IUBio Biosequences .. Software .. Molbio soft .. Network News .. FTP

Workshop Report

Tue Jan 25 16:17:12 EST 1994



                           August 16-17, 1993

                      National Institutes of Health
                           Bethesda, Maryland

Attached is a copy of the Introduction, the Table of Contents, and the
Executive Summary from this report.  If you are interested in receiving
a complete copy of the report, please contact me and provide a mailing

Sincerely yours,

Peter C. Preusch, Ph.D.
Program Administrator
Cellular and Molecular Basis
 of Disease Program
National Institute of General Medical Sciences
National Institutes of Health
Westwood Building, Room 906
5333 Westbard Avenue
Bethesda, Maryland, U.S.A.  20982

Phone:  301-594-7806
FAX:    301-594-7728
E-Mail: peter_preusch at nihgmsww.bitnet


                        NIGMS SPONSORED WORKSHOP

              Crystallography and NMR of Membrane Proteins


As of August, 1993, the Brookhaven Protein Data Bank contained
approximately 2,300 structures representing over 1,000 functionally
distinct proteins.  Only five of those structures (of only three
functionally distinct enzymes) were of intrinsic membrane proteins.
This ratio of known structures clearly fails to mirror the relative
abundance of soluble versus membrane bound enzymes, nor does it reflect
well the growing importance of membrane bound proteins in a wide range
of biological and human health-related research.

On August 16-17, 1993, NIGMS sponsored a workshop on the NIH campus,
entitled, "Structural Biology of Respiratory Enzymes: Crystallography
and NMR of Membrane Proteins".  A significant effort was made to assure
that a broad scientific community would be consulted regarding the need
to stimulate research in the topic areas.  The date was chosen to take
advantage of information reported at other meetings held during the
summer (e.g., Bioenergetics Gordon Conference, Photosynthesis Gordon
Conference, and 5th International Conference of the Crystallization of
Biological Macromolecules).

The meeting attracted over 130 participants and brought together a
significantly different grouping of researchers than has been assembled
at any other recent meeting.  A copy of the Roster of Meeting
Participants is given at the end of the report.  Correspondence was
received from a number of interested individuals who were unable to
attend the meeting.  They are listed separately before the Roster and
their opinions have been incorporated into the report.

The focus of the meeting was on the potential for determining structures
of the mitochondrial respiratory enzymes and related enzymes of plant
and microbial bioenergetics.  The rationale for this choice included:

1)  Knowledge about the structure and function of these enzymes is
relevant to general human health and a growing number of specific human
2)  The state of knowledge about the mechanisms of these enzymes is
advanced to the point that further progress requires the availability of
high resolution structural data;
3)  The proteins possess non-trivial membrane-embedded domain structures
whose solutions will challenge the state-of-the-art in technology;
4)  Many of these proteins have been cloned, sequenced, and expressed in
native and recombinant forms -- the capacity for production of pure
proteins in large quantities and in specific isotopically labeled forms
is available;
5)  The initial successes in determining structures of membrane proteins
were for members of this enzyme class, and significant work is in

These enzymes were chosen as models for membrane proteins, in general,
and much of the methodology for membrane protein isolation and
reconstitution has been developed through their study.  Investigators in
this field may be especially well poised to take advantage of an
intitative in the area of membrane protein structural biology.  However,
progress on other membrane proteins may similarly have reached the point
that detailed structural information should be pursued (e.g., other ion
pumps, ion channels, substrate transporters, neurotransmitter uptake
systems, and G-protein coupled receptors).

Major scientific points that were illustrated during the course of the
meeting included:

1)  the value of even the few known membrane protein structures in
revolutionizing several fields of research,
2)  the feasibility of obtaining atomic resolution structures for
membrane bound proteins,
3)  the ability to produce suitable material for crystallization trials
for a variety of energy transducing membrane bound proteins,
4)  progress toward understanding the intermolecular contacts that are
important for producing diffraction quality crystals,
5)  the value of pursuing both 2D crystallization and 3D crystallization
approaches in parallel,
6)  the adequacy of current methods to solve the large unit cell
structures of membrane proteins, once crystals are in hand, and
7)  the value of NMR to provide complete structural solutions for
fragments of the proteins of interest and local information for even
very large intact proteins in their native membrane environments.

