JOURNAL
OF STRUCTURAL
BIOLOGY
107,
l%--WJ
(1991)
The Hand of the Helix of Deoxyhemoglobin MICHAEL Laboratory
for Image Analysis The University
R. LEWIS, THOMAS E. LEE, AND ROBERT JOSEPHS and Electron of Chicago,
Microscopy, 920 East Received
Department 58th Street,
July
IN.
INTRODUCTION
Intracellular polymerization of deoxyhemoglobin S (HbS)l to form fibers is the cause of sickle cell anemia. In viuo this aggregation reaction initiates the cascade of events which characterize sickle cell disease. We have studied the structure and assembly of HbS fibers in order to gain a better understanding of the intermolecular interactions which are responsible for the pathological aggregation of these molecules. The fiber structure has been reconstructed by Carragher et al. (1988) and Dykes et al. (1979). A complete description of the fiber structure requires a knowledge of both the molecular coordinates and the hand of the particle. In principle the hand of a helical particle can be determined from the changes in the phases of the particle’s Fourier transform caused by tilting. Finch (1972) has demonstrated this approach by tilting lateral views of TMVperpendicular 1 Abbreviation
used:
HbS,
deoxyhemoglobin
Inc. reserved.
and
Cell Biology,
24, 1991
2.
MATERIALS
AND
METHODS
Embedded HbS fiber samples prepared according to Potel et al. (1984) and McDade et al. (1989) were provided by Dr. William McDade. Thin sections were cut using a Sorvel MTSB ultramicrotome and poststained with uranyl acetate and lead citrate. A Philips CM10 electron microscope was used to record stereo pairs of electron micrographs at an electron optical magnification of
S. 196
1047~8477/91 $3.00 Copyright 0 1991 by Academic Press, All rights of reproduction in any form
of Molecular Genetics Chicago, Illinois 60637
to the particle axis. This approach produces maximum changes in particle appearance along its short pitch helices. Dykes et al. (1979) used changes in the appearance of lateral views which occur as the particle is rotated about its axis to propose that sickle cell hemoglobin fibers are right-handed. The hand of the HbS fiber is difficult to determine from tilting lateral views for a variety of reasons. The bonding vectors in a fiber are along paired strands of hemoglobin molecules which twist about the axis with a pitch of about 3000 A. In addition it is now recognized that the fiber has a variable pitch. Consequently the Fourier transform of the fiber is extremely complicated as the unit cell consists of 14 molecules in seven quasiequivalent positions. Because of these factors, measurements of phase changes upon tilting can be subject to errors whose magnitude is uncertain, but likely to be considerably larger than in the case of well-ordered particles. These errors could lead to an incorrect assessment of the particle hand. Because a knowledge of the particle hand is an essential prerequisite to synthesizing a molecular model of the fiber we have developed a method for determining the hand from tilting cross sections. The method is based upon the principle that when a short segment of a long pitch helix is tilted so that it is roughly parallel to the direction of view it is resolved in projection as a bright spot. Knowing the tilt direction gives the hand. Finch’s method and the one proposed here are complimentary. Tilted lateral views show pronounced changes along short pitch helices, while in cross sections the most pronounced changes occur along long pitch helices. In practice our approach is simple and might readily be adopted for other helical particles.
Electron micrographs of deoxyhemoglobin S fiber cross sections provide an end-on view of the fiber whose appearance is sensitive to small changes in orientation. We have developed a procedure to exploit this sensitivity in order to determine the hand of these particles. In a sickle hemoglobin fiber the hemoglobin molecules form long pitch helical strands which twist about the particle axis with a pitch of about 3000 A. Tilting a 400~A-thick cross section by a few degrees aligns one of the long pitch helices so that it is nearly parallel to the direction of view. When a strand of hemoglobin molecules in a fiber is aligned in this manner it appears as a strongly contrasted bright spot. It is this spot, rather than the fiber axis, which appears to be the apparent center of rotation of the cross section. The direction of the displacement of the spot from the particle axis depends upon the particle hand and tilt direction. We have used this property to determine that sickle hemoglobin fibers are right-handed particles. This method may be applicable to other particles with long pitch helices as well. o 1991Academic press, 1.
