DE19727437A1 - Improving the structure determination of biological aggregates in X-ray crystallography - Google Patents

Abstract

The invention relates to novel crystalline heavy metal cluster compounds and the crystallographic data thereof, in addition to the use of said crystallographic data in x-ray structural analysis. The invention specially relates to improvements in determining the structure of biological molecules. By using said method, the identification of effectors for such biological molecules can also be improved.

Description

The invention relates to new crystalline heavy metal cluster compounds and their crystallographic data and the use of this crystallograph phical data in X-ray structure analysis, especially Verbes changes in the structure determination of biological molecules. With help This method can also be used to identify effectors for such biological molecules are improved.

Heavy metal cluster compounds are of great importance for that Analysis of biological structures, especially for structure determination large protein systems. They are used in macromolecular crystallography to assign phases for the measured structure factor amplitudes used (see e.g. Thygesen et al., Structure 4 (1996), 513-518). By Insertion of these electron-rich compounds into the protein crystals measurable changes in the reflection intensities are generated, whereby a phase determination is made possible.

Elements of the second and third transition groups, e.g. B. niobium and tantalum, in their low oxidation states form a series of metal atom cluster compounds, of which, for example, the M ₆ Hal ₁₂ ²⁺ species are easily accessible. Ta ₆ Br ₁₂ ²⁺ and Ta ₆ Cl ₁₂ ²⁺ clusters are particularly suitable for protein crystallography because they are electron-rich, stable and water-soluble. Above all, the Ta ₆ Br ₁₂ ²⁺ cluster can be easily localized at low resolution by analyzing the difference Patterson function.

A tantalum bromide compound with the formula Ta ₆ Br ₁₄ × 7 H ₂ O was already produced in 1910 (Chapin, J. Am. Chem. Soc. 32 (1910), 323-326). In 1950 Pauling and co-workers proposed a structure for the Ta ₆ Br ₁₂ ²⁺ cluster in which six tantalum atoms are positioned at the corners of a regular octahedron and bridged by 12 bromine atoms over the 12 edges of the octahedron (Vaughan et al., J. Am. Chem. Soc. 72 (1950), 5477-5486). Exact crystallographic data for the Ta ₆ Br ₁₂ ²⁺ cluster are not known, however, since the preparation of crystals required for the structure determination has so far not been successful.

The Ta ₆ Br ₁₂ ²⁺ cluster has nevertheless already been successful in dissolving the structures of large protein aggregates such as the proteasome with 673 kd (Loewe et al., Science 268 (1995), 533-539); the transketolase with 680 amino acid residues per asymmetric unit (Lindqvist et al., EMBO J. 11 (1992), 2373-2379) the ribulose-1,5-bisphosphate carboxylase / oxygenase with 550 kd (Andersson et al., Nature 337 ( 1989), 229-234), the GTP cyclohydrolase I, a 250 kd homodecamer (Nar et al., Structure 3 (1995), 459-466) and the 86 kd dimethyl sulfoxide reductase (Schneider et al., J. Mol Biol. 263 (1996), 53-69). Up to now, the cluster has been treated as a single scattering center, which, however, shows only unsatisfactory results at a higher resolution (ie higher than 0.6 nm). The use of a model in which the cluster is treated as a Ta ₆ octahedron, i.e. as several scattering centers or an ensemble of several point spreaders, in a random orientation with its focus at the position of the individual scattering center led to an improved phase determination (Schneider and Linqvist, Acta Cryst. D 50 (1994), 186-191 and Stanford, Acta Cryst. 15 (1962), 805-806). Nevertheless, this method also has the disadvantage that a precise determination of the cluster orientation has so far not been possible due to the lack of exact crystallographic data.

One object underlying the present invention was to at least partially eliminate the disadvantages of the prior art. In particular, new crystallized tantalum halide cluster compounds and improved methods for structure determination of macromolecular Connections by X-ray crystallographic methods under Ver Heavy metal cluster compounds are provided.

