X-RAY CRYSTALLOGRAPHY OF FISSION YEAST RNG2

X-R

AY

CR

YSTA

LLOG

RA

PHY

OF FISSIO

N Y

EAST R

NG

2

WA

NG

CH

ERN

HO

E 2004

WANG CHERN HOE

INSTITUTE OF MOLECULAR AND CELL BIOLOGY

2004

X-RAY CRYSTALLOGRAPHY OF FISSION YEAST RNG2

WANG CHERN HOE

B.Sc. (Hons.), NUS

A THESIS SUBMITTED

FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

INSTITUTE OF MOLECULAR AND CELL BIOLOGY

AN A*STAR BIOMEDICAL SCIENCES INSTITUTE AFFILIATED TO THE NATIONAL

UNIVERSITY OF SINGAPORE

2004

Acknowledgements

This thesis is dedicated in memory of my mother and late father. Dad,

although you left me long ago, I have not forgotten about you and your hardwork as a

chemical engineer. Mom, I know you have suffered many years in his absence and

from the sudden lost of his financial support. Working from dawn to dusk all this while,

I hope this thesis brings you some pride. 情爱的刘小姣，若没你的支持，我是不会

有今天的！

I thank Stephen Wong Yeow Whye and Wong Siew Chin from the

Bioinformatics Institute (www.bii.a-star.edu.sg) for allowing me to use their Linux

cluster for the 5 nanosecond molecular dynamics simulation.

Lastly, I would like to thank my supervisor Dr Terje Dokland for his patience

and support throughout the entire period of my candidature.

i

Table of contents

Acknowledgements.......................................................................................................... i

Table of contents............................................................................................................. ii

List of figures................................................................................................................. iv

List of tables.................................................................................................................viii

List of abbreviations ...................................................................................................... ix

List of publications ......................................................................................................... x

Abstract .......................................................................................................................... xi

Summary .......................................................................................................................xii

1. Biological background ............................................................................................... 1

1.1. Fission yeast cytokinesis.................................................................................. 1

1.2. Fission yeast Rng2 ........................................................................................... 7

1.3. The calponin-homology domain...................................................................... 8

2. Crystallographic background ................................................................................... 12

2.1. Crystal symmetry........................................................................................... 12

2.2. Reciprocal lattice ........................................................................................... 17

2.3. Bragg's law..................................................................................................... 18

2.4. Scattering by an electron................................................................................ 20

2.5. Data reduction................................................................................................ 22

2.6. Scattering by an atom .................................................................................... 27

2.7. Anomalous scattering by an atom.................................................................. 28

2.8. Location of heavy atoms and the determination of protein phase angles...... 30

2.9. Phase improvement........................................................................................ 33

2.10. Refinement..................................................................................................... 34

ii

3. Materials and methods ............................................................................................. 37

3.1. Plasmid construction...................................................................................... 37

3.2. Protein expression.......................................................................................... 38

3.3. Protein purification ........................................................................................ 39

3.4. Protein crystallization .................................................................................... 40

3.5. Data collection ............................................................................................... 41

3.6. Structure determination.................................................................................. 42

3.7. Molecular dynamics simulation in an aqueous system.................................. 48

3.8. Phylogenetic analysis..................................................................................... 49

3.9. Generation of figures and tables .................................................................... 50

4. Results and discussion.............................................................................................. 51

4.1. Overall structure ............................................................................................ 51

4.2. Comparison of CH domain structures ........................................................... 58

4.3. Intermolecular contacts.................................................................................. 62

4.4. Comparison of Br- and Hg-derivatives of Rng232-189 .................................... 72

4.5. Calponin-homology domain phylogeny ........................................................ 78

5. Concluding remarks: Rng2 structure and function .................................................. 82

6. Appendix .................................................................................................................. 85

7. References ................................................................................................................ 96

iii

List of figures

Figure 1: Domain organization within Rng2. Domain boundaries are defined by performing a global Pfam search (Eddy, 1998) across a Pfam HMM library (Bateman et al., 2002), and shown in red numerals. ....................................................................... 8 Figure 2.1: The right-handed axial system is shown on a triclinic P lattice. ................ 15 Figure 2.2: One P unit cell of a monoclinic structure, with scattering entities at the lattice points. X-ray beam are "reflected" by the (110) planes of the monoclinic structure......................................................................................................................... 20 Figure 3: Thrombin cleavage of GST-Rng21-189. Lane M, molecular weight standard; Lane 1, before cleavage; Lane 2, eluate containing Rng21-189 after 3 hr cleavage at 4 °C with 20 U of thrombin; Lane 3, eluate using glutathione buffer; Lane 4, purified Rng21-

189................................................................................................................................... 40 Figure 4: Crystals of Rng21-189. They typically measure 200 µm in the longest dimension. ..................................................................................................................... 41 Figure 5: Harker section y = ½ corresponding to the Br positions in a Br-anomalous difference Patterson map............................................................................................... 43 Figure 6: Harker section y = ½ corresponding to the Hg positions in a Hg-anomalous difference Patterson map............................................................................................... 44 Figure 7(a): Multiple sequence alignment of calponin-homology domains from proteins known to bind F-actin. The blue square scale at the top shows the relative accessibility of each residue extracted from DSSP (Kabsch and Sander, 1983) - blue with red borders indicate exposed residues with an accessibility value A > 1; blue indicates accessible residues with A > 0.4, cyan indicates intermediate accessibility with 0.1 < A < 0.4, and white indicates buried residues with A < 0.1. The secondary structure of Rng2 and its residue numbering follows below and are shown in black. The predicted secondary structure of Rng2 based on PHD is shown at the bottom in red. The residues corresponding to Rng2 core α-helices (Figure 8c) are highlighted in yellow............................................................................................................................ 53 Figure 7(b): Structural alignment of all known calponin-homology domains. In the sequence name column, the number following the four digit PDB coordinate entry refers to either a CH1 or CH2 domain; it is absent for single CH domains. The last alphabet is a chain identifier. The cyan square scale at the bottom shows the hydropathic character of each residue according to the algorithm of Kyte and Doolittle (1982) with a window size of 3 - pink indicates hydrophobic residues with a value H > 1.5; grey indicates intermediate hydrophobicity with -1.5 < H < 1.5 and cyan indicates hydrophilic residues with H < -1.5. .............................................................................. 54 Figure 7(c): Multiple sequence alignment of calponin-homology domains from the IQGAP family of proteins............................................................................................. 54

iv

Figure 8: Structure of Rng232-189. (a) Ribbon representation of the crystal structure of Rng232-189 shown in two orientations with relative rotation of 180°. Bromine atoms are represented by spheres coloured according to B factors (blue to red denoting lower to higher values). The amino- and carboxyl-terminal residues are labelled together with α-helices a1-a7 and 310-helices h1-h3. For comparison purposes, the region corresponding to actin-binding sites ABS1 (a1) and ABS2 (h2-a6) of dystrophin CH1 domain (Norwood et al., 2000) are highlighted in green and blue, respectively. The 39 residues carboxyl-terminal to the CH domain (150-189) are highlighted in red. Orientations in (a)-(c) are the same. The molecular surfaces below the ribbon diagram are in the same orientation. Surface residue conservation scores ranging from 0 to 1 are coloured based on a white (not conserved) to red (conserved) scale. Kyte-Doolittle (Kyte and Doolittle, 1982) surface hydrophobicities ranging from -4.5 to 4.5 are coloured based on a cyan (hydrophilic) to wheat (hydrophobic) scale. Electrostatic surface potentials ranging from -155.214 kT/e to 318.183 kT/e are coloured based on a red (acidic) to blue (basic) scale, with a scale factor of 1............................................. 55 Figure 8(b): Cα backbone superposition of the CH domains from the Br-derivative of Rng2 CH, Hg-derivative of Rng2 CH, dystrophin CH1, dystrophin CH2, calponin CH, EB1 CH, fimbrin CH1-1, fimbrin CH1-2, utrophin CH1, utrophin CH2, plectin CH1, plectin CH2 and beta-spectrin CH2. For clarity, only the Br-derivative of Rng232-189 CH (pink), dystrophin CH1 (red), calponin CH (orange) and plectin CH2 (blue) domains are shown here................................................................................................ 55 Figure 8(c): Cα backbone superposition of the four core α-helices belonging to the Br-derivative of Rng232-189 CH (pink), dystrophin CH1 (red), calponin CH (orange) and plectin CH2 (blue) domains. Orientations in (a), (b) and (c) are the same................... 56 Figure 9: A 310-helical structure between residues 62-64 (helix h1). Water molecules are represented by cyan spheres and dotted lines indicate hydrogen bonds. The 2mFo - DFc map is contoured at 1σ and superimposed on the coordinates. ............................. 61 Figure 10: Intermolecular contacts in a P21 crystal of the Rng2 CH domain. (a1) Molecular packing projected from the c axis (upper) and b axis (lower) directions for the Br-derivative. The unit cell (P21) and its related symmetry elements are coloured red. The macrobonds, labelled A-E are represented by dashed lines and circles for bonds parallel and perpendicular to the projection plane respectively. The colouring scheme of the ribbon model of Rng2 follows figure 8(a). (a2) Same as (a1) for the Hg-derivative. The MAD and SAD data sets for the Br- and Hg-derivatives respectively were independently solved, accounting for the different origins which are related by half a unit translation along the a axis. Blue arrows point at the loop between residues 92-99 of Rng2. .............................................................................................................. 65 Figure 10(b)-(d): The 2mFo - DFc map is contoured at 1σ and superimposed on the coordinates. Grey and green carbons indicate separate molecules. Amino-acid codes are shown in lower case. The chain number is in upper case next to the residue number. (b1) Chemical environment at the interface of two molecules along macrobond A of the Br-derivative. There is a potential salt bridge and a potential water-mediated hydrogen bond between the ζ-amino group of Arg70 and the γ-carboxyl group of Asp151 The ζ-amino group of Arg70 is linked via a potential salt bridge to the δ-carbonyl group of Gln67. .............................................................................................. 66

v

Figure 10(b2): Macrobond A of the Br-derivative. A bromide anion with a temperature factor of 24.54 Å2 forms a potential hydrogen bond to the ζ-amino group of Arg175, which is doubly hydrogen bonded to the carbonyl group of Leu152. .......................... 67 Figure 10(b3): Macrobond A of the Br-derivative. There is a nonpolar cluster (>5 Å) of residues Ile172, Leu152, Phe82, Phe110 and Ile114 in a background of specific intermolecular hydrogen bonds and salt bridges. ......................................................... 68 Figure 10(c): Macrobond C of the Br-derivative. There is a potential direct hydrogen bond between the ζ-amino group of Arg99 and the δ-carbonyl group of Glu161. It also has another potential direct hydrogen bond between the α-imino of Tyr92 and γ-carbonyl group of Asn162. ........................................................................................... 69 Figure 10(d): Macrobond D of the Br-derivative. It has a potential water-mediated hydrogen bond from the γ-carbonyl group of Asn162 to the ε-amino group of Lys131........................................................................................................................................ 70 Figure 10(e): Macrobond E of the Br-derivative. It has a potential bromide-mediated hydrogen bond between the α-imino group of Pro117 and the ζ-amino group of Arg178. ......................................................................................................................... 71 Figure 11: Comparison between Br- and Hg-derivatives of Rng232-189. (a) Ribbon representation of the Br-derivative (red helices and blue coils) superimposed on the Hg-derivative (pink helices and cyan coils). Hg atoms are shown as spheres coloured according to their B factors. .......................................................................................... 74 Figure 11(b): Environment around the Hg atom near the variable loop from residues Tyr92-Arg99 in the Hg-derivative. The 2mFo - DFc map is shown contoured at 1σ. The Hg anomalous map (red) is superimposed onto the electron density map. .................. 75 Figure 11(c): Mass weighted Cα r.m.s.d. of Rng232-189 as a function of time over the entire trajectory of a 5 ns molecular dynamics simulation for Rng232-189, in the Br-derivative (red) and Hg-derivative (black). .................................................................. 76 Figure 11(d): Main chain (N, Cα, C and O) r.m.s.d. for each residue as a function of time during the same MD. α-Helical sections of the structure are highlighted in green........................................................................................................................................ 77 Figure 12: Consensus of phylogenies of 117 individual calponin-homology (CH) sequences from 73 proteins. The species represented include Arabidopsis thaliana (at), Dictyostelium discoideum (dd), Gallus gallus (gg), Homo sapiens (hs), Meleagris gallopavo (mg), Mus musculus (mm), Rattus norvegicus (rn), Saccharomyces cerevisiae (sc) and Schizosaccharomyces pombe (sp). The number appended to the species symbol represents the order of the CH domain with 1, 3 being amino-terminal and 2, 4 being carboxyl-terminal. The CH groups were designated following Korenbaum and Rivero (2002). Rng2 (Rng2_sp) is highlighted in dark grey at the top of the figure. The protein codes for the 9 species, together with their accession numbers, are as follows: Rng2_sp (NP_593860); Iqg1_sc (NP_015082); IQGAP1_hs (NP_003861); IQGAP2 (NP_006624); CNN_sp (NP_593863); CNN1_hs (NP_001290); 1h67_gg (P26932); CNN_mg (P37803); CNN3_hs (NP_001830); CNN2_hs (NP_004359); ARHGEF7b_hs (NP_663788); ARHGEF6_hs (NP_004831); LRRN4_hs (NP_002310); CHDC1_hs (XP_166260); LMO7_hs

vi

(NP_005349); VAV1_hs (NP_005419); VAV3_hs (NP_006104); VAV2_hs (NP_003362); TAGLN_hs (NP_003177); NP25_hs (NP_037391); TAGLN2_hs (NP_003555); Scp1_sc (NP_015012); GAS2L1_hs (NP_006469); GAR17_hs (NP_644814); GAS2_hs (NP_005247); ASPM_hs (NP_060606); SMTNb_hs (NP_599031); SMTNa_hs (NP_599032); SMTNc_hs (NP_008863); SMTN_mm (NP_038898); MICAL2_hs (NP_055447); MICAL3_hs (XP_032997); TANGN_hs (XP_170658); MIRAB13_hs (NP_203744); NICAL_hs (NP_073602); NUMA1_hs (NP_006176); 1pa7_hs (NP_036457); MAPRE3_hs (NP_036458); MAPRE2_hs (NP_055083); Sac6_sc1 (NP_010414); Fim1_sp1 (NP_596289); PLS1_hs1 (NP_002661); 1aoa1_hs1 (NP_005023); LCP1_hs1 (NP_002289); AtFim1_at1 (NP_200351); ACTN3_hs1 (NP_001095); ACTN4_hs1 (NP_004915); ACTN1_hs1 (NP_001093); ACTN2_hs1 (NP_001094); FLNB_hs1 (NP_001448); FLNA_hs1 (NP_001447); FLNC_hs1 (NP_001449); FLN_gg1 (BAB63943); ctxA_dd1 (AAB62275); abpC_dd1 (P13466); SYNE1_hs1 (NP_149062); CLMN_hs1 (NP_079010); SYNE2_hs1 (NP_055995); 1dxx1_hs1 (NP_003997); 1qag1_hs1 (NP_009055); MACF1a_hs1 (NP_036222); PLEC_rn1 (NP_071796); 1mb81_hs1 (NP_000436); SPTBN1_hs1 (NP_003119); SPTB_hs1 (NP_000338); SPTBN2_hs1 (NP_008877); SPTBN4_hs1 (NP_079489); SPTBN5_hs1 (NP_057726); NAV2_hs (NP_660093); NAV3_hs (NP_055718); PARVG_hs1 (NP_071424); PARVA_hs1 (NP_060692); PARVB_hs1 (NP_037459)..................................................................................................................... 80 Figure 13: Residues shown in stick representation on the ribbon model of Rng232-189 are conserved within the IQGAP family of proteins (Figure 7c). The transparent surface electrostatic potential highlight the presence of a basic patch where Rng232-189 possibly interacts with F-actin. ..................................................................................... 83

vii

List of tables

Table 1: S. pombe genes involved in various aspects of cytokinesis.............................. 3 Table 2: The seven crystal systems, their corresponding Bravais lattices and symmetries encountered in biological crystals. ............................................................ 14 Table 3: Rng2 plasmids. Restriction sites of the primers are shown in brown, whereas insert sequences are in blue........................................................................................... 36 Table 4: Data collection, processing, phasing and refinement statistics. Numbers in parentheses refer to the highest resolution shell. .......................................................... 46 Table 5: R.m.s.d. values of the pairwise comparison of all the CH domains using the Cα belonging to the four core helices of each structure. The core helices of Rng232-189 correspond to residues 43-54, 73-81, 99-112 and 133-144 (Figure 8c). Abbreviations: DMD, dystrophin; UTRN, utrophin; PLEC, plectin; CNN, calponin; SPTB, beta-spectrin; PLS3, fimbrin ABD1; EB1, end-binding protein 1........................................ 59 Table 6: Biochemical characteristics of several CH domains. Structures deposited in RCSB are in square brackets......................................................................................... 85

viii

List of abbreviations

ABD actin-binding domain ABP-120 actin-binding protein 120 ABS actin-binding site ACTN1 actinin, alpha 1 ARHGEF6 alpha rho guanine nucleotide exchange factor 6 ASPM asp (abnormal spindle)-like, microcephaly associated ATP adenosine 5'-triphosphate BPAG1 bullous pemphigoid antigen 1 CH calponin-homology CHASM calponin homology-associated smooth muscle protein CLMN calponin like transmembrane domain protein CNN calponin Da Dalton, atomic mass unit, 1.67377 × 10-27 kg DMD dystrophin DNase deoxyribonuclease EB1 end-protein 1 EHBP1 EH domain binding protein 1 ERK extracellular regulated kinase Esu estimated standard uncertainties FLNA filamin A GAP GTPase-activating protein GAR17 GAS2-related protein GAS2 growth arrest-specific 2 GDP guanosine 5'-diphosphate GEF guanine nucleotide exchange factor GST glutathione-S-transferase GTP guanosine 5'-triphosphate hGAR22 human Gas2-related gene on chromosome 22 IQGAP IQ motif containing GTPase-activating protein K Kelvin LCP1 lymphocyte cytosolic protein-1 LMO7 LIM domain only 7 LRCH4 leucine-rich repeats and calponin-homology (CH) domain containing 4 MACF1a microtubule-actin crosslinking factor, isoform a MAPRE1 microtubule-associated protein, RP/EB family, member 1 MICAL2 molecule interacting with CasL MIRAB13 molecule interacting with Rab13 NAV2 neuron navigator 2 NICAL NEDD9 interacting protein with calponin-homology and LIM domains NP25 neuronal protein NUMA1 nuclear/mitotic apparatus protein 1 PAGE polyacrylamide gel electrophoresis PARVA parvin, alpha PBC periodic bond chain PBS phosphate buffered saline pI isoelectric point PIP2 phosphatidyl inositol-3,4-bisphosphate PLEC1 plectin 1 PLS1 plastin 1 PMSF phenylmethylsufonyl fluoride r.m.s.d. root-mean-square deviation SDS sodium dodecyl sulphate SM22 smooth muscle protein 22-alpha SMTNa smoothelin isoform a SPB spindle pole body SPTB spectrin beta erythrocytic SPTBN1 spectrin beta non-erythrocytic 1

ix

SYNE1 synaptic nuclei expressed gene 1 TAGLN transgelin TAGN tangerin UTRN utrophin

List of publications

Wang, C.H., Walsh, M., Balasubramanian, M.K. and Dokland, T. (2003). Expression, purification, crystallization and preliminary crystallographic analysis of the calponin-homology domain of Rng2. Acta Crystallogr. D59, 1809-1812. Wang, C.H., Balasubramanian, M.K. and Dokland, T. (2004). Structure, crystal packing and molecular dynamics of the calponin-homology domain of Schizosaccharomyces pombe Rng2. Acta Crystallogr. D60, 1396-1403.

x

Name: Wang Chern Hoe

Degree: Doctor of Philosophy

Department: Institute of Molecular and Cell Biology

Thesis Title: X-ray crystallography of yeast cytoskeletal proteins

Abstract

Schizosaccharomyces pombe Rng2 is an IQGAP protein essential for the assembly of

an actomyosin ring during cytokinesis. Rng2 contains an amino-terminal calponin-

homology (CH) domain, eleven IQ repeats and a RasGAP homology domain. CH

domains are known mainly for their ability to bind F-actin, although they have other

ligands in vivo and there are only few examples of actin-binding single CH domains.

The structures of several CH domains have already been reported, but this is the third

report from actin-binding proteins that contain a single CH domain (the structures of

calponin and EB1 have been reported previously). The 2.21 Å resolution crystal

structure of the amino-terminal 189 residues of Rng2 determined using Br- and Hg-

derivatives includes 39 residues (150-189) carboxyl-terminal to the CH domain that

resembles neither the extended conformation seen in utrophin nor the compact

conformation seen in fimbrin, although residues 154-160 form an unstructured coil

which adopts a substructure similar to dystrophin residues 240-246 in the carboxyl-

terminal portion of the CH2 domain. It wraps around the stretch of residues that would

be equivalent to the proposed actin-binding site ABS1 and ABS2 from dystrophin.

Another feature revealed by comparing the two derivatives is the presence of two loop

conformations between Tyr92-Arg99.

Keywords: calponin-homology, CH, Rng2, yeast cell-division proteins.

xi

Summary

Cytoskeletal proteins form the framework for all biological processes occuring

in a cell. In particular, it is involved in separating a mother cell into two daughter cells

via the actomyosin ring. This process is known as cytokinesis, and is the most directly

visible biological process in a cell, responsible for the the ability of living organisms to

reproduce. The aim of this study was to solve by means of X-ray crystallography, the

structure of cell division proteins in the fission yeast Schizosaccharomyces pombe,

which happens to be an ideal model to analyse cell division; for the purpose of making

further insights into the molecular process of cytokinesis.

Rng2 is an IQGAP protein that is essential for the assembly of an actomyosin

ring during cytokinesis. It contains a single calponin-homology (CH) domain at the

amino-terminus, followed by eleven IQ repeats and a RasGAP homology domain. CH

domains are known mainly for their ability to bind F-actin, although they have other

ligands in vivo and there are only few examples of actin-binding single CH domains.

In this study, the structure of the amino-terminal 189 residues of Rng2 (Rng232-189) to

2.21 Å resolution was solved using the anomalous dispersion technique - two

wavelength Br-MAD and Hg-SAD. Its crystal structure revealed a CH domain that is

similar to the CH domains deposited in RCSB, although a molecular replacement

solution was impossible partly because unlike other CH domains, the 39 residues (150-

189) at its carboxyl-terminus bears no sequence similarity to known proteins. The 31

residues at its amino-terminus are also disordered and invisible in the electron density

map. When the amino-terminal 189 residues were expressed in S. pombe as a GFP-

Rng21-189 fusion protein (Wachtler et al., 2003), it localized to the same site as cortical

actin patches, suggesting that residues 1-189 of Rng2 bind F-actin. Two features new

xii

to CH domains were revealed. First of all, it has a carboxyl-terminal tail that wraps

around the stretch of residues that would be equivalent to the putative actin-binding

site ABS1 and ABS2 of dystrophin. The structure of this carboxyl-terminal tail of

Rng21-189, comprising residues 150-189, resembles neither the extended conformation

seen in utrophin nor the compact conformation seen in fimbrin. However, Rng2

residues 154-160 form an unstructured coil which adopts a substructure similar to

dystrophin residues 240-246 in the carboxyl-terminal portion of the CH2 domain.

The other new feature is the presence of two possible loop conformations

between Tyr92 and Arg99. A molecular dynamics (MD) simulation indicated that

Asn94 of the Br-derivative had a higher r.m.s.d. relative to the Hg-derivative,

corresponding to higher temperature factors in the crystal structure of the Br-derivative

compared to the Hg-derivative. This shows that the MD simulations were accurate and

lend weight to the view that the loop switches between two possible conformations.

