2008 - Vol. 8

E d i t o r i a l

New FOX3D optic for protein crystallography

A revolution in X-ray optics Dear Reader, It is a pleasure to welcome Microphotonics as our distributor and partner for the United States. Microphotonics has a team of highly qualified people who are extremely experienced with X-ray analytical equipment, and we are certain that our collaboration will allow us to provide high quality and timely service to our customers in the United States. Microphotonics will distribute our GeniX line of micro-focus beam delivery systems, our FOX 2D line of optics as well as our new FOX 3D line of best-in-class X-ray optics. In this issue you will find articles showing examples of the performance achievable with each of these products. The FOX3D CU 14-39P was used on a Rigaku MM007 at Sanofi Aventis with great results, allowing screening of crystals that would otherwise have been discarded. This result is one example of the power of our new line of FOX 3D optics. Take a look inside to discover how the 3D optics are revolutionising X-ray optics.

M. Mathieu and V. Steier, Sanofi–Aventis, Vitry, France The FOX3D CU 14_39P is the first 3D single reflection optic designed for protein crystallography. Based on the latest progress in Xenocs manufacturing technology this mirror features a high precision ellipsoidal substrate (see Fig. 1) and a state-of-the-art multilayer. Thanks to this novel technology, the FOX3D CU 14_39P boasts a large collection angle and produces a highly symmetric focal spot. These features lead to an increase in flux hitting the crystal and thus delivers a significantly improved signal to noise ratio compared to the FOX2D multilayer mirrors. Dr. Magali Mathieu and Dr Valérie Steier of SanofiAventis in Vitry (France) compared the performance of a FOX3D CU 14_39P to that of a FOX2D mirror using crystals of different sizes. Both mirrors were installed on a rotating anode with a 70 µm source size. Three crystals were used : A) a large, well-diffracting crystal; B) a small, weakly diffracting crystal; C) a small, moderately diffracting crystal. The results obtained for crystal B and C are presented in the tables at page 2. A dataset is considered to be usable to a given resolution if Rmerge is below 0.35 and Mean (I/sd(I)) is greater than 2 for that resolution shell. The data were processed with Mosflm and scaled with Scala (CCP4 suite).

Fig. 1 : FOX3D CU 14_39P schematic concept showing the optic’s aspheric form (ellipsoid of revolution)

In order to better illustrate the differences between the data obtained using the two mirrors, images of crystal C (with the same grayscale setting) taken using each optic are shown in Fig.2. Both FOX2D CU 12_38P and FOX3D CU 14_39P mirrors provide good quality data for large, strongly-diffracting crystals. For medium to small size, moderately or weakly diffracting crystals, the enhanced collection angle and focusing properties of the FOX3D CU 14_39P lead to a significantly improved S/N ratio.

Peter Høghøj

Ta b l e o f c o n t e n t s New FOX3D optic for protein crystallography . . . . . . . . . . . . .1 FOX3D CU 21_21HC X-ray optic more than half a revolution . . . . . . . . . . . .2 FOX2D CU 25_25P mirror compared to Helios on Bruker Microstar . . .3 GeniX successfully tested for SAXS at UCSB, USA . . . . . . . . . . . . . . . . . . . . .4 Forthcoming conferences . . . . . . . . . . .4

with FOX2D CU 12_38P

with FOX3D CU 14_39P

Fig. 2 : Diffraction images of Crystal C acquired using the FOX2D CU 12_38P optic (on the left) and the FOX3D CU 14_39P (on the right)

Low resolution limit High resolution limit Rmerge Rmeas (within I+/I-) Rmeas (all I+ & I-) Rpim (within I+/I-) Rpim (all I+ & I-) Fractional partial bias Tot. # of observations Total number unique Mean((I)/sd(I)) Completeness Multiplicity

Overall 42.41 3.10 0.215 0.251 0.251 0.130 0.130 -0.031 35098 10121 8.3 92.9 3.5

FOX2DCU 12_38P InnerShell OuterShell 42.41 3.27 9.80 3.10 0.044 0.770 0.051 0.899 0.051 0.899 0.027 0.461 0.027 0.461 -0.007 -0.056 1254 4222 348 1322 51.8 1.7 95.5 83.1 3.6 3.2

