J. Mol. Biol. (1992) 226. 555-557

CRYSTALLIZATION NOTES

Purification

and Crystallization

of Human Cathepsin D

Martin Fusek1*2,Miroslav BaudyS2p3and Peter Metcalf’? 1European Molecular Biology Laboratory Meyerhofstrasse 1, Postfach 10.2209 09600 Heidelberg, Germany 21nstitute

of Organic CS-16610,

Chemistry and Biochemistry Prague, C.S.F.R.

‘oniver&~ityo~4~~~h,~S~lt~Lah~e , . . . (Received 18 February

City

1992; accepted 4 March

1992)

The two-chain form of human cathepsin D was purified from human spleen with a method utilizing an ion exchange chromatography step prior to the pepstatin affinity column normally used to purify aspartic proteases. The protein was crystallized from 21 y. polyethylene glycol 8000 at pH 4.0 using the hanging drop vapour diffusion method. Small crystals were used as seeds to grow crystals suitable for X-ray data collection. The crystals diffract to a resolution of 3.2 a and have space group P2,2,2, with unit cell dimensions a = 599 A, b = 99.6 A, c = 133.6 8. There are two molecules in the asymmetric unit. Keywords: cathepsin D; aspartic protease; lysosomal targeting; three-dimensional structure

polypeptide structure (Baranski et al., 1990) that is recognized by N-acetylglucosamine phosphotransferase. The phosphorylation of mannose residues of the two cathepsin D carbohydrates by the phosphotransferase serves as a signal for mannose-6phosphate receptor molecules to transport lysosomal hydrolases to lysosomes (Kornfeld, 1987). Recently the three-dimensional structure of the cysteine lysosomal protease, cathepsin B, was reported (Musil et al., 1991) and since no homology in primary structure between the two enzymes is evident, the common structural features necessary for t,he recognition by the phosphotransferase are probably to be found by a comparison of the threedimensional structures. Previous crystallization trials using bovine cathepsin D have led to inconsistent results, and the heterogeneity of different preparations resulting from the different post-translational modifications was probably the reason for this. The production of sufficient yields of recombinant human cathepsin D for crystallization trials has also proved difficult (Conner & Udey, 1990). Here, we report the purificat,ion of pure two-chain form of human spleen cat’hepsin D, its crystallization and preliminary crysta,llographic results. Human spleen (950 g) was homogenized and suspended in three litres of 2% (w/v) NaCl and the pH adjusted to pH 5.0 by the addition of 2 M-HCl

Cathepsin D (EC 3.4.23.5) is a member of the aspartic protease family and is the main aspartic protease occurring within the lysosomes in cells of most mammalian tissues (Shewale et al., 1985). Cathepsin D has been reported to play important roles in both physiological and pathological processes. It was found to be an independent prognostic factor for the development of metastases of primary breast cancer tumours (Spyratos et al., 1989; Tandon et aZ., 1990) and has been reported to be involved in muscular dystrophy (Gopalan et al., 1987) and brain changes during aging (Matus & Green, 1987; Kenessey et al., 1989). Like other soluble lysosomal enzymes, cathepsin D undergoes complex post-translational modifications including the glycosylation of two asparagine residues, the removal of the 45 residue N-terminal pro-peptide on act’iva,tion and partial endoproteolytic hydrolysis producing a mixture of single and two-chain forms of the mature glycoprotein (Hasilik & Neufeld, 1980). The determination of the three-dimensional structure of cathepsin D is likely to lead to a better understanding of the targeting of lysosomal proteases. Like other soluble lysosomal hydrolases cat,hepsin D is presumed to contain a site on its t Author to whom all correspondence should be addressed 0022-2836/92/140555~3

$03.00/O

crystal;

555

0 1992 Academic Press Limited

556

M. Fused et al.

I-

- 0.6

Na Cl gradlent

0

I I I I,, 50

I,, 100

, , ,I ,, I- , , 1 150 200 Fraction

Figure 1. DEAE Sephadex A 50 chromatography first, eluted with 10 mM-sodium phosphate (pH 63) was recovered from the pooled fractions, A, before chain cathepsin D in addition to the 2-chain form,

, ,

, 250

,

number

of the active fraction from the gel filtration column. The column was and then with a 0 to 500 mM-NaCl gradient. Two-chain cathepsin D the salt gradient. Fractions B and C contained un-processed singleas shown by SDS electrophoresis (see Fig. 2).

while stirring at 4°C. After one hour, 30 g of NaCl was added and the pH adjusted to 3.5 with 2 M-HCl. The suspension was then stirred overnight at 4°C. The supernatant after centrifugation (10,000 g,. 1 h) was adjusted to 30% (w/v) saturated ammonium sulphate. A second precipitation step at 65% saturated ammonium sulphate was then carried out. The pellet after centrifugation was dissolved in 300 ml of Hz0 and dialysed extensively against 1 mM-EDTA. After the dialysis and concentration by lyophilization to 60 ml, the material was chromatographed on a three litre Sephadex G-150 gel filtration column previously equilibrated in 10 mMsodium phosphate (pH 6%). Fractions from the column were assayed for protease activity using the haemoglobin method (Huang et al., 1979). The active fractions were pooled, concentrated and dialyzed against 10 mM sodium phosphate (pH 63). The dialysed solution (150 ml) was chromatographed using a DEAE Sephadex A50 column previously equilibrated with 10 mM-sodium phosphate (pH 6.3) as shown in Figure 1. Active fractions were pooled and, after adjusting the pH to 3.5, chromatographed on a pepstatin A Sepharose column. . The active fractions were then dialyzed against H,O and lyophilized. The procedure used here for the isolation of cathepsin D produced the fully processed two-chain form as judged by SDS electrophoresis (Fig. 2). The key step in the purification protocol is equilibrium DEAE ion exchange chromatography. The two-chain form is only slightly retained by the column and is eluted before the gradient. The other forms are then eluted by the salt gradient. This approach, followed by affinity chromatography, produced a yield of approximately 1.5 mg of pure two-chain form from 950 g of human spleen.