A number of needs and opportunities were identified to stimulate
research on the crystallography and NMR of membrane proteins.  A number
of reasonable suggestions were made that can be implemented by NIGMS
without adversely affecting the availability of funds for general
investigator-initiated research grants.  However, membrane proteins are
of trans-NIH interest and a cooperative effort involving additional
components of the NIH will be needed to stimulate research in this

The body of the report was prepared from speakers abstracts and
overheads and the reports of the session chairpersons that were prepared
during the meeting.  These have been edited to include comments that
were made during the meeting but not otherwise captured.  The final
section includes a summary of NIGMS Staff suggestions to address the
needs and opportunities that were identified during the workshop.

                            TABLE OF CONTENTS

1.  Workshop Announcement and Agenda.................................. 4

2.  Executive Summary................................................. 8
    Report of Chairpersons of the Monday Morning Session
      Norma Allewell and Ronald Kaback

3.  Keynote Speech: On the Value of Structure Determination in
    Elucidating Function.  A Case Study: the Bacterial Photoreaction
    Center.  George Feher............................................ 12

4.  Topic Overview Lectures - Monday Morning Session

    Membrane Protein Crystallography-Work in Progress: Solved
    Structures.  Johann Deisenhofer.................................. 15

    Membrane Protein Production and Purification for Crystallization
    Shelagh Ferguson-Miller.......................................... 20

    Crystallization Methodology
    Michael Garavito................................................. 24

    Electron Crystallography of Membrane Proteins
    Robert Glaeser................................................... 27

    Methods for Solving Structures of 3D Crystals
    Elinor Adman..................................................... 30

    NMR Methods for Structural Studies of Membrane Proteins
    Robert Griffin................................................... 41

5.  Reports on Monday Afternoon Panel Discussions

    Membrane Protein Crystallography-Work in Progress: Updates and
    Critique.  William Cramer........................................ 45

    Protein Production Issues-Cloning/Expression/Purification
    Robert Gennis.................................................... 49

    Crystallization of Membrane Proteins
    Enrico Stura..................................................... 52

    Solution of Protein Structures
    NIGMS Staff...................................................... 56

    Potential for NMR Methods
    Robert Fillingame................................................ 58

6.  Final Discussion Session:  Recommendations for Stimulation of
    Research on Membrane Protein Structure and Function
    Peter Pedersen................................................... 61

    a.  General Needs for Membrane Protein Structural Biology
    b.  Specific Needs with Respect to Membrane Protein Crystallography
    c.  Specific Needs with Respect to NMR of Membrane Proteins
    d.  Implementation of Recommendations

7.  Epilogue: NIGMS staff actions taken since the meeting to stimulate
    this field of research........................................... 65

8.  Meeting Preparations and Correspondence Received................. 66

9.  Roster of Meeting Participants................................... 68

EXECUTIVE SUMMARY:  Report of the Chairpersons of the Monday Morning

Norma Allewell (University of Minnesota) and H. Ronald Kaback
(University of California-Los Angeles)

Understanding the mechanisms of function of membrane proteins depends
critically upon understanding their structure.  Although the structures
of a few proteins have been determined, progress has been relatively
slow.  The purpose of this conference was to examine the state of the
art and to consider policy changes which might result in more rapid
progress.  The central role of membrane proteins in many clinical
problems, metabolism, fundamental cell and developmental biology and the
neurosciences warrants a substantial initiative and perhaps new funding
mechanisms.  Because of recent advances in determining the structures of
respiratory enzymes and related proteins, the discussion focused upon
these systems.  The spectacular results obtained with the photosynthetic
reaction centers as discussed by Drs. George Feher and Johann
Deisenhofer illustrate the impact which structural studies have in terms
of fundamental advances in understanding mechanisms.