These distinctive features were not highlighted in previously published CH domains

and may have useful implications in determining the F-actin binding orientation of

single CH domains.

xiii

Biological background

Fission yeast cytokinesis

Fission yeast cells maintain their specific sizes by coordinating their growth

and division via a cell-size checkpoint (Fantes and Nurse, 1978), which is exerted

primarily at the G2/M transition (Rupes, 2002). Wee1 delays entry into mitosis by

inhibiting the activity of Cdc2/Cdk1, the cyclin-dependent kinase that promotes entry

into mitosis (Russell and Nurse, 1987). The phosphatase Cdc25 promotes entry into

mitosis by removing the inhibitory phosphorylation of a highly conserved tyrosine

residue at the amino-terminus of Cdk1 (Dunphy and Kumagai, 1991). When Cdk1

activity reaches a critical level, it triggers early mitotic events such as chromosome

condensation, nuclear envelope breakdown and spindle assembly.

In the absence of defects at the end of mitosis, the three pairs of properly

segregated chromosomes are compartmentalized into two daughter cells during

cytokinesis, the final event of the cell cycle. It is perhaps the first major cell cycle

event that was observed directly. Schizosaccharomyces pombe cells are cylindrical

with a uniform diameter of about 3.5 µm and grow by polarized membrane and cell

wall addition at both ends to a length ranging from 7 to 16 µm. Fission yeast cells

divide by medial fission, with the haploid state being more stable. After cell division,

both daughter cells will initially grow from the end that existed in the mother cell prior

to division. Typical of all eukaryotic cells, the cell cycle has discrete G1, S, G2 and M

phases. Early in G2 phase, the new end created by the previous cell division will

initiate growth, and the cells will grow in a symmetrical way throughout the remaining

interphase. The M phase is the most dramatic, where replicated DNA become precisely

1

separated into two daughter nuclei during mitosis, and cytokinesis then divides the cell

equally into two identical daughter cells (Pringle et al., 1997).

Morphological transitions are well defined during the fission yeast cell cycle,

making it a simple and attractive model for the analysis of cytokinesis. It has the

smallest fully sequenced eukaryotic genome compartmentalized into three observable

chromosomes that clearly condense during mitosis, allowing individual chromosome

behaviour to be followed by cytological methods. S. pombe carries out mitosis and

cytokinesis much like animal and protozoan cells. The mitotic spindle is present for a

short time during mitosis and the spindle pole body (SPB) remains undivided until

mitosis, typical of higher eukaryotes (McCully and Robinow, 1971). Some differences

with animal cells are present, such as the actin ring emerging at an earlier stage of M

phase in fission yeast, and the synthesis of a division septum behind the constricting

ring. These processes are dependent on the proper organization of the actin

cytoskeleton. There are three types of filamentous actin structures in fission yeast:

cortical patches, cables and ring (Marks and Hyams, 1985). During interphase, F-actin

patches localize to both growing ends of the cell, whereas F-actin cables run

longitudinally and are linked to F-actin patches (Arai and Mabuchi, 2002). During

mitosis, the patches disappear and F-actin reappears in an actomyosin ring in the

medial region of the cell attached to F-actin cables (Arai and Mabuchi, 2002).

It is crucial that cytokinesis occurs only after mitosis to ensure proper

segregation of genetic material to each daughter cell. Since Nurse's pioneering work on

the cell cycle (Nurse et al., 1976), genetic screens have identified mutations in more

than 50 genes affecting distinct steps in cytokinesis (Rajagopalan et al., 2003; Guertin

et al., 2002). Phenotypic characterization of S. pombe cytokinesis defective mutants

2

has placed them in three categories: ring positioning, ring assembly and septation

initiation network (SIN; Table 1).

Table 1: S. pombe genes involved in various aspects of cytokinesis Ring positioning Septation initiation network

(SIN) Cell separation

mid1 dmf1

PH domain byr4 No known domain imp2 FCH and SH3 domains/PCH family protein

plo1 serine/threonine protein kinase

cdc7 serine/threonine protein kinase

bgs1 cps1

(1-3) beta-D-glucan synthase

pom1 serine/threonine protein kinase

cdc11 No known domain mok1 (1-3) alpha-D-glucan synthase

Ring assembly cdc14 clp1 flp1

Phosphatase Mitosis to interphase transition

cdc3 Profilin cdc16 Rab GTPase activating protein

sbp1 Ran GTPase activating protein

cdc4 Essential light chain mob1 No known domain mog1 Ran GTPase cdc8 Tropomyosin plo1 serine/threonine

protein kinase spi1 Ran GTPase

cdc12 FH domain spg1 Ras GTPase pim1 GEF cdc15 FCH and SH3

domains/PCH family protein

sid1 serine/threonine protein kinase

rna1 Ran GTPase activating protein

rng2 IQGAP sid2 serine/threonine protein kinase

rng3 UCS domain sid4 No known domain act1 Actin myo2 Myosin II heavy chain rlc1 Regulatory light chain myp2 Myosin II heavy chain fim1 Fimbrin ain1 Actinin crn1 Coronin

The Dyrk kinase Pom1 localizes to cell ends with the help of microtubules and

its activity is required for both symmetrical cell growth and symmetrical division

(Bahler and Nurse, 2001). The anillin-like protein Mid1 resides in the nucleus in

interphase cells.

Upon entry into mitosis, a broad equatorial band of proteins that will later form

the contractile ring is set up. Plo1, the single member of the polo family of kinases,

phosphorylates Mid1 during prophase and triggers its relocation to the medial region

of the cortex as a broad ring (Bahler et al., 1998). At the G2/M transition, this broad

3

band of Mid1 is joined by conventional myosin II (heavy chain Myo2, essential light

chain Cdc4 and regulatory light chain Rlc1) and IQGAP Rng2, followed by PCH

Cdc15 and formin Cdc12. Equatorial accumulation of these proteins depends on Mid1

but not actin. Myosin II localizes to a spot-like structure in interphase cells (Kitayama

et al., 1997) distinct from the formin Cdc12 spot (Chang, 1999). The myosin II-

containing spot seems to originate from the constricted ring of the previous mitotic

cycle. Since Myo2 is insoluble at physiological salt concentrations, the myosin II spots

may consist of myosin filaments (Bezanilla and Pollard, 2000). Rng2 and Myo51

physically interact with the essential light chain Cdc4 of myosin II (D’souza et al.,

2001). Thus, it is likely that Rng2 binds myosin II through its interactions with Cdc4.

During early anaphase, actin filaments and tropomyosin Cdc8 join the broad

band which then coalesces into a compact ring that includes alpha-actinin Ain1.

Formin Cdc12 and profilin Cdc3 cooperate to assemble the contractile ring (Kovar et

al., 2003). Profilins have been shown to bind monomeric actin, an actin-related protein

complex, phospholipid PIP2, and proline-rich peptides such as poly-L-proline and

VASP, and may regulate actin dynamics in the formation of F-actin structures (Chang

et al., 1997). Tropomyosin and alpha-actinin should bind to the actin ring as it

polymerizes. The interaction of myosin II with the actin ring provides a possible

mechanism for the contractile force. Alpha-actinin forms dynamic cross-links between

actin filaments, providing greater stiffness than actin alone at higher rates of

deformation (Xu et al., 1998).

During late anaphase, the compact contractile ring matures with the addition of

unconventional type II myosin Myp2 and septins. Myp2 is necessary under conditions

of stress and increases the efficiency of cytokinesis (Bezanilla et al., 2000). Septins are

GTPases which are not essential for fission yeast viability (Berlin et al., 2003).

4

In all eukaryotic cells examined, exit from mitosis requires inactivation of the

cyclin-dependent kinase Cdc2/Cdk1. Cdc2 is essential for both mitotic and meiotic cell

cycle progression in fission yeast. Cdc2 inactivation occurs coincidently with

chromosome segregation, ensuring that cell division does not initiate before

chromosomes have been separated (Trautmann et al., 2001). Loss of Cdc2 activity is

mediated through destruction of mitotic cyclins by ubiquitin-mediated proteolysis and

Cdc2 inhibitors. Ubiquitination of cyclin is driven by an E3 ubiquitin ligase known as

the anaphase-promoting complex/cyclosome (APC/C). Following cyclin destruction in

late mitosis, several events occur in rapid succession, such as the breaking down of the

mitotic spindle, the constriction and disassembly of the actomyosin ring, and the

formation of a septum. This pathway is carefully regulated by the septation initiation

network (SIN) so that cytokinesis and mitotic exit are prevented until the proper

segregation of chromosomes has occurred. At least eleven genes are involved in the

septation initiation network that is responsible for triggering the onset of septum

formation and cytokinesis (Table 1).

The localization of SIN components to the SPB is required for cytokinesis and

is dependent on Sid4-Cdc11, both constitutive residents of SPBs (Tomlin et al., 2002).

Cdc11 provides the physical link between Sid4 and the Ras GTPase Spg1, which

controls the onset of septation (Tomlin et al., 2002). In interphase cells, Spg1 is in the

GDP-bound form, but upon entry into mitosis it converts to the GTP-bound form

(Sohrmann et al., 1998). It localizes asymmetrically to the SPB in interphase and to

both spindle poles during mitosis (Sohrmann et al., 1998). A guanine nucleotide

exchange factor (GEF) for Spg1 has not been identified in S. pombe so far and Spg1

might prove to be negatively regulated solely by the recruitment of the GTPase-

activating protein (GAP) Cdc16-Byr4 to the SPB during interphase, where it acts to

5

restrain septum formation (Cerutti and Simanis, 1999). Cdc7 encodes a

serine/threonine protein kinase that shows no discrete localization during interphase

but is recruited to the SPB by GTP-Spg1 during mitosis (Sohrmann et al., 1998).

When the spindle separates during late anaphase, the protein kinase Sid1 functions in a

complex with phosphatase Cdc14/Clp1/Flp1 and is recruited to the SPB containing

elevated Spg1-Cdc7 (Guertin et al., 2000). Cdc2 inactivation, triggered by cyclin

proteolysis, regulates Sid1-Cdc14 localization to the SPB (Guertin et al., 2000). Mob1

localizes to the SPB and the medial ring during mitosis. It recruits the protein kinase

Sid2 to form the Sid2-Mob1 complex, which functions downstream of Sid1-Cdc14

(Hou et al., 2000). One candidate binding partner for Sid2-Mob1 is Cdc15, which is

required for Sid2 localization to the medial region (Sparks et al., 1999).

Imp2 colocalizes with the medial ring and destabilizes it after the initiation of

septum formation (Demeter and Sazer, 1998). It provides a link between SIN and the

division septum. Bgs1/Cps1 is an integral membrane protein that localizes to the cell

division site during late anaphase. Assembly of Cps1 into the actomyosin ring depends

on F-actin and SIN proteins (Liu et al., 2002). In addition, Cps1 localizes to areas of

polarized cell wall growth and might be involved in synthesizing the lineal (1-3) beta-

D-glucan of the primary septum during mating, sporulation and vegetative growth

(Cortes et al., 2002). Mok1 is a (1-3) alpha-D-glucan synthase localized on the cell

membrane and moves from the growing tips during interphase to the medial ring upon

mitosis (Katayama et al., 1999). Sterol-rich membrane domains are localized to

distinct regions of the plasma membrane in a cell cycle and secretory pathway

dependent manner. They may be involved in anchoring the actomyosin ring

components to the plasma membrane (Wachtler et al., 2003).

6

Ran GTPases Mog1 (Tatebayashi et al., 2001) and Spi1 (Matynia et al., 1996)

localize to the nucleus and are essential for the mitosis-to-interphase transition. Their

distribution between the GTP- and GDP-bound forms are regulated by GAP Rna1 and

GEF Pim1 (Matynia et al., 1996). In this class of mutants, the cells are typically

arrested in the cell cycle as septated, binucleated cells with highly condensed

chromatin, fragmented nuclear envelopes, and abnormally wide septa.

Fission yeast Rng2

Rng2 is a orthologue of mammalian IQGAPs (Eng et al., 1998; Wang et al.,

2003). It is associated with the spindle pole body during interphase and mitosis. It is

also detected in the medial actomyosin ring during mitosis and cytokinesis (Eng et al.,

1998) where its localization requires F-actin. Fission yeast has three simple F-actin

structures: cortical patches, cables and rings (Marks and Hyams, 1985). In contrast, its

Saccharomyces cerevisiae orthologue Iqg1, does not depend upon F-actin for its

localization to the ring (Epp and Chant, 1997). Rng2 contains several protein domains

and motifs. It has a calponin-homology (CH) domain in the amino-terminus, eleven IQ

repeats, two coiled-coil domains, a rasGAP domain and a rasGAP carboxyl-terminal

domain (Figure 1). IQGAP1, the human orthologue of Rng2 is rather well studied.

Residues 216-683 of IQGAP1 contain coiled-coil regions and have been shown to be

responsible for the oligomerization of IQGAP1 (Fukata et al., 1997). IQGAP1 is able

to bind Cdc42 (Ho et al., 1999) and inhibit its GTPase activity via residues 918-1657

(McCallum et al., 1996), which contains the RasGAP domain. IQGAP1 binds F-actin

and calmodulin via residues 1-216, which contains a CH domain (Fukata et al., 1997)

(Table 6). It binds calmodulin via residues 740-869 (Ho et al., 1999; Table 6), which

7

contains several IQ domains. Likewise, Cam1 and Cdc4 have been shown to bind

Rng2 (Eng et al., 1998; D'souza et al., 2001).

Figure 1: Domain organization within Rng2. Domain boundaries are defined by performing a global Pfam search (Eddy, 1998) across a Pfam HMM library (Bateman et al., 2002), and shown in red numerals.

The calponin-homology domain

The calponin-homology (CH) domain consists of about 100 residues first

identified at the amino-terminus of calponin, a protein implicated in the regulation of

smooth muscle contraction through its interaction with F-actin and inhibition of the

actin-activated MgATPase activity of phosphorylated myosin (Castresana and Saraste,

1995; Winder and Walsh, 1990; Tang et al., 1996). Three groups of CH domain

containing proteins are recognizable from sequence analysis. Human proteins

containing a single CH domain include IQGAP1, IQGAP2 and their fission yeast

orthologue Rng2; calponins CNN1, CNN2, CNN3; oncogenes VAV1, VAV2, VAV3;

ARHGEF6, ARHGEF7b; GAS2, GAS2L2, GAR17; LMO7; MAPRE1/EB1,

MAPRE2, MAPRE3; ASPM; NUMA1; NP25; SM22/TAGLN, TAGLN2; NAV2,

NAV3; LRCH1/CHDC1; LRCH2; LRCH3; LRCH4/LRRN4; SMTNa, SMTNb,

SMTNc; MICAL-L1/MIRAB13, MICAL-L2, MICAL2, MICAL3; NICAL; CHASM

and EHBP1. Human proteins containing tandem CH domains include alpha-actinins

ACTN1, ACTN2, ACTN3, ACTN4; beta-spectrins SPTB, SPTBN1, SPTBN2,

SPTBN4, SPTBN5; dystrophin DMD; utrophin UTRN; filamins FLNA, FLNB, FLNC;

plectin PLEC1; dystonin/BPAG1, MACF1a; SYNE1, SYNE2; calmin CLMN; parvins

8

PARVA, PARVB and PARVG. In addition, Dictyostelium discoideum encodes ABP-

120 (Bresnick et al., 1991) and cortexillin ctxA (Stock et al., 1999), which are

orthologous to human alpha-actinins, beta-spectrins, dystrophins and utrophins. Finally,

human proteins of the fimbrin/plastin family PLS1, LCP1, PLS3 contain tandem actin-

binding domains (ABDs).

There is abundant evidence that CH domains are able to bind F-actin in

isolation, in the form of tandem CH repeats known as actin-binding domains (ABDs).

For instance, recombinant alpha-actinin ABD (Kuhlman et al., 1992), dystrophin ABD

(Way et al., 1992, Fabbrizio et al., 1993) and utrophin ABD (Winder et al., 1995) have

been shown to bind F-actin, but not G-actin; and ABDs from filamin (Lebart et al.,

1994), fimbrin/plastin (Hanein et al., 1997,1998), MACF (Leung et al., 1999), plectin

(García-Alvarez et al., 2003) and beta-spectrin (Karinch et al., 1990) have been shown

to bind F-actin in isolation. ABDs from alpha-actinin (Leinweber et al., 1999), beta-

spectrin (Karinch et al., 1990; Raae et al., 2003), dystrophin (Renley et al., 1998),

utrophin (Winder et al., 1995), plectin (Garcia-Alvarez et al., 2003) and SYNE2 (Zhen

et al., 2002) have average dissociation constants (Kd) for ABD binding to F-actin

ranging from 2.5 to 22.3 µM (Table 6). Fimbrins form a distinct class since these

monomeric actin-bundling proteins contain two tandem ABDs. Plectin ABD is unique

in its ability to bind both F-actin and beta(4) integrin (Geerts et al., 1999; Table 6)

providing a mechanism to switch its localization from actin filaments to

hemidesmosomes when beta(4) integrin is expressed. Similarly, actopaxin/alpha-

parvin consists of a pair of CH domains that is able to bind both F-actin and the

LD1/LD4 motifs of paxillin (Nikolopoulos et al., 2000; Table 6). Since the LD1 motif

of paxillin is able to bind both actopaxin and the serine/threonine integrin-linked

kinase (ILK), this points to the role of actopaxin in integrin-dependent remodelling of

9

the actin cytoskeleton during cell adhesion. Aside from biochemical data, cryo-EM

reconstructions of F-actin decorated with fimbrin ABD (Hanein et al., 1997,1998),

utrophin ABD (Moores et al., 2000; Galkin et al., 2002) and calponin (Bramham et al.,

2002) have yielded atomic models of their respective complexes, although that of F-

actin decorated with alpha-actinin ABD (McGough et al., 1994) did not yield an

atomic model since the structure of the CH domain was unknown. From a larger

perspective, the CH domain plays an important role in developmental genetics, since

gain-of-function mutations in the ABD and rod domain repeats of filamin A are

implicated in many congenital malformations (Robertson et al., 2003).

Unlike the ABD, the single CH domain binds a wide variety of ligands, and

some single CH domain proteins are again implicated in single-gene genetic disorders.

In the case of ARHGEF6, both the amino-terminal CH and carboxyl-terminal coiled-

coil domains are necessary for binding β-parvin (Rosenberger et al., 2003; Table 6);

and a partial deletion of the single CH domain of ARHGEF6 is implicated in X-linked

mental retardation (Kutsche et al., 2000). The nuclear-mitotic apparatus protein,

NUMA, which contains a single CH domain and is localized in centrosomes, has been

correlated with abnormal mitotic spindle assembly after nuclear transfer in nonhuman

primates (Simerly et al., 2003). The CH domain acts in a regulatory fashion in the

oncogene VAV1 where its basal activity is inhibited in cis through interactions

between the CH and zinc finger region (Zugaza et al., 2002; Table 6). There is

opposing evidence regarding the F-actin binding ability of isolated CH domains. For

instance, the CH domains of calponin and SM22 are unable to bind F-actin (Gimona

and Winder, 1998; Gimona and Mital, 1998; Gimona et al., 2002; Mino et al., 1998;

Stradal et al., 1998; Table 6). Instead the CH domain of calponin was found to bind

caltropin with a slightly higher affinity than that of calmodulin (Wills et al., 1994).

10

This interaction was proposed to modulate the inhibitory activity of calponin towards

the actin activation of myosin ATPase. Calponin may also play a role in extracellular

regulated kinase (ERK) mediated signal transduction since its CH domain has been

shown to interact with ERK (Leinweber et al., 1999; Table 6).

In contrast to its regulatory role in ARHGEF6, VAV1, calponin and possibly

SM22, the single CH domain has also been shown to bind to the cytoskeleton. For

example, both the CH domain of GAR22 (Goriounov et al., 2003) and residues 1-216

of IQGAP1 which contains a CH domain (Fukata et al., 1997; Ho et al., 1999), are

able to bind F-actin in cosedimentation assays (Table 6). It is interesting to note that

residues 1-232 of IQGAP1, which contains a CH domain, is able to bind both F-actin

and calmodulin (Ho et al., 1999), whereas the CH domain of calponin is able to bind

calmodulin, caltropin and ERK, but not F-actin (Wills et al., 1994; Table 6). Residues

1-133 of EB1 contain a CH domain that has been shown to bind microtubules (Hayashi

and Ikura, 2003; Table 6). Similarly, residues 7-182 of smooth muscle calponin

encodes a CH domain and was shown to interact with desmin (Wang and Gusev, 1996;

Table 6). All these facts highlight the versatile nature of the CH fold that allows it to

bind different ligands depending on the specific residues evolved to suit the function of

the protein in question.

X-ray structures are known for several CH domains, including CH2 of beta-

spectrin (Carugo et al., 1997); the ABDs from dystrophin (Norwood et al., 2000),

utrophin (Keep, Winder et al., 1999; Keep, Norwood et al., 1999) and plectin (García-

Alvarez et al., 2003); the first ABD (ABD1) from fimbrin (Goldsmith et al., 1997);

and the single CH domain from the EB1 (Hayashi and Ikura, 2003). Well diffracting

crystals of alpha-actinin ABD (Ekström et al., 2003) have also been reported. The

NMR structure of a CH domain from calponin shows a similar fold in spite of their

11

sequence divergence (Bramham et al., 2002). However, it does not bind F-actin.

Instead, its three carboxyl-terminal calponin family repeats are essential for binding F-

actin (Gimona and Winder, 1998; Gimona and Mital, 1998; Mino et al., 1998; Table 6).

In a similar fashion, the carboxyl-terminus of SM22 which contains a calponin family

repeat is required for full F-actin affinity (Fu et al., 2000; Table 6).

Crystallographic background

Crystal symmetry

Crystals are distinguished by the presence of a regular repetition or symmetry

in the three dimensional space of an object, although in reality they often have defects

at a non zero temperature and may have impurities trapped in the lattice. Crystal

symmetry can be described in terms of symmetry operations which leave the object

unchanged. Symmetry operations encountered in biological crystals include rotations

and screw rotations. There is a corresponding symmetry element such as a point, axis

or plane with respect to which each symmetry operation is performed, with the

restriction that point group symmetry operations are compatible with infinite

translational repeats in a lattice; and cannot induce a higher symmetry than the space

group of the crystal. The permutation of all symmetry elements to three-dimensional

spatial patterns results in 230 space groups. These are in turn distributed among 14

Bravais lattices, 11 Laue groups, 32 point groups and 7 crystal systems (Table 2). The

first step in the derivation of the 230 space groups was made by Leonhard Sohncke. He

combined the 14 Bravais lattices with the 32 point groups, which involved n-fold

rotation axes, inversion centres and mirror planes; and added translational motion: n-

fold screw axes, glide lines and glide planes; where n = 1, 2, 3, 4, 6. This resulted in

12

the 65 space groups which he published in 1879. The second and final step accounted

for additional inversion centres which led to a further 165 space groups. They were

first worked out in 1885 by Evgraph Stepanovich Fedorov and independently by

Arthur Moritz Schönflies in 1891 and William Barlow in 1894. The 230 space groups

are arbitrarily numbered from 1 to 230, beginning with triclinic crystals and ending

with cubic crystals. They are systematically drawn and described in the "International

Tables for Crystallography Volume A". There are two space group symbols, one due

to Schönflies and the generally adopted one due to Carl Heinrich Hermann and Charles

Mauguin. The commonly used short Hermann-Mauguin symbol consist of a letter

indicating the centring type of the Bravais lattice (P, C, A, B, I, F, R, H); followed by

one to three characters designating the symmetry elements (m, a, b, c, e, n, d, g, 1, 2, 3,

4, 6, 1,

Bravais lattices are regular stacks of an infinite array of points in which every

point has the same environment as any other point. There are fourteen possible ways to

do this and the Bravais lattices are distributed unequally among the seven crystal

systems. Any point in the lattice may be chosen as an origin.