The object of the invention is achieved by crystalline compounds of the formula:

[Ta ₆ Hal ₁₂ (H ₂ O _n )] ²⁺ X ₁ ⁻X ₂ ⁻.m H ₂ O

wherein Hal is a halide ion selected from Cl⁻ and Br⁻,
X ₁ ⁻ and X ₂ ⁻ each represent a single negatively charged anion,
n is a number from 0 to 6 and
m represents a number from 0 to 6.

Hal preferably means in the compound Br according to the invention. X ₁ ⁻ and X ₂ ⁻ are any simply negatively charged inorganic or organic anions. X ₁ - and X ₂ - are preferably selected from Br⁻ and OH⁻. Most preferably, the crystalline compound of the invention has the formula [Ta ₆ Br ₁₂ (H ₂ O) ₆ ] ²⁺ Br⁻OH⁻ × 5 H ₂ O.

By means of the present invention, crystallization of the Ta ₆ Br ₁₂ ²⁺ cluster could be achieved for the first time, specifically in a saturated solvent atmosphere. Suitable single crystals for X-ray structure analysis were obtained in the form of green platelets with dimensions up to 300 µm, the crystal structure of which could be determined by X-ray diffraction analysis with resolution in the atmospheric region. The cluster is a regular octahedron consisting of 6 metal atoms with 12 bridging bromine atoms along the 12 edges of the octahedron. The cluster is compact, of approximately spherical shape with a radius of approximately 0.43 nm and has a high degree of symmetry. The crystal structure data are shown in Table 1.

The crystallization of the Ta ₆ Br ₁₂ ²⁺ cluster according to the invention provides precise crystallographic structural data of this compound for the first time. This structural data can be used in X-ray crystallography, in particular to improve the phase determination in crystals of biological molecules such as protein molecules. Furthermore, the improved structure determination method also makes it easier to identify effectors for biological molecules.

Using the crystallographic data of the Ta ₆ Br ₁₂ ²⁺ cluster according to the invention in investigations on a number of protein molecules, such as an antibody V _L domain and the DMSOR, a significantly simplified solution to the structure can be compared by the availability of improved phasing statistics for the initial phases known methods can be achieved. The cluster can be used as an ensemble of 18 point scatterers (6 Ta and 12 Br atoms) in the model calculation, whereby a phase determination can be carried out with a significantly better resolution than 0.6 nm. In addition, it is possible to determine the correct orientation of the cluster within biological structures using the crystal data according to the invention, whereby a further improvement of the initial phases can be achieved. Thus, when analyzing crystals of two different proteins (antibody V _L domain and DMSOR) it was shown that the phase error when using the phasing method according to the invention is more than 30 ° less than when calculating the single point using a resolution of 0.22 nm .

A further possibility of using the crystallographic data according to the invention for the Ta ₆ Br ₁₂ ²⁺ cluster is, for example, the MAD method (multiple wavelength anomalous dispersion), the multi-wavelength anomaly dispersion (MAD), since the cluster has two types of atoms with different absorption properties contains, namely Ta (L _III edge 0.125 nm) and Br (K edge 0.095 nm), each of which has a significantly abnormal signal. When using the MAD technique, a crystal is sufficient to obtain phases, provided that no radiation damage occurs during the measurement of data records at different wavelengths. For example, cryogenic cooling devices can be used for this purpose.

Another object of the present invention is a method for determining the structure of a macromolecular compound by X-ray crystallography, characterized in that

• (a) a derivative of the compound with a symmetrical heavy metal cluster creates,
• (b) the correct binding site of a heavy metal cluster in the derivative determined and
• (c) the correct orientation of a heavy metal cluster within the Derivative determined at the binding site.

The method according to the invention is particularly suitable for the phase determination of biological molecules, in particular of protein molecules. Suitable symmetrical heavy metal clusters that can be used in this process are, for example, Ta ₆ Cl ₁₂ ²⁺ or Ta ₆ Br ₁₂ ²⁺ . Ta ₆ Br ₁₂ ²⁺ is particularly preferred.

Step (b) of the method according to the invention comprises determining the correct binding site of a heavy metal cluster in the derivative. If several heavy metal clusters are specific to the one to be determined Attach connection, several could in this step Binding sites can be determined if desired. The determination of Binding site can by known methods, for example by  Analysis of the difference Patterson function done. For this, con conventional computer program packages e.g. B. the PROTEIN or CCP4 can be used.