S Fibers
HAND 46 000. The micrographs were recorded promptly in imize section collapse due to beam damage (Luther Models in Fig. 1 were calculated with the program on a DEC Microvax 3200 computer, displayed on a nologies model l/25 (512 x 512 x 24) display system on film with a Matrix 3000 graphics camera. 3.
RESULTS
AND
OF HbS
order to minet al., 1988). RASTER3D Raster Techand recorded
DISCUSSION
Figure 1 presents schematic fiber models which illustrate the effects of tilting a cross section of an HbS fiber. These surface renditions depict left- and right-handed twists and they contain the fraction of a turn contained in a typical HbS fiber cross section (ca. 45”). The electron micrographs of cross sections in Fig. 2 can be interpreted by simply comparing them to the models in Fig. 1. Below we describe in detail how this is accomplished.
FIBERS
197
The helices in the model cross sections depicted in Fig. 1 are inclined to the particle axis and their inclination increases with radius from about 2.5” near the axis to about 12” for the outermost strands. When the cross section is viewed parallel to the particle axis, as in the 0” tilt views in Fig. 1, the inclination of each strand to the axis is the same as its inclination to the direction of view. In a cross section of a fiber the long pitch helices twist through a limited arc of usually 10”-60”. The direction of twist depends upon the hand of the helix, as shown in Fig. 1 which presents a perspective view of right- and left-handed renditions of an HbS fiber. In an electron micrograph of a cross section the density is a projection rather than a perspective and thus the direction of the twist cannot be determined
gh
FIG. 1. Schematic views of HbS fiber cross sections showing the effects of tilting right-handed (upper) and by small angles. The particle axis is depicted as a thin cylinder in the center of the cross section. The direction is indicated by the arrows. The long pitch strands of HbS molecules are depicted as curved tubes. The changes tilting through a narrow angle can be fully appreciated only by viewing in stereo. The stereo pairs are the left center and right images. (This is best done by viewing the left images with the left eye and the right images figure the common practice of crossing eyes does notproduce a satisfactory stereo effect.) Note, for example, that model by +2.5” foreshortens the strand just above the axis and elongates the strand below the axis, while the for the left-handed model.
left-handed (lower) models the top of the axis is tilted in appearance caused by and center images and the with the right eye. In this tilting the 0” right-handed opposite effect is observed
198
LEWIS,
LEE,
AND
JOSEPHS
a +2.5’
d FIG. 2. Electron micrographs of cross-sectioned HbS fibers taken before and after tilting the cross sections (a and b). The views are shown as pairs whose tilt angles correspond to the end-on views in Fig. 1. The bright spots which are seen in the fiber cross sections correspond to strands which appear foreshortened as depicted in Fig. 1. The positions of the bright spots shift when the particle is tilted. The direction of the shift shows that these particles are right-handed. In c and d the sections have been tilted by + 2.5” - 2.5”. Some of the fibers show bright spots in both micrographs but the position of the spots has moved. For instance the particle indicated in c and d has a bright spot directly above its center in the left view, while a strand directly below its center appears as a bright spot in the right view, as predicted for a right-handed particle. Calibration bar equals 250 A.
from a single micrograph since it is not possible to distinguish between the top and bottom of the particle. Tilting changes the strand’s angle of inclination to the direction of view. When the view angle approaches zero degrees the strand is viewed roughly “end-on.” Consequently the molecules in the strand appear nearly superimposed (in projection). This results in the formation of a “bright spot” of high density which appears as the apparent center of rotation of the particle. The position of the spot depends on the direction of tilt and the hand of the helix. Thus tilting a particle in a known direction and noting the position of the spot provides a means of determining the particle’s hand.