Dr. Feher's presentation made the very clear point that in order to
explain electron transport and proton translocation, it is essential to
understand the molecular orientation and electronic configuration of the
reaction center components.  The question of whether or not the protein
behaves as a scaffold or contributes to the chemistry can be tested with
site-directed mutagenesis.

Dr. Deisenhofer discussed published work on high resolution structures
of membrane proteins.  It is apparent that very few structures have thus
far been solved.  Structures discussed were of photosynthetic reaction
centers, porins and prostaglandin synthetase.  In the photosynthetic
reaction centers, there are hydrophilic domains on both surfaces and the
crystal contacts are much like those formed in soluble protein crystals.
The choice of detergent to protect hydrophobic surfaces is critical and
Hartmut Michel's notion of mixing detergents turned out to be crucial.
Dr. Deisenhofer feels that the major problems are availability of pure,
stable material, collaborators willing undertake a difficult problem
that may not yield results, and stable funding.

The problem of determining the structures of membrane proteins can be
broken down into four phases: protein production, characterization,
crystallization, and solution of the structure.   Although all four
phases are difficult, the major bottlenecks appear to be the first three

Problems of protein production were reviewed by Dr. Shelagh Ferguson-
Miller.  In addition to traditional sources from animal and plant
tissues, three natural sources for production of respiratory enzymes
exist: bacteria, yeasts, and cell cultures.  The latter sources offer
the potential for application of recombinant DNA technology.

Of these, bacteria appear the most promising at the present time,
because they provide convenient sources of large quantities of material,
have simpler subunit compositions, can be engineered for better
purification, and provide the option of studying pieces and domains.
Bacterial respiratory proteins have provided the basis for many recent
advances in the field of membrane protein structure.  They have the
disadvantages of being susceptible to proteolysis and subject to
variation with growth conditions, therefore presenting purification
problems.  However, these problems are common to all sources and may be
more readily controlled in the case of bactera.

Purification relies upon column methods where dispersion (critically
dependent upon choice of detergent), choice of column matrix (highly
dependent upon the specific protein being studied), and removal of
detergent are critical.  HPLC and FPLC provide speed, resolution and
control; programmable gradient "holds"  increase resolution
substantially.  Specific affinity methods are particularly useful when
purifying low abundance proteins or when "designer" approaches can be
used.  There is considerable room for the development of new methods
with the potential to detect subtle, or not so subtle, heterogeneity.
Examples include Fourier transform infrared spectroscopy, electrospray
mass spectrometry, and dynamic light scattering.

The choice of detergents is critical.  The requirements for membrane
solubilization, lipid removal, column chromatography and crystallization
differ.  The head group appears to be important in determining ability
to displace phospholipids.  Further investment in detergent development
and characterization is well justified.

Crystallization is still rate limiting.  Impediments to crystallization
include aggregation, detergent impurity, proteolytic cleavage,
heterogeneous covalent modification, excess phospholipid, and, perhaps,
conformational equilibria.   Dr. Michael Garavito discussed these
issues, in the context of his work with OmpF Porin and Prostaglandin H
synthase.  He pointed out that the conditions used to crystallize these
proteins are very similar to those used to crystallize soluble proteins,
except for the presence of detergents, and that it is the polar regions
of the protein that initiate crystallization.  Where these regions are
not present in nature, it may be possible to create them by genetic
engineering or antibody binding, as Dr. Enrico Stura's work illustrates.
Microcrystals are easy to obtain, while obtaining diffraction quality
crystals is very difficult.  Tight binding inhibitors irreversibly
inactivate cyclooxygenase activity but stabilize prostaglandin H
synthase for crystallization.  As Dr. Gene Scarborough pointed out in a
later talk, conditions must be found where the solubility line of the
detergent passes through the protein crystallization zone.  The best
detergents for crystallization (thus far) have been small with an alkyl
tail and moderate critical micellar concentrations that form
monodisperse small micelles.  Several speakers emphasized the importance
of sensitive temperature control, because of the complex composition of
the crystallization medium.   Reduced pressure speeds crystal growth in
some cases.