Proteins are built from enantiomers of optically active L-amino acids, whereas

nucleic acids are built from D-ribose or D-deoxyribose. As a result, their crystal forms

can be either right- or left-handed; and belong to one of the eleven enantiomorphous,

non-centrosymmetric classes of point group symmetries that contain neither mirror

planes nor inversion centres (Table 2).

3 , 4 , 6 , 21, 31, 32, 41, 42, 43, 61, 62, 63, 64, 65) for the corresponding primary

(monoclinic), secondary (rhombohedral) and tertiary (orthorhombic, tetragonal,

hexagonal and cubic) directions.

13

Crystal systems Bravais lattices Constraints imposed on unit cell geometry

Characteristic symmetry

Laue groups Enantiomorphous non-centrosymmetric point groups

Non-enantiomorphous non-centrosymmetric point groups

Non-enantiomorphous centrosymmetric point groups

Cubic cP cI cF

a = b = c α = β = γ = 90°

Four 3-fold rotation axis at 109.47°

3m mm3

23 432

m34

3m mm3

Hexagonal hP a = b α = β = 90°, γ = 120°

One 6-fold rotation axis

m/6 mmm/6

6P

622 6 6mmP

26m

m/6 mmm/6

hP a = b α = β = 90°, γ = 120° (hexagonal axes)

3P

32 3mP

3 m3

Trigonal

hR a b c = = α = β = γ (rhombohedral axes)

One 3-fold rotation axis

3 m3

Tetragonal tP tI

a b = α = β = γ = 90°

One 4-fold rotation axis

m/4 mmm/4

4P

422 4 4mmP

m24

m/4 mmm/4

Orthorhombic oP oI oC oF

α = β = γ = 90° Three 2-fold rotation axis at 90°

mmm 222 mm2P mmm

Monoclinic mP mC

α = γ = 90° (unique b-axis)

One 2-fold rotation axis

m/2 2P mP m/2

Triclinic aP none None 1 1P 1 P

Table 2: The seven crystal systems, their corresponding Bravais lattices and symmetries encountered in biological crystals.

There are ten polar, non-centrosymmetric point groups which have a unique axis not relat d by symmetry. e

14

Bravais lattices of various shapes and sizes are measured in terms of their unit

cell lengths a, b, c and the angles α, β, γ between them; where α refers to the angle

between lattice vectors b and c, β the angle between a and c, and γ the angle between a

and b. The right-handed axial system is conventionally used; here the unit cell lengths

and angles are defined with respect to a common origin at the back left-hand corner of

the cell (Figure 2.1). The unit cell is chosen based on the convention that it should

have the smallest volume for which its delineating vectors are parallel or coincide with

important symmetry directions in the lattice. The monoclinic system is characterized

by a two-fold rotation or inversion, with the b axis parallel to it.

Figure 2.1: The right-handed axial system is shown on a triclinic P lattice.

The eleven Laue groups are determined from Laue images captured using the

Laue X-ray method. The Laue group assigned to a crystal describes the symmetry of

the complete X-ray diffraction pattern obtainable from the crystal. This is because a

single X-ray image is two-dimensional in nature and cannot record the complete three-

dimensional diffraction symmetry.

The Neumann principle is of basic importance to understand how the symmetry

of a crystal is related to its point group. It states that the symmetry elements of any

physical property must include the symmetry elements of the crystal point group. In

15

agreement with Neumann’s principle, physical experiments such as X-ray diffraction

may not reveal the true symmetry of the crystal. For instance, a center of symmetry is

introduced into all X-ray diffraction patterns regardless of whether the crystal is

centrosymmetric or not.

Symmetry elements may occur singly or in certain combinations in a body. A

point group is defined as a set of symmetry elements all of which pass through a single

fixed point known as the origin. The symmetry operations of a point group must leave

at least one point unmoved. The point group symmetry is independent of whether the

unit cell is primitive (P), body centred (I), C-face centred (C), all-face centred (F) or

rhombohedrally centred (R). The result of repeating a point group pattern by the

translations of a Bravais lattice is a space group. A space group consists of a infinite

set of symmetry elements, the operation with respect to any of which brings them into

a state indistinguishable from that before the operation. It can also be described in

terms of two parts, a pattern motif and a repeat mechanism. The pattern motif consists

of the asymmetric unit. It does not have to occupy the entire unit cell although it may

be placed anywhere within the unit cell. It is impossible to deduce the point group and

space group solely from the observed symmetry, although it is possible to infer the

crystal system from either the point group symmetry or unit cell dimensions.

The trigonal system can either have a rhombohedral or hexagonal Bravais

lattice. Some crystals adopt the hexagonal lattice. The rest adopt a rhombohedral

lattice and are distinguished by systematically absent X-ray reflections. In practice, the

hexagonal lattice is used because it is easier to visualize. All crystal systems with the

exception of cubic forms are anisotropic in terms of their physical properties.

If a Bravais lattice has a higher symmetry than triclinic, each particle in the

unit cell will be repeated several times as a consequence of space group symmetry.

16

Each space group has only one general Wyckoff position. It describes a set of

symmetrically equivalent independent particles, each occupying a point that is left

invariant only by the identity operation. The number of molecules in a unit cell

depends on the number of such independent particles in each asymmetric unit

multiplied by the number of such asymmetric units present in the unit cell. An

asymmetric unit of a space group is the smallest space from which an entire lattice can

be filled exactly by application of all symmetry operations of the space group. In

addition, each space group may have several special Wyckoff positions. It describes a

set of symmetrically equivalent independent particles, each occupying a point that is

mapped onto itself by at least one further symmetry operation of the space group. If a

molecule occupies a special position, which is common in small molecule crystals; it

implies the presence of a symmetry axis running through it, and that unit cell will have

less molecules than anticipated from the number of asymmetric units multiplied by the

number of general positions per asymmetric unit.

Reciprocal lattice

For each direct (Bravais) lattice, there is a corresponding reciprocal lattice. It

has the same symmetry as the direct lattice from which it was deduced. d*(100), d*(010)

and d*(001) are referred to as a*, b* and c*, respectively and define a unit cell in the

reciprocal lattice. In general,

)()(*

hkldKhkld =

where d*(hkl) is the distance of the reciprocal lattice from the origin, and K is a

constant which is usually the wavelength λ of the X-ray used. The reciprocal lattice

units are thus dimensionless.

17

The reciprocal lattice points form a true lattice with a representative unit cell

outlined by three sides a*, b* and c*, and three angles α*, β* and γ*. The size of the

reciprocal lattice is governed by the choice of the X-ray wavelength λ used. A

reciprocal lattice row h,k,l; 2h,2k,2l;… are derived from the families of planes (nh, nk,

nl) with n = 1, 2, … in the Bravais lattice.

Bragg's law

When X-rays encounter any form of matter, they are partly transmitted and

partly absorbed. Matter absorbs X-rays in two distinct ways, by scattering and by true

absorption. The scattering of X-rays by atoms is similar to the scattering of visible

light by dust particles in the air, and occurs in all directions. In some directions, the

scattered beams satisfy Bragg's law and are completely in phase, thus reinforcing each

other (constructive interference) to form a diffracted beam with an increased amplitude.

In most directions, they do not satisfy Bragg's law and annul one another (destructive

interference). With the exception of very light elements, scattering is responsible for

only a small fraction of the total absorption of X-rays with wavelengths in the usual

range used for diffraction. Thus the diffracted beam is rather strong compared to the

sum of all the rays scattered in the same direction, but is extremely weak compared to

the incident beam since the atoms in a crystal scatter only a small fraction of the

incident energy. True absorption is caused by electronic transitions within the atom

and results in fluorescent radiation.

William Lawrence Bragg's treatment of X-ray diffraction is an

oversimplification of the complete process. He described it in 1912 as a consequence

of the partial reflection of X-rays by various lattice planes of the same family, like

sheets of atomic mirrors. Although the term 'reflection' is commonly used, it differs

18

fundamentally from diffraction. Reflection of visible light takes place in a thin surface

layer, occurs at all angles of incidence with no regard for Bragg's law, and is highly

efficient. However, the intensity of a diffracted beam scattered by all atoms of the

crystal that lie in the path of the incident beam is very weak, and it occurs only at

angles of incidence that satisfies Bragg's law. Diffraction is essentially reinforced

coherent scattering. A diffraction vector or structure factor F(hkl) has the Miller

indices h, k, l; an amplitude |F(hkl)| and a phase angle α(hkl). The structure factor for a

protein is called FP and that for a heavy atom derivative is called FPH.

Bragg's approach was the result of early experiments showing that if a crystal

was turned from one diffracting position to another through an angle θ, then the

diffracted ray was rotated through an angle of 2θ. The Bragg reflection of three parallel

rays is illustrated in figure 2.2, which shows a c-axis projection of a monoclinic unit

cell. His treatment of diffraction leave implicit the requirement for the physical

superposition of interacting X-rays (i.e. waves 1 and 3 in figure 2.2). The difference in

'path' between the waves scattered at D and B is equal to AB + BC = 2d sin θ. If it is

multiple of λ then the two waves combine themselves with maximum positive

interference, resulting in Bragg's law

λθ nd =sin2

There is a special relationship between the interplanar spacing d(hkl), Bragg's angle θ,

X-ray wavelength λ, and order of diffraction n relative to the same family of lattice

planes, with all planes in the (hkl) family cooperating in the scattering process. X-rays

penetrate deeply into the crystal and are reflected by a series of lattice planes which

satisfy Bragg's law. All X-rays reflected from the same plane remain in phase after

reflection, since no path difference is introduced.

19

Figure 2.2: One P unit cell of a monoclinic structure, with scattering entities at the lattice points. X-ray beam are "reflected" by the (110) planes of the monoclinic structure.

Since sin θ cannot exceed unity,

1sin2

pθλ=

dn

Therefore nλ must be less than 2d. Since the smallest value of n is 1,

d2pλ

For most sets of crystal planes, d is of the order of 3 Ǻ or less, meaning that

electromagnetic radiation exceeding 6 Ǻ will not be diffracted.

Scattering by an electron

A tightly bound electron set into oscillation by the oscillating electric field of

X-rays is continuously accelerating and decelerating during its motion and emits an

electromagnetic wave (photon). Thus, the oscillating electron scatters X-rays; or in

other words, the scattered X-ray beam is radiated from the electron under the influence

20

of the incident X-ray beam. Unlike the electron, the nucleus has an extremely large

mass and does not oscillate significantly; as a result, X-ray scattering is proportional to

the electron density of the atoms in the crystal. Electrons can scatter X-rays coherently

and incoherently. Coherent or Raleigh scattering is the only kind capable of being

diffracted. It comes from incident X-ray photons that lack sufficient energy to eject the

scattering electrons from the atom. Since no energy is transferred, the scattered X-rays

have the same wavelength and frequency as the incident beam; although there is a

phase change of π/2. The intensity of unmodified radiation decreases with increasing

Bragg angle, i.e. higher resolution and shorter wavelengths.

Incoherent or Compton modified radiation was discovered by Arthur Holly

Compton in 1923, and occurs whenever X-rays encounter loosely bound or free

electrons. The incident X-ray beam is an electromagnetic wave of photons, each with

the energy quantum hν. When an incident photon strikes a loosely bound electron in an

elastic collision, the electron gets knocked aside and the photon deflects through an

angle 2θ. Some of the energy of the incident photon is transferred to the electron as

kinetic energy. Compton modified radiation has a slightly increased wavelength and a

random phase relative to the incident beam. In contrast to unmodified radiation, its

intensity increases with increasing Bragg angle and with shorter wavelengths. It does

not take part in diffraction due to its random phase and is responsible for increasing the

background in diffraction images. Since it is due to collisions of quanta with loosely

bound electrons, its intensity relative to coherent radiation increases as the proportion

of loosely bound electrons increases. For this reason, diffraction images of organic and

protein crystals, which consist of elements with low atomic numbers, have increased

background.

21

Data reduction

The analysis and reduction of single crystal diffraction data consists of several

steps. Initially, several diffraction spots or Bragg intensities from a single frame in the

series of oscillation images collected are chosen. It is important that the oscillation

range is small enough such that the diffraction pattern will consist of spots limited

between two ellipses that form a lune that do not overlap. If the lunes overlap, the

spots will not have a unique Miller index. An oscillation range of 1.0 and 0.3° is

usually chosen for protein (small unit cell) and virus (large unit cell) crystals

respectively, since this provides enough spots to establish the periodicity of the

diffraction pattern. All reflections within the same lune originate from the same family

of parallel planes in the reciprocal lattice layer. The lunes are most obvious with large

unit cell crystals such as those from viruses, which have their reciprocal lattice planes

perpendicular to the incident X-ray beam or parallel to the detector. Within each lune,

diffraction spots are arranged along lines that reflect the regularity of the reciprocal

lattice. As a consequence of mapping the curved Ewald sphere onto the flat detector

surface, the straight lines of reflections become hyperbolas at high diffraction angles,

i.e. high resolution, but can be approximated by a plane at low angles.

The diffraction spots chosen from an oscillation image are then mapped onto

reciprocal space, based on the the center of the oscillation range used. The real-space

indexing procedure as implemented in Denzo (Otwinowski and Minor, 1997) uses Fast

Fourier Transform (FFT) to completely search all possible indices of all reflections

that are consistent with the chosen spots. To decrease computational time, the range of

indices is restricted to those found in protein crystals. Lower resolution reflections are

used first, before extending to higher resolution after fitting more parameters. When

integer values of an index h is found for all the chosen spots, one real space vector of

22

the crystal axis is located. The remaining two indices, k and l are then located, and

three linearly independent vectors defining a minimal unit cell volume that would

index all of the observed spots is determined. The three vectors forming a primitive

unit cell is then transformed into one of the fourteen possible Bravais lattices (Table 2).

The lattice with the lowest distortion index is chosen, together with its refined unit cell,

crystal and detector orientation parameters. The crystal orientation parameters describe

the orientation of the reciprocal lattice with respect to the spindle, beam and vertical

axes of the detector. The detector parameters include the X-ray wavelength, crystal to

detector distance, precise coordinates of the direct beam, detector missetting angles

and internal scanner alignment parameters, in the case of image plates.

The process of indexing the diffraction spots provides an estimation of the

mosaicity of the crystal. It is usually of the order of more than 0.5° for protein crystals,

with higher values suggesting lower quality. Real crystals are not ideal because the

regular repetition of unit cells is interrupted by lattice defects. Its diffraction pattern is

the cumulation originating from tiny blocks with slightly different orientations. In

addition, the X-ray is monochromated to a narrow wavelength window with a

bandpass δλ/λ. This increases divergence to the total rocking curve, which is the sum

of beam divergence and crystal mosaicity. In practical terms, mosaicity is defined as

the oscillation angle of a crystal where all the observed reflections are matched with

the predicted reflections.

Bragg intensities are calculated by subtracting an estimate of the detector

background from the reflection profile. Its precise integration requires its Miller

indices and an accurate prediction of its position on the oscillation image. The position

of weak reflections are predicted from those of strong reflections. An analytical

derivation for the integrated intensity I(hkl) from the structure factor F(hkl)

23

20

2

2

2

2

3

)()( hklFELPTIVmce

VhklI rcr ×⎟⎟

⎠

⎞⎜⎜⎝

⎛=ωλ

has the terms wavelength λ, angular speed of rotation of the reciprocal lattice ω, unit

cell volume V, electron charge e, electron mass m, speed of light c, crystal volume Vcr,

intensity of the incident beam I0, Lorentz factor L, polarization factor P, transmission

factor Tr and the extinction coefficient E. The Lorentz factor depends on the geometry

of the diffraction system. It arises from the different length of time required for each

reciprocal lattice node to cross the Ewald sphere. For instance, peaks nearer to the

rotation axis spend more time passing through the Ewald sphere than those further

away. The polarization factor originates from the fact that electrons do not scatter

along its direction of oscillation, but in other directions. When unpolarized incident X-

ray beams from laboratory sources hits a plane, the scattered beams will become

polarized; its intensity is reduced at high angles of scattering. Incident beams with

electric vectors parallel to the diffraction place have little attenuation, whereas those

with electric vectors perpendicular to the diffraction plane will be significantly

weakened. The transmission factor is related to the absorption factor. The use of longer

X-ray wavelengths to diffract crystals containing more heavy atoms will result in

higher absorption of X-rays and correspondingly reduce the scattered X-ray intensity.

The extinction coefficient depends on the inherent mosaicity in crystals and has two

components. Secondary extinction plays a major role in relatively large, perfect

crystals. In these cases, the lattice planes first encountered by the incident beam will

reflect a significant fraction, and as a consequence, the deeper planes will receive less

primary radiation. It can be recognized in the final stages of crystal structure

refinement where |Fo| < |Fc| for some high intensity reflections.

24

For mosaic crystals, primary extinction plays a more significant role. For an

incident beam at a Bragg angle, there are multiple internal reflections due to coherent

or Raleigh scattering within each mosaic block of the crystal. Each reflection shifts the

phase angle by π/2. After two reflections, the beam travels in the same direction as the

incident beam but with reduced intensity and a phase difference of π.

The analytical derivation assumes that the integrated intensities measured do

not change with time. Unfortunately, exposure to X-rays triggers crystal

decomposition, particularly for protein crystals at room temperatures, where the

amount of radiation damage is significant. An efficient way of minimizing radiation

damage is by flash cooling the crystals to 100 K. This process also decreases

background scattering and reduces atomic B factors. However, it broadens the crystal

mosaicity from below 0.02° to above 0.2°, and also degrades diffraction resolution to

varying degrees. Flash cooling creates far more disorder in protein crystals due to their

high solvent content (20-90%), compared to small molecule or inorganic crystals.

After all the Bragg reflections are integrated, a single isotropic scale and B

factor for each oscillation image in a data set is calculated. One or more frames can be

chosen as the reference, and its scale and B factor will not be refined. Merging of

redundant or equivalent reflections, such as those related by space group symmetry,

occurs throughout this refinement process. It provides an opportunity to identify and

reject outliers, which are reflections with intensities that are significantly different

from their equivalents. It should not account for more than 1% of the entire data set

and comes from errors in the classification of partially and fully recorded reflections.

These are usually spots that lie close to the blind region, which refer to the reciprocal

lattice lying on either side of the spindle/Φ axis that do not cross the Ewald sphere

even after a 360° rotation. Alternatively, the reflections may be wrongly measured

25

because the CCD pixels are either shadowed, inactive or affected by 'zingers', i.e.

sparks from trace radioactivity of the fiber optic taper which transports visible light

transformed from X-rays by the phosphor.

The last step is the scaling and merging of different data sets, such as the three

data sets from a single crystal in a MAD experiment; and the global or post-refinement

of crystal parameters using the entire data set, which is more precise than the

processing of a single oscillation image.

The quality of X-ray data is usually assessed by the global Rsym factor, based

on intensities or rather F2

Rsym = ∑

∑ −

hklobs

hklobs

hklI

hklIhklI

)(

)()(

It is a ratio of the spread of intensities of multiple measurements of symmetry

equivalent reflections to an estimate of the reflection intensity. Rsym is highly

influenced by data multiplicity and is always higher for data sets with a high symmetry

space group (Diederichs and Karplus, 1997). This is a disadvantage since its value can

be manipulated. Rsym was first introduced as a reliability indicator for data collected by

precession photography, where it was summed over symmetry-related intensities on

the same film, with Rsca reporting the agreement over different films. Over time, Rsym

and Rsca were combined into the modern day Rsym which is summed over all observed

equivalent reflections. Rmerge, although used interchangeably with Rsym, is obtained

after merging and scaling symmetry equivalent reflections from several crystals and

data sets.

Complementary information about the data quality is given by the ratio of

intensities to their uncertainties,

26

∑∑

)( hkl

hkl

II

σ

Correct estimation of intensity standard uncertainties is important in subsequent

applications based on statistical treatments, such as maximum-likelihood or direct

methods of phasing and refinement.

Scattering by an atom

When an X-ray beam encounters an atom, each electron scatters part of the

radiation coherently according to the Thomson equation. However the nucleus has an

extremely large mass relative to the electron and does not oscillate. X-rays scattered by

an atom of atomic number Z containing Z electrons has an amplitude of Z times the

wave from a single electron when the scattering is in the forward direction (2θ = 0). In

other directions of scattering, the electrons are situated at different points in space and

this results in phase differences. Since the electron cloud has a radius comparable to

the X-ray wavelength, the contribution falls off at higher diffraction angles, i.e. higher

resolution. This is represented by the atomic scattering factor ),( λθf , which describes

the efficiency of scattering X-rays in a given direction by an atom compared to a single

electron. An alternative definition describes it as the ratio of the amplitude of a wave

scattered by an atom to the amplitude scattered by a single electron. The signal is

isotropic and is treated as a real number,

)(),( θλθ °= ff

Along the incident beam where sin θ = 0, ),( λθf has its maximum value because

there is no interference, and is equal to Z. Interference increases with increasing sin θ

and decreasing λ since the path differences within the atom become larger. The normal

27

scattering factor, )(θ°f applies when the scattered radiation has a wavelength much

shorter than the absorption edge of the scattering atom. However when these two

wavelengths are nearly the same, resonance occurs and the X-rays will have sufficient

energy to excite electrons from lower to higher energy shells, resulting in an

anomalous signal which is expressed as a complex number,

)(")(' λλ iff +

Anomalous scattering by an atom

Under normal circumstances when all atomic scattering factors are real with

zero )(" λf contribution, the X-ray diffraction pattern from a crystal is

centrosymmetric and obey Friedel's law, such that )()( hklFhklF = . The X-ray

diffraction pattern, which is the equivalent of a 3D weighted reciprocal lattice, will

exhibit the symmetry of the crystal's point group, with an additional center of

symmetry if not already present i.e. Laue group. However Friedel's law breaks down

when heavy atoms are present in a protein crystal. The tightly bound electrons close to

the atomic nucleus scatter X-rays anomalously, causing reflections of the Bijvoet pair

and )(hklF )(hklF to have unequal intensities, and with phases that are no longer

complementary. This is because electrons are not completely free but are bound to the

nucleus by forces which depend on the exchange of spin-1 gauge bosons (photons)

between the electron and the atomic nucleus; and on the quantum state of the electron,

a spin-½ lepton. Usually )(" λf is proportional to the atomic absorption of the X-rays

and to their fluorescence and )(' λf follows the derivative of this function, according

to the Kramer-Kronig transformation (James, 1958). It is usually expressed as a

complex quantity,

28

)(")(')(),( λλθλθ iffff ++°=

In contrast to the normal atomic scattering factor )(θ°f , the real (dispersive)

correction )(' λf and imaginary (anomalous) correction )(" λf depend only on the

wavelength of the X-rays used and do not diminish with increasing diffraction angle

(or resolution). The soft X-rays from in-house radiation sources such as CrKα (λ =

2.291 Å) and CuKα (λ = 1.542 Å) generate rather small dispersion effects for most

organic compounds and are normally insufficient for obtaining their phases. However,

anomalous dispersion induces substantial variation of the scattered intensities

depending on the chosen wavelength and is an important tool for solving crystal

structures. Thus anomalous scattering from S atoms collected from an in-house CrKα

radiation source can be exploited to phase the X-ray diffraction data from thaumatin

and trypsin crystals (Yang et al., 2003). The contribution to the imaginary component

of anomalous scattering )(" λf for sulfur of 1.14 e- is twice that from CuKα radiation

( )(" λf = 0.56 e-), resulting in reduced data redundancy required for the sulfur single

wavelength anomalous dispersion (SAD) phasing procedure compared with a highly

redundant data set required if CuKα radiation was used. Most of the atoms in

macromolecules have negligible )(' λf and )(" λf . The few anomalous scatterers that

remain result in small intensity differences which must be accurately measured for

SAD, also known as single wavelength anomalous scattering (SAS); and multiple

wavelength anomalous dispersion (MAD) phasing. In the MAD technique, changes to

the heavy atom structure factor FH are induced by tuning the X-ray wavelength to

different points in its absorption spectrum, thus changing )(' λf and )(" λf . As with

the multiple isomorphous replacement (MIR) and single isomorphous replacement

with anomalous scattering (SIRAS) phasing methods, three data sets are usually

required. In the case of MAD, the three data sets are measured at different wavelengths

29

- edge or inflection wavelength λ1 where )(' λf has its minimum, white line or peak

wavelength λ2 where )(" λf has its maximum and where the Bijvoet differences

−+ − PHPH FF are the largest, and a low or high energy remote wavelength λ3 where

)(' λf and )(" λf are small.