Figure 1 presents axial views of fiber models. The figure illustrates how the hand of the helix affects end-on views of the central strands in the HbS fiber. For a given tilt angle the strand which appears endon would be expected to be on opposite sides of the axis for left- and right-handed particles. For example, if the fiber is tilted + 2.5” as in Fig. 1, upper, the strand above the axis appears end-on in the righthanded model, while in the left-handed model in Fig. 1, lower, the strand below the axis is viewed end-on. Their positions are reversed with a tilt of - 23, but they are still on opposite sides of the particle for left- and right-handed helices. Similarly a tilt of 5” (say from + 2.5” to - 2.5”) moves the strand being viewed end-on from the bottom to the top for a
HAND
OF HbS
right-handed particle and from top to bottom for a left-handed particle. These relationships can be fully appreciated by viewing the constructions in stereo. (This is best done by viewing the left image with the left eye and the right image with the right eye. In Fig. 1 the common practice of crossing eyes does not produce a satisfactory stereo effect.) Figures 2a and 2b demonstrate this effect. Figure 2a presents an untilted micrograph and Fig. 2b presents a tilted micrograph of the same fiber. The position of the bright spot establishes that the particle is righthanded. As a practical matter fibers in cross sections are seldom cut exactly normal to the particle axis and their orientation varies by a small but unknown amount about some mean value. This element of uncertainty regarding the particle orientation could result in an incorrect assessment of the particle hand in experiments such as those described by the procedure shown in Fig. 1. This problem can be avoided by tilting a section 2.5” counterclockwise and then 5” clockwise. (These values align different long pitch helices nearly parallel to the direction of view.) The effect of this operation is schematically shown in the upper panel of Fig. 1 for a right-handed particle. The corresponding electron micrographs are shown in Fig. 2. In the first view (Fig. 2~) many of the particles display a bright spot as in the above example. In the second view (Fig. 2d) some of these same particles also show a bright spot but the position of the spot shifted to the opposite side of the particle. By considering only those particles which show a bright spot in both views we select those particles whose orientation is known thereby eliminating the uncertainty in particle orientation. The movement of the spot from top to bottom demonstrates that these particles are right-handed. In Fig. 2 we have selected data for a tilt direction which is coincident with the models presented in Fig. 1. The data we have obtained in other tilt directions, and also by tilting obliquely to the initial cross-sectional orientation, have consistently indicated that HbS fibers are right-handed particles.
199
FIBERS
These data directly and independently confirm the hand proposed by Dykes et al. (1979). We conclude, therefore, that the 14-start helices of HbS fibers are right-handed. CONCLUSION
We have presented a method for determining the hand of helical particles from electron micrographs of their cross sections. We demonstrate the application of the method using HbS fibers as an example. The basis of the procedure involves tilting a cross section of a particle so that a short segment of one of the long pitch helices is nearly parallel to the direction of view. A strand of molecules aligned in this fashion appears as a bright spot. The change in tilt angle which produces an end-on view of different helices indicates the relative tilt of these strands in the fiber and consequently the hand of the particle. This method provides a means for determining the hand of helical particles which is complimentary to that of Finch (1972). These methods are best suited to relatively different kinds of helical particles. Finch’s procedure produces pronounced parallax changes along short pitch helices. In contrast, tilting a cross section has a greater effect on the appearance of long pitch helices. The authors thank Gerald Grofman for sectioning this material, Dr. William McDade for providing embedded HbS fiber samples, and David Bacon for supplying the RASTER3D program. This work was supported by NIH Grant HL30121 (R.J.) and Training Grant GM08282 (M.R.L.). REFERENCES CARRAGHER, B., BLUEMKE, JOSEPHS, R. (1988) J. DYKES,
G., CREPEAU,
D. A., GABRIEL,
B., POTEL,
M. J., AND
Mol. Biol. 199, 315-331.
R. H., AND EDELSTEIN,
S. J. (1979)
J. Mol.
Biol. 130, 451-472. FINCH, J. T. (1972) J. Mol. Biol. 66, 291-294. LUTHER, P. K., LAWKENCE, M. C., AND CROWTHER, Ultramicroscopy 24, 7-18.
R. A. (1988)
MCDADE, W. A., CARRAGHER, B., MILLER, C. A., AND JOSEPHS, R. (1989) J. Mol. Biol. 206, 637-649. POTEL, J. J., WELLEMS, T. E., DEER, B., VASSAR, R. J., AND JoSEPHS, R. (1984) J. Mol. Biol. 177, 819-839.