FOX3DCU 14_39P Overall InnerShell OuterShell 24.49 24.49 3.27 3.10 9.80 3.10 0.147 0.033 0.446 0.174 0.039 0.525 0.174 0.039 0.525 0.092 0.021 0.274 0.092 0.021 0.274 -0.038 -0.014 -0.023 32835 1025 3958 9800 320 1250 13.9 78.4 2.4 90.9 89.2 80.1 3.4 3.2 3.2

Low resolution limit High resolution limit Rmerge Rmeas (within I+/I-) Rmeas (all I+ & I-) Rpim (within I+/I-) Rpim (all I+ & I-) Fractional partial bias Tot. # of observations Total number unique Mean((I)/sd(I)) Completeness Multiplicity

Overall 24.71 2.03 0.275 0.307 0.307 0.133 0.133 -0.046 91716 18742 6.6 96.1 4.9

FOX2DCU 12_38P InnerShell OuterShell 24.71 2.14 6.42 2.03 0.081 1.095 0.089 1.234 0.089 1.234 0.037 0.555 0.037 0.555 -0.007 -0.259 3003 11676 598 2603 26.0 1.3 88.4 92.1 5.0 4.5

FOX3DCU 14_39P Overall InnerShell OuterShell 24.69 24.69 2.14 2.03 6.42 2.03 0.128 0.045 0.455 0.144 0.051 0.514 0.144 0.051 0.514 0.064 0.023 0.233 0.064 0.023 0.233 -0.014 -0.007 -0.144 91220 2994 11383 19263 676 2529 13.5 45.4 3.0 98.3 98.1 89.4 4.7 4.4 4.5

C R Y S TA L B Size : MA345dtb slits : Distance : Dphi/image : Exposure time : Space group : # collected images :

300*60*5 µm3 1*1/0.8/0.8 250 mm 1° 20’ P21 180

Data were processed only to 3.1Å and were usable to about 3.2Å. Conclusion : the data are significantly better on the FOX 3D CU 14_39P. Signal/noise ratio has improved 40 to 67% depending on the resolution shell.

C R Y S TA L C Size : MA345dtb slits : Distance : Dphi/image : Exposure time : Space group : # collected images :

100*60*60 µm3 1*1/0.8/0.8 175 mm 1° 5’ I222 120

The data were usable to 2.15Å. Conclusion: Signal/noise ratio has improved by 75 to 130% with FOX3D CU 14_39P.

FOX3D CU 21_21 HC X-ray optic more than half a revolution One of the direct advantages of Xenocs’ single reflection optical scheme vs. the traditional double bounce geometries (so-called Montel or Side by Side) is that the collection angle is not limited in the sagittal plane by the mirror length, and the FOX3D CU 21_21HC makes maximum use of this fundamental superiority. With the appropriate multilayer coating, the active area is increased to half the ellipsoid of revolution (180°) achieving both an unprecedented solid collection angle of 10-3 sr and high reflectivity of monochromatic copper radiation (K-alpha 1 and 2).

Beam at focus (magnified)

This optic also benefits from the improved focusing properties inherent to the FOX3D technology, resulting in a small focal spot and improved flux density. Initial test results and simulations are presented in Table I.

Fig. 3 : Images of the FOX3D Cu 21-21HC beam at focus (left), at the mirror exit for the full beam (centre) and at the mirror exit for the beam limited by a slit (right).

Table I : Experimental results

As shown in Fig. 3, the native beam divergence can be tuned simply by placing a variable slit in the beam path. The slit width determines the convergence angle in both planes (at the expense of integrated flux) in order to suit your application constraints. In this example, in order to achieve a resolution/divergence of ΦL = 0.05 degrees, a slit with the appropriate opening was inserted in the beam path, limiting the divergence in the orthogonal plane down to 2.6 degrees for a total flux to 2,6x107 ph/s, with the spot size remaining unchanged.

2008 - Vol. 8

Slit-limited beam at mirror exit (magnified)

Spot size (FWHM) ~80 µm with 50 µm source ~50 µm with 30 µm source

High Convergence (HC) solutions are available as either optics alone, to be mounted on sealed or rotating sources, or as a GeniX CU High Convergence complete beam delivery system. Obviously all applications cannot benefit from such a large and asymmetric convergence angle. Typical applications for such a beam profile are small spot reflectometry (Small Spot Rapid XRR for instance) or microdiffraction where high flux density on a small spot is required and moderate to high divergence is accepted or needed in at least one plane.

