Page 133 - Macromolecular Crystallography
P. 133

122  MACROMOLECULAR CRYS TALLOGRAPHY

        Table 8.1 Observed and expected MAD signals for VCP

        Resolution (Å)              B sig  (centric)           D Sig  (R sym )
        4.20                        0.092 (0.016)               0.097 (0.011)
        3.33                        0.097 (0.019)               0.086 (0.027)
        2.91                        0.114 (0.030)               0.131 (0.024)
        2.85                        0.173 (0.047)               0.147 (0.051)
        2.46                        0.191 (0.049)               0.169 (0.066)
        2.31                        0.215 (0.060)               0.182 (0.083)

        Expected signals for N A =  1  2    3       4       (f      = 10.7, f =−18.3
                                                            N P = 1400, Z eff = 6.7)
        Bijvoet            0.037   0.052    0.064   0.074
        Dispersive         0.032   0.045    0.055   0.063
                |F  − F  |          |F  − F  |
                 +Peak  −Peak        Edge  Remote
             Nref                 Nref
        B sig =             D sig =
            1/2  (F +Peak + F −Peak )  1/2  (F  Edge  + F Remote )
               Nref                Nref

               |I − I |
             Nref
        R Sym =     where  I =  I i /Nobs
                 I         Nref
               Nref
        In the original, algebraic implementation, this was  Hendrickson (1985). It follows that:
        done by determination of the three unknown
                                                        0       √    0        √
        quantities through least squares minimization of the  | F T (h)|=  p 1 , | F A (h)|=  p 2  and
        MAD equation:                                       0      0          −1
                                                        φ = φ T (h) − φ A (h) = tan  (p 4 /p 3 )
         λ     2
        | F T (h)|                                       0
                                                     The | F A (h)| values thus derived can be used to
              0     2     0     2
           =| F T (h)| + a(λ)| F A (h)|              determine the positions of anomalous scatterers
                                                     through computation of a Patterson synthesis, or

                                           0
                  0
                         0
             + b(λ)| F T (h)|| F A (h)| cos  0 φ (h) − φ (h)  by other methods. This step leads to the knowl-
                                             A
                                     T
                                                           0
                                                                               0
                                                                                       0
                                                     edge of φ A (h), and since  φ = φ T (h) − φ A (h),
                         0
                                          0

                                   0
                  0

             + c(λ)| F T (h)|| F A |(h) sin  φ (h) − φ (h)  the φ T (h) for each reflection can be computed. All
                                     T
                                                        0
                                            A
                                                     of these steps are implemented in the MADSYS
        Making the following substitutions,
                                                     (Hendrickson, 1991) system of programs.
                            0
                     2
               0
                                  2
           p 1 =| F T (h)| , p 2 =| F A (h)| ,        In implementation of MAD as a special case of
                                                     MIR, the well developed theoretical foundation for
                               	 0     0
               0
                      0
           p 3 =| F T (h)|| F A (h)| cos  φ T (h) − φ A (h)
                                                     MIR (Blundell and Johnson, 1976), and the program
                                                     system that has long been in use can directly be
           and
                                                     used (Ramakrishnan and Biou, 1997). The anoma-
                               	 0     0
                      0
               0
           p 4 =| F T (h)|| F A (h)| sin  φ T (h) − φ A (h)  lous scatterer positions are determined either from
                                                     Patterson functions or from the use of direct meth-
        the MAD equation can written as              ods, and the phases calculated and refined using a
           λ     2                                   robust maximum likelihood target (La Fortelle and
           | F T (h)| = p 1 + a(λ)p 2 + b(λ)p 3 + c(λ)p 4 = Y c
                                                     Bricogne, 1997). For example, the CCP4 program
        From which the p’s can be determined by fitting  MLPHARE can be used for this purpose (Project,
        the measured Bijvoet and dispersive differences,  1994). Some details from the structure determina-

        along with the knowledge of the f and f      val-  tion of the dengue virus serotype 2 (Den2) pro-
        ues, as described in Krishna Murthy (1996) and  tease complexed with the mung bean Bowman–Birk
   128   129   130   131   132   133   134   135   136   137   138