OF STRUCTURAL
BIOLOGY
107,
l%--WJ
(1991)
The Hand of the Helix of Deoxyhemoglobin MICHAEL Laboratory
for Image Analysis The University
R. LEWIS, THOMAS E. LEE, AND ROBERT JOSEPHS and Electron of Chicago,
Microscopy, 920 East Received
Department 58th Street,
July
IN.
INTRODUCTION
Intracellular polymerization of deoxyhemoglobin S (HbS)l to form fibers is the cause of sickle cell anemia. In viuo this aggregation reaction initiates the cascade of events which characterize sickle cell disease. We have studied the structure and assembly of HbS fibers in order to gain a better understanding of the intermolecular interactions which are responsible for the pathological aggregation of these molecules. The fiber structure has been reconstructed by Carragher et al. (1988) and Dykes et al. (1979). A complete description of the fiber structure requires a knowledge of both the molecular coordinates and the hand of the particle. In principle the hand of a helical particle can be determined from the changes in the phases of the particle’s Fourier transform caused by tilting. Finch (1972) has demonstrated this approach by tilting lateral views of TMVperpendicular 1 Abbreviation
used:
HbS,
deoxyhemoglobin
Inc. reserved.
and
Cell Biology,
24, 1991
2.
MATERIALS
AND
METHODS
Embedded HbS fiber samples prepared according to Potel et al. (1984) and McDade et al. (1989) were provided by Dr. William McDade. Thin sections were cut using a Sorvel MTSB ultramicrotome and poststained with uranyl acetate and lead citrate. A Philips CM10 electron microscope was used to record stereo pairs of electron micrographs at an electron optical magnification of
S. 196
1047~8477/91 $3.00 Copyright 0 1991 by Academic Press, All rights of reproduction in any form
of Molecular Genetics Chicago, Illinois 60637
to the particle axis. This approach produces maximum changes in particle appearance along its short pitch helices. Dykes et al. (1979) used changes in the appearance of lateral views which occur as the particle is rotated about its axis to propose that sickle cell hemoglobin fibers are right-handed. The hand of the HbS fiber is difficult to determine from tilting lateral views for a variety of reasons. The bonding vectors in a fiber are along paired strands of hemoglobin molecules which twist about the axis with a pitch of about 3000 A. In addition it is now recognized that the fiber has a variable pitch. Consequently the Fourier transform of the fiber is extremely complicated as the unit cell consists of 14 molecules in seven quasiequivalent positions. Because of these factors, measurements of phase changes upon tilting can be subject to errors whose magnitude is uncertain, but likely to be considerably larger than in the case of well-ordered particles. These errors could lead to an incorrect assessment of the particle hand. Because a knowledge of the particle hand is an essential prerequisite to synthesizing a molecular model of the fiber we have developed a method for determining the hand from tilting cross sections. The method is based upon the principle that when a short segment of a long pitch helix is tilted so that it is roughly parallel to the direction of view it is resolved in projection as a bright spot. Knowing the tilt direction gives the hand. Finch’s method and the one proposed here are complimentary. Tilted lateral views show pronounced changes along short pitch helices, while in cross sections the most pronounced changes occur along long pitch helices. In practice our approach is simple and might readily be adopted for other helical particles.
Electron micrographs of deoxyhemoglobin S fiber cross sections provide an end-on view of the fiber whose appearance is sensitive to small changes in orientation. We have developed a procedure to exploit this sensitivity in order to determine the hand of these particles. In a sickle hemoglobin fiber the hemoglobin molecules form long pitch helical strands which twist about the particle axis with a pitch of about 3000 A. Tilting a 400~A-thick cross section by a few degrees aligns one of the long pitch helices so that it is nearly parallel to the direction of view. When a strand of hemoglobin molecules in a fiber is aligned in this manner it appears as a strongly contrasted bright spot. It is this spot, rather than the fiber axis, which appears to be the apparent center of rotation of the cross section. The direction of the displacement of the spot from the particle axis depends upon the particle hand and tilt direction. We have used this property to determine that sickle hemoglobin fibers are right-handed particles. This method may be applicable to other particles with long pitch helices as well. o 1991Academic press, 1.