News letter

Full beam at mirror

Source

Flux in vacuum

Spot Size

Divergence

50W – 50 µm

4x108 ph/s

80x80 µm²

4°x 2°

Measured

100x100 µm²

4°x 2°

Simulated

1.2KW- 70 µm

2

9

8.5x10 ph/s

FOX2D CU 25_25P mirror compared to Helios on Bruker Microstar Courtesy of I.Vetter, MPI für Molekulare Physiologie, Dortmund, Germany Xenocs FOX2D CU 25_25P is a single reflection mirror designed for high precision measurements of small crystals, and is especially advantageous when coupled with a microfocus X-ray source. In order to demonstrate this capability, lysozyme crystals of various sizes were measured by Dr. Ingrid Vetter from MPI Dortmund. Experiments were performed with a BrukerAXS MicroStar rotating anode generator with an expected focus size of 100 µm, operating at 45 kV and 60 mA. The generator was equipped with a HELIOS mirror on one side and a Xenocs FOX2D CU 25_25P mirror on the other. Both mirrors focused onto the sample position. Data were collected on MAR345 image plate detectors. Three different lysozyme crystals were tested on both the HELIOS and FOX2D CU 25_25P sides:

- A, a very small crystal (< 50 micrometers); - B, a medium-sized crystal (110x110x50 µm3); - C, a larger crystal (200x200x100µm3). On both sides the Mar 345 slit settings were 0.6x0.6 (front) and 0.3x0.3 (back), crystal-todetector distance was 175 mm, and ∆Phi was 1°. The exposure time was 5 min for crystal A and 1 min for crystals B and C. Special care was taken to ensure that the crystals were measured in the same orientations on the two sides. The results are shown in Table II. The data labelled with * and # are from the same data collections. For lines 2 and 3 only the first 27 frames were processed to make them comparable with the previously collected 27 frames on the Helios mirror to check the reproducibility after moving the crystal to the Xenocs side and back. Values in parentheses represent the highest resolution shell. Values highlighted in red represent the better data quality when

Table II : Data collection results

1

compared between the two mirror systems. It is clear that the FOX2D CU 25_25P optic results in far better data quality for the smallest crystal (A) than the Helios optic. The FOX2D CU 25_25P data gave a dataset useful to at least 2.8 Å, whereas the Helios data are of a comparable quality (R-factor, I/sigma) only up to 3.2 Å. This can also be clearly seen from the two diffraction images (shown in Fig.4), representing the same crystal orientation. For small crystals and a standard MAR345 setup on a Bruker-AXS MicroStar X-ray generator, the FOX2D CU 25_25P mirror (which was designed for measuring small crystals) yields significantly better data than the HELIOS optics. The results obtained confirm this capability of the mirror. For larger crystals, the HELIOS mirror yields better high-resolution data. It should be noted that Xenocs offers other mirrors (i.e. the FOX2D CU 12_38P) that are designed for use with larger crystals.

: xenocs1 stands for FOX2D CU 25_25P optic

xtal

mirror

frms

resolution

complete.

Rsym

Rmeas

I/σ

wilsonB

30µ 30µ* 30µ#

helios xenocs1 helios

27 27 27

2.8(2.9-2.8) dto. dto.

82.5(82.5) 84.6(87.9) 82.1(82.2)

13.8(41.5) 7.4(27.2) 13.7(40.7)

17.9(54.1) 9.6(35.9) 17.8(53.5)

6.05(2.25) 10.07(3.47) 6.08(2.28)

36.91 38.69 37.67

30µ* 30µ#

xenocs1 helios

43 43

dto. dto.

94.8(95.6) 93.4(94.6)

8.0(28.2) 15.1(47.9)

9.6(33.6) 18.1(56.9)

12.84(4.36) 7.66(2.72)

29.41 37.86

110µ 110µ

helios xenocs1

60 60

2.1(2.2-2.1) dto.

97.7(93.5) 97.1(89.4)

16.6(34.1) 14.0(41.2)

18.8(38.7) 15.8(46.9)

8.85(4.46) 10.80(3.76)

24.44 24.48

200µ 200µ

helios xenocs1

60 60

dto. dto.