‘4! j kDa

31 kDa

14 kDa

Figure 2. SDS gel electrophoresis (15% (w/v) acrylamide) of cathepsin D after ion exchange and pepstatin A Sepharose affinity chromatography. Lanes A, B and C correspond to pooled fractions as shown in Fig. 1. Lane A contains 2-chain cathepsin D (14 and 31 kDa) as used for crystallization trials; lane B contains a mixture of clipped and unclipped cathepsin D and lower molecular mass contaminants; lane C contains predominantly intact cathepsin D (45 kDa). The heterogeneity seen is probably a result of incomplete processing and non-uniform glycosylation.

Crystallization

Notes

processed using the BUDDHA software package (Blum et al., 1987). The unit cell was determined by autoindexing and the space group by the symmetry and systematic absences of the data. A 95% complete 3.2 A (1 A = 0.1 nm) data set was collected from a single crystal and a molecular replacement search using a pepsin model revealed a dimer in the asymmetric unit. The R-factor for the current partially refined model is 24 o/O. The structure will be described separately. The space group is P2,2,2, with unit cell dimensions a = 599 8, b = 99.6 8, c = 133.6 8.

References

Figure 3. Still photograph from a human cathepsin D crystal showing diffraction to 3.2 A (arrows). The exposure time was 2 h, the camera length 10 cm, and the radiation source an Enraf-Nonius GX21 rotating anode generator operated at 40 kV, 70 mA with a 300 pm focussing cup and equipped with a Ni filter.

Only the two-chain form of cathepsin D was used for the crystallization trials. The crystallization conditions used were similar to those previously determined for bovine liver cathepsin D (R. Arni & P. Metcalf, unpublished results). Lyophilized protein was dissolved in 10 mm-Tris . HCl (pH 7-O) to a Drops 25 mg/ml. concentration of protein containing 1 ~1 of protein solution and 1 ~1 of well buffer (21 y. (w/v) polyethylene glycol 8000) in 50 mM-sodium citrate (pH 4.0) buffer containing 6 mM-KSCN) were made on silanized cover slips and equilibrated against 1 ml of well buffer. Small crystals appeared within two weeks and did not grow further. The small crystals prepared this way were washed by sequential transfer through drops containing 1@5, 9, 9 and 8.25% polyethylene glycol 8000 in 50 mlrr-sodium citrate (pH 4.0) buffer. The then added to drops washed crystals were containing 4 ~1 of a 25 mg protein/ml solution and 4 ~1 of 16% polyethylene glycol 8000 in 50 mMsodium citrate (pH 4.0) buffer. The drops were equilibrated against the same well solution as used for growing the seed crystals. The crystals grew to a maximum dimension of 0.25 mm within two weeks. Crystals were harvested in well buffer and mounted in 0.5 mm quartz capillaries for X-ray data collection. Data were collected with a rotating anode source and a Siemens XlOOA area detector and

Baranski, T. ,J., Faust, P. L. & Kornfeld, S. (1990). Generation of a lysosomal enzyme targeting signal in the secretory protein pepsinogen. Cell, 63. 281-291. Blum, M., Metcalf, P., Harrison, S. C. &. Wiley, D. C. (1987). A system for collection and on-line integration of X-ray diffraction data from a multiwire area detector. J. Appl. Crystallogr. 20, 235-242. Connor, G. E. & Udey, J. A. (1990). Expression and refolding of recombinant human fibroblast procathepsin D. DNA Cell Biol. 9, l-9. Gopalan, P., Dufrense, M. J. & Warner, A. H. (1987). Thiol protease and cathepsin D activities in selected tissues and cultured cells from normal and dystrophic mice. Can. J. Physiol. Pharmacol. 65, 124-129. Hasilik, A. & Neufeld, E. F. (1980). Biosynthesis of lyso-

somal enzymes in fibroblasts. J. Biol. Chem. 255. 49374945. Huang. J. S., Huang, S. S. & Tang, J. (1979). (lathepsin D isozymes from porcine spleen. J. Biol. Chrm. 254. 11405-11417. Kenessey, A., Banay-Schwartz, M., DeGuzman, T. & Lajtha, A. J. (1989). Increase in cathepsin D activity in rat brain in aging. Neurosci. Res. 23, 454-456. Kornfeld, 8. (1987). Trafficking of lysosomal enz.ymes.

FASEB J. 1, 462468. Matus, A. & Green, G. D. J. (1987). Age-related

increase

in cathepsin D like protease that degrades brain microtubule-associated proteins. Biochrmistry, 26, 8083-8086. Musil, D., Zucic, D., Turk. D.. Engh. R. 4.. Mayr, I..

Huber, R., Popovic? T., Turk, I’.. Towatari. T.. Katunuma, N. & Bode, W. (1991). The refined 2.15 A X-ray crystal the structural

structure of human liver cathepsin B: basis for its specificity. EMBO J. 10,

2321-2330. Shewale, J. G., Takahashi,

T. & Tang. J. (1985). The primary structure of cathepsin D and the implications for its biological functions. In Aspartic Proteinases and their Inhibitors (Kostka, V., ed.), pp. 101-l 16, Walter de Gruyter & Co., Berlin, New York. Spyratos, F., Brouillet,