S Fibers
HAND 46 000. The micrographs were recorded promptly in imize section collapse due to beam damage (Luther Models in Fig. 1 were calculated with the program on a DEC Microvax 3200 computer, displayed on a nologies model l/25 (512 x 512 x 24) display system on film with a Matrix 3000 graphics camera. 3.
RESULTS
AND
OF HbS
order to minet al., 1988). RASTER3D Raster Techand recorded
DISCUSSION
Figure 1 presents schematic fiber models which illustrate the effects of tilting a cross section of an HbS fiber. These surface renditions depict left- and right-handed twists and they contain the fraction of a turn contained in a typical HbS fiber cross section (ca. 45”). The electron micrographs of cross sections in Fig. 2 can be interpreted by simply comparing them to the models in Fig. 1. Below we describe in detail how this is accomplished.
FIBERS
197
The helices in the model cross sections depicted in Fig. 1 are inclined to the particle axis and their inclination increases with radius from about 2.5” near the axis to about 12” for the outermost strands. When the cross section is viewed parallel to the particle axis, as in the 0” tilt views in Fig. 1, the inclination of each strand to the axis is the same as its inclination to the direction of view. In a cross section of a fiber the long pitch helices twist through a limited arc of usually 10”-60”. The direction of twist depends upon the hand of the helix, as shown in Fig. 1 which presents a perspective view of right- and left-handed renditions of an HbS fiber. In an electron micrograph of a cross section the density is a projection rather than a perspective and thus the direction of the twist cannot be determined
gh
FIG. 1. Schematic views of HbS fiber cross sections showing the effects of tilting right-handed (upper) and by small angles. The particle axis is depicted as a thin cylinder in the center of the cross section. The direction is indicated by the arrows. The long pitch strands of HbS molecules are depicted as curved tubes. The changes tilting through a narrow angle can be fully appreciated only by viewing in stereo. The stereo pairs are the left center and right images. (This is best done by viewing the left images with the left eye and the right images figure the common practice of crossing eyes does notproduce a satisfactory stereo effect.) Note, for example, that model by +2.5” foreshortens the strand just above the axis and elongates the strand below the axis, while the for the left-handed model.
left-handed (lower) models the top of the axis is tilted in appearance caused by and center images and the with the right eye. In this tilting the 0” right-handed opposite effect is observed
198
LEWIS,
LEE,
AND
JOSEPHS
a +2.5’
d FIG. 2. Electron micrographs of cross-sectioned HbS fibers taken before and after tilting the cross sections (a and b). The views are shown as pairs whose tilt angles correspond to the end-on views in Fig. 1. The bright spots which are seen in the fiber cross sections correspond to strands which appear foreshortened as depicted in Fig. 1. The positions of the bright spots shift when the particle is tilted. The direction of the shift shows that these particles are right-handed. In c and d the sections have been tilted by + 2.5” - 2.5”. Some of the fibers show bright spots in both micrographs but the position of the spots has moved. For instance the particle indicated in c and d has a bright spot directly above its center in the left view, while a strand directly below its center appears as a bright spot in the right view, as predicted for a right-handed particle. Calibration bar equals 250 A.
from a single micrograph since it is not possible to distinguish between the top and bottom of the particle. Tilting changes the strand’s angle of inclination to the direction of view. When the view angle approaches zero degrees the strand is viewed roughly “end-on.” Consequently the molecules in the strand appear nearly superimposed (in projection). This results in the formation of a “bright spot” of high density which appears as the apparent center of rotation of the particle. The position of the spot depends on the direction of tilt and the hand of the helix. Thus tilting a particle in a known direction and noting the position of the spot provides a means of determining the particle’s hand.
