation
Lys-27 Lys-50 Lys-74 Lys-124 Lys-117

WT (R) +  -a + + 12.6
1  -  -  -  - + 2.3
2 + + + + + 79.2
3 + + + +  - 94.9

a -, indicates a Lys right-arrow Ala mutation.

Using the data in Fig. 2 and Table I, it can be calculated that the effects of the lysines at positions 27 and 74 are almost additive. Thus the R27K construct is phosphorylated 41.3% over the basal level (53.9-12.6%) whereas the N74K construct is phosphorylated 37.7% over the basal value (50.3-12.6%). If the effects are independent, the double mutant should be phosphorylated 91.6% (12.6% + 41.3% + 37.7%) and the observed value was 79.2%.

Effect of Lysines on Phosphorylation at Specific Glycosylation Sites

Bovine DNase I has two Asn-linked glycosylation sites, one at position 18 and the other at position 106. To analyze the effects of the various lysines on phosphorylation at these sites, constructs were prepared in which one or the other glycosylation site was deleted. This provided a convenient way to determine whether the lysines were influencing phosphorylation at the individual glycosylation sites or at both sites. The results of these experiments are summarized in Table II. It is apparent that the R27K substitution selectively stimulated phosphorylation of the oligosaccharide at position 18 whereas the N74K substitution caused the selective phosphorylation of the oligosaccharide at position 106. Mutation of the lysine at position 117 to an alanine resulted in enhanced phosphorylation at both sites. Mutation of Lys-50 to alanine caused a selective loss of phosphorylation of the oligosaccharide at position 106. These experiments also revealed that the oligosaccharide at position 106 mostly acquired two Man-6-P residues whereas the oligosaccharide at position 18 acquired only one Man-6-P residue, regardless of the extent of phosphorylation.

Table II. Effect of lysine mutations on phosphorylation at specific glycosylation sites

Constructs were prepared in which either Asn-18 or Asn-106 was changed to a Gln and the other mutations were made as noted. COS-1 cells were transfected with these constructs and labeled with [2-3H] mannose. The secreted DNase I was immunoprecipitated and the labeled oligosaccharides were analyzed to determine the percentage of the Asn-linked oligosaccharide that was phosphorylated and the content of high mannose oligosaccharides with one phosphomonoester (HM + 1PM) and two phosphomonoesters (HM + 2PM).

Glycosylation site % Oligosaccharide phosphorylation
Construct HM + 1PM HM + 2PM Total

Asn-18 WT 13.1 1.4 14.5
R27K 57.4 3.7 61.1
N74K 13.0 0.6 13.6
K117A 33.4 4.1 37.5
R27K/K50A 57.5 4.3 61.8
Asn-106 WT 12.9 7.5 20.4
R27K 11.4 9.2 20.6
N74K 7.2 81.6 88.8
K117A 18.6 22.8 41.4
N74K/K50A 13.3 49.2 62.5


DISCUSSION

DNase I is a major secretory protein of the pancreas and parotid glands, but it is also present in other tissues, plasma, and urine (20). The DNase I produced in the bovine pancreas contains a single oligosaccharide located at Asn-18 with the potential Asn-linked glycosylation site at position 106 not being utilized (21). In this regard Shakin-Eshleman and co-workers (22) have reported that the sequence Asn-Asp-Ser (which is the sequence at positions 106-108 in bovine DNase I) is poorly glycosylated in a cell-free translation system supplemented with canine pancreas microsomes whereas the sequence Asn-Ala-Thr (which is the sequence at positions 18-20 in DNase I) is efficiently glycosylated. Since the bovine DNase I expressed in COS-1 cells is well glycosylated at both sites, there must be cellular differences in the preference for the X residue in the sequence Asn-X-(Ser/Thr). The fact that the COS-1 cells utilized both glycosylation sites on DNase I made this a useful cell line for this study. It is also advantageous that the COS-1 cells secrete the majority of DNase I under the conditions of the assay.

The initial indication that DNase I was a substrate for phosphotransferase came from the work of Cacia et al. (14) who showed that DNase I isolated from human urine contained Man-6-P residues on its Asn-linked oligosaccharides. In preliminary experiments we found that about 4% of the oligosaccharides of human DNase I expressed in COS-1 cells contain Man-6-P residues, compared with the 12.6% value that was obtained with bovine DNase I. Bovine DNase I was therefore chosen for use in the experiments to define the phosphotransferase recognition domain.