Location of heavy atoms and the determination of protein

phase angles

The purpose of crystallographic analysis is to obtain the atomic positions

starting from the diffraction data. However, from the experimental data, only

integrated intensities I(hkl) can be calculated after applying the correction factors L, P,

Tr and E. Since I(hkl) = |F(hkl)|2, the amplitudes |F(hkl)| can be found, but not the

phase angles α(hkl). There are four principle techniques to solve the phase problem -

isomorphous replacement, anomalous dispersion, molecular replacement and direct

methods. It is pointless to implement techniques that will allow us to recover the phase

of individual reflections since this quantity changes continually during the experiment.

Thus, the values of some phases, such as those from anomalous scatterers need to be

solved before the origin of the crystal can be established as a reference for evaluating

the remaining phases. This leads to the fact that the recovered phases of the protein

structure are actually relative phases, since they are calculated relative to the phase of

other scattered amplitudes.

The anomalous dispersion technique will be the focus here since it was used

to solve the structure. The main problem with this technique is noise because it is

based on small differences between observed structure factors, thus necessitating the

collection of highly redundant data (Dauter and Adamiak, 2001). MAD is quickly

30

becoming the most popular method to solve protein crystal structures. In the beginning,

the positions of anomalous substructures were calculated using Patterson and

reciprocal space direct methods based on the tangent formula. The drawback was that

as the number of atom increases, the phase probability distributions become more

easily invalidated by the few aberrant reflections, putting an upper limit of about 20

anomalous scatterers that can be solved because conventional direct methods tend to

become weaker as the number of atom increases. A significant advance was made with

Shake and Bake (Weeks et al., 1993), an algorithm for solving crystal structures by ab

initio, dual space direct methods. It increased the size of the substructure that can be

solved to as many as 160 Se sites, by incorporating information from both strong and

weak reflections. For substructure applications, the resolution of the FA data is not

critical (FA refers to the anomalous contribution to the structure factor F, where F = FP

+ FA). This is because the distances between heavy atoms are usually greater than the

resolution of the data. In fact, data used to a higher resolution without significant

dispersive and anomalous information will only add noise. However, it is essential that

the data is collected with a high redundancy to optimize the anomalous signal to noise

ratio and to eliminate the outliers properly. An effective way to decide the resolution

where data is truncated is to calculate the correlation coefficient (CC) between the

signed anomalous differences −+ − PHPH FF at different wavelengths as a function of

resolution, and truncate the data where the CC falls below 25% (Schneider and

Sheldrick, 2002). A 3.5 Å anomalous wavelength data set is usually sufficient,

although this will result in a poor quality electron density map if the crystal only

diffracted to this resolution (Brodersen et al., 2000). This procedure was first

implemented in SnB (Miller et al., 1994) and then in SHELXD (Sheldrick, 1998). They

31

use the strongest normalized structure factors E, which are usually the most significant

15-20% of the F in each resolution shell. o

After the heavy atom substructure has been derived using Patterson or direct

methods, the heavy atom positions can be refined by allowing |FPH(obs)| to approach

|FPH(calc)| using the least squares or maximum likelihood method. The interatomic

vectors of these sites should be checked for correspondence with peaks in the

difference Patterson map, made using differences between the acentric reflections,

such as −+ − PHPH FF for anomalous difference Patterson (or Bijvoet Patterson) maps.

The Harker peaks on the Harker section should be more than 5σ.

With the refined heavy atom positions, the protein phase angles can be

determined. However, it is important to note that it is not formally possible to

determine protein phases exactly from two measurements, such as when the data is

restricted to a single wavelength in the case of SAD or with only one native and heavy

atom derivative in the case of SIR. There is a twofold ambiguity in the estimation of

the protein phase, which presents itself as either the correct protein structure or its

enantiomorph. One method of distinguishing between them requires a high resolution

electron density map to check for the presence of L-amino acids and right handed α-

helices. An alternative method uses the fact that since anomalous scatterers are part of

the structure, there is a higher probability that the protein phases will be similar to

those from the anomalous substructure. Modern crystallographic software such as

SHARP (de La Fortelle and Bricogne, 1997), SOLVE (Terwilliger and Berendzen,

1999) and MLPHARE (Otwinowski, 1991) make use of this approach to provide

realistic estimates of the initial phases and figures of merit. An advantage of using

SAD is that there is no problem of non-isomorphism between different crystals or data

sets. Since radiation damage is a problem in modern synchrotrons during a MAD

32

experiment, it is beneficial to first try a SAD on the peak wavelength, and then add a

second wavelength if the increased non-isomorphism does not negate the benefits of

using multiple wavelengths.

Phase improvement

The first electron density map inspected by crystallographers is usually the

unmodified Fo map obtained directly from phasing the positions of the heavy atoms.

However, for purposes of interpreting and building the structure, several steps of phase

improvement are required. A powerful approach to improve phase distributions

depends on density modification procedures such as those implemented in SOLOMON

(Abrahams and Leslie, 1996), DM (Cowtan, 1999) and RESOLVE (Terwilliger, 2000).

These methods repeatedly modify the initial density and use it to generate a new set of

phases which are combined with the experimental ones. In spite of this, they still

require a complete and accurately measured anomalous data set, especially for

improving the phases of SAD data sets.

There are several density modification tools. The most common calculation

involves solvent flattening with various solvent contents to determine its optimal value.

It can be applied to resolutions down to 5 Å. The principle is based on the localization

of static solvent molecules as a monolayer or double layer adjacent to the protein

surface. Solvent molecules positioned further away are dynamic and their electron

density can be reset to a low constant value. The solvent flattened phases can be used

to detect minor heavy atom sites with low occupancies through difference Fourier

techniques.

The next method requires an estimate of Z, the number of molecules per unit

cell, from the formula proposed by Matthews (1968)

33

MWZVVm ×

=

For most protein crystals, Vm, which is the ratio of unit cell volume V to the molecular

weight MW in Daltons, is about 2.3 Å3 Da-1. The number of molecules in each

asymmetric unit can then be derived from Z and the crystal space group. It is quite

common for protein molecules in each asymmetric unit to be related by non-

crystallographic symmetry (NCS). Once the NCS averaging operators are located, a

NCS mask can be applied and NCS averaging is then performed between symmetry-

related areas of density in the unit cell.

Histogram modification is a technique borrowed from 2D image processing

where the histogram of an electron density map of the protein region is matched to the

expected histogram computed from the map of a well refined protein. The method

works down to 4.5 Å and is weaker than solvent flattening although it performs better

at extending phases to higher resolutions for structures that do not have a large

proportion of heavy metal atoms. The topology of the protein is not important but the

resolution range and solvent content needs to be matched to the computed histogram.

Lastly, multi-resolution modification is an extension of solvent flattening and

histogram matching that allows them to be applied at two resolutions simultaneously.

Refinement

Refinement is the process of improving the fit between the calculated protein

model and its observed structure factors. The R factor is typically 50% for a starting

model obtained from tracing the electron density, rather close to the value of 59% for a

random acentric structure (Wilson, 1950). The R factor is used to estimate errors in a

34

data set. It is usually expressed as the percentage agreement between the observed

F(hkl)calc and calculated structure factors F(hkl)obs.

R = ∑

∑ −

hklobs

hklcalcobs

hklF

hklFhklF

)(

)()(

Refinement lowers the R factor to 10-20% for protein crystals. It consists of

changing the three positional parameters (x, y, z) and temperature factor B, for all

atoms in the structure, except hydrogen atoms. This is because hydrogen atoms have

only one electron and consequently have little influence on X-ray scattering. Small

molecule crystal structure analysis is always overdetermined since there are more

observations than unknowns. However, this ratio is very low in protein crystal

structure refinement, and stereochemical parameters obtained from amino acid and

small peptide structures have to be incorporated in the refinement process as

'observations'. The data from small molecule structures have precise bond lengths,

bond angles, torsion angles and van der Waals contacts; and are assumed to be valid in

proteins. The stereochemical information can be applied as rigid constraints, allowing

only the dihedral angles to vary. These constraints make it difficult to move small parts

of the structure to an ideal position because many angular motions are involved. On the

other hand, the stereochemical parameters can be restrained around a standard value,

controlled by an energy term. Restraints make it easy to move small parts of the

structure, but not entire molecules or domains.

Low resolution Bragg reflections (d ≥8-10 Å), are usually omitted from the

refinement process and bulk solvent correction is applied, since their intensities are

seriously affected by the disordered solvent. Besides, the configuration for MAD and

MIR data collection is not ideal for measuring such data, because the beamstop is

positioned close to the crystal to block the low order reflections, and the shortest

35

possible crystal to detector distance is used to avoid using a helium path. MAD and

MIR data collection techniques are optimized for collecting medium to high resolution

data (d ≤8 Å) where their effects are dominant. Ordered solvent molecules contribute

to high resolution diffraction and are introduced during the refinement process if the

protein crystal diffracts to better than 2.5 Å resolution.

36

37

Materials and methods

Plasmid construction

The relevant DNA segments coding for CHDK, CHDT, CHDv2, CHDv3,

CHDv4, IQ, GRD1, GRD2, GRD2v3, GRDv4 and GRDv5 of rng2 (gene

SPAC4F8.13c, GenBank accession number NP_593860.1) (Table 3) was amplified

from the pREP4X-GST-rng2 (pCDL222.1) construct containing the full length rng2

sequence using the corresponding 5' forward primers and the 3' reverse complement

primers. Residues 1-189 of Rng2 contain the calponin-homology domain (Rng21-189).

The PCR product of CHDK (Table 3) was cloned into the pGEX-KG bacterial

expression vector (Guan and Dixon, 1991) for producing proteins with an amino-

terminal fusion to a Glutathione S-Transferase (GST) tag; whereas the PCR products

of CHDT, CHDv2, CHDv3, CHDv4, IQ, GRD1, GRD2, GRD2v3, GRDv4 and

GRDv5 (Table 3) were cloned into the pRSETC bacterial expression vector

(Invitrogen) for producing amino-terminal polyhistidine (6×His) fusion proteins. In the

case of the GST-fusion protein, there is a 26 amino acid residue linker

(SDLVPR^GSPGISGGGGGINSRLHGST) separating the GST domain from the

protein. It has a glycine rich sequence PGISGGGGG immediately after the thrombin

cleavage site that increases the thrombin cleavage efficiency (Guan and Dixon, 1991).

The linker is cleaved at the indicated site between arginine and glycine by thrombin. In

the His6-CHDT fusion protein, there is an amino-terminal fusion of

MRGSHHHHHHGMASMTGGQQMGRDLYDDDDKDRWIP preceding the insert;

whereas the His6-CHDv2, His6-CHDv3, His6-CHDv4, His6-IQ, His6-GRD1, His6-

GRD2, His6-GRD2v2, His6-GRD2v3 and His6-GRD2v4 fusion proteins have an

amino-terminal fusion of MRGSHHHHHHGMAS due to the 5' NheI restriction site.

Construct 5' forward primer 3' reverse complement primer

Rng2 residue number pI/MW (after cleavage for GST fusion proteins)

Expression

pGEX-KG-CHDK (SalI) 5' GACGAC GTCGAC AATGGACGTAAATGTGGG 3' (HindIII) 5' GCGCCG AAGCTT TGA AGTTAGGAAGGATTAC 3'

1-189 6.66/23951.3 +

pRSETC-CHDT (EcoRI) 5' GACGAC GAATTC ATGGACGTAAATGTGGGA 3' (SacI) 5' GCGCCG GAGCTC TCA TAAAGCTTTGAAGTTAGG 3'

1-190 6.19/25788.39 low

pRSETC-CHDv2 (NheI) 5' GACGAC GCTAGC AGAGAAACCCTTCAAGCA 3' (EcoRI) 5' GCGCCG GAATTC TCA CATAGAAAGGAAATAAGA 3'

32-149 6.34/15407.68 -

pRSETC-CHDv3 (NheI) 5' GACGAC GCTAGC GACGTAAATGTGGGATTA 3' (EcoRI) 5' GCTCCG GAATTC TCA TAAAGCTTTGAAGTTAGG 3'

2-190 6.79/23062.39 low

pRSETC-CHDv4 (NheI) 5' GACGAC GCTAGC GACGTAAATGTGGGATTA 3' (EcoRI) 5' GCGCCG GAATTC TCA TAAAGTAGGAGAAGTATA 3'

2-238 8.38/28293.19 low

pRSETC-IQ (NheI)5' GACGAC GCTAGC GAACAAAGTTCTTCTGTC 3' (EcoRI) 5' TCGACC GAATTC TCA CTTATGAAAATTTTGAAG 3'

358-684 10.29/39910.3 -

pRSETC-GRD1 (NheI) 5' GATGAT GCTAGC GAAAATCCCGTTGCTCAG 3' (EcoRI) 5' TCATCA GAATTC TCA GATTGAAAAGATCTCAGA 3'

867-1078 9.22/25804.99 -

pRSETC-GRD2 (NheI) 5' GATGAT GCTAGC GAAAATCCCGTTGCTCAG 3' (NcoI) 5' TCATCA CCATGG TCA AAATTTGCCAACTTCATC 3'

1202-1431 9.69/28594.80 low

pRSETC-GRD2v2 (NheI) 5' GACGAC GCTAGC AAGCTTATTCCTTTTTCT 3' (EcoRI) 5' GCTCCG GAATTC TCA TGAATAAAATTTGGAGAA 3'

1368-1489 9.81/15712.34 -

pRSETC-GRD2v3 (NheI) 5' GACGAC GCTAGC CGGAATAATTACCAAGAT 3' (EcoRI) 5' GCTCCG GAATTC TCA TGAATAAAATTTGGAGAA 3'

1297-1489 10.0/24236.01 low

pRSETC-GRD2v4 (NheI) 5' GACGAC GCTAGC GTTACAGATTCAGACGAA 3' (EcoRI) 5' GCTCCG GAATTC TCA TGAATAAAATTTGGAGAA 3'

1239-1489 9.45/30857.29 low

Table 3: Rng2 plasmids. Restriction sites of the primers are shown in brown, whereas insert sequences are in blue.

37

Protein expression

The plasmids were transformed into Escherichia coli BL21-CodonPlus-RIL (F-

ompT hsdS(rB- mB

-) dcm+ Tetr gal endA Hte [argU ileY leuW Camr]) from Stratagene

and sequenced. A 0.5 ml aliquot of freshly saturated cells was mixed with an equal

portion of a cryoprotectant consisting of 65% glycerol, 0.1 M MgSO4, 25 mM Tris pH

8 and frozen away at -80 °C.

To check the solubiblity of the various recombinant proteins, freshly thawed

cells were inoculated into 2 ml of Luria-Bertani (LB) media supplemented with 100

µg/ml ampicillin and grown in a shaker overnight at 37 °C. The overnight culture was

then inoculated into 50 ml of 2×YT media in a 250 ml conical flask and induced for

three hours with 0.5 mM IPTG at A600 ≈ 0.6. A 25 ml aliquot of the bacterial culture

was harvested by centrifugation (4000 g for 15 mins at 4°C), resuspended in 1 ml of

cold PBS pH 7.4, 2 mM DTT and homogenized in a cold pressure cell with a

FRENCH press (Thermo Spectronic) operated at a gauge pressure of 1500 psi. A 100

µl aliquot of the cell lysate was set aside, and the remainder was centrifuged (10000 g

for 20 mins at 4 °C) to separate the supernatant containing the soluble fraction from

the pellet containing the insoluble fraction. An analyis using SDS-PAGE indicated that

only the pGEX-KG-CHDK plasmid expressing the GST-CHDK (GST-Rng21-189)

fusion protein was both soluble and had sufficient yield (Table 3).

For the purpose of GST-Rng21-189 production, freshly thawed cells were

inoculated into 50 ml of Luria-Bertani (LB) media supplemented with 100 µg/ml

ampicillin and grown in a shaker for seven hours at 37 °C. A 10 ml aliquot was then

used to inoculate 1 liter of 2×YT media in a 5 liter Erlenmeyer flask and induced with

0.5 mM IPTG at A600 = 0.6. The cells were harvested 3 hours later by centrifugation

(4500 g for 10 mins), resuspended in 5 ml of cold PBS pH 7.4, 2 mM DTT and

38

homogenized in a cold pressure cell with a FRENCH press (Thermo Spectronic)

operated at a gauge pressure of 1500 psi. 10 µg ml-1 of DNase and RNase was added to

the lysate and incubated for 30 mins at 4°C on a rocking platform. Centrifugation at

10000 g for 20 mins was used to separate the supernatant from the pellet, which was

discarded. The supernatant was clarified using a 0.45 µm filter.

Protein purification

The GST-Rng21-189 fusion protein was bound to a 5 ml GSTrap FF column

(Pharmacia), washed with 20 mM Tris pH 8, 0.5 M NaCl and eluted with 50 mM Tris,

10 mM reduced glutathione pH 8. Twenty units of thrombin were added to the eluate

and incubated overnight at 4 °C. The fusion protein was cleaved efficiently (Figure 3).

The eluate was dialyzed into 20 mM Tris, 0.5 M NaCl using a 3500 MW cutoff

dialysis tubing (Pierce). Thrombin and the cleaved GST were removed by passing

through a 5 ml HiTrap Benzamidine FF (Pharmacia) column and a 5 ml GSTrap FF

column that was connected in series. Rng21-189 was concentrated to 15 mg ml-1 in 50

mM Tris pH 8, 0.15 M NaCl, 2.5 mM CaCl2 with centrifugal filter devices from

Amicon.

39

Figure 3: Thrombin cleavage of GST-Rng21-189. Lane M, molecular weight standard; Lane 1, before cleavage; Lane 2, eluate containing Rng21-189 after 3 hr cleavage at 4 °C with 20 U of thrombin; Lane 3, eluate using glutathione buffer; Lane 4, purified Rng21-

189.

Protein crystallization

Crystallization screens of Rng21-189 were initially performed with commercially

available solutions (Wizard I and II, deCODE genetics) using the sitting drop vapour

diffusion method. Crystallization setups contained 2 µl of Rng21-189 protein solution

(15 mg ml-1 in 50 mM Tris pH 8, 0.15 M NaCl, 2.5 mM CaCl2) and 2 ul of Wizard I/II

solution. The wells were sealed with clear sealing tape (Hampton Research). A rod

shape crystal was observed in one well from the Wizard II screen. This condition was

slightly optimized by exploring different precipitant concentrations (18-21% PEG),

precipitants (PEG 3000 and 4000), salt concentrations (0.2-0.3 M calcium acetate) and

buffer pHs (pH 7-7.5). Up to five monoclinic crystals of Rng21-189 appeared in each

40

sitting drop well with 21% PEG 3000, 0.3 M Calcium Acetate and 0.1 M Tris pH 7

after incubation at room temperature (22 °C) for eight days. They grew to maximal

dimensions (0.3 × 0.2 × 0.05 mm) within two weeks (Figure 4).

Figure 4: Crystals of Rng21-189. They typically measure 200 µm in the longest dimension.

Data collection

The crystals were first transferred into the reservoir buffer solution to remove

the precipitate. This was followed by 15 mins fixation in 21% PEG 3000, 3%

glutaraldehyde at 15 ºC in a sitting drop tray. This brief fixation prevents crystals with

dimensions greater than 0.2 mm from cracking across the unique axis while soaking in

cryobuffer or in derivatization solutions (Fitzpatrick et al., 1994). My fixation protocol

was modified by doubling the glutaraldehyde concentration and shortening the fixation

times since my crystals did not diffract after overnight fixation. In contrast, extended

cross-linking times of a day or more with lysozyme crystals did not make the crystals

disordered (Wang et al., 1998).

The native crystals were soaked in a cryobuffer containing 21% PEG 3000,

40% glycerol. Bromide derivatives were soaked for 30 mins in 21% PEG 3000, 25%

glycerol, 1 M NaBr. Mercury derivatives were soaked for 2 h in saturated HgCl2, 40%

glycerol. The crystals were mounted on nylon loops, frozen in liquid nitrogen and

41

transported to the synchrotron in a dry shipping dewar (Taylor-Wharton CP-100) for

data collection by the European Synchrotron Radiation Facility (ESRF).

X-ray diffraction data from Rng21-189 were measured under cryogenic

conditions at the UK/EMBL MAD beamline BM14 at the ESRF using a MarCCD 165

mm detector (Mar USA). The data set was collected as a continuous series of 2°

oscillation images covering a rotation range of 200°. Indexing and integration of the

images was performed in DENZO and scaling of the intensity data was performed in

SCALEPACK from the HKL package (Otwinowski and Minor, 1997; Table 4).

Structure determination

A three wavelength MAD data set for the Br-derivative was collected.

However, the inflection and remote wavelengths had sufficient anomalous signal to do

a two wavelength MAD (Gonzalez, 2003). This served as input data to SHELXD

(Sheldrick, 1998) after extracting Bijvoet differences by XPREP (Bruker). SHELXD

was run with 388 Es greater than 1.5 looking for 4 anomalous sites. The six highest

peaks with a correlation coefficient (CC) of 38.41 and a figure-of-merit (PATFOM) of

46.95 were accepted after the Harker peaks (2x, 2z) at the Harker section y = ½

corresponding to the Br positions in a Br-anomalous difference Patterson (or Bijvoet

Patterson) map were examined for verification (Figure 5).

42

Figure 5: Harker section y = ½ corresponding to the Br positions in a Br-anomalous difference Patterson map.

SHARP (de La Fortelle and Bricogne, 1997) was used to refine the six Br sites,

from which the protein phases were derived and used to trace an initial model with

wARP (Perrakis et al., 1999). The map obtained after density modification was readily

interpretable; 147 out of 158 residues could be built automatically using wARP

(Perrakis et al., 1999). The incomplete polyalanine model was visualized with O

(Jones et al., 1991) to fill in missing portions of the main chain and to delete

symmetry-related main chains. At this point side chains were not added owing to

ambiguity while matching the sequence to the Fo map due to the poor side chain

densities along the random coils interspersed between the α-helices.

43

A SAD data set was also collected for the Hg-derivative at a wavelength of

1.653 Å. It had a strong anomalous signal even though the L-III wavelength is around

1 Å. SHELXD was run with 391 Es greater than 1.5 looking for 3 anomalous sites. The

four highest peaks with a CC of 46.67 and a PATFOM of 100.97 were accepted after

the Harker peaks at Harker section y = ½ of the Hg Bijvoet Patterson map were

checked to verify the four Hg sites (Figure 6).

Figure 6: Harker section y = ½ corresponding to the Hg positions in a Hg-anomalous difference Patterson map.