Methods for solving the structures of 3D and 2D crystals were discussed
by Dr. Elinor Adman and Dr. Robert Glaeser, respectively.  When crystals
of high quality have been obtained, standard crystallographic methods
can be used to determine the structure.  The greatest advances in the
past twenty years have been in the area of data collection and
processing.  The relatively large size of most membrane proteins makes
data collection, obtaining good heavy atom derivatives, phase
determination and computation more complex, but not qualitatively
different from the problems faced in determining the structures of
soluble proteins.  Well-maintained, accessible synchrotron sources are
particularly important because of the instability of many membrane
protein crystals and the size of the asymmetric unit.  Methods for
dealing with disorder are not well-developed; the most successful
approach is to begin with well-ordered crystals.  Computational problems
are not rate limiting, although regional supercomputer facilities
continue to be needed.

As Dr. Glaeser, and, later, Dr. Mary Lyon discussed, 2D crystals have
the advantage that relatively high resolution structures can be obtained
with very small amounts of material.  Electron scattering has the
advantage that the scattering cross section for electrons is 106 that of
X-rays and that electrons can be focused.   Differences in conformation
in bacteriorhodopsin were detected by trapping M1 and M2 intermediates
during proton translocation.  2D and 3D trials can be launched
simultaneously and may feed each other, as recent advances with
actomyosin illustrate.  Development of this field in turn depends upon
protein production, detergent research, and second generation electron
crystallographic facilities.  Specimen flatness, short range order,
"transparent" data collection and analysis, and further reduction of
systematic defects in the image require further attention.  Ultimately,
it should be possible to utilize 10 x 10 unit cells and to achieve
atomic resolution with samples that form regular structures other than
flat sheets.

The use of NMR was reviewed by Dr. Robert Griffin.  When small membrane
proteins can be solubilized, their structures can be determined through
the use of distance constraints by standard methods.  Molecular weight
restrictions often dictate "divide and conquer" approaches despite the
potential limitations of this approach.  Structures of oriented samples
can be determined through the use of orientation constraints established
by dipolar and quadrupolar splitting.  Magic angle dipolar recoupling
provides a means to establish distance constraints through homo- and
heteronuclear coupling.  Examples of systems in which NMR has been
successfully utilized include gramicidin A (Timothy Cross) and
bacteriorhodopsin (Robert Griffin), where the use of magic angle dipolar
recoupling has allowed interatomic distances at the retinal binding site
to be determined.  The importance of maintaining centers where state-of-
the-art instruments and methodology are readily available was emphasized
by several speakers.  Broad-based education of life scientists who would
benefit from utilizing recent advances, but are probably not aware of
them, is also required to maximize the benefits of these centers.

Basic problems in the field of membrane protein structural biology lie
in the area of infrastructure (facilities, instrumentation, and
training) and in the need for long term support.  There is no "magic
bullet" that will solve the structures of membrane proteins; they will
yield only to sustained effort and even then at unpredictable rates.
The existence of large labs with stable funding and professional staff
in Europe has provided the stability needed to make incremental
progress.  Group efforts by experienced professionals with complementary
expertise is needed to move the field ahead.  Maintaining the physical
infrastructure is also critical, as the frequent references to
temperature control illustrate.  Insight into protein, detergent, and
lipid chemistry is the key to success, and maintaining our strength in
these fields through adequate funding of training programs is essential.

It is clearly impossible to understand function without structure.  In
the past 10 years, the explosion in molecular biology has enabled the
manipulation of membrane proteins to an extent undreamed of earlier.
Thus, one can make almost any modification to a protein, but, without a
structure, it is virtually impossible to interpret the results of the
manipulation.  In order for the U.S. to maintain a serious effort in
this modern aspect of biochemistry and molecular biology, it will be
essential to provide funding for the crystallization of membrane
proteins--not only funds for structure determination after
crystallization, as is presently the case.  Relatively few membrane
proteins have been sufficiently purified to be ready for
crystallization.  Thus it would be relatively inexpensive for NIH to
increase funding for those interested in these endeavors.  Further
development and more wide-spread application of NMR approaches will also
be fruitful.

More information about the Xtal-log mailing list

Send comments to us at biosci-help [At] net.bio.net