Our data indicate that most of the base-line phosphorylation of bovine DNase I depends on three residues, Lys-50, Lys-124, and Arg-27. When these residues were mutated to alanines, the extent of mannose phosphorylation dropped from 12.6 to 2.3%. Substitution of these residues individually showed that each contributes about equally to the phosphorylation. Since replacement of all three residues failed to decrease phosphorylation to the level observed in a standard secretory glycoprotein (<0.1%), we conclude that the DNase I must have additional residues that contribute, albeit weakly, to the phosphotransferase-binding site. It is important to note, however, that not every lysine exhibits activity in this assay. Thus mutation of the lysines at positions 2, 15, 20, 77, 80, 126, and 157 to alanines did not impair mannose phosphorylation.

The most dramatic effects were observed when lysine residues were placed at positions 27 and 74 (the R27K and N74K constructs). Individually these mutations stimulated phosphorylation about 4-fold and the double mutant had 79% of its oligosaccharides phosphorylated. Substitution of Asn-74 with an arginine (N74R) caused a 2.5-fold enhancement of phosphorylation while the N74A mutation had no effect on this process. This latter result excludes the possibility that the stimulatory effect of the lysine (and arginine) substitution at position 74 is actually due to relief from an inhibitory effect of the asparagine at this position. At both positions 27 and 74 it is apparent that lysine is much more stimulatory than arginine, indicating that phosphotransferase interacts best with lysines while having some affinity for arginines. These findings confirm results obtained with cathepsin D where a lysine at position 203 is much more potent than an arginine at that position in stimulating oligosaccharide phosphorylation (5). When all the lysine mutagenesis data is considered together, it is apparent that lysines at four positions (27, 50, 74, and 124) determine the level of oligosaccharide phosphorylation between 2.3 and 79.2%. The other lysines, with the exception of Lys-117 (see below), have little or no effect on this process. This is analogous to findings that have been made with several lysosomal hydrolases. In the case of procathepsin L, two lysine residues in the propeptide are required for efficient phosphorylation while mutation of 19 other lysines does not impair this process (10). Similarly, while certain lysine residues of cathepsin D contribute to its phosphorylation (4, 5, 10), others are not involved in its targeting to lysosomes (23). Thus in the case of DNase I, as in acid hydrolases, lysine residues at particular locations are key elements of the phosphotransferase recognition domain (4-10).

Another striking finding was that the R27K substitution selectively stimulated phosphorylation of the oligosaccharide at Asn-18 whereas the N74K substitution resulted in phosphorylation of the oligosaccharide at Asn-106. Furthermore, the Asn-18 oligosaccharide primarily acquired a single Man-6-P residue while the Asn-106 mostly acquired two such residues. These results raise two fundamental issues concerning phosphotransferase action. First, how does a lysine residue stimulate phosphorylation at one glycosylation site while having no effect on phosphorylation at another site on the same protein? And second, why is it that the oligosaccharide at Asn-18 only acquires a single phosphate? To answer the first question, it is useful to examine the location of the critical lysines relative to the two Asn-linked oligosaccharides (Fig. 3). It is apparent that Lys-27 and Lys-74 are located relatively close to the oligosaccharide whose phosphorylation they affect and a considerable distance from the other oligosaccharide. Therefore an attractive possibility is that phosphotransferase can bind DNase I in two orientations, with Lys-27 and Lys-74 determining which orientation is achieved. When bound in one orientation, phosphotransferase can reach one of the oligosaccharides but not the other, and when bound in the other orientation the opposite would occur. This model suggests that DNase I molecules containing both Lys-27 and Lys-74 will first bind to phosphotransferase in one orientation to allow phosphorylation of the accessible oligosaccharide. The DNase I will then dissociate and undergo rotational and short range translational diffusion in the vicinity of phosphotransferase resulting in rebinding in the alternate orientation. This would be followed by phosphorylation of the other oligosaccharide. These results, together with our previous findings with cathepsin D, point out the complexity of the phosphotransferase recognition domain. It is possible, however, that in some proteins, such as cathepsin L, the phosphotransferase recognition element may be a simpler structure (10).