148 out of 158 residues could be built with wARP. The three Hg sites with the

highest occupancies were used to determine the positions of the three cysteine residues

present in the recombinant protein. The fourth residue had a lower occupancy and was

44

not covalently bonded to any residue (Figure 11b). This information allowed me to

match the sequence to the Br and Hg Fo maps with confidence. A Br anomalous

difference Fourier map was used to locate five additional Br sites.

The two structures containing either eleven Br or four Hg atoms but devoid of

water were refined initially with CNS (Brünger et al., 1998). 5% of randomly selected

reflections from the data set were used for evaluating Rfree. The refinement was

continued with SHELXL (Sheldrick and Schneider, 1997) using the same set of

reflections for calculating Rfree and with default restraints applied to geometrical and

atomic displacement parameters. Each cycle of SHELXL refinement consists of ten

steps of conjugate-gradient least-squares minimization followed by inspection of the

model and electron density maps with O. The water molecules added during SHELXL

refinement were processed with unit occupancies and inspected for the presence of

spherical electron density (1σ) in the 2mFo - DFc map, at <3.5 Å (Ippolito et al., 1990)

from suitable hydrogen-bond acceptors or donors. The process was repeated to add,

delete and refine the water positions until no further interpretation of the difference (Fo

– Fc) density contoured at 3σ and -3σ was possible. At the end of refinement, when the

process reached convergence, three cycles of tightly restrained (weight matrix 0.1),

isotropic refinement with REFMAC (Murshudov et al., 1997) was performed using all

reflections, including those previously used for Rfree, to obtain as accurate a model as

possible (Dauter et al., 2001). The refined structure was analyzed with PROCHECK

(Laskowski et al., 1993), and the coordinates of the Br- and Hg-derivatives of Rng232-

189 are deposited in RCSB Protein Data Bank with the respective accession codes,

1P2X and 1P5S. Its crystallographic data are listed in Table 4.

45

Table 4: Data collection, processing, phasing and refinement statistics. Numbers in parentheses refer to the highest resolution shell. Data collection statistics Crystal Rng21-189 + 1 M NaBr Rng21-189 +

saturated HgCl2Temperature (K) 100 100 Wavelength (Å) peak, 0.919087; inflexion, 0.919338; remote,

0.855057 1.653

X-ray source ESRF BM14 ESRF BM14 Data processing statistics Resolution range (Å) processed

Peak, 23-2.21 (2.29-2.21); inflexion, 23-2.21 (2.29-2.21); remote, 23-2.06 (2.13-2.06)

28-2.21 (2.29-2.21)

Spacegroup P21 P21Unit-cell parameters peak wavelength a (Å) 31.266 30.884 b (Å) 68.836 68.666 c (Å) 39.811 35.309 α = γ (°) 90 90 β (°) 105.544 102.514 VM (Å3·Da-1) 2.3 2.0 Solvent content (%) 46.3 39.3 Observed reflections 25342 130550 Unique reflections 7347 (202) 7034 (570) Completeness (%) 89.6 (24.8) 97.4 (81.5) Rmerge (%) a 5.3 (7.7) 8.4 (14.5) I/σ(I) 20.43 (9.39) 32.70 (9.93) MAD/SAD phasing statistics Resolution range (Å) processed

22.667-2.062 27.606-2.216

Accepted reflections 9130 6895 Isomorphous RCullis (centric) b 0.533 - Anomalous RCullis (acentric) c 0.772 0.625 Isomorphous phasing power (acentric) d 1.43 - Anomalous phasing power (acentric) e 1.248 1.915 Figure of merit (acentric) 0.51757 0.50549 Refinement statistics Resolution (Å) range processed

23-2.21 21-2.21

Reflections in working set 7173 6571 Reflections in test set 369 319 Rcryst (%) f 20.7 (27.5) 18.6 (22.0) Rfree (%)

g 26.9 (29.3) 24.4 (30.1) Protein atoms 1303 1303 Derivative atoms 11 4 Water atoms 157 145 R.m.s.d. bond lengths (Å) 0.008 0.007 R.m.s.d. bond angles (°) 1.0 1.2 R.m.s.d. dihedral angles (°) 19.9 23.0 R.m.s.d. improper angles (°) 1.01 1.78 Mean B factor (Å2) 17.6 22.2

46

a Rmerge = ∑

∑ −

hklobs

hklobs

hklI

hklIhklI

)(

)()(

b Isomorphous RCullis (centric) = ∑

∑±

−±

hklPPH

hklcalcHPPH

FF

FFF )(

c Anomalous RCullis (acentric) = ∑

∑±

±±

∆

∆−∆

hklobsPH

hklcalcPHobsPH

F

FF

)(

)()(

d Isomorphous phasing power = ∑

∑−

hklcalcPHobsPH

hklcalcH

FF

F

)()(

)(

e Anomalous phasing power (acentric) = ∑

∑±± ∆−∆

hklcalcPHobsPH

hklifH

FF

F

)()(

")(

f Rcryst = ∑

∑ −

hklobs

hklcalcobs

hklF

hklFhklF

)(

)()(

g Rfree = Rfactor for 5% of reflections not used in refinement (test set).

47

Molecular dynamics simulation in an aqueous system

Production-phase molecular dynamics was performed with the AMBER 7.0

suite of programs (Case et al., 2002) and PMEMD (Duke and Pedersen, 2003), with a

2.0 fs time step under the isothermal-isobaric ensemble (300 K and 1 atm) with the

ff99 force field (Wang et al., 2000), the TIP3P (Jorgensen et al., 1983) model for water,

periodic boundary conditions, the particle mesh Ewald method (PME) (Darden et al.,

1993) for electrostatics, a 10 Å cutoff for Lennard-Jones interactions, and the use of

SHAKE (Ryckaert et al., 1977) for restricting motion of all covalent bonds involving

hydrogen. In an iterative manner, five sodium cations were placed around Rng232-189

near residues with the highest electrostatic potential to compensate the overall negative

charge of the molecule. A total of 8763 TIP3P water molecules were then added

around residues 32-189 of Rng2 in order to end up with a buffer of about 10 Å from

the edge of the periodic box, which resulted in a box size of 72.535 × 70.410 × 72.229

Å. Temperature was maintained by the Berendsen coupling algorithm (Berendsen et al.,

1984) using separate τ coupling constants of 1.0 for the protein and solvent, and

pressure was maintained with isotropic molecule-based scaling (Berendsen et al.,

1984), also with a τ coupling constant of 1.0. The PME grid spacing was about 1.0 Å

and was interpolated on a cubic B-spline, with the direct sum tolerance set to 10-5. The

net center of velocity was removed every 100 ps to correct for the small energy drains

that resulted from the use of SHAKE, discontinuity in the potential energy near the

Lennard-Jones cutoff value and constant pressure conditions.

For equilibration, the solute (Rng232-189) was minimized using the steepest

descent method for the first 200 steps, followed by the conjugate-gradient method until

the root-mean-square (r.m.s.) of the Cartesian elements of the gradient was <0.4 kcal

mol-1 Å-1. Water molecules and sodium ions were then minimized in the same way

48

until the RMS was <0.1 kcal mol-1 Å-1 and then slowly heated, while allowing them to

move unrestrained for 25 ps (with a 1.0 fs time step) in order to fill in any vacuum

pockets. The solute atoms alone were then minimized in the presence of ever

decreasing positional restraints, thereby allowing them to slowly feel the forces of the

equilibrated solvent, until the positional restraints reached zero. Finally, a temperature

ramp was used to gradually raise the temperature of the whole system to 300 K over 20

ps. Coordinates from the trajectories were saved every 5 ps and analyzed using

AMBER 7.0.

Phylogenetic analysis

The boundaries of CH domains within each protein were located using profile

hidden Markov models (profile HMM) (Eddy, 1998) and a multiple sequence

alignment was obtained by running the list of CH domains through CLUSTALX

(Thompson et al., 1997). For the phlygenetic analysis, sites where gaps were

introduced during the multiple sequence alignment were excluded from all pairwise

comparisons. This is because there is no consensus to the problem of judging the order

and number of insertions, deletions or rearrangements of genetic material. To

reconstruct phylogenetic trees from the 117 individual CH domains, the consensus

from five common tree-building methods was used. Neighbour joining (NJ) (Saitou

and Nei, 1987), maximum parsimony (MP) (Eck and Dayhoff, 1966) and maximum

likelihood (ML) (Felsenstein 1981) trees were produced using PHYLIP (Felsenstein,

1989). The minimum evolution (ME) (Rzhetsky and Nei, 1992) tree was constructed

using MEGA2 (Kumar et al., 2001), whereas the quartet maximum likelihood (QML)

(Strimmer and von Haeseler, 1996) tree was constructed using TreePuzzle-5.1

(Schmidt et al., 2002). NJ and ME trees do not assume an evolutionary clock and are

49

unrooted. Default amino acid substitution matrices were used for trees constructed

using PHYLIP and MEGA2. QML trees were based on the VT model of amino acid

sequence evolution (Müller and Vingron, 2000) with a uniform rate heterogeneity.

The significance of internal branches in ME trees was tested by the standard

error test, and the standard errors of branch length were computed by the bootstrap

method implemented in the MEGA2 program. The reliability of each internal branch in

NJ, ME, MP and ML trees was assessed by bootstrapping: 1000 bootstrap samples

were used with the exception of ML where 100 bootstrap samples were used. The

reliability of each internal branch in QML trees was assessed by the proportion of

10000 puzzling steps supporting the branch (Strimmer and von Haeseler 1996).

Generation of figures and tables

The r.m.s.d. values listed in Table 5 were made with the help of Stralign

(http://www.hgc.ims.u-tokyo.ac.jp/service/tooldoc/stralign/intro.html). Other r.m.s.d.

values were calculated with the help of the MMTSB Tool Set (Feig et al., 2001).

Figures 7(a)-7(c) were prepared using ESPript (Gouet et al., 1999). The structural

alignment shown in Figure 7(b) was prepared with STAMP (Russell and Barton, 1992).

Figures 8(a)-8(c), 9, 10(a)-10(e), 11(a)-11(b) and 13 were prepared using the programs

POVScript+ (Fenn et al., 2003) and Raster3D (Merritt and Bacon, 1997). The multiple

structure superposition of CH domains shown in Figures 8(b) and 8(c) were prepared

using Indonesia (http://xray.bmc.uu.se/dennis/), using the brute-force method with a

cutoff of 10 Å and a fragment length of 100 residues. The molecular surface

representation of Rng232-189 sequence conservation as shown in Figure 8(a) was

prepared using both PyMOL (http://pymol.sourceforge.net/) and ProtSkin

(http://www.mcgnmr.ca/ProtSkin/). The surface hydrophobicities and surface

50

electrostatic potential as shown in Figure 8(a) were prepared using PyMOL and DelPhi

(Rocchia et al., 2001). DelPhi uses kT at 298 K as the energy unit and kT/e as the

electrostatic potential unit, where 1 eV = 38.94 kT (298 K). The crystal contacts shown

in Figures 10(a1) and 10(a2) were visualized with the help of Swiss-PdbViewer (Guex

and Peitsch, 1997). Figure 12 was prepared with Phylodendron

(http://iubio.bio.indiana.edu/treeapp/treeprint-form.html), using the phenogram mode.

Results and discussion

Overall structure

Residues 1-189 of S. pombe Rng2, which contain a single calponin-homology

domain, has been crystallized and the structures of a Br- and a Hg-derivative were

determined using multi-wavelength anomalous diffraction (MAD) methods (Wang et

al., 2003; Table 4). There is a single molecule per asymmetric unit in space group P21,

and interpretable electron density was observed from residues 32-189. Residues 1-31

are disordered and no electron density can be seen. The overall structure is compact

and globular. 100% of the residues in both derivatives were either in the 'most

favoured' or 'additional allowed' regions of Ramachandran space (Laskowski et al.,

1993). The overall mean B values are 17.51 and 19.89 Å2 for the Br- and Hg-derivative,

respectively, in reasonable agreement with the Wilson B values of 21.4 and 18.3 Å2.

The respective estimated standard uncertainties (esu) for the B values based on

maximum likelihood are 5.58 and 5.51 Å2.

A multiple sequence alignment of all CH domains known to bind F-actin is

shown in Figure 7(a). There are clearly no conserved residues although hydrophobic

residues are seen to alternate with polar/acidic/basic residues between helices a1-a3

51

and h2-a6. The CH domain resides between residues 40-150, and consists of four core

α-helices (a1, a3, a4, a6) connected by long loops (Figure 8a). There are two short α-

helices, a2 and a5. There are also three short stretches of 310-helices, h1, h2 and h3. α-

Helices a3 and a6 are almost parallel to one another, while a1 and a4 are perpendicular

to each other. The CH domain fold is held together by extensive hydrophobic contacts

among the α-helices in the core helical structure (Figure 8c).

Residues 150-189 do not have a defined secondary structure, except for the α-

helical stretch (a7) between Asp167 to Gln179. It wraps around helix a1, forming a

substructure that resembles neither the extended conformation seen in utrophin, nor the

compact conformation seen in fimbrin, although residues 154-160 forms an

unstructured coil which adopts a substructure similar to dystrophin residues 240-246 in

the carboxyl-terminal portion of the CH2 domain. The primary sequence within

residues 150-189 is unique to Rng2 and cannot be found elsewhere.

Data from fission yeast showed that Rng2 localizes to the F-actin ring which

forms during mitosis (Eng et al., 1998). A F-actin binding function probably exists

within the first 189 residues of Rng2, since an amino-terminal fusion of GFP to Rng21-

189 localizes to cortical actin patches within the growing cell tips (Wachtler et al.,

2003). F-actin cosedimentation assays indicate a direct interaction with residues 1-216

of human IQGAP1 (Fukata et al., 1997), which contain a single CH domain.

Interestingly, residues 150-189 of Rng2 wraps around the proposed actin-binding site

ABS1 and ABS2 (Corrado et al., 1994; Norwood et al., 2000; Figure 2a) via mainly

hydrophobic interactions, suggesting the putative ABS1 and ABS2 in Rng2 are not

involved in binding F-actin. However, this does not rule out the possibility of a direct

interaction between Rng21-189 and F-actin.

52

Figure 7(a): Multiple sequence alignment of calponin-homology domains from proteins known to bind F-actin. The blue square scale at the top shows the relative accessibility of each residue extracted from DSSP (Kabsch and Sander, 1983) - blue with red borders indicate exposed residues with an accessibility value A > 1; blue indicates accessible residues with A > 0.4, cyan indicates intermediate accessibility with 0.1 < A < 0.4, and white indicates buried residues with A < 0.1. The secondary structure of Rng2 and its residue numbering follows below and are shown in black. The predicted secondary structure of Rng2 based on PHD is shown at the bottom in red. The residues corresponding to Rng2 core α-helices (Figure 8c) are highlighted in yellow.

53

Figure 7(b): Structural alignment of all known calponin-homology domains. In the sequence name column, the number following the four digit PDB coordinate entry refers to either a CH1 or CH2 domain; it is absent for single CH domains. The last alphabet is a chain identifier. The cyan square scale at the bottom shows the hydropathic character of each residue according to the algorithm of Kyte and Doolittle (1982) with a window size of 3 - pink indicates hydrophobic residues with a value H > 1.5; grey indicates intermediate hydrophobicity with -1.5 < H < 1.5 and cyan indicates hydrophilic residues with H < -1.5.

Figure 7(c): Multiple sequence alignment of calponin-homology domains from the IQGAP family of proteins.

54

Figure 8: Structure of Rng232-189. (a) Ribbon representation of the crystal structure of Rng232-189 shown in two orientations with relative rotation of 180°. Bromine atoms are represented by spheres coloured according to B factors (blue to red denoting lower to higher values). The amino- and carboxyl-terminal residues are labelled together with α-helices a1-a7 and 310-helices h1-h3. For comparison purposes, the region corresponding to actin-binding sites ABS1 (a1) and ABS2 (h2-a6) of dystrophin CH1 domain (Norwood et al., 2000) are highlighted in green and blue, respectively. The 39 residues carboxyl-terminal to the CH domain (150-189) are highlighted in red. Orientations in (a)-(c) are the same. The molecular surfaces below the ribbon diagram are in the same orientation. Surface residue conservation scores ranging from 0 to 1 are coloured based on a white (not conserved) to red (conserved) scale. Kyte-Doolittle (Kyte and Doolittle, 1982) surface hydrophobicities ranging from -4.5 to 4.5 are coloured based on a cyan (hydrophilic) to wheat (hydrophobic) scale. Electrostatic surface potentials ranging from -155.214 kT/e to 318.183 kT/e are coloured based on a red (acidic) to blue (basic) scale, with a scale factor of 1.

55

Figure 8(b): Cα backbone superposition of the CH domains from the Br-derivative of Rng2 CH, Hg-derivative of Rng2 CH, dystrophin CH1, dystrophin CH2, calponin CH, EB1 CH, fimbrin CH1-1, fimbrin CH1-2, utrophin CH1, utrophin CH2, plectin CH1, plectin CH2 and beta-spectrin CH2. For clarity, only the Br-derivative of Rng232-189 CH (pink), dystrophin CH1 (red), calponin CH (orange) and plectin CH2 (blue) domains are shown here.

56

Figure 8(c): Cα backbone superposition of the four core α-helices belonging to the Br-derivative of Rng232-189 CH (pink), dystrophin CH1 (red), calponin CH (orange) and plectin CH2 (blue) domains. Orientations in (a), (b) and (c) are the same.

57

58

Comparison of CH domain structures

A superposition of the CH domains from Rng2, calponin, EB1, fimbrin,

dystrophin, utrophin, beta-spectrin and plectin (Figure 8b) shows that most of the

structural variations are concentrated at the amino- and carboxyl-terminus; i.e. before

residue 43 in helix a1 of Rng2 and beyond residue 148 in helix a6 of Rng2. The loops

between helices a1-a2, a3-a4 and h2-a6 also had considerable variation, consistent with

the alternative loop conformation found in Rng2 residues 92-99 (Figure 4a).

The Cα backbone for the four core helices containing 47 equivalent atoms of

other solved structures can be superimposed on those of Rng2 with up to 2.33 Å root-

mean-square deviation (r.m.s.d.) (Table 5). The closest match of Rng2 CH domain is

with the CH1 of dystrophin with a r.m.s.d. of 1.19 Å. In contrast, the second CH (CH2)

domain of plastin 3/fimbrin has a r.m.s.d. of 1.75 Å when compared with Rng2. The

CH1 domains of dystrophin, utrophin and plectin are all similar with a r.m.s.d. of

about 0.5 Å. This is consistent with their known actin-binding properties.

Rng2 CH DMD CH1 UTRN CH1 PLEC CH1 CNN CH PLEC CH2 SPTB CH2 DMD CH2 PLS3 CH1 UTRN CH2 PLS3 CH2 EB1 CH

DMD CH1 1.19 -

UTRN CH1 1.28 0.47 -

PLEC CH1 1.30 0.52 0.48 -

CNN CH 1.45 1.54 1.54 1.54 -

PLEC CH2 1.53 1.25 1.18 1.11 1.33 -

SPTB CH2 1.56 1.23 1.16 1.09 1.40 0.43 -

DMD CH2 1.63 1.19 1.14 1.01 1.47 0.52 0.60 -

PLS3 CH1 1.72 1.45 1.29 1.27 1.63 1.41 1.31 1.30 -

UTRN CH2 1.74 1.44 1.35 1.27 1.46 0.74 0.67 0.82 1.33 -

PLS3 CH2 1.75 1.33 1.38 1.20 1.59 1.08 1.02 0.92 1.44 1.25 -

EB1 CH 2.28 2.25 2.36 2.33 2.15 2.19 2.12 2.32 1.84 2.11 2.33 -

Table 5: R.m.s.d. values of the pairwise comparison of all the CH domains using the Cα belonging to the four core helices of each structure. The core helices of Rng232-189 correspond to residues 43-54, 73-81, 99-112 and 133-144 (Figure 8c). Abbreviations: DMD, dystrophin; UTRN, utrophin; PLEC, plectin; CNN, calponin; SPTB, beta-spectrin; PLS3, fimbrin ABD1; EB1, end-binding protein 1.

59

Superposition of the Rng2 CH domain with the calponin CH domain reveals a

relatively large r.m.s.d. of 1.45 Å. The largest deviation is found between the CH

domains of Rng2 and end-binding protein 1 (EB1) (Hayashi and Ikura, 2003), with a

r.m.s.d. of 2.33 Å. It is apparent from Table 5 that the CH domains from Rng2,

calponin and EB1 are quite dissimilar to each other and do not cluster together like

CH1 and CH2 domains. The Rng2 CH domain is more similar to the F-actin binding

CH1 domains, whereas the calponin CH domain is more similar to CH2 domains. In

contrast, the EB1 CH domain is dissimilar to the rest of the CH domains. This agrees

with the different functional roles they play. For instance, Rng2 CH domain

localization correlates with cortical actin patches in fission yeast (Wachtler et al.,

2003), and the CH domain from its mammalian orthologue IQGAP binds F-actin and

calmodulin (Ho et al., 1999); calponin CH domain binds calmodulin, caltropin and

ERK (Wills et al., 1994; Leinweber et al., 1999); and EB1 associates with microtubule

filaments (Hayashi and Ikura, 2003).

The Rng2 CH domain has three 310 helices (h1-h3) that are uniformly three

residues long and thus only one turn in length (Figure 8B). The dipoles are less well

aligned here compared to regular α-helices (Figure 9). The C=O groups are hydrogen

bonded to crystallographic waters and point away from the helix axis. The 310-helices

have less favourable side chain packing and backbone conformation and are usually

found in solvent accessible areas. In the case of Rng2, all the 310-helices are solvent

accessible. They link random coils or helices together at abrupt angles. This is most

clearly seen in helix h2 (Figure 8a).

It has been reported that the crystal structure of proteins subjected to light

cross-linking with glutaraldehyde were identical to those without such treatment

(Fitzpatrick et al., 1994). Thus, no data were collected from a native crystal. No

60

glutaraldehyde cross-links were observed in the electron density map. They form

randomly and should not occur in high densities if the crystals were cross-linked

briefly (Fitzpatrick et al., 1993).

Figure 9: A 310-helical structure between residues 62-64 (helix h1). Water molecules are represented by cyan spheres and dotted lines indicate hydrogen bonds. The 2mFo - DFc map is contoured at 1σ and superimposed on the coordinates.

61

Intermolecular contacts

In the monoclinic crystal of Rng21-189, each molecule is surrounded by ten

symmetry-related molecules within an intermolecular distance of 4 Å (Figure 10a1 and

10a2). The direct intermolecular contacts and those mediated via crystallographic

waters are named macrobonds (Oki et al., 1999), with symmetrically equivalent

macrobonds forming the same interatomic interactions. There are five groups of

macrobonds, labeled A-E. The contacting molecules between macrobond A are related

by a two-fold screw axis along the b axis. Those between macrobond B are related by a

unit translation of the repeating lattice unit along the a axis. Macrobonds A and B

consist of numerous non-covalent interactions. Figures 10(b1) to 10(b3) shows the

different types of interactions in macrobond A. There is a potential salt bridge and a

potential water-mediated hydrogen bond between the ζ-amino group of Arg70 and the

γ-carboxyl group of Asp151 (Figure 10b1). However, the temperature factor of the

water molecule is large (49.65 Å2) suggesting local disorder and a weaker hydrogen

bond. The ζ-amino group of Arg70 is linked via a potential salt bridge to the δ-

carbonyl group of Gln67. A bromide anion with a temperature factor of 24.54 Å2

forms a potential hydrogen bond to the ζ-amino group of Arg175, which is doubly

hydrogen bonded to the carbonyl group of Leu152 (Figure 10b2). There is a nonpolar

cluster (>5 Å) of residues Ile172, Leu152, Phe82, Phe110 and Ile114 in a background

of specific intermolecular hydrogen bonds and salt bridges (Figure 10b3). The inter-

residue distances are longer than the usual distance of 3.7 Å (Matsuura and Chernov,

2003) and are unlikely to be van der Waals contacts.