Step (c) of the method according to the invention comprises determining the correct orientation of a heavy metal cluster at the binding site, what the use of the cluster as a multi-point ensemble spreading requires. The cluster orientation is determined with the help of "Molecular Replacement" methods using differences of the Structure factor amplitudes from the native and derivative data set certainly. Preferably, all heavy atoms of the cluster are in the Calculation included.

The optimal orientation is preferably determined by Calculation of structure factor amplitudes in several possible spatial Orientations of the cluster model and determination of the optimal spatial Orientation by minimizing the difference between the dates of the theoretically resulting structures and the practically measured ones Structure. For this purpose, the cluster is preferably increased by at least one out of three Space axes, particularly preferably around all three space axes, in several Steps turned. For example, the rotation of the cluster can be in steps of 10 to 30 °, preferably in steps of about 15 °. Structure factor am is calculated at several of these stages flounder. The optimal spatial can then be obtained from the resulting data Orientation of the cluster can be determined, preferably by determination of the R factor and / or the Patterson correlation value. A maximum of Patterson correlation value, which is a difference between theoretical and expressed practically measured structure corresponds to that Orientie the real orientation of the cluster in the compound derivative corresponds. Alternatively or additionally, the real orientation of the Heavy metal clusters also determined via a minimum of the R factor will. Suitable calculations exist for performing such calculations  Computer program packages, e.g. B. the program packages X-PLOR, AMoRe or GLRF. The X-PLOR program package is particularly preferred.

When several heavy metal cluster molecules bind to the one to be determined Macromolecular connection can help determine cluster orientation for each cluster either independently or in correlation with each other be performed.

The present invention is further intended to be illustrated by the following examples are explained in more detail.

Examples example 1 Preparation of crystalline [Ta ₆ Br ₁₂ (H ₂ O) ₆ ] ²⁺ Br⁻OH⁻ × 5H ₂ O

A quartz reaction vessel was degassed in vacuo and filled with 16.25 g (28 mmol) tantalum pentabromide, 4.76 g (40 mmol) potassium bromide and 18.1 g (100 mmol) tantalum powder under nitrogen. After thorough mixing, the reaction vessel was vacuum sealed, heated to 700 ° C within 10 h and held at 700 ° C for a further 15 h. The product of this reaction, K ₄ Ta ₆ Br ₁₈ , was extracted with four 250 ml portions of water, filtered and treated with an equal volume of concentrated hydrobromic acid with stirring and heating to 80 ° C.

To air oxidation of the cluster in the following steps prevent, the dark green solution was saturated with argon, sealed and stored at 4 ° C for two hours. The precipitate was at Centrifuged 11,000 g in a JA10 rotor (Beckmann) and the obtained Product dried in a vacuum over phosphorus pentoxide.  

An aqueous solution of the dark green powder was used to suitable conditions for crystallization by evaporation determine. However, all crystals decomposed on drying.

In contrast, crystallization could be achieved using the Sitztropfen vapor diffusion method from drops which equilibrated 3 μl Ta ₆ Br ₁₄ solution (2 mg / ml in water) against a 1 ml of an aqueous Ta ₆ Br ₁₄ reservoir solution (50 mg / ml). Attempts to analyze these crystals by standard crystallographic techniques for inorganic compounds led to decomposition. To overcome this problem, the crystals were analyzed in a glass capillary in the presence of the reservoir buffer.

Example 2 Crystallographic data

The crystallographic data were obtained on a Nonius CAD4 diffractometer using graphite-monochromatized CuK _α radiation from a rotating anode generator RU200 (Rigaku, Tokyo, Japan), which with 5.4 kW with a crystal of the dimensions 0.15 mm × 0, 11 mm × 0.08 mm was operated. The space group of the crystals is P1 with a = 0.9475 nm, b = 0.9631 nm, c = 1.0571 nm, α = 69.14 °, β = 82.42 °, γ = 62.69 °. 2717 independent reflections in the range from 4 <<65 ° C. to a resolution of 0.0851 nm were obtained. An empirical absorption correction using the psi-scan method was applied to all reflexes.

