核磁共振

NOTE [1]: GENERAL EXPLANATION OF HOW TO SET UP EXPERIMENTS

Here I describe some of Lewis's frequently-used pulse sequences for the
structure determination of proteins.  There are many other sequences which
are not mentioned here.
____________________________________________________________________________

(1)  Experiments performed using N15-labelled samples:

N15 HSQC   : hsqc_gd_sl_seduce_500.c
N15 TOCSY  : dipsi3dn_pfg_enhanced_sel_v1_600.c
              Note that t1 dimension is not tppi-states
N15 NOESY  : noesy3dn_pfg_enhanced+_sel_600.c
HNHA       : hnha_jr_600.c
              This is analyzed together with NHT1_enhanced.c
HNHB       : hnhb_3c_pfg_sel_600.c

               
NOTE       : There are variations of HSQC pulse sequence.
              For the protein sample with broad signals,
              hsqc_unenh_sel_exchg_600.c or hsqc_gd_seduce_ri_cpmg41.c
              may be good.  N15 CT_HSQC: CT_hsqc_600_ch3_ri.c is useful
              to check overlapped signals in hsqc_gd_sl_seduce_500.c
              MAKE NOTE that the extension "ri" in the pulse sequence
              indicates modification or creation of the sequence by ri.
              The sequences by ri are not so much kind as Lewis. (? -ed.)
              TAKE CARE IN THE SETTING OF PARAMETERS.      



(2.1)  Experiments performed using double-labelled samples in H2O

Constant time HSQC for carbon: CT_hsqc_600.c
(Non constant time HSQC for carbon: hsqc_c13_600.c)
hnco_3c_pfg_lek+_sel11_600.c   1.5 days
hbcbcaconnh_3c_pfg1_sel_600.c  2.0 days
hncacb_3c_pfg_sel_600.c        3.5 days

hncoca_D_sel_pfg_600.c         2.0 days
hnca_D_sel_pfg_600.c           3.5 days

hahbcbcaconnh_3c_pfg1_sel_600.c 3 days
   
hbcbcacocaha_3c_pfg1+_600.c     2 days

noesyc_pfg_h2o_NC+_600.c        4 days
(or noesyc_pfg_h2o_NC+_600_v2.c - this is with chirp)

ccctocsy_nnh_3c_pfg1_sel_500_v2.c  3~4 days
hcctocsy_nnh_3c_pfg1_sel_500_v2.c  3~4 days
     NB:  these two work well for small proteins.  If they do not work for
    large proteins, try hcchtocsy or try the following:
  ccctocsy_nnh_3c_pfg1_sel_600_v3.c
  hcctocsy_nnh_3c_pfg1_sel_600_v3.c
       These versions have beta-carbon decoupling and therefore work
   relatively well for large proteins.  These are useful for backbone
   assignment as well.  See details in nmrexp_ccctocsy_nnh_chirp.txt)


(2.2)  Experiments performed using double-labelled samples in D2O
       (Also possible to do in H2O)

CT_hsqc_600.c
     NB:  you can try this twice - once setting dof and taua for aliphatic,
          and the second time for aromatic.

hbcbcacocaha_3c_pfg1+_600.c     2 days
     NB:  take care with the Bloch-Siegert shift

hcchtocsy_3c_pfg_600.c          4~5 days
     or hcchflopsy_3c_pfg_600.c, flopsy is used instead of dipsi

noesy_C_chirp_purge_lek_v2.c    4~5 days



(3)  ASSIGNMENT Procedure

There might be several strategies to assign signals.  Here are a few examples:

               
   (1) connection by carbon
       hncacb + cbcaconnh (+ hnco, etc)
       variation of them (cbcacocaha + ccctocsy_nnh + hncacb)

   (2) connection by carbon, large molecules
       hnca   + hncoca (+ hnco, + hcaco etc)

   (3) connection by proton
       hn(cacb)hahb + hahb(cbcaco)nnh
       variation of them (hacahb_cosy + hnha + ...)
   (4) ...