Figure 1 presents axial views of fiber models. The figure illustrates how the hand of the helix affects end-on views of the central strands in the HbS fiber. For a given tilt angle the strand which appears endon would be expected to be on opposite sides of the axis for left- and right-handed particles. For example, if the fiber is tilted + 2.5” as in Fig. 1, upper, the strand above the axis appears end-on in the righthanded model, while in the left-handed model in Fig. 1, lower, the strand below the axis is viewed end-on. Their positions are reversed with a tilt of - 23, but they are still on opposite sides of the particle for left- and right-handed helices. Similarly a tilt of 5” (say from + 2.5” to - 2.5”) moves the strand being viewed end-on from the bottom to the top for a
HAND
OF HbS
right-handed particle and from top to bottom for a left-handed particle. These relationships can be fully appreciated by viewing the constructions in stereo. (This is best done by viewing the left image with the left eye and the right image with the right eye. In Fig. 1 the common practice of crossing eyes does not produce a satisfactory stereo effect.) Figures 2a and 2b demonstrate this effect. Figure 2a presents an untilted micrograph and Fig. 2b presents a tilted micrograph of the same fiber. The position of the bright spot establishes that the particle is righthanded. As a practical matter fibers in cross sections are seldom cut exactly normal to the particle axis and their orientation varies by a small but unknown amount about some mean value. This element of uncertainty regarding the particle orientation could result in an incorrect assessment of the particle hand in experiments such as those described by the procedure shown in Fig. 1. This problem can be avoided by tilting a section 2.5” counterclockwise and then 5” clockwise. (These values align different long pitch helices nearly parallel to the direction of view.) The effect of this operation is schematically shown in the upper panel of Fig. 1 for a right-handed particle. The corresponding electron micrographs are shown in Fig. 2. In the first view (Fig. 2~) many of the particles display a bright spot as in the above example. In the second view (Fig. 2d) some of these same particles also show a bright spot but the position of the spot shifted to the opposite side of the particle. By considering only those particles which show a bright spot in both views we select those particles whose orientation is known thereby eliminating the uncertainty in particle orientation. The movement of the spot from top to bottom demonstrates that these particles are right-handed. In Fig. 2 we have selected data for a tilt direction which is coincident with the models presented in Fig. 1. The data we have obtained in other tilt directions, and also by tilting obliquely to the initial cross-sectional orientation, have consistently indicated that HbS fibers are right-handed particles.
199
FIBERS
These data directly and independently confirm the hand proposed by Dykes et al. (1979). We conclude, therefore, that the 14-start helices of HbS fibers are right-handed. CONCLUSION
We have presented a method for determining the hand of helical particles from electron micrographs of their cross sections. We demonstrate the application of the method using HbS fibers as an example. The basis of the procedure involves tilting a cross section of a particle so that a short segment of one of the long pitch helices is nearly parallel to the direction of view. A strand of molecules aligned in this fashion appears as a bright spot. The change in tilt angle which produces an end-on view of different helices indicates the relative tilt of these strands in the fiber and consequently the hand of the particle. This method provides a means for determining the hand of helical particles which is complimentary to that of Finch (1972). These methods are best suited to relatively different kinds of helical particles. Finch’s procedure produces pronounced parallax changes along short pitch helices. In contrast, tilting a cross section has a greater effect on the appearance of long pitch helices. The authors thank Gerald Grofman for sectioning this material, Dr. William McDade for providing embedded HbS fiber samples, and David Bacon for supplying the RASTER3D program. This work was supported by NIH Grant HL30121 (R.J.) and Training Grant GM08282 (M.R.L.). REFERENCES CARRAGHER, B., BLUEMKE, JOSEPHS, R. (1988) J. DYKES,
G., CREPEAU,
D. A., GABRIEL,
B., POTEL,
M. J., AND
Mol. Biol. 199, 315-331.
R. H., AND EDELSTEIN,
S. J. (1979)
J. Mol.
Biol. 130, 451-472. FINCH, J. T. (1972) J. Mol. Biol. 66, 291-294. LUTHER, P. K., LAWKENCE, M. C., AND CROWTHER, Ultramicroscopy 24, 7-18.
R. A. (1988)
MCDADE, W. A., CARRAGHER, B., MILLER, C. A., AND JOSEPHS, R. (1989) J. Mol. Biol. 206, 637-649. POTEL, J. J., WELLEMS, T. E., DEER, B., VASSAR, R. J., AND JoSEPHS, R. (1984) J. Mol. Biol. 177, 819-839.