The chemical composition of the crystalline material was derived from the Derived electron density and, as discussed below, is not complete clearly. The structure was determined by direct methods using the SHELXS-86 program (Sheldrick, Acta Cryst. A. 46 (1990), 467-473) dissolved and refined by further development of this program. The final R factor, taking into account all reflections, was 0.0763.  

In a unit cell there are two copies of the [Ta ₆ Br ₁₂ aq ₆ ] ²⁺ cluster, the centers of which coincide with an inversion center of space group P1. The 6 tantalum atoms are arranged at the tips of an octahedron. Each of the 12 Br atoms is coordinated with two tantalum atoms along the 12 edges of the octahedron. In addition, a water ligand is bound to each tantalum atom.

The arrangement of the two cluster molecules is such that they are located around the same inversion center. The occupancies were determined more precisely during the crystallographic refinement and converged to an average value of 0.43 for all atoms of the two cluster molecules. Furthermore, the unit cell has an electron density which can be interpreted as two bromine and twelve oxygen atoms with degrees of occupancy of about 0.5 each. These Br ions are arranged around another inversion center of the unit cell and their position is not in spatial conflict with the two Ta ₆ Br ₁₂ (aq) ₆ ²⁺ clusters. Two oxygen atoms are in spatial conflict with the bromine atoms, which indicates alternative assignments. The remaining 10 oxygen positions are occupied by disordered water molecules, which can each be assigned to two sets of 4 water molecules for each of the Ta ₆ Br ₁₂ ²⁺ clusters. The two isolated, half-occupied positions of the bromine atoms and the four half-occupied positions of the oxygen atoms can compensate for the positive charges of the two Ta ₆ Br ₁₂ ²⁺ clusters if the two oxygen atoms represent a pair of hydroxide-water in hydrogen-bond distance. This leads to a formula for the crystalline material of [Ta ₆ Br ₁₂ (H ₂ O) ₆ ] ²⁺ Br⁻OH⁻ × 5 H ₂ O.

Example 3 Use of Ta ₆ Br ₁₂ ²⁺ clusters for X-ray structure analysis

The binding of Ta ₆ Br ₁₂ ²⁺ clusters to an antibody V _L domain (Steipe et al., J. Mol. Biol. 225 (1992), 739-753) and to the dimethyl sulfoxide reductase (DMSOR) from Rhodobacter capsulatus (Schneider et al., J. Mol. Biol. 263 (1996), 53-69). The cluster compound binds to a single position in both proteins. The cluster position could be localized by analyzing the difference Patterson function.

The shortest distance (0.23 nm) between the cluster and the protein for a tantalum atom (TA1) and the carboxyl oxygen of an aspartic acid residue (ASP1) was found in the antibody V _L domain. A second tantalum atom (TA6) interacts with a main chain carbonyl oxygen of a serine residue (SER 62) from an adjacent protein molecule at a distance of 0.31 nm. In DMSOR, two tantalum atoms (TA6 and TA4) are 0.29 nm and 0.31, respectively nm from the side chain carboxylate oxygen atoms of two different glutamic acid residues (GLU 659 and GLU 43). The carbonyl oxygen of a valine residue (VAL 656) could also have a stabilizing effect on the binding of the cluster. Two symmetrically oriented DMSOR molecules are arranged around the cluster, forming a large pocket.

Furthermore, the binding of the Ta ₆ Br ₁₂ ²⁺ cluster to the 72-symmetrical proteasome (Loewe et al., Science 268 (1995), 533-539), which consists of 14α and 14β subunits, was determined, a Ta ₆ Br ₁₂ binds ²⁺ clusters to each β subunit. The cluster is surrounded by three amino acid residues, namely lysine, glutamic acid and arginine. In GTP cyclohydrolase I (Nar et al., Structure 3 (1995), 459-466), a cluster in the center of each of the four pentamers is bound in the asymmetric unit and surrounded by histidine and arginine residues.