               


(4)  Experiments for aromatic assignment

Proton homonuclear TOCSY and NOESY (dipsif2n_h2o_pfg_600.c and
noesyh2o_pfg_600_v1.c, 50, 100 ms) are mainly used to assign signals of
aromatic side chain.  Furthermore, hbcbcgcdhd_aro_pfg_500.c and
hbcbcgcdcehe_aro_pfg_500.c are used using C13-labelled sample in D2O.  These
two experiments do not work in large proteins as well.

dipsif2n_h2o_pfg_600.c and noesyh2o_pfg_600_v1.c  12 ~24 h
hbcbcgcdhd_aro_pfg_500.c and hbcbcgcdcehe_aro_pfg_500.c  12~36 h
      (See detail of the section of nmrexp_aromatic.txt)



(5)  About the sample preparation

USUAL Wilmad nmr tube:  This is a 5mm diameter tube.
500ul solution corresponding to 30 mm height is required.
                        
SHIGEMI tube:  5mm diameter tube. 17.0 mm height is required.
(this corresponds to ~280 ul.  However, total volume depends on the skill of
how to push in the inner tube to the outer tube.  If you are beginner, prepare
~300 ul.)
COATING OF SHIGEME TUBE (IWAKI brand Labware, Siliconization, SIL-COAT 5, 10
pcs for 1 pack):  The protocol was originally written in Japanese.  Here is a
translation:   
1.  Wash the tube well.
2.  40 ~ 50 times dilute the solution in the ampul by water at 70 C.
            (Once the ampul is open, it can survive only 2 days.)
3.  Soak the tube 10 second and soon rince it by water.
     (You dont need to soak it.  Put the solution to the tube)
4.  Dry 24 hours.  Then, the tube is coated.      
         (In a hurry, dry it in an oven at 100 C ~ 150 C for 30 min.)

*  Coating is not removed by detergent or alcohol, but is removed by
    conc.HNO3, conc.HCl, NaCl, or KCl.  When you want to remove the
    coating, wash it with 1N NaCl.

               
ISOTOPE LABELLED BUFFERS.  Use deuterated materials in the sample buffer, such
as deuterated DTT, deuterated glucose, so on.   



(6)  Pulse sequences used for calibration of pulse width

Details of how to calibrate are described later.  Here is an introduction to
the name of pulse sequences.

1H 90-degree pulse:  t2pul_lek_600.c
      Observe protein signal
      When you meausre pulse width weaker than 30 us or so,
           (1) Set on-resonance to protein (not 4.7 ppm) and observe protein
               signal.
           (2) Or, set on-resonance to water signal and observe water signal
                without presat (in this case, you need a very long delay time).

1H 90-degree water pulse with shape (selective 'water flip back' pulse):
      mainly for "seduce" or "hard"
                       t2pul_lek_sel3.c (without phase correction) or
                       t2pul_lek_sel3_600.c (with phase correction).

      NOTE:  difference in the rf-power, such as between a high-power-90-
      degree pulse and a weak 90-degree pulse for water causes phase
      modulation.

13C 90-degree pulse:   hetcal_echorm_600.c or hetcal_echorm.c
                        or hetcal_echo_dec1(_ri).c

13C shaped pulse:      mainly for seduce
                        hetcal_echolek_shape_600.c or hetcal_echolek_shape.c

15N 90-degree pulse:   hetcal_echolek2_600.c or hetcal_echolek2.c

13C and 15N decoupling profile: t2pul_lek_600.c
      Array dof2 or dof under proper condition for decoupling.  Used wft(nodc)
      for the processing.

Calibration of chirp, dual_S-S, g3, and so on. -> use pulsetool
      For further details, look at these other notes:
            chirp    -> nmrexp_cnoesy_chirp.txt
            dual_S-S -> nmrexp_N15_relaxation.txt
            g3       -> nmrexp_aromatic.txt  


NOTE:  Difference of calibration pulse sequences between that of Lewis and
that of Varian:
      In the hetcal (Lewis's one), a pulse is written as sim3pulse
      (2*pw, pwx1, 0.0, phase1, phase2, phase3, 0.0,0.0)
      In pwxcal (Varian's), sim3pulse
      (2*pw, pwx1, pwx2, phase1, phase2, phase3, 0.0,0.0)
      Therefore, in pwxcal, N15 pulse width has to be set zero in order to
      measure carbon pulse, while in hetcal, N15 pulse width is automoatically
      set to zero.

NOTE:  Different sign of offset in the Lewis's sequence and the one in the
pulsetool.
      When you create seduce pulse with offset 20000 Hz to the lower field,
      RF_seduce 333 0 333 20000 1023 0
      you put the value positive 20000.  In fact this corresponds to -20000 Hz
      offset from the on-resonance.  In the pulsetool the center of excitation
      becomes -20000 as you created.  In the usage of the RF_seduce, it is
      commented as well.