The contacting molecules between macrobond C are related by a unit

translation along the c axis followed by another unit translation along the a axis

(Figure 10c). There is a potential direct hydrogen bond between the ζ-amino group of

62

Arg99 and the δ-carbonyl group of Glu161. It also has another potential direct

hydrogen bond between the α-imino of Tyr92 and γ-carbonyl group of Asn162.

Macrobond D comprises contacts made by molecules related by a unit

translation along the c axis (Figure 10d). It has a potential water-mediated hydrogen

bond from the γ-carbonyl group of Asn162 to the ε-amino group of Lys131. However

the sigma-A weighted map (Read, 1986) suggests the side chain of Lys131 is

disordered.

Molecules between macrobond E are related by a lattice translation of one unit

along the a axis followed by a two-fold screw axis along the b axis (Figure 10e). It has

a unique occurrence of a potential bromide-mediated hydrogen bond between the α-

imino group of Pro117 and the ζ-amino group of Arg178.

The total numbers of potential direct hydrogen bonds in the Br-derivative with

a length of less than 3.5 Å (Ippolito et al., 1990) between amino acid residue atoms are

5, 2, 3, 0, and 0, respectively, for macrobonds A, B, C, D and E. The numbers of bound

water molecules involved in the intermolecular hydrogen bonds are 24, 24, 8, 1 and 0,

respectively. In line with the relatively high solvent content of protein crystals, water-

mediated hydrogen bonds are clearly responsible for holding the crystal together. It is

interesting to note that the longest dimension of the unit cell extends in the b axis and

involves macrobonds A and A' which has the most direct hydrogen bonds and potential

water-mediated hydrogen bonds. The favourable formation of the periodic bond chain

(PBC) A-A' results in a tendency to elongate in this direction (Hartman, 1973). PBCs

refer to the intermolecular interactions or bonds occuring periodically across a crystal

that contribute to its growth habit. This was indeed observed during the crystal

screening process when many conditions resulted in crystals with a fibrous

63

morphology. The c axis and a axis have about the same length and involve

macrobonds C and D, and B respectively.

The Hg-derivative shows a similar arrangement of intermolecular contacts

(Figure 10a2). This is to be expected since the crystals were derivatized using crystals

grown under the same conditions. The total numbers of potential direct hydrogen

bonds in the Hg-derivative with a length of less than 3.5 Å between amino acid residue

atoms are 8, 6, 6, 6, and 1, respectively for macrobonds A, B, C, D and E. The numbers

of bound water molecules involved in the intermolecular hydrogen bonds are 13, 20, 9,

20 and 4, respectively. The significant increase in the number of potential direct and

water-mediated hydrogen bonds in macrobond D relative to the Br-derivative

correlates with a significant decrease in unit cell dimensions along the c axis.

64

Figure 10: Intermolecular contacts in a P21 crystal of the Rng2 CH domain. (a1) Molecular packing projected from the c axis (upper) and b axis (lower) directions for the Br-derivative. The unit cell (P21) and its related symmetry elements are coloured red. The macrobonds, labelled A-E are represented by dashed lines and circles for bonds parallel and perpendicular to the projection plane respectively. The colouring scheme of the ribbon model of Rng2 follows figure 8(a). (a2) Same as (a1) for the Hg-derivative. The MAD and SAD data sets for the Br- and Hg-derivatives respectively were independently solved, accounting for the different origins which are related by half a unit translation along the a axis. Blue arrows point at the loop between residues 92-99 of Rng2.

65

Figure 10(b)-(d): The 2mFo - DFc map is contoured at 1σ and superimposed on the coordinates. Grey and green carbons indicate separate molecules. Amino acid codes are shown in lower case. The chain number is in upper case next to the residue number. (b1) Chemical environment at the interface of two molecules along macrobond A of the Br-derivative. There is a potential salt bridge and a potential water-mediated hydrogen bond between the ζ-amino group of Arg70 and the γ-carboxyl group of Asp151 The ζ-amino group of Arg70 is linked via a potential salt bridge to the δ-carbonyl group of Gln67.

66

Figure 10(b2): Macrobond A of the Br-derivative. A bromide anion with a temperature factor of 24.54 Å2 forms a potential hydrogen bond to the ζ-amino group of Arg175, which is doubly hydrogen bonded to the carbonyl group of Leu152.

67

Figure 10(b3): Macrobond A of the Br-derivative. There is a nonpolar cluster (>5 Å) of residues Ile172, Leu152, Phe82, Phe110 and Ile114 in a background of specific intermolecular hydrogen bonds and salt bridges.

68

Figure 10(c): Macrobond C of the Br-derivative. There is a potential direct hydrogen bond between the ζ-amino group of Arg99 and the δ-carbonyl group of Glu161. It also has another potential direct hydrogen bond between the α-imino of Tyr92 and γ-carbonyl group of Asn162.

69

Figure 10(d): Macrobond D of the Br-derivative. It has a potential water-mediated hydrogen bond from the γ-carbonyl group of Asn162 to the ε-amino group of Lys131.

70

Figure 10(e): Macrobond E of the Br-derivative. It has a potential bromide-mediated hydrogen bond between the α-imino group of Pro117 and the ζ-amino group of Arg178.

71

Comparison of Br- and Hg-derivatives of Rng232-189

The Br- and Hg-derivatives of Rng21-189 are somewhat different, with a r.m.s.d.

of 2.14 Å between all 159 atoms and 1.37 Å between the Cα atoms. This is mainly

owing to residues Tyr92 to Arg99, which form a loop that points away from helix a2 in

the Hg-derivative, but points towards helix a2 in the Br-derivative (Figure 11a). The

r.m.s.d. between Cα atoms dropped to 0.52 Å after the loop was omitted from the

calculation. The largest deviation among the backbone atoms was seen at Leu96 with

an overall r.m.s.d. of 11.53 Å. This contrasts with a comparison of the orthorhombic

and tetragonal crystal forms of lysozyme, which have r.m.s.d.s that do not exceed 2 Å

(Oki et al., 1999).

There are only three cysteine residues stretching between residues 32-189 of

Rng2. In the Hg-derivative, they are all found to be associated with a Hg atom within

3.2 Å. They have occupancies between 45-64%, low temperature factors not exceeding

9.51 Å2 and show strong density exceeding 5σ in the 2mFo - DFc map (Read, 1986)

(Figure 11b). The fourth Hg atom (Figure 11b, Hg194) is responsible for the

maintaining the loop in an 'open' conformation relative to the Br-derivative. The

residues surrounding the fourth Hg atom are Asn71 (4.36 Å relative to Cβ), Phe98

(3.69 Å to Cζ) and His100 (2.66 Å to Nε). The bond length between Nε of His100 and

Hg194 is short (2.66 Å), suggesting the presence of a metal-ligand bond in some of the

protein molecules in the crystal. This is consistent with Hg194 having a low occupancy

of 30%, a high temperature factor of 35.7 Å2, together with weak electron density not

exceeding 2σ. The Hg anomalous map (Figure 11b, red density) shows a weak peak

and confirms its presence.

In order to compare the crystal structure of both derivatives, molecular

dynamics (MD) simulation was performed for the aqueous solution system of Rng232-

72

189 in the absence of Br and Hg atoms. After the initial relaxation procedure for 45 ps,

the system representing the conformation seen in the Hg-derivative reached a

stationary state within 50 ps after starting the MD simulations, whereas the Br-

derivative took about 1.9 ns (Figure 11c). Furthermore, the Hg-derivative had a r.m.s.d.

of 1.5 Å whereas the Br-derivative had a r.m.s.d. of about 2.5 Å over the 5 ns

simulation time. Although both derivatives had some conformational differences in the

loop between residues 92-99, both structures remained stable and retained their

differences around residues 92-99 throughout the simulation. The average r.m.s.d. of

each residue over the 5 ns simulation relative to the starting point is presented in

Figure 11(d). The loops clearly had more motion than the α-helices. The largest

motions are seen in the loop between Tyr92-Arg99 of the Br-derivative of Rng2,

whereas the Hg-derivative remained relatively stable. The higher r.m.s.d. seen in the

Br-derivative relative to the Hg-derivative is consistent both with the slightly higher

temperature factors (~ 32-51 Å2) at Asn94 of the Br-derivative, compared to the Hg-

derivative (~ 19-24 Å2); and with funnel diagrams depicting the energetic drive to the

native state (Pande et al., 1998). The crystallographic and MD data both suggest that

residues 92-99 switches between the Br- and Hg-derivative conformations, and that

Hg194 had trapped it in the alternative conformation. The thirty-nine residues (150-

189) carboxyl-terminal to the CH domain maintained its conformation throughout the

simulation. The thirty-one residues at the amino-terminus are disordered and were not

included in the simulation.

73

Figure 11: Comparison between Br- and Hg-derivatives of Rng232-189. (a) Ribbon representation of the Br-derivative (red helices and blue coils) superimposed on the Hg-derivative (pink helices and cyan coils). Hg atoms are shown as spheres coloured according to their B factors.

74

Figure 11(b): Environment around the Hg atom near the variable loop from residues Tyr92-Arg99 in the Hg-derivative. The 2mFo - DFc map is shown contoured at 1σ. The Hg anomalous map (red) is superimposed onto the electron density map.

75

Figure 11(c): Mass weighted Cα r.m.s.d. of Rng232-189 as a function of time over the entire trajectory of a 5 ns molecular dynamics simulation for Rng232-189, in the Br-derivative (red) and Hg-derivative (black).

76

Figure 11(d): Main chain (N, Cα, C and O) r.m.s.d. for each residue as a function of time during the same MD. α-Helical sections of the structure are highlighted in green.

77

Calponin-homology domain phylogeny

A phylogenetic tree of 117 CH domains shared by 73 proteins from human,

chicken, mouse, rat, turkey, slime mold, budding yeast, fission yeast and Arabidopsis

was calculated from a consensus of several tree-building methods (Figure 12). The tree

was rooted on the basis of twelve groups of CH domains and is similar to Elena

Korenbaum and Rivero (2002). However, they reported the presence of a novel group

of CH domains (CHc) found in carnitine palmitoyltransferases (CPT) and carnitine

acetyltransferases (CAT) which could not be verified using either BLAST (Altschul et

al., 1990) or hmmsearch (Eddy, 1998) with a CH HMM (Hidden Markov model)

constructed from my multiple sequence alignment of CH domains. Furthermore, there

seems to be no biological basis for the presence of a CH domain in either CAT or CPT

and they were omitted in my analysis. The CH domains from ASPM, NICAL, NUMA

and SPTBN5, which were omitted from previous phylogenetic analyses (Banuelos et

al., 1998; Stradal et al., 1998; Gimona et al., 2002; Korenbaum and Rivero, 2002),

each formed a separate group.

For those proteins that have tandem CH domains (ABDs), the amino-terminal

CH1 domains are more similar to each other than the carboxyl-terminal CH2 domains.

The CHf1 (CH1.1) and CHf3 (CH2.1) domains of fimbrin clustered close to the F-

actin binding CH1 domains. This agrees with data from a genetic analysis of

Saccharomyces cerevisiae Sac6 which provided indirect evidence for the F-actin

binding ability of fimbrin ABD1 and ABD2 (Brower et al., 1995). Similarly, direct

evidence from experiments on AtFim1 showed that the ABD2 binds F-actin (Nakano

et al., 2001). Thus it makes sense that the CHf1 and CHf3 domain of fimbrins

clustered close to the CH1 domains. The CHf2 (CH1.2) domain of fimbrins grouped

with the CH2 domains, whereas the CHf4 (CH2.2) domains grouped close to the

78

79

microtubule-binding MAPRE CH domains (CHe) (Hayashi and Ikura, 2003). This

agrees with experiments showing that CH2 domains generally bind F-actin weakly

(Corrado et al., 1994; Winder et al., 1995).

Most of the proteins that contain a single CH domain cluster together (Figure

12, CH3). In contrast to Gimona et al., 2002, no conserved residues were observed

although an alternating pattern of hydrophobic and polar/ionic residues was present

(Figure 7a). The Rng2 CH domain grouped together with the F-actin binding IQGAP

CH domains (Fukata et al., 1997) in a separate branch distinct from calponin, VAV,

SM22/TAGLN, smoothelin/SMTN and ARHGEF CH domains that do not bind F-

actin. Similarly, residues 1-200 of the human Gas2-related gene on chromosome 22

(hGAR22), which formed a separate branch, contains a CH domain and cosediments

with F-actin (Goriounov et al., 2003). The fact that F-actin binding CH domains from

IQGAP and Gas2 proteins formed two separate branches can be rationalized from the

ability of the CH domain from IQGAP to bind both F-actin and calmodulin (Ho et al.,

1999), whereas the equivalent domain from the Gas2 orthologue hGAR22 only binds

F-actin (Goriounov et al., 2003). The phylogenetic analysis shows that proteins with

single CH domains are a heterogenous group (CH3) with different ligands and

biological functions.

80

Figure 12: Consensus of phylogenies of 117 individual calponin-homology (CH) sequences from 73 proteins. The species represented include Arabidopsis thaliana (at), Dictyostelium discoideum (dd), Gallus gallus (gg), Homo sapiens (hs), Meleagris gallopavo (mg), Mus musculus (mm), Rattus norvegicus (rn), Saccharomyces cerevisiae (sc) and Schizosaccharomyces pombe (sp). The number appended to the species symbol represents the order of the CH domain with 1, 3 being amino-terminal and 2, 4 being carboxyl-terminal. The CH groups were designated following Korenbaum and Rivero (2002). Rng2 (Rng2_sp) is highlighted in dark grey at the top of the figure. The protein codes for the 9 species, together with their accession numbers, are as follows: Rng2_sp (NP_593860); Iqg1_sc (NP_015082); IQGAP1_hs (NP_003861); IQGAP2 (NP_006624); CNN_sp (NP_593863); CNN1_hs (NP_001290); 1h67_gg (P26932); CNN_mg (P37803); CNN3_hs (NP_001830); CNN2_hs (NP_004359); ARHGEF7b_hs (NP_663788); ARHGEF6_hs (NP_004831); LRRN4_hs (NP_002310); CHDC1_hs (XP_166260); LMO7_hs (NP_005349); VAV1_hs (NP_005419); VAV3_hs (NP_006104); VAV2_hs (NP_003362); TAGLN_hs (NP_003177); NP25_hs (NP_037391); TAGLN2_hs (NP_003555); Scp1_sc (NP_015012); GAS2L1_hs (NP_006469); GAR17_hs (NP_644814); GAS2_hs (NP_005247); ASPM_hs (NP_060606); SMTNb_hs (NP_599031); SMTNa_hs (NP_599032); SMTNc_hs (NP_008863); SMTN_mm (NP_038898); MICAL2_hs (NP_055447); MICAL3_hs (XP_032997); TANGN_hs (XP_170658); MIRAB13_hs (NP_203744); NICAL_hs (NP_073602); NUMA1_hs (NP_006176); 1pa7_hs (NP_036457); MAPRE3_hs (NP_036458); MAPRE2_hs (NP_055083); Sac6_sc1 (NP_010414); Fim1_sp1 (NP_596289); PLS1_hs1 (NP_002661); 1aoa1_hs1 (NP_005023); LCP1_hs1 (NP_002289); AtFim1_at1 (NP_200351); ACTN3_hs1 (NP_001095); ACTN4_hs1 (NP_004915); ACTN1_hs1 (NP_001093); ACTN2_hs1 (NP_001094); FLNB_hs1 (NP_001448); FLNA_hs1 (NP_001447); FLNC_hs1 (NP_001449); FLN_gg1 (BAB63943); ctxA_dd1 (AAB62275); abpC_dd1 (P13466); SYNE1_hs1 (NP_149062); CLMN_hs1 (NP_079010); SYNE2_hs1 (NP_055995); 1dxx1_hs1 (NP_003997); 1qag1_hs1 (NP_009055); MACF1a_hs1 (NP_036222); PLEC_rn1 (NP_071796); 1mb81_hs1 (NP_000436); SPTBN1_hs1 (NP_003119); SPTB_hs1 (NP_000338); SPTBN2_hs1 (NP_008877); SPTBN4_hs1 (NP_079489); SPTBN5_hs1 (NP_057726); NAV2_hs (NP_660093); NAV3_hs (NP_055718); PARVG_hs1 (NP_071424); PARVA_hs1 (NP_060692); PARVB_hs1 (NP_037459).

81

82

Concluding remarks: Rng2 structure and function

Rng2 and its mammalian IQGAP orthologues are multifunctional, multidomain

proteins that link the signal transduction pathways with cytoskeletal reorganization

through many ligands, such as F-actin, EF-hand proteins like calmodulin and Cdc4 and

the GTPase Cdc42 (Eng et al., 1998; Ho et al., 1999; Fukata et al., 1997; D’souza et

al., 2001; Epp et al., 1997; McCallum et al., 1996). The structural and phylogenetic

evidence presented here indicates that Rng2 CH domain is most similar to the F-actin

binding CH1 domain of ABD-containing proteins (Table 5, Figures 7, 8 and 12).

Coupled with the in vivo localisation of GFP-Rng21-189 at the same site as cortical actin

patches (Wachtler et al., 2003), it is likely that residues 1-189 of Rng2 are involved in

binding F-actin structures.

The Rng232-189 structure has two flexible regions: residues 1-31, which are

disordered in the crystal structure; and residues 92-99, which switches between two

stable conformations (Figure 11A). Flexible loops are often found at the interface

between a protein and its respective ligand, which can be a small molecule, nucleic

acid or another protein. In the case of Rng2 with F-actin being the in vivo ligand, it is

possible that a molecular surface inclusive of the two flexible regions would represent

the area where they interact. In this scenario, residues 1-31 might become ordered

upon binding F-actin.

Residues 150-189 of Rng2 have no known homologue or orthologue and adopt

a conformation that resembles neither the extended conformation seen in utrophin, nor

the compact conformation seen in fimbrin, although residues 154-160 forms an

unstructured coil which adopts a substructure similar to dystrophin residues 240-246 in

the carboxyl-terminal portion of the CH2 domain. Residues 150-189 wrap around the

proposed actin-binding site ABS1 and ABS2 (Corrado et al., 1994; Norwood et al.

2000) (Figure 2A), suggesting the putative ABS1 and ABS2 in Rng2 are not involved

in binding F-actin. The Rng2 CH domain within residues 43-148 is also unlikely to

bind F-actin independently, since data from the single CH domain of calponin

expressed without its neighbouring residues showed a lack of F-actin binding capacity

(Gimona and Mital, 1998; Mino et al., 1998; Table 6). In the absence of similar data

from the IQGAP family of proteins, it seems likely that the stretch of residues from

150 to 189 of Rng2 may be important for the F-actin binding ability of Rng21-189. The

flexible loop at residues 92-99 and the α-helical stretch within residues 150-189 face

the same direction, potentially defining the molecular surface where Rng232-189

interacts with F-actin. To test its biological importance, mutations to four residues

conserved within the IQGAP family of proteins (Figure 7c) can be generated and

tested for their ability to disrupt Rng2 localisation to the actomyosin ring - L96D

within the flexible loop, R177D within α-helix a7, R70E and K131E within the basic

patch (Figure 13).

Figure 13: Residues shown in stick representation on the ribbon model of Rng232-189 are conserved within the IQGAP family of proteins (Figure 7c). The transparent surface electrostatic potential highlight the presence of a basic patch where Rng232-189 possibly interacts with F-actin.

83

84

A possible model of Rng2 function derived from currently available evidence

has at least one actin-binding domain within residues 1-189. Previous studies on

calponin (Gimona and Mital, 1998; Gimona and Winder, 1998; Mino et al., 1998) and

SM22 (Fu et al., 2000) showed that residues C-terminal to the CH domain were

essential for their localization to F-actin (Table 6). This suggests the presence of

another actin-binding region downstream of the CH domain. A possible location

comes from the following data. Firstly, the EF-hand proteins Cdc4 (D’souza et al.,

2001) and Cam1 (Eng et al., 1998), have been shown to physically interact with Rng2.

In the case of Mlc1, which is another EF-hand protein, crystal structures of complexes

of Mlc1 with IQ motifs from Myo2 and Iqg1 and known (Terrak et al., 2003). Thus,

the second actin-binding region probably resides within the IQ repeats, and should act

in an indirect fashion via Cdc4 or Cam1. Oligomerisation of Rng2 via its downstream

coiled coil domain might result in the generation of an F-actin binding interface via the

dimerisation of its CH domain. Once the medial F-actin spot has formed during mitosis,

the carboxyl-terminal RasGAP domain potentially initiates the assembly of a broad

actomyosin ring through as yet unidentified effectors. Rng2 CH domain probably

serves as vehicle to bring its carboxy-terminal RasGAP domain to the actomyosin ring

during cell division, and activate its signalling function during cytokinesis.