Fig. 3. Space filling model of bovine DNase I. Two views are shown to illustrate the location of Asn-74, Lys-50, Lys-117, and Arg-27 relative to the glycosylation sites at Asn-18 (with core portion of oligosaccharide visible) and Asn-106. This model was developed by Suck and co-workers (16, 17).
[View Larger Version of this Image (103K GIF file)]

In a previous study we demonstrated that while there is considerable flexibility in the placement of glycosylation sites on cathepsin D in terms of the ability of oligosaccharides to be phosphorylated by phosphotransferase, oligosaccharides located closer to the protein recognition domain were preferentially phosphorylated (7, 24). We suggest that DNase I may represent the extreme example of this process and to test this we plan to move the glycosylation sites to different regions on the DNase I molecule and determine how far phosphotransferase can "reach" when it interacts with either Lys-27 or Lys-74. We predict that phosphotransferase will only be able to phosphorylate oligosaccharides in the general region of the critical lysine.

The reason why the oligosaccharide at position 18 only acquires a single Man-6-P residue is unclear at this time. Lazzarino and Gabel (25) showed that the terminal mannose on the alpha 1,6-branch of the core beta -linked mannose of the high mannose oligosaccharide must be removed in order for the penultimate mannose to be phosphorylated (the first phosphate is usually transferred to a mannose on the alpha 1,3-branch of the oligosaccharide) (25). Thus one possibility is that the terminal mannose residue on the alpha 1,6-branch is not removed on the oligosaccharide at Asn-18, thereby preventing the acquisition of the second Man-6-P residue. Alternatively, the oligosaccharide at Asn-18 may be oriented relative to the phosphotransferase-binding site in such a way that only one arm of the high mannose unit can enter the catalytic site of phosphotransferase. It is of interest that a similar phenomenon occurs on cathepsin D where the oligosaccharide at position 70 only receives one phosphate while the oligosaccharide at position 199 usually receives two phosphates (7, 8, 26).

An unexpected result was that replacement of lysine at position 117 with an alanine caused a 2-3-fold stimulation of phosphorylation at both glycosylation sites. This indicates that a lysine residue at this particular location actually impairs phosphorylation, perhaps by serving as a binding site for phosphotransferase and orienting the enzyme in a manner that is unfavorable for phosphorylation of the existing oligosaccharides. Regardless of the explanation, this finding raises the possibility of "inhibitory" amino acids that function to prevent phosphorylation. Since mouse DNase I, which contains both Lys-27 and Lys-74, is phosphorylated only slightly better than bovine DNase I when expressed in COS-1 cells,2 it seems likely that this protein contains inhibitory residues in addition to Lys-117. It may turn out that there are a number of mechanisms for preventing phosphotransferase from acting on secretory glycoproteins that otherwise might express functional binding sites for this enzyme.

In summary, the productive interaction of bovine DNase I with phosphotransferase requires critical lysine and arginine residues at specific locations on the surface of the protein, similar to the case of authentic lysosomal hydrolases. However, bovine DNase I is a relatively weak substrate for phosphotransferase and consequently the oligosaccharides in the majority of the molecules are not phosphorylated. Therefore most of the newly synthesized DNase I is secreted rather than being diverted to lysosomes. This process is enhanced further in the pancreas where the potential glycosylation site at Asn-106 is not utilized.


FOOTNOTES

*   This work was supported in part by United States Public Health Service Grant CA08759 and by a grant from Genentech.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Dagger    Current address: Dept. of Biochemistry, Okayama University of Science, Okayama 700, Japan.
   To whom correspondence should be addressed: Dept. of Medicine, Washington University School of Medicine, 660 S. Euclid Ave., Box 8125, St. Louis, MO 63110. Tel.: 314-362-8803; Fax: 314-362-8826.
1   The abbreviations used are: phosphotransferase, UDP-GlcNAc:lysosomal enzyme N-acetylglucosamine-1-phosphotransferase; Man-6-P, mannose 6-phosphate.
2   A. Nishikawa, W. Gregory, J. Frenz, and S. Kornfeld, unpublished results.