NOTE:  water saturation in t2pul_lek_600.c (the sequence for 1D proton
spectrum).
           tsatpwr     excited spectral width
             0         plus-minus 250 Hz
             2         plus-minus 330 Hz
             4         plus-minus 420 Hz   
      This means, signals in the region from 4 ppm to 6 ppm are attenuated
      at tsatpwr = 4.  I recommend to use tsatpwr=0.



(7)  Channels

Following is the typical usage of the channels to run Lewis's pulse sequence.

Channel 1 : H1 observation.  
      (In our Unity+ 500MHz, C and N can be directly observed while in our
      Unity 600MHz, change of connection is required to observe C and N.  To
      observe N15 directly, a filter to remove 2H frequency is required.)

Channel 2 : C13.  
      If you want to set channel 2 to 15N, change the cable and a filter coming
      in to a proble as well as the change in dof.

Channel 3 : N15.

Channel 4 : H2 (Although we don't have the deuterium channel yet.)
      On our unity 600 (victor), 4th channel was used or carbon before.  In
      this case, channels 1,2,3 and 4 correspond to 1H, 13C, 15N, and 13C.  On
      the other hand, in our unity+ 500MHz (miryl), channel 1,2,3, and 4 are
      often assigned to H1, N15, C13, and C13.  Therefore, there are three
      types of experiments using 4 channels: the one for (1) victor, (2) miryl,
      and (3) with 2H.  We don't use (1) and (2) anymore.  

      Our Unity+ 500 MHz (miryl) share channel 4's hardware with that of
      channel 3.  You can not use channels 3 and 4 for different nuclei.  In
      Lewis's sequence with 2H option (there is an extention _D_ in the name
      of the pulse sequences), the third channel is assigned to N15 while the
      fourth is 2H.  Therefore, set dn3='' (channel 4 equals non) and set
      ampmode='dddd' when you use such pulse sequence currently.  Ampmode is
      sometimes defined to 'dddp' in the Lewis's parameter file.  'd' means
      default mode while 'p' means pulse mode.



(8)  How to get pulse sequences.

I strongly recommend you to get pulse sequences for 600MHz and for 500MHz each
from Lewis's group together with parameters.  This is essential not to miss
experiments for you since parameters are different between 500MHz and 600MHz.

The pulse sequences available listed on cuturl('http://abragam.med.utoronto.ca.')
Send e-mail to ranjith@bloch.med.utoronto.ca (Ranjith Muhandiram).



(9)  Difference of shaped pulses created by Lewis and Pbox.

Lewis's shaped pulse is created by executing C programs installed in the
shapelib.  Varian's original shaped pulses are in the ./nmr/wavelib and can
be modified further using Pbox.  Follows is a comparison of near-identical
seduce pulses for carbon created by both methods.

      1) Inversion seduce.  These two creates the seduce pulse using almost
         the same parameters as each other.

         Pbox:  { seduce 2000/0.000240 }
                pw=240 us
                power= 4.53 kHz

         Lewis: RF_seduce 240 0 240 0 1023 0
                pw=240
                power= 4.53 kHz (by pulse tool)

      2) seduce decoupling.  The values of paramters used to create the pulse
         are different between Pbox and Lewis.  Especially, dmf in the Pbox
         looks to be smaller than Lewis's one because of difference of dres.

              Pbox: " { SEDUCE1 2000.0/0.0002400 }"
                    power= ~ 47  (power of 54 us as a hard 90 pulse)
                    B1 = 4.59
                    dmf=20833
                    dres =9
                    dmf*90/9=20833*90/9=208330

                    "DEC_seduce_cos 0 240"
                    -> pw_shape =  240 us
                    -> dres = 2
                    -> power = ~39 + 6 in our machine.
                    -> dres = 2
                    1/pw_shape=4166
                    dmf*90/dres=4166*90/2=187470.

      NOTE:  what's dres?
      example:
           1.  dres=9, and dmf=40000
           2.  dres=90, and dmf=400000
      1. and 2. are the same as each other.  If you set dres = 90 then
      dmf = 1/pw90, which makes it 1/stepsize.  When dres is not 90, dmf is
      not just 1/pw90.  For example, when dres = 9 and dmf=1/pw90,
      1/stepsize = dmf*90/9.  See the manual for details.

               


** In the following explanation of individual experiments, I may not
   particularly describe dres, dres2, and dressed.  Typical values for dres
   used in the Lewis's sequence are written in the nmrexp_hnco.txt.