Appendix

Table 6: Biochemical characteristics of several CH domains. Structures deposited in RCSB are in square brackets. CH domains pI [RCSB]

Binding constants (Kd) F-actin bundling Types of interaction Actin-binding sites (ABS)

Mutants (residues changed)

ACTN CH1 (gg) 9.78 ACTN CH2 (gg) 6.07

120-134 Kuhlman et al., 1992

ACTN1 CH1 9.78 ACTN1 CH2 6.07

actinin 1-269 4 µM Way et al., 1992 actinin 25-257 3 µM Leinweber et al., 1999

ABD binds actin hydrophobically. Kuhlman et al., 1992 actinin 25-257 binds ERK Leinweber et al., 1999

ACTN2 CH1 9.86 ACTN2 CH2 5.92

ACTN3 CH1 9.93 ACTN3 CH2 6.07

ACTN4 CH1 9.78 ACTN4 CH2 5.78

mutants had increased F-actin affinity. Kaplan et al., 2001

228, 232, 235Kaplan et al., 2001

85

86

CH domains pI [RCSB]

Binding constants (Kd) F-actin bundling Types of interaction Actin-binding sites (ABS)

Mutants (residues changed)

FLN CH1 (gg) 9.83 FLN CH2 (gg) 4.73

hydrophobic and electrostatic. Lebart et al., 1994

121-147 Lebart et al., 1994

FLNA CH1 9.76 FLNA CH2 4.93

1-1761 (187kD) 3µM Gorlin et al., 1990

filamin C-terminal (2551-2647) needed to crosslink actin. Gorlin et al., 1990

170, 172, 196, 200, 207, 254, 273 Robertson et al., 2003

FLNB CH1 9.69 FLNB CH2 4.77

FLNC CH1 9.86 FLNC CH2 4.53

SPTB 9.8 SPTB CH2 6.04

47-186 5 µM Karinch et al., 1990

182, 220Hassoun et al., 1997

87

CH domains pI [RCSB]

Binding constants (Kd) F-actin bundling Types of interaction Actin-binding sites (ABS)

Mutants (residues changed)

SPTBN1 CH1 9.93 SPTBN1 CH2 7.22 [1AA2/1BKR]

beta-spectrin 1-313 26 µM Li and Bennett, 1996 beta-spectrin 1-281 20 µM CH2 (173-182) does not bind actin. Banuelos et al., 1998 minispectrin 2.5 µM Raae et al., 2003

SPTBN2 CH1 9.75 SPTBN2 CH2 6.56

SPTBN4 CH1 9.73 SPTBN4 CH2 7.09

SPTBN5 CH1 9.51 SPTBN5 CH2 5.13

88

CH domains pI [RCSB]

Binding constants (Kd) F-actin bundling Types of interaction Actin-binding sites (ABS)

Mutants (residues changed)

DMD CH1 9.63 [1DXX] DMD CH2 [1DXX] 5.84

dystrophin 2-90 100 nM dystrophin 2-246 150 nM Corrado et al., 1994 dystrophin 2-246 13.7 µM Renley et al., 1998

ABD (1-337) can form dimers and bundle F-actin.Fontao et al., 2001

dystrophin 1-246 binds actin hydrophobically. Amann et al., 1998

dystrophin 2-90 Corrado et al., 1994

UTRN CH1 9.70 [1QAG] UTRN CH2 8.22 [1BHD/1QAG]

utrophin 28-261 67 µM Keep et al., 1999a utrophin 1-261 19 µM utrophin 113-371 275 µM Winder et al., 1995

ABD (28-261) is a monomer in solution. Keep, Winder et al., 1999 Unable to crosslink actin. Winder et al., 1995

utrophin 1-261 binds actin hydrophobically. Winder et al., 1995

PLEC1 CH1 10.06 [1MB8] PLEC1 CH2 [1MB8] 4.95

plectin 1-339 0.3 µM Fontao et al., 2001 plectin 59-293 22.3 µM (open) 59.3 µM (closed) Garcia-Alvarez et al., 2003

ABD (1-339) can form dimers and bundle F-actin.Fontao et al., 2001

plectin 65-302 binds β-4 integrin. Geerts et al., 1999

plectin CH1 10.06 plectin CH2 (rn) 4.95

plectin 1-1129 0.32 µM Andra et al., 1998

89

CH domains pI [RCSB]

Binding constants (Kd) F-actin bundling Types of interaction Actin-binding sites (ABS)

Mutants (residues changed)

BPAG1n CH1 10.13 BPAG1n CH2 4.68

full length BPAG1n 0.21 µM Yang et al., 1996

ABD (1-336) can form dimers and bundle F-actin.Fontao et al., 2001

MACF1a CH1 10.43 MACF1a CH2 4.82

MACF1 ABD (74-293)Leung et al., 1999

SYNE1 CH1 10.14 SYNE1 CH2 6.15

SYNE2 CH1 8.76 SYNE2 CH2 6.14

SYNE2 1-285 3.8 µM Zhen et al., 2002

ABD (1-285) bundles F-actin. Zhen et al., 2002

ABP-120 CH1 (dd) 9.40 ABP-120 CH2 4.31

ABP-120 89-115 Bresnick et al., 1991

PARVA CH1 5.36 PARVA CH2 5.63

full length alpha-parvin. 8.4 µM Olski et al., 2001

actopaxin 223-372 (CH2 domain) binds paxillin LD1 and LD4 motifs. Full length actopaxin binds actin. Nikolopoulos and Turner, 2000

PARVB CH1 6.08 PARVB CH2 6.39

90

CH domains pI [RCSB]

Binding constants (Kd) F-actin bundling Types of interaction Actin-binding sites (ABS)

Mutants (residues changed)

PARVG CH1 5.47 PARVG CH2 5.00

cortexillin I CH1 9.30 cortexillin I CH2 4.74

cortexillin C-terminal(422-444) bundles F-actin.Stock et al., 1999

cortexillin I (1-233)binds F-actin very weakly. Stock et al., 1999

PLS1 CH1 5.08 PLS1 CH2 5.85

PLS1 CH3 9.37 PLS1 CH4 8.74

LCP1 CH1 5.06 LCP1 CH2 5.34

L-plastin binds unpolymerized vimentin subunits. 0.25 µM Correia et al., 1999

L-plastin 143-188 (CH1 domain) binds vimentin. Correia et al., 1999

L-plastin binds vimentin hydrophobically. Correia et al., 1999

LCP1 CH3 9.72

PLS3 CH1 5.53 [1AOA] PLS3 CH2 6.12 [1AOA]

PLS3 CH3 9.37 PLS3 CH4 5.75

91

CH domains pI [RCSB]

Binding constants (Kd) F-actin bundling Types of interaction Actin-binding sites (ABS)

Mutants (residues changed)

Fim1 CH1 5.17 Fim1 CH2 9.07

fimbrin 1-429 showsweak binding. Nakano et al., 2001

Fim1 CH3 6.87 Fim1 CH4 6.94

fimbrin 369-614 shows strong binding. Nakano et al., 2001

AtFim1 CH1 (at) 5.87 AtFim1 CH2 6.08

full length fimbrin 0.45 µM (high affinity) 1.9 µM (low affinity) Kovar et al., 2000

AtFim1 CH3 (at) 9.47 AtFim1 CH4 8.80

Sac6 CH1 (sc) 4.86 Sac6 CH2 8.68

electrostaticHonts et al., 1994

fimbrin 140, 145, 252, 259, 264, 268, 379, 380, 383 Brower et al., 1995

Sac6 CH3 (sc) 5.81 Sac6 CH4 5.73

414, 484, 560, 610 Brower et al., 1995

Rng2 CH (sp) 5.75 [1P2X/1P5S]

Rng2 1-189Wachtler et al., 2003

92

CH domains pI [RCSB]

Binding constants (Kd) F-actin bundling Types of interaction Actin-binding sites (ABS)

Mutants (residues changed)

IQGAP1 CH 6.47

IQGAP1 918-1657 binds Cdc42. McCallum et al., 1996 IQGAP1 740-869 binds calmodulin. Ho et al., 1999

IQGAP1 1-216 binds F-actin Fukata et al., 1997 IQGAP1 1-232 binds F-actin and calmodulin. Ho et al., 1999

IQGAP2 CH 7.89

GAR22 CH 4.63

GAR22 1-200 Goriounov et al., 2003

MAPRE1 CH 9.33 [1PA7/1UEG]

MAPRE1 binds microtubules. Hayashi and Ikura, 2003

MAPRE1 1-133 Bu and Su, 2003

93

CH domains pI [RCSB]

Binding constants (Kd) F-actin bundling Types of interaction Actin-binding sites (ABS)

Mutants (residues changed)

basic calponin CH (gg) 6.07 [1H67]

full length calponin binds ERK6.4 µM Leinweber et al., 1999 full length calponin binds desmin tetramers 3-15 µM Wang et al., 1996 Mabuchi et al., 1997 full length calponin binds the regulatory light chain of myosin. 4.4 µM Szymanski and Goyal, 1999 calponin 2-51 binds calmodulin. 1.8 µM calponin 2-51 binds caltropin. 1.2 µM calponin 45-228 binds calmodulin and caltropin. 0.07 µM Wills et al., 1994 calponin 145-182 binds myosin HMM - 4 µM; S2 - 4.7 µM; LMM - 2.6 µM Szymanski and Tao, 1997

calponin binds F-actin through electrostatic interactions Tang et al., 1997 calponin 7-182 binds desmin. Wang et al., 1996

94

CH domains pI [RCSB]

Binding constants (Kd) F-actin bundling Types of interaction Actin-binding sites (ABS)

Mutants (residues changed)

calponin CH (mg)7.46

calponin 7-91 binds ERK Leinweber et al., 1999 calponin 145-163 binds F-actin. calponin 153-163 binds calmodulin caltropin and tropomyosin. Mezgueldi et al., 1995

acidic calponin CH (rn) 8.83

full length calponin binds microtubules 0.3 µM Fattoum et al., 2003

calponin 145-182 and 183-292 binds microtubules. Fattoum et al., 2003

CNN1 CH 6.46

full length calponin 0.24 µM Mino et al., 1998

calponin 141-160 and172-187 Mino et al., 1998

CNN1 CH (mm) calponin 164-273 Gimona et al., 1998

CNN2 CH (mm) calponin 166-275 Gimona et al., 1998

Scp1 CH (sp) 5.37

full length Scp1 0.7 µM Goodman et al., 2003

Scp1 136-200 crosslinks actin via two distinct sites.Goodman et al., 2003

SM22/TAGLN 4.78

170-186 Fu et al., 2000

SMTN (mm) 5.89

1-134535-682 Quensel et al., 2002

ARHGEF6 4.59

ARHGEF6 56-83interacts with PARVB. Rosenberger et al., 2003

95

CH domains pI [RCSB]

Binding constants (Kd) F-actin bundling Types of interaction Actin-binding sites (ABS)

Mutants (residues changed)

VAV 7.59

Vav 1-144interacts with zinc-finger (ZF) domain of VAV. Zugaza et al., 2002

References

Abrahams, J.P. and Leslie, A.G.W. (1996). Methods used in the structure determination of bovine mitochondrial F1 ATPase. Acta Crystallogr. D52, 30-42. Altschul, S.F., Gish, W., Miller, W., Myers, E.W. and Lipman, D.J. (1990). Basic local alignment search tool. J. Mol. Biol. 215, 403-410. Amann, K.J., Renley, B.A. and Ervasti, J.M. (1998). A cluster of basic repeats in the dystrophin rod domain binds F-actin through an electrostatic interaction. J. Biol Chem. 273, 28419-28423. Andra, K., Nikolic, B., Stocher, M., Drenckhahn, D. and Wiche, G. (1998). Not just scaffolding: plectin regulates actin dynamics in cultured cells. Genes Dev. 12, 3442-3451. Arai, R. and Mabuchi, I. (2002). F-actin ring formation and the role of F-actin cables in the fission yeast Schizosaccharomyces pombe. J. Cell Sci. 115, 887-898. Bahler, J. and Nurse, P. (2001). Fission yeast Pom1p kinase activity is cell cycle regulated and essential for cellular symmetry during growth and division. EMBO J. 20, 1064-1073. Bahler, J., Steever, A.B., Wheatley, S., Wang, Y., Pringle, J.R., Gould, K.L. and McCollum, D. (1998). Role of polo kinase and Mid1p in determining the site of cell division in fission yeast. J. Cell Biol. 143, 1603-1616. Banuelos, S., Saraste, M. and Carugo, K.D. (1998). Structural comparisons of calponin homology domains: implications for actin binding. Structure 6, 1419-1431. Bashour, A.M., Fullerton, A.T., Hart, M.J. and Bloom, G.S. (1997). IQGAP1, a Rac- and Cdc42-binding protein, directly binds and cross-links microfilaments. J. Cell Biol. 137, 1555-1566. Bateman, A., Birney, E., Cerruti, L., Durbin, R., Etwiller, L., Eddy, S.R., Griffiths-Jones, S., Howe, K.L., Marshall, M. and Sonnhammer, E.L. (2002). The Pfam Protein Families Database. Nucleic Acids Res. 30, 276-280. Berendsen, H.J.C., Postma, J.P.M., van Gunsteren, W.F., DiNola, A. and Haak, J.R. (1984). Molecular dynamics with coupling to an external bath. J. Chem. Phys. 81, 3684-3690. Berlin, A., Paoletti, A. and Chang, F. (2003). Mid2p stabilizes septin rings during cytokinesis in fission yeast. J. Cell Biol. 160, 1083-1092. Bezanilla, M., Wilson, J.M. and Pollard, T.D. (2000). Fission yeast myosin-II isoforms assemble into contractile rings at distinct times during mitosis. Curr. Biol. 10, 397-400.

96

Bezanilla, M. and Pollard, T.D. (2000). Myosin-II tails confer unique functions in Schizosaccharomyces pombe: characterization of a novel myosin-II tail. Mol. Biol. Cell. 11, 79-91. Bramham, J., Hodgkinson, J.L., Smith, B.O., Uhrin, D., Barlow, P.N. and Winder, S.J. (2002). Solution structure of the calponin CH domain and fitting to the 3D-helical reconstruction of F-actin:calponin. Structure 10, 249-258. Bresnick, A.R., Janmey, P.A. and Condeelis, J. (1991). Evidence that a 27-residue sequence is the actin-binding site of ABP-120. J. Biol. Chem. 266, 12989-12993. Brodersen, D.E., de La Fortelle, E., Vonrhein, C., Bricogne, G., Nyborg, J. and Kjeldgaard, M. (2000). Applications of single-wavelength anomalous dispersion at high and atomic resolution. Acta Crystallogr. D56, 431-441. Brower, S.M., Honts, J.E. and Adams, A.E. (1995). Genetic analysis of the fimbrin-actin binding interaction in Saccharomyces cerevisiae. Genetics 140, 91-101. Brünger, A.T., Adams, P.D., Clore, G.M., DeLano, W.L., Gros, P., Grosse-Kunstleve, R.W., Jiang, J.S., Kuszewski, J., Nilges, M., Pannu, N.S., Read, R.J., Rice, L.M., Simonson, T. and Warren, G.L. (1998). Crystallography & NMR system: A new software suite for macromolecular structure determination. Acta Crystallogr. D54, 905-921. Bu, W. and Su, L.K. (2003). Characterization of functional domains of human EB1 family proteins. J. Biol. Chem. 278, 49721-49731. Carugo, K.D., Banuelos, S. and Saraste, M. (1997). Crystal structure of a calponin homology domain. Nature Struct. Biol. 4, 175-179. Case, D.A., Pearlman, D.A., Caldwell, J.W., Cheatham III, T.E., Wang, J., Ross, W.S., Simmerling, C.L., Darden, T.A., Merz, K.M., Stanton, R.V., Cheng, A.L., Vincent, J.J., Crowley, M., Tsui, V., Gohlke, H., Radmer, R.J., Duan, Y., Pitera, J., Massova, I., Seibel, G.L., Singh, U.C., Weiner, P.K., and Kollman, P.A. (2002). AMBER 7 (San Francisco: University of California). Castresana, J. and Saraste, M. (1995). Does Vav bind to F-actin through a CH domain? FEBS Lett. 374, 149-151. Cerutti, L. and Simanis, V. (1999). Asymmetry of the spindle pole bodies and spg1p GAP segregation during mitosis in fission yeast. J. Cell Sci. 112, 2313-2321. Chang, F. (1999). Movement of a cytokinesis factor cdc12p to the site of cell division. Curr. Biol. 9, 849-852. Chang, F., Drubin, D. and Nurse, P. (1997). cdc12p, a protein required for cytokinesis in fission yeast, is a component of the cell division ring and interacts with profilin. J. Cell Biol. 137, 169-182.

97

Corrado, K., Mills, P.L. and Chamberlain, J.S. (1994). Deletion analysis of the dystrophin-actin binding domain. FEBS Lett. 344, 255-260. Correia, I., Chu, D., Chou, Y.H., Goldman, R.D. and Matsudaira, P. (1999). Integrating the actin and vimentin cytoskeletons: Adhesion-dependent formation of fimbrin-vimentin complexes in macrophages. J. Cell Biol. 146, 831-842. Cortes, J.C., Ishiguro, J., Duran, A. and Ribas, J.C. (2002). Localization of the (1,3)beta-D-glucan synthase catalytic subunit homologue Bgs1p/Cps1p from fission yeast suggests that it is involved in septation, polarized growth, mating, spore wall formation and spore germination. J. Cell Sci. 115, 4081-4096. Cowtan, K. (1999). Error estimation and bias correction in phase-improvement calculations. Acta Crystallogr. D55, 1555-1567. D'souza, V.M., Naqvi, N.I., Wang, H. and Balasubramanian, M.K. (2001). Interactions of Cdc4p, a myosin light chain, with IQ-domain containing proteins in Schizosaccharomyces pombe. Cell Struct. Funct. 26, 555-565. Darden, T., York, D. and Pedersen, L. (1993). Particle mesh Ewald: An N·log(N) method for Ewald sums in large systems. J. Chem. Phys. 98, 10089-10092. Dauter, Z., Li, M. and Wlodawer, A. (2001). Practical experience with the use of halides for phasing macromolecular structures: a powerful tool for structural genomics. Acta Crystallogr. D57, 239-249. Dauter, Z. and Adamiak, D.A. (2001). Anomalous signal of phosphorus used for phasing DNA oligomer: importance of data redundancy. Acta Crystallogr. D57, 990-995. Demeter, J. and Sazer, S. (1998). imp2, a new component of the actin ring in the fission yeast Schizosaccharomyces pombe. J. Cell Biol. 143, 415-427. Diederichs, K. and Karplus, P.A. (1997). Improved R-factors for diffraction data analysis in macromolecular crystallography. Nature Struct. Biol. 4, 269-275. Duke, R.E. and Pedersen, L.G. (2003). PMEMD 3 (University of North Carolina-Chapel Hill). Dunphy, W.G. and Kumagai, A. (1991). The cdc25 protein contains an intrinsic phosphatase activity. Cell 67, 189-196. Eck, R.V. and Dayhoff, M.O. (1966). Atlas of protein sequence and structure (Silverspring, Maryland: Natl Biomed. Res. Foundation). Eddy, S.R. (1998). Profile hidden Markov models. Bioinformatics 14, 755-763. Ekström, F., Stier, G. and Sauer, U.H. (2003). Crystallization of the actin-binding domain of human α-actinin: analysis of microcrystals of SeMet-labelled protein. Acta Crystallogr. D59, 724-726.

98

Eng, K., Naqvi, N.I., Wong, K.C.Y. and Balasubramanian, M.K. (1998). Rng2p, a protein required for cytokinesis in fission yeast, is a component of the actomyosin ring and the spindle pole body. Curr. Biol. 8, 611-621. Epp, J.A. and Chant, J. (1997). An IQGAP-related protein controls actin-ring formation and cytokinesis in yeast. Curr. Biol. 7, 921-929. Fabbrizio, E., Bonet-Kerrache, A., Leger, J.J. and Mornet, D. (1993). Actin-dystrophin interface. Biochemistry 32, 10457-10463. Fantes, P.A. and Nurse, P. (1978). Control of the timing of cell division in fission yeast. Cell size mutants reveal a second control pathway. Exp. Cell Res. 115, 317-329. Fattoum, A., Roustan, C., Smyczynski, C., Der Terrossian, E. and Kassab, R. (2003). Mapping the microtubule binding regions of calponin. Biochemistry 42, 1274-1282 Feig, M., Karanicolas, J. and Brooks III, C.L. (2001). MMTSB Tool Set, MMTSB (NIH Research Resource, The Scripps Research Institute). Felsenstein, J. (1989). PHYLIP -- Phylogeny Inference Package (Version 3.2). Cladistics 5, 164-166. Felsenstein, J. (1981). Evolutionary trees from DNA sequences: a maximum likelihood approach. J. Mol. Evol. 17, 368-376. Fenn, T.D., Ringe, D. and Petsko, G.A. (2003). POVScript+: a program for model and data visualization using persistence of vision ray-tracing. J. Appl. Cryst. 36, 944-947. Fitzpatrick, P.A., Ringe, D. and Klibanov, A.M. (1994). X-ray crystal structure of cross-linked subtilisin Carlsberg in water vs. acetonitrile. Biochem Biophys Res Commun. 198, 675-681. Fitzpatrick, P.A., Steinmetz, A.C., Ringe, D. and Klibanov, A.M. (1993). Enzyme crystal structure in a neat organic solvent. Proc. Natl Acad. Sci. U S A. 90, 8653-8657. Fontao, L., Geerts, D., Kuikman, I., Koster, J., Kramer, D. and Sonnenberg, A. (2001). The interaction of plectin with actin: evidence for cross-linking of actin filaments by dimerization of the actin-binding domain of plectin. J. Cell Sci. 114, 2065-2076. Fu, Y., Liu, H.W., Forsythe, S.M., Kogut, P., McConville, J.F., Halayko, A.J., Camoretti-Mercado, B. and Solway, J. (2000). Mutagenesis analysis of human SM22: characterization of actin binding. J. Appl. Physiol. 89, 1985-1990. Fukata, M., Kuroda, S., Fujii, K., Nakamura, T., Shoji, I., Matsuura, Y., Okawa, K., Iwamatsu, A., Kikuchi, A. and Kaibuchi, K. (1997). Regulation of cross-linking of actin filament by IQGAP1, a target for Cdc42. J. Biol. Chem. 272, 29579-29583.

99

Galkin, V.E., Orlova, A., VanLoock, M.S., Rybakova, I.N., Ervasti, J.M. and Egelman, E.H. (2002). The utrophin actin-binding domain binds F-actin in two different modes: implications for the spectrin superfamily of proteins. J. Cell Biol. 157, 243-251. García-Alvarez, B., Bobkov, A., Sonnenberg, A. and de Pereda, J.M. (2003). Structural and functional analysis of the actin binding domain of plectin suggests alternative mechanisms for binding to F-actin and integrin beta4. Structure 11, 615-625. Geerts, D., Fontao, L., Nievers, M.G., Schaapveld, R.Q., Purkis, P.E., Wheeler, G.N., Lane, E.B., Leigh, I.M. and Sonnenberg, A. (1999). Binding of integrin alpha6beta4 to plectin prevents plectin association with F-actin but does not interfere with intermediate filament binding. J. Cell Biol. 147, 417-434. Gimona, M., Djinovic-Carugo, K., Kranewitter, W.J. and Winder, S.J. (2002). Functional plasticity of CH domains. FEBS Lett. 513, 98-106. Gimona, M. and Winder, S.J. (1998). Single calponin homology domains are not actin-binding domains. Curr Biol. 8, R674-R675. Gimona, M. and Mital, R. (1998). The single CH domain of calponin is neither sufficient nor necessary for F-actin binding. J. Cell Sci. 111, 1813-1821. Goldsmith, S.C., Pokala, N., Shen, W., Fedorov, A.A., Matsudaira, P. and Almo, S.C. (1997). The structure of an actin-crosslinking domain from human fimbrin. Nature Struct. Biol. 4, 708-712. Gonzalez, A. (2003). Optimizing data collection for structure determination. Acta. Crystallogr. D59, 1935-1942. Goodman, A., Goode, B.L., Matsudaira, P. and Fink, G.R. (2003). The Saccharomyces cerevisiae calponin/transgelin homolog Scp1 functions with fimbrin to regulate stability and organization of the actin cytoskeleton. Mol. Biol. Cell 14, 2617-2629. Goriounov, D., Leung, C.L. and Liem, R.K. (2003). Protein products of human Gas2-related genes on chromosomes 17 and 22 (hGAR17 and hGAR22) associate with both microfilaments and microtubules. J Cell Sci. 116, 1045-1058. Gorlin, J.B., Yamin, R., Egan, S., Stewart, M., Stossel, T.P., Kwiatkowski, D.J. and Hartwig, J.H. (1990). Human endothelial actin-binding protein (ABP-280, nonmuscle filamin): a molecular leaf spring. J. Cell Biol. 111, 1089-1105. Gouet, P., Courcelle, E., Stuart, D.I. and Metoz, F. (1999). ESPript: multiple sequence alignments in PostScript. Bioinformatics 15, 305-308. Guan, K.L. and Dixon, J.E. (1991). Eukaryotic proteins expressed in Escherichia coli: an improved thrombin cleavage and purification procedure of fusion proteins with glutathione S-transferase. Anal. Biochem. 192, 262-267.