REFERENCES

  1. Kornfeld, S. (1986) J. Clin. Invest. 77, 1-6 [Medline] [Order article via Infotrieve]
  2. von Figura, K., and Hasilik, A. (1986) Annu. Rev. Biochem. 55, 167-193 [CrossRef][Medline] [Order article via Infotrieve]
  3. Reitman, M. L., and Kornfeld, S. (1981) J. Biol. Chem. 256, 11977-11980 [Abstract/Free Full Text]
  4. Baranski, T. J., Faust, P. L., and Kornfeld, S. (1990) Cell 63, 281-291 [Medline] [Order article via Infotrieve]
  5. Baranski, T. J., Koelsch, G., Hartsuck, J. A., and Kornfeld, S. (1991) J. Biol. Chem. 266, 23365-23372 [Abstract/Free Full Text]
  6. Baranski, T. J., Cantor, A. B., and Kornfeld, S. (1992) J. Biol. Chem. 267, 23342-23348 [Abstract/Free Full Text]
  7. Cantor, A. B., Baranski, T. J., and Kornfeld, S. (1992) J. Biol. Chem. 267, 23349-23356 [Abstract/Free Full Text]
  8. Dustin, M. L., Baranski, T. J., Sampath, D., and Kornfeld, S. (1995) J. Biol. Chem. 270, 170-179 [Abstract/Free Full Text]
  9. Cuozzo, J. W., and Sahagian, G. G. (1994) J. Biol. Chem. 269, 14490-14496 [Abstract/Free Full Text]
  10. Cuozzo, J. W., Tao, K., Wu, Q., Young, W., and Sahagian, G. G. (1995) J. Biol. Chem. 270, 15611-15619 [Abstract/Free Full Text]
  11. Herzog, V., Neumuller, W., and Holzmann, B. (1987) EMBO J. 6, 555-560 [Abstract]
  12. Lee, S., and Nathans, D. (1988) J. Biol. Chem. 263, 3521-3527 [Abstract/Free Full Text]
  13. Kovacina, K. S., Steele-Perkins, G., Purchio, A. F., Lioubin, M., Miyazona, K., Heldin, C. H., and Roth, R. A. (1989) Biochem. Biophys. Res. Commun. 160, 393-403 [Medline] [Order article via Infotrieve]
  14. Cacia, J., Quan, C., and Frenz, J. (1995) Glycobiology 4, 99
  15. Shak, S., Capon, D. J., Hellmiss, R., Marsters, S. A., and Baker, C. L. (1990) Proc. Natl. Acad. Sci. U. S. A. 87, 9188-9192 [Abstract]
  16. Suck, D., Oefner, C., and Kabsch, W. (1984) EMBO J. 3, 2423-2430 [Abstract]
  17. Suck, D., Lahm, A., and Oefner, C. (1988) Nature 332, 464-468 [CrossRef][Medline] [Order article via Infotrieve]
  18. Worrall, A. F., and Connolly, B. A. (1990) J. Biol. Chem. 265, 21889-21895 [Abstract/Free Full Text]
  19. Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith, J. A., and Struhl, K. (1989) Current Protocols in Molecular Biology, Greene Publishing Associates, Brooklyn, NY
  20. Mannherz, H. G., Peitsch, M. C., Zanotti, S., Paddenberg, R., and Polzar, B. (1995) Curr. Top. Microbiol. Immunol. 198, 161-174 [Medline] [Order article via Infotrieve]
  21. Paudel, H. K., and Liao, T. H. (1986) J. Biol. Chem. 261, 16012-16017 [Abstract/Free Full Text]
  22. Shakin-Eshleman, S. H., Spitalnik, S. L., and Kasturi, L. (1996) J. Biol. Chem. 271, 6363-6366 [Abstract/Free Full Text]
  23. Schorey, J. S., Fortenberry, S. C., and Chirgwin, J. M. (1995) J. Cell Sci. 108, 2007-2015 [Abstract/Free Full Text]
  24. Cantor, A. B., and Kornfeld, S. (1992) J. Biol. Chem. 267, 23357-23363 [Abstract/Free Full Text]
  25. Lazzarino, D., and Gabel, C. (1989) EMBO J. 3, 5015-5023
  26. Dittmer, F., Pohlmann, R., and von Figura, K. (1997) J. Biol. Chem. 272, 852-858 [Abstract/Free Full Text]

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.