Example 4 Optimization of cluster orientation

A quick and easy procedure to determine the correct orientation of the cluster using high-resolution data from DMSOR and the V _L domain was used. For this purpose, the position of the center of gravity of the Ta ₆ Br ₁₂ ²⁺ cluster was first determined using the difference Patterson function using the PROTEIN program (Steige mann (1992), in: Crystallographic computing 5 (HRSG: Moras, D., Podjarny , AD, Thiery, JC), Oxford University Press, Oxford, 115-125). The orientation of the cluster was then determined in an analysis using the "rigid body" refinement routine of the computer program X-PLOR (Brünger et al., Science 35 (1987), 458-460). By rotating the cluster model in steps of 15 ° around each of its three axes, different starting positions were created while the center of gravity was left at its specific position. Then there was a refinement against F- _Deri -F _Nati structure factor amplitudes. Both the R factor and the Patterson correlation value were optimized and a clear clustering of solutions was found. The determined orientations for the cluster agree with the difference Fourier density maps of the final models.

For comparison, phase calculations were carried out in which the cluster was used as a single scattering center or as an ensemble of 18 scattering centers. The results of this comparison are shown in Table 3. The very significant improvements that were obtained using the 18 scattering center model can be seen from this table.

In all test calculations, the clustered solutions with the lowest R factors and the highest Patterson correlation values corresponded to the correct orientation of the cluster in accordance with the difference Fourier electron density map. In all cases, several starting positions converged to identical solutions that differed significantly from the next best solutions. Using the Patterson correlation as the target function, the highest and lowest values for DMSOR in the resolution range from 0.22 nm to 0.8 nm were 15.9 and 49.6% and for the R factor 48.9 and 61, 3 %. In the resolution range from 0.3 nm to 0.8 nm, the values were 16.1 and 35.5% for the Patterson correlation and 59.5 and 67.9% for the R factor. For the antibody V _L domain, the values in the resolution range from 0.2 nm to 0.8 nm for the Patterson correlation were from 14.3 to 43.7% and from 59.5 to 67.9% for the R -Factor. In the resolution range from 0.3 nm to 0.8 nm, the values for the Patterson correlation were from 13.2 to 35.7% and for the R factor from 60.4 to 62.3%. The Patterson correlation values for the correct solution and for the next best incorrect solution were 49.5% and 35.4% for DMSOR and 44.6% and 39.8% for the antibody V _L domain, respectively. The values were obtained after refinement to convergence using data up to the maximum resolution.

Claims (12)

1. Crystalline compound of the formula:
[Ta ₆ Hal ₁₂ (H ₂ O _n )] ²⁺ X ₁ ⁻X ₂ ⁻.m H ₂ O
wherein Hal is a halide ion selected from Cl⁻ and Br⁻,
X ₁ ⁻ and X ₂ ⁻ each represent a single negatively charged anion,
n is a number from 0 to 6 and
m represents a number from 0 to 6. 2. A compound according to claim 1, wherein Hal is Br. 3. A compound according to any one of claims 1 or 2, wherein X ₁ ⁻ and X ₂ ⁻ are selected from Br⁻ and OH⁻. 4. Crystalline compound of the formula:
[Ta ₆ Br ₁₂ (H ₂ O) ₆ ] ²⁺ Br⁻OH⁻.5 H ₂ O. 5. Using crystallographic data for a compound one of claims 1 to 4 in X-ray crystallography. 6. Use according to claim 5 for the structural determination of biological macromolecules. 7. Use according to claim 5 for the structure determination of protein molecules. 8. Using crystallographic data for a compound one of claims 1 to 4, for identifying effectors for biological molecules.   9. A method for determining the structure of a compound by X-ray crystallography, characterized in that

• (a) produces a derivative of the compound with a symmetrical heavy metal cluster,
• (b) determines the correct binding site of a heavy metal cluster in the derivative and
• (c) determines the correct orientation of a heavy metal cluster within the derivative at the binding site.

10. The method according to claim 9, characterized, that step (b) involves analyzing the difference Patterson function. 11. The method according to claim 9 or 10, characterized, that step (c) rotates the heavy metal cluster by at least at least one of the 3 spatial axes in several stages, a calculation of structural factor amplitudes at several of these stages and the Determination of the optimal spatial orientation of the cluster the resulting data. 12. The method according to claim 11, characterized, that the optimal spatial orientation by determining a minimum R-factor and / or a maximum Patterson correla determined value.