100

Guertin, D.A., Trautmann, S. and McCollum, D. (2002). Cytokinesis in eukaryotes. Microbiol. Mol. Biol. Rev. 66, 155-178. Guertin, D.A., Chang, L., Irshad, F., Gould, K.L. and McCollum, D. (2000). The role of the sid1p kinase and cdc14p in regulating the onset of cytokinesis in fission yeast. EMBO J. 19, 1803-1815. Guex, N. and Peitsch, M.C. (1997). SWISS-MODEL and the Swiss-PdbViewer: An environment for comparative protein modeling. Electrophoresis 18, 2714-2723. Hanein, D., Volkmann, N., Goldsmith, S., Michon, A.M., Lehman, W., Craig, R., DeRosier, D., Almo, S. and Matsudaira, P. (1998). An atomic model of fimbrin binding to F-actin and its implications for filament crosslinking and regulation. Nature Struct. Biol. 5, 787-792. Erratum in: (1998). 5, 924. Hanein, D, Matsudaira, P. and DeRosier, D.J. (1997). Evidence for a conformational change in actin induced by fimbrin (N375) binding. J. Cell Biol. 139, 387-396. Hartman, P. (1973). Ch. 14. In Crystal Growth: An Introduction, P. Hartman, ed. (Amsterdam: North-Holland). Hassoun, H., Vassiliadis, J.N., Murray, J., Njolstad, P.R., Rogus, J.J., Ballas, S.K., Schaffer, F., Jarolim, P., Brabec, V. and Palek, J. (1997). Characterization of the underlying molecular defect in hereditary spherocytosis associated with spectrin deficiency. Blood 90, 398-406. Hayashi, I. and Ikura, M. (2003). Crystal structure of the amino-terminal microtubule-binding domain of end-binding protein 1 (EB1). J. Biol Chem. 278, 36430-36434. Ho, Y.D., Joyal, J.L., Li, Z. and Sacks, D.B. (1999). IQGAP1 integrates Ca2+/calmodulin and Cdc42 signalling. J. Biol. Chem. 274, 464-470. Honts, J.E., Sandrock, T.S., Brower, S.M., O'Dell, J.L. and Adams, A.E. (1994). Actin mutations that show suppression with fimbrin mutations identify a likely fimbrin-binding site on actin. J. Cell Biol. 126, 413-422. Hou, M.C., Salek, J. and McCollum, D. (2000). Mob1p interacts with the Sid2p kinase and is required for cytokinesis in fission yeast. Curr. Biol. 10, 619-622. Ippolito, J.A., Alexander, R.S. and Christianson, D.W. (1990). Hydrogen bond stereochemistry in protein structure and function. J. Mol. Biol. 215, 457-471. James, R.W. (1958). The Optical Principles of the Diffraction by X-rays (London: Bell & Sons). Jones, T.A., Zou, J.Y., Cowan, S.W. and Kjeldgaard, M. (1991). Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Cryst. A47, 110-119.

101

Jorgensen, W.L., Chandrasekhar, J., Madura, J. and Klein, M.L. (1983). Comparison of simple potential functions for simulating liquid water. J. Chem. Phys. 79, 926-935. Kabsch, W. and Sander, C. (1983). Dictionary of protein secondary structure: pattern recognition of hydrogen-bonded and geometrical features. Biopolymers 22, 2577-2637. Kaplan, J.M., Kim, S.H., North, K.N., Rennke, H., Correia, L.A., Tong, H.Q., Mathis, B.J., Rodriguez-Perez, J.C., Allen, P.G., Beggs, A.H. and Pollak, M.R. (2001). Mutations in ACTN4, encoding alpha-actinin-4, cause familial focal segmental glomerulosclerosis. Nature Genet. 24, 251-256. Karinch, A.M., Zimmer, W.E. and Goodman, S.R. (1990). The identification and sequence of the actin-binding domain of human red blood cell beta-spectrin. J. Biol. Chem. 265, 11833-11840. Katayama, S., Hirata, D., Arellano, M., Perez, P. and Toda, T. (1999). Fission yeast alpha-glucan synthase Mok1 requires the actin cytoskeleton to localize the sites of growth and plays an essential role in cell morphogenesis downstream of protein kinase C function. J. Cell Biol. 144, 1173-1186. Keep, N.H., Winder, S.J., Moores, C.A., Walke, S., Norwood, F.L. and Kendrick-Jones, J. (1999). Crystal structure of the actin-binding region of utrophin reveals a head-to-tail dimer. Structure 7, 1539-1546. Keep, N.H., Norwood, F.L.M., Moores, C.A., Winder, S.J. and Kendrick-Jones, J. (1999). The 2.0 Å structure of the second calponin homology domain from the actin-binding region of the dystrophin homologue utrophin. J. Mol. Biol. 285, 1257-1264. Kitayama, C., Sugimoto, A. and Yamamoto, M. (1997). Type II myosin heavy chain encoded by the myo2 gene composes the contractile ring during cytokinesis in Schizosaccharomyces pombe. J. Cell Biol. 137, 1309-1319. Korenbaum, E. and Rivero, F. (2002). Calponin homology domains at a glance. J. Cell Sci. 115, 3543-3545. Kovar, D.R., Kuhn, J.R., Tichy, A.L. and Pollard, T.D. (2003). The fission yeast cytokinesis formin Cdc12p is a barbed end actin filament capping protein gated by profilin. J. Cell Biol. 161, 875-887. Kovar, D.R., Staiger, C.J., Weaver, E.A. and McCurdy, D.W. (2000). AtFim1 is an actin filament crosslinking protein from Arabidopsis thaliana. Plant J. 24, 625-636. Kuhlman, P.A., Hemmings, L. and Critchley, D.R. (1992). The identification and characterisation of an actin-binding site in alpha-actinin by mutagenesis. FEBS Lett. 304, 201-206. Kumar, S., Tamura, K., Jakobsen, I.B. and Nei, M. (2001). MEGA2: molecular evolutionary genetics analysis software. Bioinformatics 17, 1244-1245.

102

Kutsche, K., Yntema, H., Brandt, A., Jantke, I., Nothwang, H.G., Orth, U., Boavida, M.G., David, D., Chelly, J., Fryns, J.P., Moraine, C., Ropers, H.H., Hamel, B.C., van Bokhoven, H. and Gal, A. (2000). Mutations in ARHGEF6, encoding a guanine nucleotide exchange factor for Rho GTPases, in patients with X-linked mental retardation. Nature Genet. 26, 247-250. Kyte, J. and Doolittle, R. (1982). A simple method for displaying the hydropathic character of a protein. J. Mol. Biol. 157, 105-132. La Fortelle, E. de and Bricogne, G. (1997). Maximum-likelihood heavy-atom parameter refinement for multiple isomorphous replacement and multiwavelength anomalous diffraction methods. Methods Enzymol. 276, 472-494. Laskowski, R.A., MacArthur, M.W., Moss, D.S. and Thornton, J.M. (1993). PROCHECK: a program to check the stereochemical quality of protein structures. J. Appl. Crystallogr. 26, 283-291. Lebart, M.C., Mejean, C., Casanova, D., Audemard, E., Derancourt, J., Roustan, C. and Benyamin, Y. (1994). Characterization of the actin binding site on smooth muscle filamin. J. Biol. Chem. 269, 4279-4284. Leinweber, B.D., Leavis, P.C., Grabarek, Z., Wang, C.L. and Morgan, K.G. (1999). Extracellular regulated kinase (ERK) interaction with actin and the calponin homology (CH) domain of actin-binding proteins. Biochem. J. 344, 117-123. Leung, C.L., Sun, D. and Liem, R.K. (1999). The intermediate filament protein peripherin is the specific interaction partner of mouse BPAG1-n (dystonin) in neurons. J. Cell Biol. 144, 435-446. Li, X. and Bennett, V. (1996). Identification of the spectrin subunit and domains required for formation of spectrin/adducin/actin complexes. J. Biol. Chem. 271, 15695-15702. Liu, J., Tang, X., Wang, H., Oliferenko, S. and Balasubramanian, M.K. (2002). The localization of the integral membrane protein Cps1p to the cell division site is dependent on the actomyosin ring and the septation-inducing network in Schizosaccharomyces pombe. Mol. Biol. Cell. 13, 989-1000. Mabuchi, K., Li, B., Ip, W. and Tao, T. (1997). Association of calponin with desmin intermediate filaments. J. Biol. Chem. 272, 22662-22666. Marks, J. and Hyams, J.S. (1985). Localization of F-actin through the cell division cycle of Schizosaccharomyces pombe. Eur. J. Cell Biol. 39, 27-32. Matsuura, Y. and Chernov, A.A. (2003). Morphology and the strength of intermolecular contacts in protein crystals. Acta Crystallogr. D59, 1347-1356. Matthews, B.W. (1968). Solvent content of protein crystals. J. Mol. Biol. 33, 491-497.

103

Matynia, A., Dimitrov, K., Mueller, U., He, X. and Sazer, S. (1996). Perturbations in the spi1p GTPase cycle of Schizosaccharomyces pombe through its GTPase-activating protein and guanine nucleotide exchange factor components result in similar phenotypic consequences. Mol. Cell Biol. 16, 6352-6362. Maundrell, K. (1990). nmt1 of fission yeast. A highly transcribed gene completely repressed by thiamine. J. Biol. Chem. 265, 10857-10864. McCallum, S.J., Wu, W.J. and Cerione, R.A. (1996). Identification of a putative effector for Cdc42Hs with high sequence similarity to the RasGAP-related protein IQGAP1 and a Cdc42Hs binding partner with similarity to IQGAP2. J. Biol. Chem. 271, 21732-21737. McCully, E.K. and Robinow, C.F. (1971). Mitosis in the fission yeast Schizosaccharomyces pombe: a comparative study with light and electron microscopy. J. Cell Sci. 9, 475-507. McGough, A., Way, M. and DeRosier, D. (1994). Determination of the alpha-actinin-binding site on actin filaments by cryoelectron microscopy and image analysis. J. Cell Biol. 126, 433-443. Merritt, E.A. and Bacon, D.J. (1997). Raster3D: Photorealistic Molecular Graphics. Methods Enzymol. 277, 505-524. Mezgueldi, M., Mendre, C., Calas, B., Kassab, R. and Fattoum, A. (1995). Characterization of the regulatory domain of gizzard calponin. Interactions of the 145-163 region with F-actin, calcium-binding proteins, and tropomyosin. J. Biol. Chem. 270, 8867-8876. Miller, R., Gallo, S.M., Khalak, H.G. and Weeks, C.M. (1994). SnB: crystal structure determination via Shake-and-Bake. J. Appl. Cryst. 27, 613-621. Mino, T., Yuasa, U., Nakamura, F., Naka, M. and Tanaka, T. (1998). Two distinct actin-binding sites of smooth muscle calponin. Eur. J. Biochem. 251, 262-268. Moores, C.A., Keep, N.H. and Kendrick-Jones, J. (2000). Structure of the utrophin actin-binding domain bound to F-actin reveals binding by an induced fit mechanism. J. Mol. Biol. 297, 465-480. Moreno, S., Klar, A. and Nurse, P. (1991). Molecular genetic analysis of fission yeast Schizosaccharomyces pombe. Methods Enzymol. 194, 795-823. Motegi, F., Nakano, K. and Mabuchi, I. (2000). Molecular mechanism of myosin-II assembly at the division site in Schizosaccharomyces pombe. J. Cell Sci. 113, 1813-1825. Müller, T. and Vingron, M. (2000). Modeling Amino Acid Replacement. J. Comp. Biol. 7, 761-776.

104

Murshudov, G.N., Vagin, A.A. and Dodson, E.J. (1997). Refinement of Macromolecular Structures by the Maximum-Likelihood Method. Acta Crystallogr. D53, 240-255. Nakano, K., Satoh, K., Morimatsu, A., Ohnuma, M. and Mabuchi, I. (2001). Interactions among a fimbrin, a capping protein, and an actin-depolymerizing factor in organization of the fission yeast actin cytoskeleton. Mol. Biol. Cell. 12, 3515-3526. Nikolopoulos, S.N. and Turner, C.E. (2000). Actopaxin, a new focal adhesion protein that binds paxillin LD motifs and actin and regulates cell adhesion. J. Cell Biol. 151, 1435-1448. Norwood, F.L.M., Sutherland-Smith, A.J., Keep, N.H. and Kendrick-Jones, J. (2000). The structure of the N-terminal actin-binding domain of human dystrophin and how mutations in this domain may cause Duchenne or Becker muscular dystrophy. Structure 8, 481-491. Nurse, P., Thuriaux, P. and Nasmyth, K. (1976). Genetic control of the cell division cycle in the fission yeast Schizosaccharomyces pombe. Mol. Gen. Genet. 146, 167-178. Oki, H., Matsuura, Y., Komatsu, H. and Chernov, A.A. (1999). Refined structure of orthorhombic lysozyme crystallized at high temperature: correlation between morphology and intermolecular contacts. Acta Crystallogr. D55, 114-21. Olski, T.M., Noegel, A.A. and Korenbaum, E. (2001). Parvin, a 42 kDa focal adhesion protein, related to the alpha-actinin superfamily. J. Cell Sci. 114, 525-538. Otwinowski, Z. and Minor, W. (1997). Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol. 276, 307-326. Otwinowski, Z. (1991). Isomorphous Replacement and Anomalous Scattering. In Proceedings of the CCP4 Study Weekend, nos. DL/SCI/R32, ISSN 0144-5677. W. Wolf, P.R. Evans and A.G.W. Leslie, eds. (Warrington: Daresbury Laboratory), pp. 80-86. Pande, V.S., Grosberg, A.Y., Tanaka, T. and Rokhsar, D.S. (1998). Pathways for protein folding: is a new view needed? Curr. Opin. Struct. Biol. 8, 68-79. Perrakis, A., Morris, R. and Lamzin, V.S. (1999). Automated protein model building combined with iterative structure refinement. Nature Struct. Biol. 6, 458-463. Prentice, H.L. (1992). High efficiency transformation of Schizosaccharomyces pombe by electroporation. Nucleic Acids Res. 20, 621. Pringle, J.R., Broach, J.R. and Elizabeth, W.J. (1997). The Molecular and Cellular Biology of the Yeast Saccharomyces Vol. 3, Cell Cycle and Cell Biology (New York: Cold Spring Harbor Laboratory Press).

105

Quensel, C., Kramer, J., Cardoso, M.C. and Leonhardt, H. (2002). Smoothelin contains a novel actin cytoskeleton localization sequence with similarity to troponin T. J. Cell Biochem. 85, 403-409. Raae, A.J., Banuelos, S., Ylanne, J., Olausson, T., Goldie, K.N., Wendt, T., Hoenger, A. and Saraste, M. (2003). Actin binding of a minispectrin. Biochim. Biophys. Acta. 1646, 67-76. Rajagopalan, S., Wachtler, V. and Balasubramanian, M. (2003). Cytokinesis in fission yeast: a story of rings, rafts and walls. Trends Genet. 19, 403-408. Read, R.J. (1986). Improved Fourier coefficients for maps using phases from partial structures with errors. Acta Crystallogr. A42, 140-149. Renley, B.A., Rybakova, I.N., Amann, K.J. and Ervasti, J.M. (1998). Dystrophin binding to nonmuscle actin. Cell Motil. Cytoskeleton. 41, 264-270. Robertson, S.P., Twigg, S.R., Sutherland-Smith, A.J., Biancalana, V., Gorlin, R.J., Horn, D., Kenwrick, S.J., Kim, C.A., Morava, E., Newbury-Ecob, R., Orstavik, K.H., Quarrell, O.W., Schwartz, C.E., Shears, D.J., Suri, M., Kendrick-Jones, J., Wilkie, A.O.; OPD-spectrum Disorders Clinical Collaborative Group. (2003). Localized mutations in the gene encoding the cytoskeletal protein filamin A cause diverse malformations in humans. Nature Genet. 33, 487-491. Rosenberger, G., Jantke, I., Gal, A. and Kutsche, K. (2003). Interaction of alphaPIX (ARHGEF6) with beta-parvin (PARVB) suggests an involvement of alphaPIX in integrin-mediated signaling. Hum. Mol. Genet. 12, 155-167. Rupes, I. (2002). Checking cell size in yeast. Trends Genet. 18, 479-485. Russell, P. and Nurse, P. (1987). Negative regulation of mitosis by wee1+, a gene encoding a protein kinase homolog. Cell 49, 559-567. Russell, R.B. and Barton, G.J. (1992). Multiple protein sequence alignment from tertiary structure comparison. PROTEINS: Struct. Funct. Genet. 14, 309-323. Ryckaert, J.P., Ciccotti, G. and Berendsen, H.J.C. (1977). Numerical integration of the cartesian equations of motion of a system with constraints: Molecular dynamics of n-alkanes. J. Computat. Phys. 23, 327-341. Rzhetsky, A. and Nei, M. (1992). A simple method for estimating and testing minimum evolution trees. Mol. Biol. Evol. 9, 945-967. Saitou, N. and Nei, M. (1987). The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4, 406-425. Schmidt, H.A., Strimmer, K., Vingron, M. and von Haeseler, A. (2002). TREE-PUZZLE: maximum likelihood phylogenetic analysis using quartets and parallel computing. Bioinformatics 18, 502-504.

106

Schneider, T.R. and Sheldrick, G.M. (2002). Substructure solution with SHELXD. Acta Crystallogr. D58, 1772-1779. Sheldrick, G.M. (1998). SHELX: Applications to Macromolecules. In Direct Methods for Solving Macromolecular Structures, S. Fortier, ed. (Dordrecht: Kluwer Academic Publishers), pp. 401-411. Sheldrick, G.M. and Schneider, T.R. (1997). SHELXL: high resolution refinement. Methods Enzymol. 277, 319–343. Simerly, C., Dominko, T., Navara, C., Payne, C., Capuano, S., Gosman, G., Chong, K.Y., Takahashi, D., Chace, C., Compton, D., Hewitson, L. and Schatten, G. (2003). Molecular correlates of primate nuclear transfer failures. Science 300, 297. Sohrmann, M., Schmidt, S., Hagan, I. and Simanis, V. (1998). Asymmetric segregation on spindle poles of the Schizosaccharomyces pombe septum-inducing protein kinase Cdc7p. Genes Dev. 12, 84-94. Sparks, C.A., Morphew, M. and McCollum, D. (1999). Sid2p, a spindle pole body kinase that regulates the onset of cytokinesis. J. Cell Biol. 146, 777-790. Stock, A., Steinmetz, M.O., Janmey, P.A., Aebi, U., Gerisch, G., Kammerer, R.A., Weber, I. and Faix, J. (1999). Domain analysis of cortexillin I: actin-bundling, PIP(2)-binding and the rescue of cytokinesis. EMBO J. 18, 5274-5284. Stradal, T., Kranewitter, W., Winder, S.J. and Gimona, M. (1998). CH domains revisited. FEBS Lett. 431, 134-137. Strimmer, K. and von Haeseler, A. (1996). Quartet puzzling: A quartet maximum likelihood method for reconstructing tree topologies. Mol. Biol. Evol. 13, 964-969. Szymanski, P.T. and Goyal, R.K. (1999). Calponin binds to the 20-kilodalton regulatory light chain of myosin. Biochemistry 38, 3778-3784. Szymanski, P.T. and Tao, T. (1997). Localization of protein regions involved in the interaction between calponin and myosin. J. Biol. Chem. 272, 11142-11146. Tang, D.C., Kang, H.M., Jin, J.P., Fraser, E.D. and Walsh, M.P. (1996). Structure-function relations of smooth muscle calponin. The critical role of serine 175. J. Biol. Chem. 271, 8605-8611. Tatebayashi, K., Tani, T. and Ikeda, H. (2001). Fission yeast Mog1p homologue, which interacts with the small GTPase Ran, is required for mitosis-to-interphase transition and poly(A)(+) RNA metabolism. Genetics 157, 1513-1522. Terrak, M., Wu, G., Stafford, W.F., Lu, R.C. and Dominguez, R. (2003). Two distinct myosin light chain structures are induced by specific variations within the bound IQ motifs-functional implications. EMBO J. 22, 362-371.

107

Terwilliger, T.C. (2000). Maximum likelihood density modification. Acta Crystallogr. D56, 965-972. Terwilliger, T.C. and Berendzen, J. (1999). Automated MAD and MIR structure solution. Acta Crystallogr. D55, 849-861. Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F. and Higgins, D.G. (1997). The ClustalX windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. 24, 4876-4882. Tomlin, G.C., Morrell, J.L. and Gould, K.L. (2002). The spindle pole body protein Cdc11p links Sid4p to the fission yeast septation initiation network. Mol. Biol. Cell. 13, 1203-1214. Trautmann, S., Wolfe, B.A., Jorgensen, P., Tyers, M., Gould, K.L. and McCollum, D. (2001). Fission yeast Clp1p phosphatase regulates G2/M transition and coordination of cytokinesis with cell cycle progression. Curr. Biol. 11, 931-940. Wachtler, V., Rajagopalan, S. and Balasubramanian, M.K. (2003). Sterol-rich plasma membrane domains in the fission yeast Schizosaccharomyces pombe. J. Cell Sci. 116, 867-874. Wang, C.H., Walsh, M., Balasubramanian, M.K. and Dokland, T. (2003). Expression, purification, crystallization and preliminary crystallographic analysis of the calponin-homology domain of Rng2. Acta Crystallogr. D59, 1809-1812. Wang, J., Cieplak, P. and Kollman, P.A. (2000). How well does a restrained electrostatic potential (RESP) model perform in calculating conformational energies of organic and biological molecules? J. Comput. Chem. 21, 1049-1074. Wang, Z., Zhu, G., Huang, Q., Qian, M., Shao, M., Jia, Y. and Tang, Y. (1998). X-ray studies on cross-linked lysozyme crystals in acetonitrile-water mixture. Biochim. Biophys. Acta. 1384, 335-344. Wang, P. and Gusev, N.B. (1996). Interaction of smooth muscle calponin and desmin. FEBS Lett. 392, 255-258. Way, M., Pope, B. and Weeds, A.G. (1992). Evidence for functional homology in the F-actin binding domains of gelsolin and alpha-actinin: implications for the requirements of severing and capping. J. Cell Biol. 119, 835-842. Weeks, C.M., DeTitta, G.T., Miller, R. and Hauptman, H.A. (1993). Applications of the minimal principle to peptide structures, Acta Crystallogr. D49, 179-181. Wills, F.L., McCubbin, W.D., Gimona, M., Strasser, P. and Kay, C.M. (1994). Two domains of interaction with calcium binding proteins can be mapped using fragments of calponin. Protein Sci. 3, 2311-2321. Wilson, A.J.C. (1950). Largest likely values for the reliability index. Acta Cryst. 3, 397-398.

108

Winder, S.J., Hemmings, L., Maciver, S.K., Bolton, S.J., Tinsley, J.M., Davies, K.E., Critchley, D.R. and Kendrick-Jones J. (1995). Utrophin actin binding domain: analysis of actin binding and cellular targeting. J. Cell Sci. 108, 63-71. Winder, S.J. and Walsh, M.P. (1990). Smooth muscle calponin. Inhibition of actomyosin MgATPase and regulation by phosphorylation. J. Biol. Chem. 265, 10148-10155. Wu, J.Q., Kuhn, J.R., Kovar, D.R. and Pollard, T.D. (2003). Spatial and temporal pathway for assembly and constriction of the contractile ring in fission yeast cytokinesis. Dev. Cell. 5, 723-734. Xu, J., Wirtz, D. and Pollard, T.D. (1998). Dynamic cross-linking by alpha-actinin determines the mechanical properties of actin filament networks. J. Biol. Chem. 273, 9570-9576. Yang, C., Pflugrath, J.W., Courville, D.A., Stence, C.N. and Ferrara, J.D. (2003). Away from the edge: SAD phasing from the sulfur anomalous signal measured in-house with chromium radiation. Acta Crystallogr. D59, 1943-1957. Yang, Y., Dowling, J., Yu, Q.C., Kouklis, P., Cleveland, D.W. and Fuchs, E. (1996). An essential cytoskeletal linker protein connecting actin microfilaments to intermediate filaments. Cell 86, 655-665. Zhen, Y.Y., Libotte, T., Munck, M., Noegel, A.A. and Korenbaum, E. (2002). NUANCE, a giant protein connecting the nucleus and actin cytoskeleton. J. Cell Sci. 115, 3207-3222. Zugaza, J.L., Lopez-Lago, M.A., Caloca, M.J., Dosil, M., Movilla, N. and Bustelo, X.R. (2002). Structural determinants for the biological activity of Vav proteins. J. Biol. Chem. 277, 45377-45392.

109