97.4(96.3) 96.1(85.0)

6.6(30.8) 6.5(39.8)

7.5(34.9) 7.3(45.4)

18.99(4.96) 18.46(3.89)

33.31 35.48

Fig. 4 : Diffraction images acquired using the Helios optic (left panel) and the FOX2D CU 25_25P optic (right panel) for the same crystal under identical orientation, and with the same grey-scale.

3

News letter 2008 - Vol. 8

GeniX successfully tested for SAXS

Forthcoming Conferences 2008

University of California, Santa Barbara, USA Small Angle X-ray scattering (SAXS) is a powerful X-ray analytical technique for studying systems on the micro or nanoscale. It finds use in many fields, including biology, physics, and materials science. Due to its small 50 µm source size, the GeniX provides a beam with a small natural divergence, making it an efficient source for SAXS systems where the source-to-detector distance is generally large (typically > 2 meters).

Apr 08-10, York, UK BCA Spring Meeting British Crystallographic Association

To characterise the performance of the GeniX for SAXS applications, XENOCS recently provided a demo unit to Pr. Youli Li at the University of California, Santa Barbara. The GeniX was incorporated into an existing 4-meter long SAXS instrument and the instrument was used to collect high quality SAXS data from multiple samples. The results show that the GeniX is an excellent source for SAXS.

May 31-Jun 04, Knoxville, TN, USA ACA 2008 American Crystallographic Association

Fig.5 : XS data collected with GeniX-powered SAXS system from a Silver Behenate (TOP) and a hydrated chicken leg tendon samples (BOTTOM).

In a typical SAXS system, a series of apertures are used to control the beam size and divergence so that the desired low angle scattering data can be spatially separated from the main beam at the detector. A simple way to think about this configuration is to consider the set up as effectively imaging th source through successive apertures onto the detector. A SAXS system with a conventional source (typically 300 µm) generally sacrifices a large fraction of the X-ray photon flux in order to achieve the required angular resolution. A small source coupled to a high efficiency longfocusing optic is highly effective for SAXS because more flux can be squeezed through the tight pinholes. Consequently, although the GeniX operates at low total power, it can provide usable flux levels for SAXS comparable to many conventional high-powered generators. The experiments presented here convincingly demonstrate the advantages of this unique approach. The SAXS instrument was originally constructed using a rotating anode generator with a 0.2 x 0.2 mm² source and a long-focus double multilayer mirror primary monochromator (for details see www.mrl.ucsb.edu/mrl/centralfacilities/xray/index.html). For these experiments, the GeniX was inserted into the optical path near the current source to replace both the rotating anode source and the monochromator, with all other optical components and settings unchanged. The compactness of the GeniX (which includes both the source and monochromator) greatly eased the modification : the entire set up time took less than a day. Table II shows some of key parameters of this GeniX-powered SAXS instrument, which provided a flux level comparable to the same set up with the rotating anode operating at approximately 30x higher power.

Published by : Xenocs SA - Marketing & Sales Dpt. - 19 Rue F. Blumet - 38360 Sassenage (France)

Aug 04-08, Denver, CO, USA DXC 2008 Denver X-ray Conference

Aug 23-31, Osaka, Japan IUCR 2008 International Union of Crystallography

XENOCS SA 19 Rue F. Blumet 38360 Sassenage (France) Phone: +33 (0)4 76 26 95 40 F a x : +33 (0)4 76 26 95 49 [email protected] - www.xenocs.com A spin-off company from Institut Laue Langevin

Table III : parameters of SAXS setup with GeniX Source and monochromator

GeniX

Source-to-Sample Distance

1.8 meter

Sample-to-Detector Distance

1.7 meter

Beam Collimation Detector

The high usable flux level permitted data collection with fairly short exposure times. In Fig. 5 SAXS data collected on Silver Behenate (3 min. exposure time) and chicken leg tendon (60 min. exposure time) are shown. These data demonstrate the superior performance characteristics of the GeniX for SAXS. In conclusion, the high beam delivery efficiency combined with ease of use and maintenance demonstrates that the GeniX beam delivery system is an excellent choice for SAXS applications.

News letter 2008 - Vol. 8

4

Beam size at sample position Qmin Beam Flux at sample position (50kV x 1 mA)

3 sets of motorized slits Bruker HI-STAR 1 mm x 1 mm 1.0x10-2 Å-1 1.4 x 107/s *

* Beam flux was measured with calibrated PIN diode during initial set up. Normal flux could be higher with further optimization.